From ac5b7505de7013a4c572aa80d5b43db287fa0161 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:56:33 -0700 Subject: [PATCH 001/123] scsi: mpi3mr: struct mpi3_event_data_sas_topology_change_list: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct mpi3_event_data_sas_topology_change_list with a modern flexible array. Additionally add __counted_by annotation since phy_entry is only ever accessed in loops controlled by num_entries. For example: for (i = 0; i < event_data->num_entries; i++) { ... handle = le16_to_cpu(event_data->phy_entry[i].attached_dev_handle); No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155637.3757036-1-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi/mpi30_ioc.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h index 028784949873..ae74fccc65b8 100644 --- a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h +++ b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h @@ -453,9 +453,6 @@ struct mpi3_event_data_sas_notify_primitive { #define MPI3_EVENT_NOTIFY_PRIMITIVE_POWER_LOSS_EXPECTED (0x02) #define MPI3_EVENT_NOTIFY_PRIMITIVE_RESERVED1 (0x03) #define MPI3_EVENT_NOTIFY_PRIMITIVE_RESERVED2 (0x04) -#ifndef MPI3_EVENT_SAS_TOPO_PHY_COUNT -#define MPI3_EVENT_SAS_TOPO_PHY_COUNT (1) -#endif struct mpi3_event_sas_topo_phy_entry { __le16 attached_dev_handle; u8 link_rate; @@ -496,7 +493,7 @@ struct mpi3_event_data_sas_topology_change_list { u8 start_phy_num; u8 exp_status; u8 io_unit_port; - struct mpi3_event_sas_topo_phy_entry phy_entry[MPI3_EVENT_SAS_TOPO_PHY_COUNT]; + struct mpi3_event_sas_topo_phy_entry phy_entry[] __counted_by(num_entries); }; #define MPI3_EVENT_SAS_TOPO_ES_NO_EXPANDER (0x00) From 0e11f97bfddc608fd31d4fb7b8feaf74755d6372 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:56:34 -0700 Subject: [PATCH 002/123] scsi: mpi3mr: struct mpi3_event_data_pcie_topology_change_list: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct mpi3_event_data_pcie_topology_change_list with a modern flexible array. Additionally add __counted_by annotation since port_entry is only ever accessed in loops controlled by num_entries. For example: for (i = 0; i < event_data->num_entries; i++) { handle = le16_to_cpu(event_data->port_entry[i].attached_dev_handle); No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155637.3757036-2-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi/mpi30_ioc.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h index ae74fccc65b8..c9fa0d69b75f 100644 --- a/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h +++ b/drivers/scsi/mpi3mr/mpi/mpi30_ioc.h @@ -542,9 +542,6 @@ struct mpi3_event_data_pcie_enumeration { #define MPI3_EVENT_PCIE_ENUM_ES_MAX_SWITCHES_EXCEED (0x40000000) #define MPI3_EVENT_PCIE_ENUM_ES_MAX_DEVICES_EXCEED (0x20000000) #define MPI3_EVENT_PCIE_ENUM_ES_RESOURCES_EXHAUSTED (0x10000000) -#ifndef MPI3_EVENT_PCIE_TOPO_PORT_COUNT -#define MPI3_EVENT_PCIE_TOPO_PORT_COUNT (1) -#endif struct mpi3_event_pcie_topo_port_entry { __le16 attached_dev_handle; u8 port_status; @@ -585,7 +582,7 @@ struct mpi3_event_data_pcie_topology_change_list { u8 switch_status; u8 io_unit_port; __le32 reserved0c; - struct mpi3_event_pcie_topo_port_entry port_entry[MPI3_EVENT_PCIE_TOPO_PORT_COUNT]; + struct mpi3_event_pcie_topo_port_entry port_entry[] __counted_by(num_entries); }; #define MPI3_EVENT_PCIE_TOPO_SS_NO_PCIE_SWITCH (0x00) From 41bb96296f9d3a92f3b52bcf502d6c0e86fa84a0 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:56:35 -0700 Subject: [PATCH 003/123] scsi: mpi3mr: struct mpi3_sas_io_unit_page0: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct mpi3_sas_io_unit_page0 with a modern flexible array. No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155637.3757036-3-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h index 6a19e17eb1a7..66cca35d8e52 100644 --- a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h +++ b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h @@ -1565,16 +1565,13 @@ struct mpi3_sas_io_unit0_phy_data { __le32 reserved10; }; -#ifndef MPI3_SAS_IO_UNIT0_PHY_MAX -#define MPI3_SAS_IO_UNIT0_PHY_MAX (1) -#endif struct mpi3_sas_io_unit_page0 { struct mpi3_config_page_header header; __le32 reserved08; u8 num_phys; u8 init_status; __le16 reserved0e; - struct mpi3_sas_io_unit0_phy_data phy_data[MPI3_SAS_IO_UNIT0_PHY_MAX]; + struct mpi3_sas_io_unit0_phy_data phy_data[]; }; #define MPI3_SASIOUNIT0_PAGEVERSION (0x00) From a62193abae75f3c3f68bbc7d3343751644140556 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:56:36 -0700 Subject: [PATCH 004/123] scsi: mpi3mr: struct mpi3_sas_io_unit_page1: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct mpi3_sas_io_unit_page1 with a modern flexible array. No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155637.3757036-4-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h index 66cca35d8e52..4b7a8f6314a3 100644 --- a/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h +++ b/drivers/scsi/mpi3mr/mpi/mpi30_cnfg.h @@ -1603,9 +1603,6 @@ struct mpi3_sas_io_unit1_phy_data { __le32 reserved08; }; -#ifndef MPI3_SAS_IO_UNIT1_PHY_MAX -#define MPI3_SAS_IO_UNIT1_PHY_MAX (1) -#endif struct mpi3_sas_io_unit_page1 { struct mpi3_config_page_header header; __le16 control_flags; @@ -1615,7 +1612,7 @@ struct mpi3_sas_io_unit_page1 { u8 num_phys; u8 sata_max_q_depth; __le16 reserved12; - struct mpi3_sas_io_unit1_phy_data phy_data[MPI3_SAS_IO_UNIT1_PHY_MAX]; + struct mpi3_sas_io_unit1_phy_data phy_data[]; }; #define MPI3_SASIOUNIT1_PAGEVERSION (0x00) From ed8ab02c85b3494ec81f35f12e01f6678af09f6a Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:58:23 -0700 Subject: [PATCH 005/123] scsi: megaraid_sas: struct MR_LD_VF_MAP: Replace 1-element arrays with flexible arrays Replace the deprecated[1] use of a 1-element array in struct MR_LD_VF_MAP with a modern flexible array. No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155823.work.778-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/megaraid/megaraid_sas.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/megaraid/megaraid_sas.h b/drivers/scsi/megaraid/megaraid_sas.h index 5680c6cdb221..84cf77c48c0d 100644 --- a/drivers/scsi/megaraid/megaraid_sas.h +++ b/drivers/scsi/megaraid/megaraid_sas.h @@ -2473,7 +2473,7 @@ struct MR_LD_VF_MAP { union MR_LD_REF ref; u8 ldVfCount; u8 reserved[6]; - u8 policy[1]; + u8 policy[]; }; struct MR_LD_VF_AFFILIATION { From 29b4a49750776925ea48ef440da6aa75828913db Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 08:58:42 -0700 Subject: [PATCH 006/123] scsi: megaraid_sas: struct MR_HOST_DEVICE_LIST: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct MR_HOST_DEVICE_LIST with a modern flexible array. One binary difference appears in megasas_host_device_list_query(): struct MR_HOST_DEVICE_LIST *ci; ... ci = instance->host_device_list_buf; ... memset(ci, 0, sizeof(*ci)); The memset() clears only the non-flexible array fields. Looking at the rest of the function, this appears to be fine: firmware is using this region to communicate with the kernel, so it likely never made sense to clear the first MR_HOST_DEVICE_LIST_ENTRY. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711155841.work.839-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/megaraid/megaraid_sas.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/megaraid/megaraid_sas.h b/drivers/scsi/megaraid/megaraid_sas.h index 84cf77c48c0d..088cc40ae866 100644 --- a/drivers/scsi/megaraid/megaraid_sas.h +++ b/drivers/scsi/megaraid/megaraid_sas.h @@ -814,12 +814,12 @@ struct MR_HOST_DEVICE_LIST { __le32 size; __le32 count; __le32 reserved[2]; - struct MR_HOST_DEVICE_LIST_ENTRY host_device_list[1]; + struct MR_HOST_DEVICE_LIST_ENTRY host_device_list[] __counted_by_le(count); } __packed; #define HOST_DEVICE_LIST_SZ (sizeof(struct MR_HOST_DEVICE_LIST) + \ (sizeof(struct MR_HOST_DEVICE_LIST_ENTRY) * \ - (MEGASAS_MAX_PD + MAX_LOGICAL_DRIVES_EXT - 1))) + (MEGASAS_MAX_PD + MAX_LOGICAL_DRIVES_EXT))) /* From 575b9be63684600e7e02517f0b647e5cb759120c Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:48:23 -0700 Subject: [PATCH 007/123] scsi: aacraid: union aac_init: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in union aac_init with a modern flexible array. Additionally add __counted_by annotation since rrq is only ever accessed after rr_queue_count has been set (with the same value used to control the loop): init->r8.rr_queue_count = cpu_to_le32(dev->max_msix); ... for (i = 0; i < dev->max_msix; i++) { addr = (u64)dev->host_rrq_pa + dev->vector_cap * i * sizeof(u32); init->r8.rrq[i].host_addr_high = cpu_to_le32( upper_32_bits(addr)); No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711174815.work.689-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/aacraid/aacraid.h | 2 +- drivers/scsi/aacraid/src.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/aacraid/aacraid.h b/drivers/scsi/aacraid/aacraid.h index 7d5a155073c6..659e393c1033 100644 --- a/drivers/scsi/aacraid/aacraid.h +++ b/drivers/scsi/aacraid/aacraid.h @@ -873,7 +873,7 @@ union aac_init __le16 element_count; __le16 comp_thresh; __le16 unused; - } rrq[1]; /* up to 64 RRQ addresses */ + } rrq[] __counted_by_le(rr_queue_count); /* up to 64 RRQ addresses */ } r8; }; diff --git a/drivers/scsi/aacraid/src.c b/drivers/scsi/aacraid/src.c index 11ef58204e96..28115ed637e8 100644 --- a/drivers/scsi/aacraid/src.c +++ b/drivers/scsi/aacraid/src.c @@ -410,7 +410,7 @@ static void aac_src_start_adapter(struct aac_dev *dev) lower_32_bits(dev->init_pa), upper_32_bits(dev->init_pa), sizeof(struct _r8) + - (AAC_MAX_HRRQ - 1) * sizeof(struct _rrq), + AAC_MAX_HRRQ * sizeof(struct _rrq), 0, 0, 0, NULL, NULL, NULL, NULL, NULL); } else { init->r7.host_elapsed_seconds = From 2e35b43bc9a82fde4e7aebe5d8331e1158374d5c Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:50:55 -0700 Subject: [PATCH 008/123] scsi: aacraid: struct aac_ciss_phys_luns_resp: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct aac_ciss_phys_luns_resp with a modern flexible array. No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711175055.work.928-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Reviewed-by: Kees Cook Signed-off-by: Martin K. Petersen --- drivers/scsi/aacraid/aachba.c | 2 +- drivers/scsi/aacraid/aacraid.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/aacraid/aachba.c b/drivers/scsi/aacraid/aachba.c index b22857c6f3f4..497c6dd5df91 100644 --- a/drivers/scsi/aacraid/aachba.c +++ b/drivers/scsi/aacraid/aachba.c @@ -1833,7 +1833,7 @@ static int aac_get_safw_ciss_luns(struct aac_dev *dev) struct aac_ciss_phys_luns_resp *phys_luns; datasize = sizeof(struct aac_ciss_phys_luns_resp) + - (AAC_MAX_TARGETS - 1) * sizeof(struct _ciss_lun); + AAC_MAX_TARGETS * sizeof(struct _ciss_lun); phys_luns = kmalloc(datasize, GFP_KERNEL); if (phys_luns == NULL) goto out; diff --git a/drivers/scsi/aacraid/aacraid.h b/drivers/scsi/aacraid/aacraid.h index 659e393c1033..6f0417f6f8a1 100644 --- a/drivers/scsi/aacraid/aacraid.h +++ b/drivers/scsi/aacraid/aacraid.h @@ -322,7 +322,7 @@ struct aac_ciss_phys_luns_resp { u8 level3[2]; u8 level2[2]; u8 node_ident[16]; /* phys. node identifier */ - } lun[1]; /* List of phys. devices */ + } lun[]; /* List of phys. devices */ }; /* From c72e13cf820bc79fffd5ffaaa0aaab11269027d1 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 11:07:06 -0700 Subject: [PATCH 009/123] scsi: ipr: Replace 1-element arrays with flexible arrays Replace the deprecated[1] use of a 1-element arrays in struct ipr_hostrcb_fabric_desc and struct ipr_hostrcb64_fabric_desc with modern flexible arrays. No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711180702.work.536-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/scsi/ipr.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/ipr.h b/drivers/scsi/ipr.h index c77d6ca1a210..b2b643c6dbbe 100644 --- a/drivers/scsi/ipr.h +++ b/drivers/scsi/ipr.h @@ -1030,7 +1030,7 @@ struct ipr_hostrcb_fabric_desc { #define IPR_PATH_FAILED 0x03 __be16 num_entries; - struct ipr_hostrcb_config_element elem[1]; + struct ipr_hostrcb_config_element elem[]; }__attribute__((packed, aligned (4))); struct ipr_hostrcb64_fabric_desc { @@ -1044,7 +1044,7 @@ struct ipr_hostrcb64_fabric_desc { u8 res_path[8]; u8 reserved3[6]; __be16 num_entries; - struct ipr_hostrcb64_config_element elem[1]; + struct ipr_hostrcb64_config_element elem[]; }__attribute__((packed, aligned (8))); #define for_each_hrrq(hrrq, ioa_cfg) \ From 8e76c9c9dd117c95c0ae248a217197228000912c Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:15 -0700 Subject: [PATCH 010/123] scsi: message: fusion: struct _RAID_VOL0_SETTINGS: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _RAID_VOL0_SETTINGS with a modern flexible array. Additionally add __counted_by annotation since PhysDisk is only ever accessed via a loops bounded by NumPhysDisks: lsi/mpi_cnfg.h: RAID_VOL0_PHYS_DISK PhysDisk[] __counted_by(NumPhysDisks); /* 28h */ mptbase.c: for (i = 0; i < buffer->NumPhysDisks; i++) { mptbase.c: buffer->PhysDisk[i].PhysDiskNum, &phys_disk) != 0) mptsas.c: for (i = 0; i < buffer->NumPhysDisks; i++) { mptsas.c: buffer->PhysDisk[i].PhysDiskNum, &phys_disk) != 0) mptsas.c: for (i = 0; i < buffer->NumPhysDisks; i++) { mptsas.c: buffer->PhysDisk[i].PhysDiskNum, &phys_disk) != 0) No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-1-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index 3770cb1cff7d..f59a741ef21c 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -2295,14 +2295,6 @@ typedef struct _RAID_VOL0_SETTINGS #define MPI_RAID_HOT_SPARE_POOL_6 (0x40) #define MPI_RAID_HOT_SPARE_POOL_7 (0x80) -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength at runtime. - */ -#ifndef MPI_RAID_VOL_PAGE_0_PHYSDISK_MAX -#define MPI_RAID_VOL_PAGE_0_PHYSDISK_MAX (1) -#endif - typedef struct _CONFIG_PAGE_RAID_VOL_0 { CONFIG_PAGE_HEADER Header; /* 00h */ @@ -2321,7 +2313,7 @@ typedef struct _CONFIG_PAGE_RAID_VOL_0 U8 DataScrubRate; /* 25h */ U8 ResyncRate; /* 26h */ U8 InactiveStatus; /* 27h */ - RAID_VOL0_PHYS_DISK PhysDisk[MPI_RAID_VOL_PAGE_0_PHYSDISK_MAX];/* 28h */ + RAID_VOL0_PHYS_DISK PhysDisk[] __counted_by(NumPhysDisks); /* 28h */ } CONFIG_PAGE_RAID_VOL_0, MPI_POINTER PTR_CONFIG_PAGE_RAID_VOL_0, RaidVolumePage0_t, MPI_POINTER pRaidVolumePage0_t; From 14c1f88c7f625dc12fb4a6f18154bcd6f7cd021f Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:16 -0700 Subject: [PATCH 011/123] scsi: message: fusion: struct _CONFIG_PAGE_SAS_IO_UNIT_0: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _CONFIG_PAGE_SAS_IO_UNIT_0 with a modern flexible array. Additionally add __counted_by annotation since PhyData is only ever accessed via a loops bounded by NumPhys: lsi/mpi_cnfg.h: MPI_SAS_IO_UNIT0_PHY_DATA PhyData[] __counted_by(NumPhys); /* 10h */ mptsas.c: port_info->num_phys = buffer->NumPhys; mptsas.c: for (i = 0; i < port_info->num_phys; i++) { mptsas.c: mptsas_print_phy_data(ioc, &buffer->PhyData[i]); mptsas.c: port_info->phy_info[i].phy_id = i; mptsas.c: port_info->phy_info[i].port_id = mptsas.c: buffer->PhyData[i].Port; mptsas.c: port_info->phy_info[i].negotiated_link_rate = mptsas.c: buffer->PhyData[i].NegotiatedLinkRate; No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-2-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index f59a741ef21c..c7997e32e82e 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -2547,14 +2547,6 @@ typedef struct _MPI_SAS_IO_UNIT0_PHY_DATA } MPI_SAS_IO_UNIT0_PHY_DATA, MPI_POINTER PTR_MPI_SAS_IO_UNIT0_PHY_DATA, SasIOUnit0PhyData, MPI_POINTER pSasIOUnit0PhyData; -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength at runtime. - */ -#ifndef MPI_SAS_IOUNIT0_PHY_MAX -#define MPI_SAS_IOUNIT0_PHY_MAX (1) -#endif - typedef struct _CONFIG_PAGE_SAS_IO_UNIT_0 { CONFIG_EXTENDED_PAGE_HEADER Header; /* 00h */ @@ -2563,7 +2555,7 @@ typedef struct _CONFIG_PAGE_SAS_IO_UNIT_0 U8 NumPhys; /* 0Ch */ U8 Reserved2; /* 0Dh */ U16 Reserved3; /* 0Eh */ - MPI_SAS_IO_UNIT0_PHY_DATA PhyData[MPI_SAS_IOUNIT0_PHY_MAX]; /* 10h */ + MPI_SAS_IO_UNIT0_PHY_DATA PhyData[] __counted_by(NumPhys); /* 10h */ } CONFIG_PAGE_SAS_IO_UNIT_0, MPI_POINTER PTR_CONFIG_PAGE_SAS_IO_UNIT_0, SasIOUnitPage0_t, MPI_POINTER pSasIOUnitPage0_t; From dc8932fbf6a9cad4cf6dd312d115a26d90facb7e Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:17 -0700 Subject: [PATCH 012/123] scsi: message: fusion: struct _CONFIG_PAGE_RAID_PHYS_DISK_1: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _CONFIG_PAGE_RAID_PHYS_DISK_1 with a modern flexible array. Additionally add __counted_by annotation since Path is only ever accessed via a loops bounded by NumPhysDiskPaths: lsi/mpi_cnfg.h: RAID_PHYS_DISK1_PATH Path[] __counted_by(NumPhysDiskPaths);/* 0Ch */ mptbase.c: phys_disk->NumPhysDiskPaths = buffer->NumPhysDiskPaths; mptbase.c: for (i = 0; i < phys_disk->NumPhysDiskPaths; i++) { mptbase.c: phys_disk->Path[i].PhysDiskID = buffer->Path[i].PhysDiskID; mptbase.c: phys_disk->Path[i].PhysDiskBus = buffer->Path[i].PhysDiskBus; No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-3-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index c7997e32e82e..e30132b57ae7 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -2447,14 +2447,6 @@ typedef struct _RAID_PHYS_DISK1_PATH #define MPI_RAID_PHYSDISK1_FLAG_INVALID (0x0001) -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength or NumPhysDiskPaths at runtime. - */ -#ifndef MPI_RAID_PHYS_DISK1_PATH_MAX -#define MPI_RAID_PHYS_DISK1_PATH_MAX (1) -#endif - typedef struct _CONFIG_PAGE_RAID_PHYS_DISK_1 { CONFIG_PAGE_HEADER Header; /* 00h */ @@ -2462,7 +2454,7 @@ typedef struct _CONFIG_PAGE_RAID_PHYS_DISK_1 U8 PhysDiskNum; /* 05h */ U16 Reserved2; /* 06h */ U32 Reserved1; /* 08h */ - RAID_PHYS_DISK1_PATH Path[MPI_RAID_PHYS_DISK1_PATH_MAX];/* 0Ch */ + RAID_PHYS_DISK1_PATH Path[] __counted_by(NumPhysDiskPaths);/* 0Ch */ } CONFIG_PAGE_RAID_PHYS_DISK_1, MPI_POINTER PTR_CONFIG_PAGE_RAID_PHYS_DISK_1, RaidPhysDiskPage1_t, MPI_POINTER pRaidPhysDiskPage1_t; From de80fe29ab53dc9efde5c773fadebd3e547ad785 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:18 -0700 Subject: [PATCH 013/123] scsi: message: fusion: struct _CONFIG_PAGE_IOC_2: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _CONFIG_PAGE_IOC_2 with a modern flexible array. Additionally add __counted_by annotation since RaidVolume is only ever accessed from loops controlled by NumActiveVolumes: lsi/mpi_cnfg.h: CONFIG_PAGE_IOC_2_RAID_VOL RaidVolume[] __counted_by(NumActiveVolumes); /* 0Ch */ mptbase.c: for (i = 0; i < pIoc2->NumActiveVolumes ; i++) mptbase.c: pIoc2->RaidVolume[i].VolumeBus, mptbase.c: pIoc2->RaidVolume[i].VolumeID); mptsas.c: for (i = 0; i < ioc->raid_data.pIocPg2->NumActiveVolumes; i++) { mptsas.c: RaidVolume[i].VolumeID) { mptsas.c: RaidVolume[i].VolumeBus; mptsas.c: for (i = 0; i < ioc->raid_data.pIocPg2->NumActiveVolumes; i++) { mptsas.c: ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID, 0); mptsas.c: ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID); mptsas.c: ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID, 0); mptsas.c: for (i = 0; i < ioc->raid_data.pIocPg2->NumActiveVolumes; i++) { mptsas.c: if (ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID == mptsas.c: for (i = 0; i < ioc->raid_data.pIocPg2->NumActiveVolumes; i++) mptsas.c: if (ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID == id) mptspi.c: for (i=0; i < ioc->raid_data.pIocPg2->NumActiveVolumes; i++) { mptspi.c: if (ioc->raid_data.pIocPg2->RaidVolume[i].VolumeID == id) { No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-4-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index e30132b57ae7..7713c74e515b 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -1018,14 +1018,6 @@ typedef struct _CONFIG_PAGE_IOC_2_RAID_VOL #define MPI_IOCPAGE2_FLAG_VOLUME_INACTIVE (0x08) -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength at runtime. - */ -#ifndef MPI_IOC_PAGE_2_RAID_VOLUME_MAX -#define MPI_IOC_PAGE_2_RAID_VOLUME_MAX (1) -#endif - typedef struct _CONFIG_PAGE_IOC_2 { CONFIG_PAGE_HEADER Header; /* 00h */ @@ -1034,7 +1026,7 @@ typedef struct _CONFIG_PAGE_IOC_2 U8 MaxVolumes; /* 09h */ U8 NumActivePhysDisks; /* 0Ah */ U8 MaxPhysDisks; /* 0Bh */ - CONFIG_PAGE_IOC_2_RAID_VOL RaidVolume[MPI_IOC_PAGE_2_RAID_VOLUME_MAX];/* 0Ch */ + CONFIG_PAGE_IOC_2_RAID_VOL RaidVolume[] __counted_by(NumActiveVolumes); /* 0Ch */ } CONFIG_PAGE_IOC_2, MPI_POINTER PTR_CONFIG_PAGE_IOC_2, IOCPage2_t, MPI_POINTER pIOCPage2_t; From 70631322dbab5b48f49f142dd9aef768c0fb077c Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:19 -0700 Subject: [PATCH 014/123] scsi: message: fusion: struct _CONFIG_PAGE_IOC_3: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _CONFIG_PAGE_IOC_3 with a modern flexible array. Additionally add __counted_by annotation since PhysDisk is only ever accessed via a loops bounded by NumPhysDisks: lsi/mpi_cnfg.h: IOC_3_PHYS_DISK PhysDisk[] __counted_by(NumPhysDisks); /* 08h */ mptscsih.c: for (i = 0; i < ioc->raid_data.pIocPg3->NumPhysDisks; i++) { mptscsih.c: if ((id == ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskID) && mptscsih.c: (channel == ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskBus)) { mptscsih.c: for (i = 0; i < ioc->raid_data.pIocPg3->NumPhysDisks; i++) { mptscsih.c: ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskNum); mptscsih.c: ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskNum, mptscsih.c: for (i = 0; i < ioc->raid_data.pIocPg3->NumPhysDisks; i++) { mptscsih.c: if ((id == ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskID) && mptscsih.c: (channel == ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskBus)) { mptscsih.c: rc = ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskNum; mptscsih.c: for (i = 0; i < ioc->raid_data.pIocPg3->NumPhysDisks; i++) { mptscsih.c: ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskNum); mptscsih.c: ioc->raid_data.pIocPg3->PhysDisk[i].PhysDiskNum, No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-5-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index 7713c74e515b..bac49c162165 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -1056,21 +1056,13 @@ typedef struct _IOC_3_PHYS_DISK } IOC_3_PHYS_DISK, MPI_POINTER PTR_IOC_3_PHYS_DISK, Ioc3PhysDisk_t, MPI_POINTER pIoc3PhysDisk_t; -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength at runtime. - */ -#ifndef MPI_IOC_PAGE_3_PHYSDISK_MAX -#define MPI_IOC_PAGE_3_PHYSDISK_MAX (1) -#endif - typedef struct _CONFIG_PAGE_IOC_3 { CONFIG_PAGE_HEADER Header; /* 00h */ U8 NumPhysDisks; /* 04h */ U8 Reserved1; /* 05h */ U16 Reserved2; /* 06h */ - IOC_3_PHYS_DISK PhysDisk[MPI_IOC_PAGE_3_PHYSDISK_MAX]; /* 08h */ + IOC_3_PHYS_DISK PhysDisk[] __counted_by(NumPhysDisks); /* 08h */ } CONFIG_PAGE_IOC_3, MPI_POINTER PTR_CONFIG_PAGE_IOC_3, IOCPage3_t, MPI_POINTER pIOCPage3_t; From f296cc1d7f5aeccd0d87ec167e0aea05acb8a022 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 10:28:20 -0700 Subject: [PATCH 015/123] scsi: message: fusion: struct _CONFIG_PAGE_IOC_4: Replace 1-element array with flexible array Replace the deprecated[1] use of a 1-element array in struct _CONFIG_PAGE_IOC_4 with a modern flexible array. Additionally add __counted_by annotation since SEP is only ever accessed after updating ACtiveSEP: lsi/mpi_cnfg.h: IOC_4_SEP SEP[] __counted_by(ActiveSEP); /* 08h */ mptsas.c: ii = IOCPage4Ptr->ActiveSEP++; mptsas.c: IOCPage4Ptr->SEP[ii].SEPTargetID = id; mptsas.c: IOCPage4Ptr->SEP[ii].SEPBus = channel; No binary differences are present after this conversion. Link: https://github.com/KSPP/linux/issues/79 [1] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711172821.123936-6-kees@kernel.org Reviewed-by: Gustavo A. R. Silva Signed-off-by: Martin K. Petersen --- drivers/message/fusion/lsi/mpi_cnfg.h | 10 +--------- 1 file changed, 1 insertion(+), 9 deletions(-) diff --git a/drivers/message/fusion/lsi/mpi_cnfg.h b/drivers/message/fusion/lsi/mpi_cnfg.h index bac49c162165..1167a16d8fb4 100644 --- a/drivers/message/fusion/lsi/mpi_cnfg.h +++ b/drivers/message/fusion/lsi/mpi_cnfg.h @@ -1077,21 +1077,13 @@ typedef struct _IOC_4_SEP } IOC_4_SEP, MPI_POINTER PTR_IOC_4_SEP, Ioc4Sep_t, MPI_POINTER pIoc4Sep_t; -/* - * Host code (drivers, BIOS, utilities, etc.) should leave this define set to - * one and check Header.PageLength at runtime. - */ -#ifndef MPI_IOC_PAGE_4_SEP_MAX -#define MPI_IOC_PAGE_4_SEP_MAX (1) -#endif - typedef struct _CONFIG_PAGE_IOC_4 { CONFIG_PAGE_HEADER Header; /* 00h */ U8 ActiveSEP; /* 04h */ U8 MaxSEP; /* 05h */ U16 Reserved1; /* 06h */ - IOC_4_SEP SEP[MPI_IOC_PAGE_4_SEP_MAX]; /* 08h */ + IOC_4_SEP SEP[] __counted_by(ActiveSEP); /* 08h */ } CONFIG_PAGE_IOC_4, MPI_POINTER PTR_CONFIG_PAGE_IOC_4, IOCPage4_t, MPI_POINTER pIOCPage4_t; From 6e5860b0ad4934baee8c7a202c02033b2631bb44 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 14:57:37 -0700 Subject: [PATCH 016/123] scsi: aacraid: Rearrange order of struct aac_srb_unit struct aac_srb_unit contains struct aac_srb, which contains struct sgmap, which ends in a (currently) "fake" (1-element) flexible array. Converting this to a flexible array is needed so that runtime bounds checking won't think the array is fixed size (i.e. under CONFIG_FORTIFY_SOURCE=y and/or CONFIG_UBSAN_BOUNDS=y), as other parts of aacraid use struct sgmap as a flexible array. It is not legal to have a flexible array in the middle of a structure, so it either needs to be split up or rearranged so that it is at the end of the structure. Luckily, struct aac_srb_unit, which is exclusively consumed/updated by aac_send_safw_bmic_cmd(), does not depend on member ordering. The values set in the on-stack struct aac_srb_unit instance "srbu" by the only two callers, aac_issue_safw_bmic_identify() and aac_get_safw_ciss_luns(), do not contain anything in srbu.srb.sgmap.sg, and they both implicitly initialize srbu.srb.sgmap.count to 0 during memset(). For example: memset(&srbu, 0, sizeof(struct aac_srb_unit)); srbcmd = &srbu.srb; srbcmd->flags = cpu_to_le32(SRB_DataIn); srbcmd->cdb[0] = CISS_REPORT_PHYSICAL_LUNS; srbcmd->cdb[1] = 2; /* extended reporting */ srbcmd->cdb[8] = (u8)(datasize >> 8); srbcmd->cdb[9] = (u8)(datasize); rcode = aac_send_safw_bmic_cmd(dev, &srbu, phys_luns, datasize); During aac_send_safw_bmic_cmd(), a separate srb is mapped into DMA, and has srbu.srb copied into it: srb = fib_data(fibptr); memcpy(srb, &srbu->srb, sizeof(struct aac_srb)); Only then is srb.sgmap.count written and srb->sg populated: srb->count = cpu_to_le32(xfer_len); sg64 = (struct sgmap64 *)&srb->sg; sg64->count = cpu_to_le32(1); sg64->sg[0].addr[1] = cpu_to_le32(upper_32_bits(addr)); sg64->sg[0].addr[0] = cpu_to_le32(lower_32_bits(addr)); sg64->sg[0].count = cpu_to_le32(xfer_len); But this is happening in the DMA memory, not in srbu.srb. An attempt to copy the changes back to srbu does happen: /* * Copy the updated data for other dumping or other usage if * needed */ memcpy(&srbu->srb, srb, sizeof(struct aac_srb)); But this was never correct: the sg64 (3 u32s) overlap of srb.sg (2 u32s) always meant that srbu.srb would have held truncated information and any attempt to walk srbu.srb.sg.sg based on the value of srbu.srb.sg.count would result in attempting to parse past the end of srbu.srb.sg.sg[0] into srbu.srb_reply. After getting a reply from hardware, the reply is copied into srbu.srb_reply: srb_reply = (struct aac_srb_reply *)fib_data(fibptr); memcpy(&srbu->srb_reply, srb_reply, sizeof(struct aac_srb_reply)); This has always been fixed-size, so there's no issue here. It is worth noting that the two callers _never check_ srbu contents -- neither srbu.srb nor srbu.srb_reply is examined. (They depend on the mapped xfer_buf instead.) Therefore, the ordering of members in struct aac_srb_unit does not matter, and the flexible array member can moved to the end. (Additionally, the two memcpy()s that update srbu could be entirely removed as they are never consumed, but I left that as-is.) Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711215739.208776-1-kees@kernel.org Signed-off-by: Martin K. Petersen --- drivers/scsi/aacraid/aacraid.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/aacraid/aacraid.h b/drivers/scsi/aacraid/aacraid.h index 7d5a155073c6..9b66fa29fb05 100644 --- a/drivers/scsi/aacraid/aacraid.h +++ b/drivers/scsi/aacraid/aacraid.h @@ -2029,8 +2029,8 @@ struct aac_srb_reply }; struct aac_srb_unit { - struct aac_srb srb; struct aac_srb_reply srb_reply; + struct aac_srb srb; }; /* From fdb1db6ea7f66cad970b19b5cd341b8386350bca Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Thu, 11 Jul 2024 14:57:38 -0700 Subject: [PATCH 017/123] scsi: aacraid: struct {user,}sgmap{,64,raw}: Replace 1-element arrays with flexible arrays Replace the deprecated[1] use of 1-element arrays in struct sgmap, struct sgmap64, struct sgmapraw, struct user_sgmap, and struct user_sgmap64 with modern flexible arrays. Additionally remove struct user_sgmapraw as it is unused. The resulting binary output differences from this change are limited only to stack space consumption of the smaller "srbu" variable in aac_issue_safw_bmic_identify() and aac_get_safw_ciss_luns(), as well as the smaller associated pair of memcpy()s in aac_send_safw_bmic_cmd(). Artificially growing the size of srbu back to its prior size removes all binary differences[2]. As an aside, after studying the aacraid driver code I wonder how aac_send_wellness_command() ever works. It is reporting a size 4 bytes too small for what it has constructed in memory in the DMA region: sgentry64 is size 12, whereas sgentry is size 8. Perhaps the hardware doesn't care. (Regardless, it is unchanged by this patch.) Link: https://github.com/KSPP/linux/issues/79 [1] Link: https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/commit/?h=dev/v6.10-rc2/1-element&id=45e6226bcbc5e982541754eca7ac29f403e82f5e [2] Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20240711215739.208776-2-kees@kernel.org Signed-off-by: Martin K. Petersen --- drivers/scsi/aacraid/aachba.c | 26 ++++++++++++-------------- drivers/scsi/aacraid/aacraid.h | 15 +++++---------- drivers/scsi/aacraid/commctrl.c | 4 ++-- drivers/scsi/aacraid/comminit.c | 3 +-- drivers/scsi/aacraid/commsup.c | 5 +++-- 5 files changed, 23 insertions(+), 30 deletions(-) diff --git a/drivers/scsi/aacraid/aachba.c b/drivers/scsi/aacraid/aachba.c index b22857c6f3f4..6a7159badebb 100644 --- a/drivers/scsi/aacraid/aachba.c +++ b/drivers/scsi/aacraid/aachba.c @@ -1267,7 +1267,7 @@ static int aac_read_raw_io(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u3 return ret; command = ContainerRawIo; fibsize = sizeof(struct aac_raw_io) + - ((le32_to_cpu(readcmd->sg.count)-1) * sizeof(struct sgentryraw)); + (le32_to_cpu(readcmd->sg.count) * sizeof(struct sgentryraw)); } BUG_ON(fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1302,7 +1302,7 @@ static int aac_read_block64(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u if (ret < 0) return ret; fibsize = sizeof(struct aac_read64) + - ((le32_to_cpu(readcmd->sg.count) - 1) * + (le32_to_cpu(readcmd->sg.count) * sizeof (struct sgentry64)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1337,7 +1337,7 @@ static int aac_read_block(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u32 if (ret < 0) return ret; fibsize = sizeof(struct aac_read) + - ((le32_to_cpu(readcmd->sg.count) - 1) * + (le32_to_cpu(readcmd->sg.count) * sizeof (struct sgentry)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1401,7 +1401,7 @@ static int aac_write_raw_io(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u return ret; command = ContainerRawIo; fibsize = sizeof(struct aac_raw_io) + - ((le32_to_cpu(writecmd->sg.count)-1) * sizeof (struct sgentryraw)); + (le32_to_cpu(writecmd->sg.count) * sizeof(struct sgentryraw)); } BUG_ON(fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1436,7 +1436,7 @@ static int aac_write_block64(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, if (ret < 0) return ret; fibsize = sizeof(struct aac_write64) + - ((le32_to_cpu(writecmd->sg.count) - 1) * + (le32_to_cpu(writecmd->sg.count) * sizeof (struct sgentry64)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1473,7 +1473,7 @@ static int aac_write_block(struct fib * fib, struct scsi_cmnd * cmd, u64 lba, u3 if (ret < 0) return ret; fibsize = sizeof(struct aac_write) + - ((le32_to_cpu(writecmd->sg.count) - 1) * + (le32_to_cpu(writecmd->sg.count) * sizeof (struct sgentry)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1592,9 +1592,9 @@ static int aac_scsi_64(struct fib * fib, struct scsi_cmnd * cmd) /* * Build Scatter/Gather list */ - fibsize = sizeof (struct aac_srb) - sizeof (struct sgentry) + + fibsize = sizeof(struct aac_srb) + ((le32_to_cpu(srbcmd->sg.count) & 0xff) * - sizeof (struct sgentry64)); + sizeof(struct sgentry64)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1624,7 +1624,7 @@ static int aac_scsi_32(struct fib * fib, struct scsi_cmnd * cmd) * Build Scatter/Gather list */ fibsize = sizeof (struct aac_srb) + - (((le32_to_cpu(srbcmd->sg.count) & 0xff) - 1) * + ((le32_to_cpu(srbcmd->sg.count) & 0xff) * sizeof (struct sgentry)); BUG_ON (fibsize > (fib->dev->max_fib_size - sizeof(struct aac_fibhdr))); @@ -1693,8 +1693,7 @@ static int aac_send_safw_bmic_cmd(struct aac_dev *dev, fibptr->hw_fib_va->header.XferState &= ~cpu_to_le32(FastResponseCapable); - fibsize = sizeof(struct aac_srb) - sizeof(struct sgentry) + - sizeof(struct sgentry64); + fibsize = sizeof(struct aac_srb) + sizeof(struct sgentry64); /* allocate DMA buffer for response */ addr = dma_map_single(&dev->pdev->dev, xfer_buf, xfer_len, @@ -2267,7 +2266,7 @@ int aac_get_adapter_info(struct aac_dev* dev) dev->a_ops.adapter_bounds = aac_bounds_32; dev->scsi_host_ptr->sg_tablesize = (dev->max_fib_size - sizeof(struct aac_fibhdr) - - sizeof(struct aac_write) + sizeof(struct sgentry)) / + sizeof(struct aac_write)) / sizeof(struct sgentry); if (dev->dac_support) { dev->a_ops.adapter_read = aac_read_block64; @@ -2278,8 +2277,7 @@ int aac_get_adapter_info(struct aac_dev* dev) dev->scsi_host_ptr->sg_tablesize = (dev->max_fib_size - sizeof(struct aac_fibhdr) - - sizeof(struct aac_write64) + - sizeof(struct sgentry64)) / + sizeof(struct aac_write64)) / sizeof(struct sgentry64); } else { dev->a_ops.adapter_read = aac_read_block; diff --git a/drivers/scsi/aacraid/aacraid.h b/drivers/scsi/aacraid/aacraid.h index 9b66fa29fb05..63e74266d7a2 100644 --- a/drivers/scsi/aacraid/aacraid.h +++ b/drivers/scsi/aacraid/aacraid.h @@ -507,32 +507,27 @@ struct sge_ieee1212 { struct sgmap { __le32 count; - struct sgentry sg[1]; + struct sgentry sg[]; }; struct user_sgmap { u32 count; - struct user_sgentry sg[1]; + struct user_sgentry sg[]; }; struct sgmap64 { __le32 count; - struct sgentry64 sg[1]; + struct sgentry64 sg[]; }; struct user_sgmap64 { u32 count; - struct user_sgentry64 sg[1]; + struct user_sgentry64 sg[]; }; struct sgmapraw { __le32 count; - struct sgentryraw sg[1]; -}; - -struct user_sgmapraw { - u32 count; - struct user_sgentryraw sg[1]; + struct sgentryraw sg[]; }; struct creation_info diff --git a/drivers/scsi/aacraid/commctrl.c b/drivers/scsi/aacraid/commctrl.c index e7cc927ed952..68240d6f27ab 100644 --- a/drivers/scsi/aacraid/commctrl.c +++ b/drivers/scsi/aacraid/commctrl.c @@ -523,7 +523,7 @@ static int aac_send_raw_srb(struct aac_dev* dev, void __user * arg) goto cleanup; } - if ((fibsize < (sizeof(struct user_aac_srb) - sizeof(struct user_sgentry))) || + if ((fibsize < sizeof(struct user_aac_srb)) || (fibsize > (dev->max_fib_size - sizeof(struct aac_fibhdr)))) { rcode = -EINVAL; goto cleanup; @@ -561,7 +561,7 @@ static int aac_send_raw_srb(struct aac_dev* dev, void __user * arg) rcode = -EINVAL; goto cleanup; } - actual_fibsize = sizeof(struct aac_srb) - sizeof(struct sgentry) + + actual_fibsize = sizeof(struct aac_srb) + ((user_srbcmd->sg.count & 0xff) * sizeof(struct sgentry)); actual_fibsize64 = actual_fibsize + (user_srbcmd->sg.count & 0xff) * (sizeof(struct sgentry64) - sizeof(struct sgentry)); diff --git a/drivers/scsi/aacraid/comminit.c b/drivers/scsi/aacraid/comminit.c index bd99c5492b7d..fee857236991 100644 --- a/drivers/scsi/aacraid/comminit.c +++ b/drivers/scsi/aacraid/comminit.c @@ -522,8 +522,7 @@ struct aac_dev *aac_init_adapter(struct aac_dev *dev) spin_lock_init(&dev->iq_lock); dev->max_fib_size = sizeof(struct hw_fib); dev->sg_tablesize = host->sg_tablesize = (dev->max_fib_size - - sizeof(struct aac_fibhdr) - - sizeof(struct aac_write) + sizeof(struct sgentry)) + - sizeof(struct aac_fibhdr) - sizeof(struct aac_write)) / sizeof(struct sgentry); dev->comm_interface = AAC_COMM_PRODUCER; dev->raw_io_interface = dev->raw_io_64 = 0; diff --git a/drivers/scsi/aacraid/commsup.c b/drivers/scsi/aacraid/commsup.c index 25cee03d7f97..47287559c768 100644 --- a/drivers/scsi/aacraid/commsup.c +++ b/drivers/scsi/aacraid/commsup.c @@ -2327,8 +2327,9 @@ static int aac_send_wellness_command(struct aac_dev *dev, char *wellness_str, sg64->sg[0].addr[0] = cpu_to_le32((u32)(addr & 0xffffffff)); sg64->sg[0].count = cpu_to_le32(datasize); - ret = aac_fib_send(ScsiPortCommand64, fibptr, sizeof(struct aac_srb), - FsaNormal, 1, 1, NULL, NULL); + ret = aac_fib_send(ScsiPortCommand64, fibptr, + sizeof(struct aac_srb) + sizeof(struct sgentry), + FsaNormal, 1, 1, NULL, NULL); dma_free_coherent(&dev->pdev->dev, datasize, dma_buf, addr); From 88e6804fb323d1dbe477f002a89755f8b034e890 Mon Sep 17 00:00:00 2001 From: "Bao D. Nguyen" Date: Tue, 23 Jul 2024 20:49:32 -0700 Subject: [PATCH 018/123] scsi: ufs: core: Support Updating UIC Command Timeout The default UIC command timeout still remains 500ms. Allow platform drivers to override the UIC command timeout if desired. In a real product, the 500ms timeout value is probably good enough. However, during the product development where there are a lot of logging and debug messages being printed to the UART console, interrupt starvations happen occasionally because the UART may print long debug messages from different modules in the system. While printing, the UART may have interrupts disabled for more than 500ms, causing UIC command timeout. The UIC command timeout would trigger more printing from the UFS driver, and eventually a watchdog timeout may occur unnecessarily. Add support for overriding the UIC command timeout value with the newly created uic_cmd_timeout kernel module parameter. Default value is 500ms. Supported values range from 500ms to 2 seconds. Signed-off-by: Bao D. Nguyen Link: https://lore.kernel.org/r/e4e1c87f3f867f270a3d4b5d57a00139ff0e9741.1721792309.git.quic_nguyenb@quicinc.com Suggested-by: Bart Van Assche Reviewed-by: Bart Van Assche Reviewed-by: Manivannan Sadhasivam Reviewed-by: Peter Wang Signed-off-by: Martin K. Petersen --- drivers/ufs/core/ufshcd.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index dc757ba47522..0dd26059f5d7 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -51,8 +51,10 @@ /* UIC command timeout, unit: ms */ -#define UIC_CMD_TIMEOUT 500 - +enum { + UIC_CMD_TIMEOUT_DEFAULT = 500, + UIC_CMD_TIMEOUT_MAX = 2000, +}; /* NOP OUT retries waiting for NOP IN response */ #define NOP_OUT_RETRIES 10 /* Timeout after 50 msecs if NOP OUT hangs without response */ @@ -116,6 +118,23 @@ static bool is_mcq_supported(struct ufs_hba *hba) module_param(use_mcq_mode, bool, 0644); MODULE_PARM_DESC(use_mcq_mode, "Control MCQ mode for controllers starting from UFSHCI 4.0. 1 - enable MCQ, 0 - disable MCQ. MCQ is enabled by default"); +static unsigned int uic_cmd_timeout = UIC_CMD_TIMEOUT_DEFAULT; + +static int uic_cmd_timeout_set(const char *val, const struct kernel_param *kp) +{ + return param_set_uint_minmax(val, kp, UIC_CMD_TIMEOUT_DEFAULT, + UIC_CMD_TIMEOUT_MAX); +} + +static const struct kernel_param_ops uic_cmd_timeout_ops = { + .set = uic_cmd_timeout_set, + .get = param_get_uint, +}; + +module_param_cb(uic_cmd_timeout, &uic_cmd_timeout_ops, &uic_cmd_timeout, 0644); +MODULE_PARM_DESC(uic_cmd_timeout, + "UFS UIC command timeout in milliseconds. Defaults to 500ms. Supported values range from 500ms to 2 seconds inclusively"); + #define ufshcd_toggle_vreg(_dev, _vreg, _on) \ ({ \ int _ret; \ @@ -2438,7 +2457,7 @@ static inline bool ufshcd_ready_for_uic_cmd(struct ufs_hba *hba) { u32 val; int ret = read_poll_timeout(ufshcd_readl, val, val & UIC_COMMAND_READY, - 500, UIC_CMD_TIMEOUT * 1000, false, hba, + 500, uic_cmd_timeout * 1000, false, hba, REG_CONTROLLER_STATUS); return ret == 0; } @@ -2498,7 +2517,7 @@ ufshcd_wait_for_uic_cmd(struct ufs_hba *hba, struct uic_command *uic_cmd) lockdep_assert_held(&hba->uic_cmd_mutex); if (wait_for_completion_timeout(&uic_cmd->done, - msecs_to_jiffies(UIC_CMD_TIMEOUT))) { + msecs_to_jiffies(uic_cmd_timeout))) { ret = uic_cmd->argument2 & MASK_UIC_COMMAND_RESULT; } else { ret = -ETIMEDOUT; @@ -4266,7 +4285,7 @@ static int ufshcd_uic_pwr_ctrl(struct ufs_hba *hba, struct uic_command *cmd) } if (!wait_for_completion_timeout(hba->uic_async_done, - msecs_to_jiffies(UIC_CMD_TIMEOUT))) { + msecs_to_jiffies(uic_cmd_timeout))) { dev_err(hba->dev, "pwr ctrl cmd 0x%x with mode 0x%x completion timeout\n", cmd->command, cmd->argument3); From 5b8963c53de11dcef2b6241452988427db5773ec Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:05 -0700 Subject: [PATCH 019/123] scsi: lpfc: Change diagnostic log flag during receipt of unknown ELS cmds During diagnostics, it has been determined that the 0115 log message for receipt of unknown ELS cmds does not benefit from trace buffer dumps. The trace buffer dump floods the console with unnecessary information, and the singular LOG_ELS flag has proven more beneficial in debugging efforts when dealing with unknown ELS cmds. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-2-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_els.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 929cbfc95163..50c0c0c91fdc 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -10742,7 +10742,7 @@ lpfc_els_unsol_buffer(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, rjt_exp = LSEXP_NOTHING_MORE; /* Unknown ELS command received from NPORT */ - lpfc_printf_vlog(vport, KERN_ERR, LOG_TRACE_EVENT, + lpfc_printf_vlog(vport, KERN_ERR, LOG_ELS, "0115 Unknown ELS command x%x " "received from NPORT x%x\n", cmd, did); if (newnode) From f1bfe32073964372c1dfa572480eb43d44e34909 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:06 -0700 Subject: [PATCH 020/123] scsi: lpfc: Remove redundant vport assignment when building an abort request The lpfc_sli_issue_abort_iotag() routine has a redundant assignment of abtsiocbp->vport = vport; The duplicate lines are from a previous refactoring, and this patch removes the accidental redundancy. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-3-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_sli.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index 88debef2fb6d..d240bbded4c8 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -12473,8 +12473,6 @@ lpfc_sli_issue_abort_iotag(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, cmdiocb->iocb.ulpClass, LPFC_WQE_CQ_ID_DEFAULT, ia, false); - abtsiocbp->vport = vport; - /* ABTS WQE must go to the same WQ as the WQE to be aborted */ abtsiocbp->hba_wqidx = cmdiocb->hba_wqidx; if (cmdiocb->cmd_flag & LPFC_IO_FCP) From 2be1d4f11944cd6283cb97268b3e17c4424945ca Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:07 -0700 Subject: [PATCH 021/123] scsi: lpfc: Validate hdwq pointers before dereferencing in reset/errata paths When the HBA is undergoing a reset or is handling an errata event, NULL ptr dereference crashes may occur in routines such as lpfc_sli_flush_io_rings(), lpfc_dev_loss_tmo_callbk(), or lpfc_abort_handler(). Add NULL ptr checks before dereferencing hdwq pointers that may have been freed due to operations colliding with a reset or errata event handler. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-4-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_hbadisc.c | 3 ++- drivers/scsi/lpfc/lpfc_scsi.c | 13 +++++++++++-- drivers/scsi/lpfc/lpfc_sli.c | 11 +++++++++++ 3 files changed, 24 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c index 6943f6c6395c..f21c5993e8d7 100644 --- a/drivers/scsi/lpfc/lpfc_hbadisc.c +++ b/drivers/scsi/lpfc/lpfc_hbadisc.c @@ -175,7 +175,8 @@ lpfc_dev_loss_tmo_callbk(struct fc_rport *rport) ndlp->nlp_state, ndlp->fc4_xpt_flags); /* Don't schedule a worker thread event if the vport is going down. */ - if (test_bit(FC_UNLOADING, &vport->load_flag)) { + if (test_bit(FC_UNLOADING, &vport->load_flag) || + !test_bit(HBA_SETUP, &phba->hba_flag)) { spin_lock_irqsave(&ndlp->lock, iflags); ndlp->rport = NULL; diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index 98ce9d97a225..60cd60ebff38 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -5555,11 +5555,20 @@ lpfc_abort_handler(struct scsi_cmnd *cmnd) iocb = &lpfc_cmd->cur_iocbq; if (phba->sli_rev == LPFC_SLI_REV4) { - pring_s4 = phba->sli4_hba.hdwq[iocb->hba_wqidx].io_wq->pring; - if (!pring_s4) { + /* if the io_wq & pring are gone, the port was reset. */ + if (!phba->sli4_hba.hdwq[iocb->hba_wqidx].io_wq || + !phba->sli4_hba.hdwq[iocb->hba_wqidx].io_wq->pring) { + lpfc_printf_vlog(vport, KERN_WARNING, LOG_FCP, + "2877 SCSI Layer I/O Abort Request " + "IO CMPL Status x%x ID %d LUN %llu " + "HBA_SETUP %d\n", FAILED, + cmnd->device->id, + (u64)cmnd->device->lun, + test_bit(HBA_SETUP, &phba->hba_flag)); ret = FAILED; goto out_unlock_hba; } + pring_s4 = phba->sli4_hba.hdwq[iocb->hba_wqidx].io_wq->pring; spin_lock(&pring_s4->ring_lock); } /* the command is in process of being cancelled */ diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c index d240bbded4c8..332b8d2348e9 100644 --- a/drivers/scsi/lpfc/lpfc_sli.c +++ b/drivers/scsi/lpfc/lpfc_sli.c @@ -4687,6 +4687,17 @@ lpfc_sli_flush_io_rings(struct lpfc_hba *phba) /* Look on all the FCP Rings for the iotag */ if (phba->sli_rev >= LPFC_SLI_REV4) { for (i = 0; i < phba->cfg_hdw_queue; i++) { + if (!phba->sli4_hba.hdwq || + !phba->sli4_hba.hdwq[i].io_wq) { + lpfc_printf_log(phba, KERN_ERR, LOG_SLI, + "7777 hdwq's deleted %lx " + "%lx %x %x\n", + phba->pport->load_flag, + phba->hba_flag, + phba->link_state, + phba->sli.sli_flag); + return; + } pring = phba->sli4_hba.hdwq[i].io_wq->pring; spin_lock_irq(&pring->ring_lock); From 3976beb1b410441bab9c3726e2ba76cc7a4c0b2d Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:08 -0700 Subject: [PATCH 022/123] scsi: lpfc: Fix unintentional double clearing of vmid_flag The vport->vmid_flag is unintentionally cleared twice after an issue_lip via the lpfc_reinit_vmid routine(). The first call to lpfc_reinit_vmid() is in lpfc_cmpl_els_flogi(). Then lpfc_cmpl_els_flogi_fabric() calls lpfc_register_new_vport(), which calls lpfc_cmpl_reg_new_vport() when the mbox command completes and calls lpfc_reinit_vmid() a second time. Fix by moving the vmid_flag clear outside of the lpfc_reinit_vmid() routine so that vmid_flag is only cleared once upon FLOGI completion. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-5-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_els.c | 4 +++- drivers/scsi/lpfc/lpfc_vmid.c | 1 - 2 files changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 50c0c0c91fdc..6d49e23f6a62 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -1099,8 +1099,10 @@ stop_rr_fcf_flogi: sp->cmn.priority_tagging, kref_read(&ndlp->kref)); /* reinitialize the VMID datastructure before returning */ - if (lpfc_is_vmid_enabled(phba)) + if (lpfc_is_vmid_enabled(phba)) { lpfc_reinit_vmid(vport); + vport->vmid_flag = 0; + } if (sp->cmn.priority_tagging) vport->phba->pport->vmid_flag |= (LPFC_VMID_ISSUE_QFPA | LPFC_VMID_TYPE_PRIO); diff --git a/drivers/scsi/lpfc/lpfc_vmid.c b/drivers/scsi/lpfc/lpfc_vmid.c index 773e02ae20c3..cf8ba840d0ea 100644 --- a/drivers/scsi/lpfc/lpfc_vmid.c +++ b/drivers/scsi/lpfc/lpfc_vmid.c @@ -321,6 +321,5 @@ lpfc_reinit_vmid(struct lpfc_vport *vport) if (!hash_empty(vport->hash_table)) hash_for_each_safe(vport->hash_table, bucket, tmp, cur, hnode) hash_del(&cur->hnode); - vport->vmid_flag = 0; write_unlock(&vport->vmid_lock); } From b5c18c9dd138733c16893613345af44deadcf05e Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:09 -0700 Subject: [PATCH 023/123] scsi: lpfc: Fix unsolicited FLOGI kref imbalance when in direct attached topology In direct attached topology, certain target vendors that are quick to issue FLOGI followed by a cable pull for more than dev_loss_tmo may result in a kref imbalance for the remote port ndlp object. Add an nlp_get when the defer_flogi_acc flag is set. This is expected to balance the nlp_put in the defer_flogi_acc clause in the lpfc_issue_els_flogi() routine. Because we need to retain the ndlp ptr, reorganize all of the defer_flogi_acc information into one lpfc_defer_flogi_acc struct. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-6-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc.h | 12 ++++++--- drivers/scsi/lpfc/lpfc_els.c | 46 +++++++++++++++++++------------- drivers/scsi/lpfc/lpfc_hbadisc.c | 11 ++++++-- 3 files changed, 46 insertions(+), 23 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc.h b/drivers/scsi/lpfc/lpfc.h index 7c147d6ea8a8..e5a9c5a323f8 100644 --- a/drivers/scsi/lpfc/lpfc.h +++ b/drivers/scsi/lpfc/lpfc.h @@ -306,6 +306,14 @@ struct lpfc_stats { struct lpfc_hba; +/* Data structure to keep withheld FLOGI_ACC information */ +struct lpfc_defer_flogi_acc { + bool flag; + u16 rx_id; + u16 ox_id; + struct lpfc_nodelist *ndlp; + +}; #define LPFC_VMID_TIMER 300 /* timer interval in seconds */ @@ -1430,9 +1438,7 @@ struct lpfc_hba { uint16_t vlan_id; struct list_head fcf_conn_rec_list; - bool defer_flogi_acc_flag; - uint16_t defer_flogi_acc_rx_id; - uint16_t defer_flogi_acc_ox_id; + struct lpfc_defer_flogi_acc defer_flogi_acc; spinlock_t ct_ev_lock; /* synchronize access to ct_ev_waiters */ struct list_head ct_ev_waiters; diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index 6d49e23f6a62..b5a8d050419a 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -1392,7 +1392,7 @@ lpfc_issue_els_flogi(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, phba->link_flag &= ~LS_EXTERNAL_LOOPBACK; /* Check for a deferred FLOGI ACC condition */ - if (phba->defer_flogi_acc_flag) { + if (phba->defer_flogi_acc.flag) { /* lookup ndlp for received FLOGI */ ndlp = lpfc_findnode_did(vport, 0); if (!ndlp) @@ -1406,34 +1406,38 @@ lpfc_issue_els_flogi(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, if (phba->sli_rev == LPFC_SLI_REV4) { bf_set(wqe_ctxt_tag, &defer_flogi_acc.wqe.xmit_els_rsp.wqe_com, - phba->defer_flogi_acc_rx_id); + phba->defer_flogi_acc.rx_id); bf_set(wqe_rcvoxid, &defer_flogi_acc.wqe.xmit_els_rsp.wqe_com, - phba->defer_flogi_acc_ox_id); + phba->defer_flogi_acc.ox_id); } else { icmd = &defer_flogi_acc.iocb; - icmd->ulpContext = phba->defer_flogi_acc_rx_id; + icmd->ulpContext = phba->defer_flogi_acc.rx_id; icmd->unsli3.rcvsli3.ox_id = - phba->defer_flogi_acc_ox_id; + phba->defer_flogi_acc.ox_id; } lpfc_printf_vlog(vport, KERN_INFO, LOG_ELS, "3354 Xmit deferred FLOGI ACC: rx_id: x%x," " ox_id: x%x, hba_flag x%lx\n", - phba->defer_flogi_acc_rx_id, - phba->defer_flogi_acc_ox_id, phba->hba_flag); + phba->defer_flogi_acc.rx_id, + phba->defer_flogi_acc.ox_id, phba->hba_flag); /* Send deferred FLOGI ACC */ lpfc_els_rsp_acc(vport, ELS_CMD_FLOGI, &defer_flogi_acc, ndlp, NULL); - phba->defer_flogi_acc_flag = false; - vport->fc_myDID = did; + phba->defer_flogi_acc.flag = false; - /* Decrement ndlp reference count to indicate the node can be - * released when other references are removed. + /* Decrement the held ndlp that was incremented when the + * deferred flogi acc flag was set. */ - lpfc_nlp_put(ndlp); + if (phba->defer_flogi_acc.ndlp) { + lpfc_nlp_put(phba->defer_flogi_acc.ndlp); + phba->defer_flogi_acc.ndlp = NULL; + } + + vport->fc_myDID = did; } return 0; @@ -8456,9 +8460,9 @@ lpfc_els_rcv_flogi(struct lpfc_vport *vport, struct lpfc_iocbq *cmdiocb, /* Defer ACC response until AFTER we issue a FLOGI */ if (!test_bit(HBA_FLOGI_ISSUED, &phba->hba_flag)) { - phba->defer_flogi_acc_rx_id = bf_get(wqe_ctxt_tag, + phba->defer_flogi_acc.rx_id = bf_get(wqe_ctxt_tag, &wqe->xmit_els_rsp.wqe_com); - phba->defer_flogi_acc_ox_id = bf_get(wqe_rcvoxid, + phba->defer_flogi_acc.ox_id = bf_get(wqe_rcvoxid, &wqe->xmit_els_rsp.wqe_com); vport->fc_myDID = did; @@ -8466,11 +8470,17 @@ lpfc_els_rcv_flogi(struct lpfc_vport *vport, struct lpfc_iocbq *cmdiocb, lpfc_printf_vlog(vport, KERN_INFO, LOG_ELS, "3344 Deferring FLOGI ACC: rx_id: x%x," " ox_id: x%x, hba_flag x%lx\n", - phba->defer_flogi_acc_rx_id, - phba->defer_flogi_acc_ox_id, phba->hba_flag); + phba->defer_flogi_acc.rx_id, + phba->defer_flogi_acc.ox_id, phba->hba_flag); - phba->defer_flogi_acc_flag = true; + phba->defer_flogi_acc.flag = true; + /* This nlp_get is paired with nlp_puts that reset the + * defer_flogi_acc.flag back to false. We need to retain + * a kref on the ndlp until the deferred FLOGI ACC is + * processed or cancelled. + */ + phba->defer_flogi_acc.ndlp = lpfc_nlp_get(ndlp); return 0; } @@ -10506,7 +10516,7 @@ lpfc_els_unsol_buffer(struct lpfc_hba *phba, struct lpfc_sli_ring *pring, lpfc_els_rcv_flogi(vport, elsiocb, ndlp); /* retain node if our response is deferred */ - if (phba->defer_flogi_acc_flag) + if (phba->defer_flogi_acc.flag) break; if (newnode) lpfc_disc_state_machine(vport, ndlp, NULL, diff --git a/drivers/scsi/lpfc/lpfc_hbadisc.c b/drivers/scsi/lpfc/lpfc_hbadisc.c index f21c5993e8d7..35c9181c6608 100644 --- a/drivers/scsi/lpfc/lpfc_hbadisc.c +++ b/drivers/scsi/lpfc/lpfc_hbadisc.c @@ -1255,7 +1255,14 @@ lpfc_linkdown(struct lpfc_hba *phba) lpfc_scsi_dev_block(phba); offline = pci_channel_offline(phba->pcidev); - phba->defer_flogi_acc_flag = false; + /* Decrement the held ndlp if there is a deferred flogi acc */ + if (phba->defer_flogi_acc.flag) { + if (phba->defer_flogi_acc.ndlp) { + lpfc_nlp_put(phba->defer_flogi_acc.ndlp); + phba->defer_flogi_acc.ndlp = NULL; + } + } + phba->defer_flogi_acc.flag = false; /* Clear external loopback plug detected flag */ phba->link_flag &= ~LS_EXTERNAL_LOOPBACK; @@ -1377,7 +1384,7 @@ lpfc_linkup_port(struct lpfc_vport *vport) (vport != phba->pport)) return; - if (phba->defer_flogi_acc_flag) { + if (phba->defer_flogi_acc.flag) { clear_bit(FC_ABORT_DISCOVERY, &vport->fc_flag); clear_bit(FC_RSCN_MODE, &vport->fc_flag); clear_bit(FC_NLP_MORE, &vport->fc_flag); From 1f0f7679ad8942f810b0f19ee9cf098c3502d66a Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:10 -0700 Subject: [PATCH 024/123] scsi: lpfc: Update PRLO handling in direct attached topology A kref imbalance occurs when handling an unsolicited PRLO in direct attached topology. Rework PRLO rcv handling when in MAPPED state. Save the state that we were handling a PRLO by setting nlp_last_elscmd to ELS_CMD_PRLO. Then in the lpfc_cmpl_els_logo_acc() completion routine, manually restart discovery. By issuing the PLOGI, which nlp_gets, before nlp_put at the end of the lpfc_cmpl_els_logo_acc() routine, we are saving us from a final nlp_put. And, we are still allowing the unreg_rpi to happen. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-7-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_els.c | 27 ++++++++++++++++----------- drivers/scsi/lpfc/lpfc_nportdisc.c | 22 ++++++++++++++++++++-- 2 files changed, 36 insertions(+), 13 deletions(-) diff --git a/drivers/scsi/lpfc/lpfc_els.c b/drivers/scsi/lpfc/lpfc_els.c index b5a8d050419a..de0ec945d2f1 100644 --- a/drivers/scsi/lpfc/lpfc_els.c +++ b/drivers/scsi/lpfc/lpfc_els.c @@ -5246,9 +5246,10 @@ lpfc_cmpl_els_logo_acc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, /* ACC to LOGO completes to NPort */ lpfc_printf_vlog(vport, KERN_INFO, LOG_ELS, "0109 ACC to LOGO completes to NPort x%x refcnt %d " - "Data: x%x x%x x%x\n", - ndlp->nlp_DID, kref_read(&ndlp->kref), ndlp->nlp_flag, - ndlp->nlp_state, ndlp->nlp_rpi); + "last els x%x Data: x%x x%x x%x\n", + ndlp->nlp_DID, kref_read(&ndlp->kref), + ndlp->nlp_last_elscmd, ndlp->nlp_flag, ndlp->nlp_state, + ndlp->nlp_rpi); /* This clause allows the LOGO ACC to complete and free resources * for the Fabric Domain Controller. It does deliberately skip @@ -5260,18 +5261,22 @@ lpfc_cmpl_els_logo_acc(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb, goto out; if (ndlp->nlp_state == NLP_STE_NPR_NODE) { - /* If PLOGI is being retried, PLOGI completion will cleanup the - * node. The NLP_NPR_2B_DISC flag needs to be retained to make - * progress on nodes discovered from last RSCN. - */ - if ((ndlp->nlp_flag & NLP_DELAY_TMO) && - (ndlp->nlp_last_elscmd == ELS_CMD_PLOGI)) - goto out; - if (ndlp->nlp_flag & NLP_RPI_REGISTERED) lpfc_unreg_rpi(vport, ndlp); + /* If came from PRLO, then PRLO_ACC is done. + * Start rediscovery now. + */ + if (ndlp->nlp_last_elscmd == ELS_CMD_PRLO) { + spin_lock_irq(&ndlp->lock); + ndlp->nlp_flag |= NLP_NPR_2B_DISC; + spin_unlock_irq(&ndlp->lock); + ndlp->nlp_prev_state = ndlp->nlp_state; + lpfc_nlp_set_state(vport, ndlp, NLP_STE_PLOGI_ISSUE); + lpfc_issue_els_plogi(vport, ndlp->nlp_DID, 0); + } } + out: /* * The driver received a LOGO from the rport and has ACK'd it. diff --git a/drivers/scsi/lpfc/lpfc_nportdisc.c b/drivers/scsi/lpfc/lpfc_nportdisc.c index f6a53446e57f..4574716c8764 100644 --- a/drivers/scsi/lpfc/lpfc_nportdisc.c +++ b/drivers/scsi/lpfc/lpfc_nportdisc.c @@ -2652,8 +2652,26 @@ lpfc_rcv_prlo_mapped_node(struct lpfc_vport *vport, struct lpfc_nodelist *ndlp, /* flush the target */ lpfc_sli_abort_iocb(vport, ndlp->nlp_sid, 0, LPFC_CTX_TGT); - /* Treat like rcv logo */ - lpfc_rcv_logo(vport, ndlp, cmdiocb, ELS_CMD_PRLO); + /* Send PRLO_ACC */ + spin_lock_irq(&ndlp->lock); + ndlp->nlp_flag |= NLP_LOGO_ACC; + spin_unlock_irq(&ndlp->lock); + lpfc_els_rsp_acc(vport, ELS_CMD_PRLO, cmdiocb, ndlp, NULL); + + /* Save ELS_CMD_PRLO as the last elscmd and then set to NPR. + * lpfc_cmpl_els_logo_acc is expected to restart discovery. + */ + ndlp->nlp_last_elscmd = ELS_CMD_PRLO; + ndlp->nlp_prev_state = ndlp->nlp_state; + + lpfc_printf_vlog(vport, KERN_INFO, LOG_NODE | LOG_ELS | LOG_DISCOVERY, + "3422 DID x%06x nflag x%x lastels x%x ref cnt %u\n", + ndlp->nlp_DID, ndlp->nlp_flag, + ndlp->nlp_last_elscmd, + kref_read(&ndlp->kref)); + + lpfc_nlp_set_state(vport, ndlp, NLP_STE_NPR_NODE); + return ndlp->nlp_state; } From 62b52495e6a12b37ba3f790f815e717737643bf7 Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:11 -0700 Subject: [PATCH 025/123] scsi: lpfc: Update lpfc version to 14.4.0.4 Update lpfc version to 14.4.0.4 Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-8-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_version.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/lpfc/lpfc_version.h b/drivers/scsi/lpfc/lpfc_version.h index 7ac9ef281881..2fe0386a1fee 100644 --- a/drivers/scsi/lpfc/lpfc_version.h +++ b/drivers/scsi/lpfc/lpfc_version.h @@ -20,7 +20,7 @@ * included with this package. * *******************************************************************/ -#define LPFC_DRIVER_VERSION "14.4.0.3" +#define LPFC_DRIVER_VERSION "14.4.0.4" #define LPFC_DRIVER_NAME "lpfc" /* Used for SLI 2/3 */ From 5b247f03779d3601dcb080446bd7d680149a2aee Mon Sep 17 00:00:00 2001 From: Justin Tee Date: Fri, 26 Jul 2024 16:15:12 -0700 Subject: [PATCH 026/123] scsi: lpfc: Copyright updates for 14.4.0.4 patches Update copyrights to 2024 for files modified in the 14.4.0.4 patch set. Signed-off-by: Justin Tee Link: https://lore.kernel.org/r/20240726231512.92867-9-justintee8345@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/lpfc/lpfc_vmid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/lpfc/lpfc_vmid.c b/drivers/scsi/lpfc/lpfc_vmid.c index cf8ba840d0ea..cc3e4736f2fe 100644 --- a/drivers/scsi/lpfc/lpfc_vmid.c +++ b/drivers/scsi/lpfc/lpfc_vmid.c @@ -1,7 +1,7 @@ /******************************************************************* * This file is part of the Emulex Linux Device Driver for * * Fibre Channel Host Bus Adapters. * - * Copyright (C) 2017-2023 Broadcom. All Rights Reserved. The term * + * Copyright (C) 2017-2024 Broadcom. All Rights Reserved. The term * * “Broadcom” refers to Broadcom Inc. and/or its subsidiaries. * * Copyright (C) 2004-2016 Emulex. All rights reserved. * * EMULEX and SLI are trademarks of Emulex. * From 0e21e73384d324f75ea16f3d622cfc433fa6209b Mon Sep 17 00:00:00 2001 From: David Strahan Date: Thu, 11 Jul 2024 14:47:00 -0500 Subject: [PATCH 027/123] scsi: smartpqi: Add new controller PCI IDs All PCI ID entries in hex. Add new inagile PCI IDs: VID / DID / SVID / SDID ---- ---- ---- ---- SMART-HBA 8242-24i 9005 / 028f / 1ff9 / 0045 RAID 8236-16i 9005 / 028f / 1ff9 / 0046 RAID 8240-24i 9005 / 028f / 1ff9 / 0047 SMART-HBA 8238-16i 9005 / 028f / 1ff9 / 0048 PM8222-SHBA 9005 / 028f / 1ff9 / 004a RAID PM8204-2GB 9005 / 028f / 1ff9 / 004b RAID PM8204-4GB 9005 / 028f / 1ff9 / 004c PM8222-HBA 9005 / 028f / 1ff9 / 004f MT0804M6R 9005 / 028f / 1ff9 / 0051 MT0801M6E 9005 / 028f / 1ff9 / 0052 MT0808M6R 9005 / 028f / 1ff9 / 0053 MT0800M6H 9005 / 028f / 1ff9 / 0054 RS0800M5H24i 9005 / 028f / 1ff9 / 006b RS0800M5E8i 9005 / 028f / 1ff9 / 006c RS0800M5H8i 9005 / 028f / 1ff9 / 006d RS0804M5R16i 9005 / 028f / 1ff9 / 006f RS0800M5E24i 9005 / 028f / 1ff9 / 0070 RS0800M5H16i 9005 / 028f / 1ff9 / 0071 RS0800M5E16i 9005 / 028f / 1ff9 / 0072 RT0800M7E 9005 / 028f / 1ff9 / 0086 RT0800M7H 9005 / 028f / 1ff9 / 0087 RT0804M7R 9005 / 028f / 1ff9 / 0088 RT0808M7R 9005 / 028f / 1ff9 / 0089 RT1608M6R16i 9005 / 028f / 1ff9 / 00a1 Add new h3c pci_id: VID / DID / SVID / SDID ---- ---- ---- ---- UN RAID P4408-Mr-2 9005 / 028f / 193d / 1110 Add new powerleader pci ids: VID / DID / SVID / SDID ---- ---- ---- ---- PL SmartROC PM8204 9005 / 028f / 1f3a / 0104 Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Signed-off-by: David Strahan Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240711194704.982400-2-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 104 ++++++++++++++++++++++++++ 1 file changed, 104 insertions(+) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 24c7cb285dca..9166dfa1fedc 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -9472,6 +9472,10 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x193d, 0x110b) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x193d, 0x1110) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x193d, 0x8460) @@ -9588,6 +9592,14 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1bd4, 0x0089) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x00a1) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1f3a, 0x0104) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x19e5, 0xd227) @@ -10180,6 +10192,98 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1137, 0x02fa) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0045) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0046) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0047) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0048) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x004a) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x004b) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x004c) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x004f) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0051) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0052) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0053) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0054) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x006b) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x006c) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x006d) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x006f) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0070) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0071) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0072) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0086) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0087) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0088) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x0089) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1e93, 0x1000) From bb0f5445b27f4a8f8359cc0f36a59a397b4b5e0c Mon Sep 17 00:00:00 2001 From: Kevin Barnett Date: Thu, 11 Jul 2024 14:47:01 -0500 Subject: [PATCH 028/123] scsi: smartpqi: Improve accuracy/performance of raid-bypass-counter The original implementation of this counter used an atomic variable. However, this implementation negatively impacted performance in some configurations. Switch to using per_cpu variables. Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Co-developed-by: Mahesh Rajashekhara Signed-off-by: Mahesh Rajashekhara Signed-off-by: Kevin Barnett Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240711194704.982400-3-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi.h | 2 +- drivers/scsi/smartpqi/smartpqi_init.c | 30 +++++++++++++++++++++++---- 2 files changed, 27 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi.h b/drivers/scsi/smartpqi/smartpqi.h index cdedc271857a..023fbce04e7a 100644 --- a/drivers/scsi/smartpqi/smartpqi.h +++ b/drivers/scsi/smartpqi/smartpqi.h @@ -1158,7 +1158,7 @@ struct pqi_scsi_dev { struct pqi_stream_data stream_data[NUM_STREAMS_PER_LUN]; atomic_t scsi_cmds_outstanding[PQI_MAX_LUNS_PER_DEVICE]; - unsigned int raid_bypass_cnt; + u64 __percpu *raid_bypass_cnt; struct pqi_tmf_work tmf_work[PQI_MAX_LUNS_PER_DEVICE]; }; diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 9166dfa1fedc..eaebe3cc00aa 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -1508,6 +1508,12 @@ static int pqi_get_raid_map(struct pqi_ctrl_info *ctrl_info, if (rc) goto error; + device->raid_bypass_cnt = alloc_percpu(u64); + if (!device->raid_bypass_cnt) { + rc = -ENOMEM; + goto error; + } + device->raid_map = raid_map; return 0; @@ -2099,6 +2105,10 @@ static void pqi_scsi_update_device(struct pqi_ctrl_info *ctrl_info, /* To prevent this from being freed later. */ new_device->raid_map = NULL; } + if (new_device->raid_bypass_enabled && existing_device->raid_bypass_cnt == NULL) { + existing_device->raid_bypass_cnt = new_device->raid_bypass_cnt; + new_device->raid_bypass_cnt = NULL; + } existing_device->raid_bypass_configured = new_device->raid_bypass_configured; existing_device->raid_bypass_enabled = new_device->raid_bypass_enabled; } @@ -2121,6 +2131,7 @@ static void pqi_scsi_update_device(struct pqi_ctrl_info *ctrl_info, static inline void pqi_free_device(struct pqi_scsi_dev *device) { if (device) { + free_percpu(device->raid_bypass_cnt); kfree(device->raid_map); kfree(device); } @@ -6007,6 +6018,7 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm u16 hw_queue; struct pqi_queue_group *queue_group; bool raid_bypassed; + u64 *raid_bypass_cnt; u8 lun; scmd->host_scribble = PQI_NO_COMPLETION; @@ -6053,7 +6065,8 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm rc = pqi_raid_bypass_submit_scsi_cmd(ctrl_info, device, scmd, queue_group); if (rc == 0 || rc == SCSI_MLQUEUE_HOST_BUSY) { raid_bypassed = true; - device->raid_bypass_cnt++; + raid_bypass_cnt = per_cpu_ptr(device->raid_bypass_cnt, smp_processor_id()); + (*raid_bypass_cnt)++; } } if (!raid_bypassed) @@ -7350,7 +7363,9 @@ static ssize_t pqi_raid_bypass_cnt_show(struct device *dev, struct scsi_device *sdev; struct pqi_scsi_dev *device; unsigned long flags; - unsigned int raid_bypass_cnt; + u64 raid_bypass_cnt; + int cpu; + u64 *per_cpu_bypass_cnt_ptr; sdev = to_scsi_device(dev); ctrl_info = shost_to_hba(sdev->host); @@ -7366,11 +7381,18 @@ static ssize_t pqi_raid_bypass_cnt_show(struct device *dev, return -ENODEV; } - raid_bypass_cnt = device->raid_bypass_cnt; + raid_bypass_cnt = 0; + + if (device->raid_bypass_cnt) { + for_each_online_cpu(cpu) { + per_cpu_bypass_cnt_ptr = per_cpu_ptr(device->raid_bypass_cnt, cpu); + raid_bypass_cnt += *per_cpu_bypass_cnt_ptr; + } + } spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); - return scnprintf(buffer, PAGE_SIZE, "0x%x\n", raid_bypass_cnt); + return scnprintf(buffer, PAGE_SIZE, "0x%llx\n", raid_bypass_cnt); } static ssize_t pqi_sas_ncq_prio_enable_show(struct device *dev, From f1393d52e6cda9c20f12643cbecf1e1dc357e0e2 Mon Sep 17 00:00:00 2001 From: Gilbert Wu Date: Thu, 11 Jul 2024 14:47:02 -0500 Subject: [PATCH 029/123] scsi: smartpqi: revert propagate-the-multipath-failure-to-SML-quickly Correct a rare multipath failure issue by reverting commit 94a68c814328 ("scsi: smartpqi: Quickly propagate path failures to SCSI midlayer") [1]. Reason for revert: The patch propagated the path failure to SML quickly when one of the path fails during IO and AIO path gets disabled for a multipath device. But it created a new issue: when creating a volume on an encryption-enabled controller, the firmware reports the AIO path is disabled, which cause the driver to report a path failure to SML for a multipath device. There will be a new fix to handle "Illegal request" and "Invalid field in parameter list" on RAID path when the AIO path is disabled on a multipath device. [1] https://lore.kernel.org/all/164375209313.440833.9992416628621839233.stgit@brunhilda.pdev.net/ Fixes: 94a68c814328 ("scsi: smartpqi: Quickly propagate path failures to SCSI midlayer") Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Signed-off-by: Gilbert Wu Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240711194704.982400-4-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 20 ++------------------ 1 file changed, 2 insertions(+), 18 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index eaebe3cc00aa..d8df7440bbe1 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -2365,14 +2365,6 @@ static inline void pqi_mask_device(u8 *scsi3addr) scsi3addr[3] |= 0xc0; } -static inline bool pqi_is_multipath_device(struct pqi_scsi_dev *device) -{ - if (pqi_is_logical_device(device)) - return false; - - return (device->path_map & (device->path_map - 1)) != 0; -} - static inline bool pqi_expose_device(struct pqi_scsi_dev *device) { return !device->is_physical_device || !pqi_skip_device(device->scsi3addr); @@ -3269,14 +3261,12 @@ static void pqi_process_aio_io_error(struct pqi_io_request *io_request) int residual_count; int xfer_count; bool device_offline; - struct pqi_scsi_dev *device; scmd = io_request->scmd; error_info = io_request->error_info; host_byte = DID_OK; sense_data_length = 0; device_offline = false; - device = scmd->device->hostdata; switch (error_info->service_response) { case PQI_AIO_SERV_RESPONSE_COMPLETE: @@ -3301,14 +3291,8 @@ static void pqi_process_aio_io_error(struct pqi_io_request *io_request) break; case PQI_AIO_STATUS_AIO_PATH_DISABLED: pqi_aio_path_disabled(io_request); - if (pqi_is_multipath_device(device)) { - pqi_device_remove_start(device); - host_byte = DID_NO_CONNECT; - scsi_status = SAM_STAT_CHECK_CONDITION; - } else { - scsi_status = SAM_STAT_GOOD; - io_request->status = -EAGAIN; - } + scsi_status = SAM_STAT_GOOD; + io_request->status = -EAGAIN; break; case PQI_AIO_STATUS_NO_PATH_TO_DEVICE: case PQI_AIO_STATUS_INVALID_DEVICE: From 57abab70a5e0179b5fff7c4a6dfdbcbcc97ca910 Mon Sep 17 00:00:00 2001 From: Kevin Barnett Date: Thu, 11 Jul 2024 14:47:03 -0500 Subject: [PATCH 030/123] scsi: smartpqi: Improve handling of multipath failover Improve multipath failovers by mapping firmware errors into I/O errors. In some rare instances, firmware does not return the proper error code for I/O errors caused by a multipath path failure. Map I/O errors returned by firmware into errors that help the multipath layer to detect the failure of a path. Reviewed-by: Gerry Morong Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Signed-off-by: Kevin Barnett Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240711194704.982400-5-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index d8df7440bbe1..c116d542ac2b 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -3247,6 +3247,20 @@ static void pqi_process_raid_io_error(struct pqi_io_request *io_request) sense_data_length); } + if (pqi_cmd_priv(scmd)->this_residual && + !pqi_is_logical_device(scmd->device->hostdata) && + scsi_status == SAM_STAT_CHECK_CONDITION && + host_byte == DID_OK && + sense_data_length && + scsi_normalize_sense(error_info->data, sense_data_length, &sshdr) && + sshdr.sense_key == ILLEGAL_REQUEST && + sshdr.asc == 0x26 && + sshdr.ascq == 0x0) { + host_byte = DID_NO_CONNECT; + pqi_take_device_offline(scmd->device, "AIO"); + scsi_build_sense_buffer(0, scmd->sense_buffer, HARDWARE_ERROR, 0x3e, 0x1); + } + scmd->result = scsi_status; set_host_byte(scmd, host_byte); } @@ -6021,7 +6035,7 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm ctrl_info = shost_to_hba(shost); - if (pqi_ctrl_offline(ctrl_info) || pqi_device_in_remove(device)) { + if (pqi_ctrl_offline(ctrl_info) || pqi_device_offline(device) || pqi_device_in_remove(device)) { set_host_byte(scmd, DID_NO_CONNECT); pqi_scsi_done(scmd); return 0; From 5b4ded3f35d5b2c077e4035b66af2d04668d7ee5 Mon Sep 17 00:00:00 2001 From: Don Brace Date: Thu, 11 Jul 2024 14:47:04 -0500 Subject: [PATCH 031/123] scsi: smartpqi: Update driver version to 2.1.28-025 Update driver version to 2.1.28-025. Reviewed-by: Mike Tran Reviewed-by: Gerry Morong Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240711194704.982400-6-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index c116d542ac2b..b5396d722d52 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -33,11 +33,11 @@ #define BUILD_TIMESTAMP #endif -#define DRIVER_VERSION "2.1.26-030" +#define DRIVER_VERSION "2.1.28-025" #define DRIVER_MAJOR 2 #define DRIVER_MINOR 1 -#define DRIVER_RELEASE 26 -#define DRIVER_REVISION 30 +#define DRIVER_RELEASE 28 +#define DRIVER_REVISION 25 #define DRIVER_NAME "Microchip SmartPQI Driver (v" \ DRIVER_VERSION BUILD_TIMESTAMP ")" From 0c150b30d3d51a8c2e09fadd004a640fa12985c6 Mon Sep 17 00:00:00 2001 From: John Garry Date: Mon, 5 Aug 2024 11:33:14 +0000 Subject: [PATCH 032/123] scsi: sd: Don't check if a write for REQ_ATOMIC Flag REQ_ATOMIC can only be set for writes, so don't check if the operation is also a write in sd_setup_read_write_cmnd(). Fixes: bf4ae8f2e640 ("scsi: sd: Atomic write support") Signed-off-by: John Garry Link: https://lore.kernel.org/r/20240805113315.1048591-2-john.g.garry@oracle.com Reviewed-by: Kanchan Joshi Signed-off-by: Martin K. Petersen --- drivers/scsi/sd.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c index adeaa8ab9951..e90a2e4ab318 100644 --- a/drivers/scsi/sd.c +++ b/drivers/scsi/sd.c @@ -1382,7 +1382,7 @@ static blk_status_t sd_setup_read_write_cmnd(struct scsi_cmnd *cmd) if (protect && sdkp->protection_type == T10_PI_TYPE2_PROTECTION) { ret = sd_setup_rw32_cmnd(cmd, write, lba, nr_blocks, protect | fua, dld); - } else if (rq->cmd_flags & REQ_ATOMIC && write) { + } else if (rq->cmd_flags & REQ_ATOMIC) { ret = sd_setup_atomic_cmnd(cmd, lba, nr_blocks, sdkp->use_atomic_write_boundary, protect | fua); From ea6787c695ab7595d851c3506f67c157f3b593c0 Mon Sep 17 00:00:00 2001 From: John Garry Date: Mon, 5 Aug 2024 11:33:15 +0000 Subject: [PATCH 033/123] scsi: block: Don't check REQ_ATOMIC for reads We check in submit_bio_noacct() if flag REQ_ATOMIC is set for both read and write operations, and then validate the atomic operation if set. Flag REQ_ATOMIC can only be set for writes, so don't bother checking for reads. Fixes: 9da3d1e912f3 ("block: Add core atomic write support") Signed-off-by: John Garry Link: https://lore.kernel.org/r/20240805113315.1048591-3-john.g.garry@oracle.com Reviewed-by: Kanchan Joshi Signed-off-by: Martin K. Petersen --- block/blk-core.c | 1 + 1 file changed, 1 insertion(+) diff --git a/block/blk-core.c b/block/blk-core.c index 1217c2cd66dd..bc5e8c5eaac9 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -799,6 +799,7 @@ void submit_bio_noacct(struct bio *bio) switch (bio_op(bio)) { case REQ_OP_READ: + break; case REQ_OP_WRITE: if (bio->bi_opf & REQ_ATOMIC) { status = blk_validate_atomic_write_op_size(q, bio); From fd9cb9615fca4fa322a43eeecbd71d4de1ba146a Mon Sep 17 00:00:00 2001 From: "Rob Herring (Arm)" Date: Thu, 8 Aug 2024 11:06:44 -0600 Subject: [PATCH 034/123] scsi: ufs: ufshcd-pltfrm: Use of_property_present() Use of_property_present() to test for property presence rather than of_find_property(). This is part of a larger effort to remove callers of of_find_property() and similar functions. of_find_property() leaks the DT struct property and data pointers which is a problem for dynamically allocated nodes which may be freed. Signed-off-by: Rob Herring (Arm) Link: https://lore.kernel.org/r/20240808170644.1436991-1-robh@kernel.org Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen --- drivers/ufs/host/ufshcd-pltfrm.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/ufs/host/ufshcd-pltfrm.c b/drivers/ufs/host/ufshcd-pltfrm.c index a3e69ecafd27..2e1eb898a27c 100644 --- a/drivers/ufs/host/ufshcd-pltfrm.c +++ b/drivers/ufs/host/ufshcd-pltfrm.c @@ -272,10 +272,10 @@ static int ufshcd_parse_operating_points(struct ufs_hba *hba) const char **clk_names; int cnt, i, ret; - if (!of_find_property(np, "operating-points-v2", NULL)) + if (!of_property_present(np, "operating-points-v2")) return 0; - if (of_find_property(np, "freq-table-hz", NULL)) { + if (of_property_present(np, "freq-table-hz")) { dev_err(dev, "%s: operating-points and freq-table-hz are incompatible\n", __func__); return -EINVAL; From eab0dce11dd91410204aae39d33124d9c16ab5f0 Mon Sep 17 00:00:00 2001 From: "Rob Herring (Arm)" Date: Thu, 8 Aug 2024 11:07:03 -0600 Subject: [PATCH 035/123] scsi: ufs: ufshcd-pltfrm: Use of_property_count_u32_elems() to get property length Replace of_get_property() with the type specific of_property_count_u32_elems() to get the property length. This is part of a larger effort to remove callers of of_get_property() and similar functions. of_get_property() leaks the DT property data pointer which is a problem for dynamically allocated nodes which may be freed. Signed-off-by: Rob Herring (Arm) Link: https://lore.kernel.org/r/20240808170704.1438658-1-robh@kernel.org Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen --- drivers/ufs/host/ufshcd-pltfrm.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) diff --git a/drivers/ufs/host/ufshcd-pltfrm.c b/drivers/ufs/host/ufshcd-pltfrm.c index 2e1eb898a27c..0c9b303ccfa0 100644 --- a/drivers/ufs/host/ufshcd-pltfrm.c +++ b/drivers/ufs/host/ufshcd-pltfrm.c @@ -31,7 +31,6 @@ static int ufshcd_parse_clock_info(struct ufs_hba *hba) const char *name; u32 *clkfreq = NULL; struct ufs_clk_info *clki; - int len = 0; size_t sz = 0; if (!np) @@ -50,15 +49,12 @@ static int ufshcd_parse_clock_info(struct ufs_hba *hba) if (cnt <= 0) goto out; - if (!of_get_property(np, "freq-table-hz", &len)) { + sz = of_property_count_u32_elems(np, "freq-table-hz"); + if (sz <= 0) { dev_info(dev, "freq-table-hz property not specified\n"); goto out; } - if (len <= 0) - goto out; - - sz = len / sizeof(*clkfreq); if (sz != 2 * cnt) { dev_err(dev, "%s len mismatch\n", "freq-table-hz"); ret = -EINVAL; From ff30732014f5a54f0f97fb7ae9259da833c468d0 Mon Sep 17 00:00:00 2001 From: Pedro Falcato Date: Wed, 7 Aug 2024 10:57:09 +0100 Subject: [PATCH 036/123] scsi: snic: Avoid creating two slab caches with the same name In the spirit of [1], fix the copy-paste typo and use unique names for both caches. [1]: https://lore.kernel.org/all/20240807090746.2146479-1-pedro.falcato@gmail.com/ Signed-off-by: Pedro Falcato Link: https://lore.kernel.org/r/20240807095709.2200728-1-pedro.falcato@gmail.com Signed-off-by: Martin K. Petersen --- drivers/scsi/snic/snic_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/snic/snic_main.c b/drivers/scsi/snic/snic_main.c index cc824dcfe7da..abc78320c66d 100644 --- a/drivers/scsi/snic/snic_main.c +++ b/drivers/scsi/snic/snic_main.c @@ -873,7 +873,7 @@ snic_global_data_init(void) snic_glob->req_cache[SNIC_REQ_CACHE_MAX_SGL] = cachep; len = sizeof(struct snic_host_req); - cachep = kmem_cache_create("snic_req_maxsgl", len, SNIC_SG_DESC_ALIGN, + cachep = kmem_cache_create("snic_req_tm", len, SNIC_SG_DESC_ALIGN, SLAB_HWCACHE_ALIGN, NULL); if (!cachep) { SNIC_ERR("Failed to create snic tm req slab\n"); From 6dc7050d46713591f4d9e1030b92b964287037ec Mon Sep 17 00:00:00 2001 From: Ranjan Kumar Date: Thu, 8 Aug 2024 18:24:16 +0530 Subject: [PATCH 037/123] scsi: mpi3mr: Return complete ioc_status for ioctl commands The driver masked the loginfo available bit in the iocstatus before passing it to the applications, causing a mismatch in error messages between Linux and other operating systems. Modify driver to return unmasked (complete) iocstatus, including the loginfo available bit, for the MPI commands sent through the ioctl interface. Co-developed-by: Sathya Prakash Signed-off-by: Sathya Prakash Signed-off-by: Ranjan Kumar Link: https://lore.kernel.org/r/20240808125418.8832-2-ranjan.kumar@broadcom.com Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi3mr_fw.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c index c196dc14ad20..169850393580 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c @@ -345,6 +345,7 @@ static void mpi3mr_process_admin_reply_desc(struct mpi3mr_ioc *mrioc, { u16 reply_desc_type, host_tag = 0; u16 ioc_status = MPI3_IOCSTATUS_SUCCESS; + u16 masked_ioc_status = MPI3_IOCSTATUS_SUCCESS; u32 ioc_loginfo = 0, sense_count = 0; struct mpi3_status_reply_descriptor *status_desc; struct mpi3_address_reply_descriptor *addr_desc; @@ -366,8 +367,8 @@ static void mpi3mr_process_admin_reply_desc(struct mpi3mr_ioc *mrioc, if (ioc_status & MPI3_REPLY_DESCRIPT_STATUS_IOCSTATUS_LOGINFOAVAIL) ioc_loginfo = le32_to_cpu(status_desc->ioc_log_info); - ioc_status &= MPI3_REPLY_DESCRIPT_STATUS_IOCSTATUS_STATUS_MASK; - mpi3mr_reply_trigger(mrioc, ioc_status, ioc_loginfo); + masked_ioc_status = ioc_status & MPI3_IOCSTATUS_STATUS_MASK; + mpi3mr_reply_trigger(mrioc, masked_ioc_status, ioc_loginfo); break; case MPI3_REPLY_DESCRIPT_FLAGS_TYPE_ADDRESS_REPLY: addr_desc = (struct mpi3_address_reply_descriptor *)reply_desc; @@ -380,7 +381,7 @@ static void mpi3mr_process_admin_reply_desc(struct mpi3mr_ioc *mrioc, if (ioc_status & MPI3_REPLY_DESCRIPT_STATUS_IOCSTATUS_LOGINFOAVAIL) ioc_loginfo = le32_to_cpu(def_reply->ioc_log_info); - ioc_status &= MPI3_REPLY_DESCRIPT_STATUS_IOCSTATUS_STATUS_MASK; + masked_ioc_status = ioc_status & MPI3_IOCSTATUS_STATUS_MASK; if (def_reply->function == MPI3_FUNCTION_SCSI_IO) { scsi_reply = (struct mpi3_scsi_io_reply *)def_reply; sense_buf = mpi3mr_get_sensebuf_virt_addr(mrioc, @@ -393,7 +394,7 @@ static void mpi3mr_process_admin_reply_desc(struct mpi3mr_ioc *mrioc, sshdr.asc, sshdr.ascq); } } - mpi3mr_reply_trigger(mrioc, ioc_status, ioc_loginfo); + mpi3mr_reply_trigger(mrioc, masked_ioc_status, ioc_loginfo); break; case MPI3_REPLY_DESCRIPT_FLAGS_TYPE_SUCCESS: success_desc = (struct mpi3_success_reply_descriptor *)reply_desc; @@ -408,7 +409,10 @@ static void mpi3mr_process_admin_reply_desc(struct mpi3mr_ioc *mrioc, if (cmdptr->state & MPI3MR_CMD_PENDING) { cmdptr->state |= MPI3MR_CMD_COMPLETE; cmdptr->ioc_loginfo = ioc_loginfo; - cmdptr->ioc_status = ioc_status; + if (host_tag == MPI3MR_HOSTTAG_BSG_CMDS) + cmdptr->ioc_status = ioc_status; + else + cmdptr->ioc_status = masked_ioc_status; cmdptr->state &= ~MPI3MR_CMD_PENDING; if (def_reply) { cmdptr->state |= MPI3MR_CMD_REPLY_VALID; From 199510e33dea357e48389241a28b8d43762394e8 Mon Sep 17 00:00:00 2001 From: Ranjan Kumar Date: Thu, 8 Aug 2024 18:24:17 +0530 Subject: [PATCH 038/123] scsi: mpi3mr: Update consumer index of reply queues after every 100 replies Instead of updating the ConsumerIndex of the Admin and Operational ReplyQueues after processing all replies in the queue, the index will now be periodically updated after processing every 100 replies. Co-developed-by: Sathya Prakash Signed-off-by: Sathya Prakash Signed-off-by: Ranjan Kumar Link: https://lore.kernel.org/r/20240808125418.8832-3-ranjan.kumar@broadcom.com Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi3mr.h | 1 + drivers/scsi/mpi3mr/mpi3mr_fw.c | 18 ++++++++++++++++-- 2 files changed, 17 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h index dc2cdd5f0311..cbb6e4b2d447 100644 --- a/drivers/scsi/mpi3mr/mpi3mr.h +++ b/drivers/scsi/mpi3mr/mpi3mr.h @@ -213,6 +213,7 @@ extern atomic64_t event_counter; #define MPI3MR_HDB_QUERY_ELEMENT_TRIGGER_FORMAT_INDEX 0 #define MPI3MR_HDB_QUERY_ELEMENT_TRIGGER_FORMAT_DATA 1 +#define MPI3MR_THRESHOLD_REPLY_COUNT 100 /* SGE Flag definition */ #define MPI3MR_SGEFLAGS_SYSTEM_SIMPLE_END_OF_LIST \ diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c index 169850393580..6eb5bcd8e757 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c @@ -443,6 +443,7 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc) u32 admin_reply_ci = mrioc->admin_reply_ci; u32 num_admin_replies = 0; u64 reply_dma = 0; + u16 threshold_comps = 0; struct mpi3_default_reply_descriptor *reply_desc; if (!atomic_add_unless(&mrioc->admin_reply_q_in_use, 1, 1)) @@ -466,6 +467,7 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc) if (reply_dma) mpi3mr_repost_reply_buf(mrioc, reply_dma); num_admin_replies++; + threshold_comps++; if (++admin_reply_ci == mrioc->num_admin_replies) { admin_reply_ci = 0; exp_phase ^= 1; @@ -476,6 +478,11 @@ int mpi3mr_process_admin_reply_q(struct mpi3mr_ioc *mrioc) if ((le16_to_cpu(reply_desc->reply_flags) & MPI3_REPLY_DESCRIPT_FLAGS_PHASE_MASK) != exp_phase) break; + if (threshold_comps == MPI3MR_THRESHOLD_REPLY_COUNT) { + writel(admin_reply_ci, + &mrioc->sysif_regs->admin_reply_queue_ci); + threshold_comps = 0; + } } while (1); writel(admin_reply_ci, &mrioc->sysif_regs->admin_reply_queue_ci); @@ -529,7 +536,7 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc, u32 num_op_reply = 0; u64 reply_dma = 0; struct mpi3_default_reply_descriptor *reply_desc; - u16 req_q_idx = 0, reply_qidx; + u16 req_q_idx = 0, reply_qidx, threshold_comps = 0; reply_qidx = op_reply_q->qid - 1; @@ -560,6 +567,7 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc, if (reply_dma) mpi3mr_repost_reply_buf(mrioc, reply_dma); num_op_reply++; + threshold_comps++; if (++reply_ci == op_reply_q->num_replies) { reply_ci = 0; @@ -581,13 +589,19 @@ int mpi3mr_process_op_reply_q(struct mpi3mr_ioc *mrioc, break; } #endif + if (threshold_comps == MPI3MR_THRESHOLD_REPLY_COUNT) { + writel(reply_ci, + &mrioc->sysif_regs->oper_queue_indexes[reply_qidx].consumer_index); + atomic_sub(threshold_comps, &op_reply_q->pend_ios); + threshold_comps = 0; + } } while (1); writel(reply_ci, &mrioc->sysif_regs->oper_queue_indexes[reply_qidx].consumer_index); op_reply_q->ci = reply_ci; op_reply_q->ephase = exp_phase; - + atomic_sub(threshold_comps, &op_reply_q->pend_ios); atomic_dec(&op_reply_q->in_use); return num_op_reply; } From f856e57d6138a2dcbe4493c7333f0efdb535c19c Mon Sep 17 00:00:00 2001 From: Ranjan Kumar Date: Thu, 8 Aug 2024 18:24:18 +0530 Subject: [PATCH 039/123] scsi: mpi3mr: Driver version update to 8.10.0.5.50 Update driver version to 8.10.0.5.50. Signed-off-by: Ranjan Kumar Link: https://lore.kernel.org/r/20240808125418.8832-4-ranjan.kumar@broadcom.com Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi3mr.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h index cbb6e4b2d447..875bad7538f2 100644 --- a/drivers/scsi/mpi3mr/mpi3mr.h +++ b/drivers/scsi/mpi3mr/mpi3mr.h @@ -57,8 +57,8 @@ extern struct list_head mrioc_list; extern int prot_mask; extern atomic64_t event_counter; -#define MPI3MR_DRIVER_VERSION "8.9.1.0.51" -#define MPI3MR_DRIVER_RELDATE "29-May-2024" +#define MPI3MR_DRIVER_VERSION "8.10.0.5.50" +#define MPI3MR_DRIVER_RELDATE "08-Aug-2024" #define MPI3MR_DRIVER_NAME "mpi3mr" #define MPI3MR_DRIVER_LICENSE "GPL" From 5ec4f820cb9766e4583df947150a6febce8da794 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 040/123] scsi: mac_scsi: Revise printk(KERN_DEBUG ...) messages After a bus fault, capture and log the chip registers immediately, if the NDEBUG_PSEUDO_DMA macro is defined. Remove some printk(KERN_DEBUG ...) messages that aren't needed any more. Don't skip the debug message when bytes == 0. Show all of the byte counters in the debug messages. Cc: stable@vger.kernel.org # 5.15+ Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/7573c79f4e488fc00af2b8a191e257ca945e0409.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mac_scsi.c | 42 +++++++++++++++++++++-------------------- 1 file changed, 22 insertions(+), 20 deletions(-) diff --git a/drivers/scsi/mac_scsi.c b/drivers/scsi/mac_scsi.c index 53ee8f84d094..e67b038a3577 100644 --- a/drivers/scsi/mac_scsi.c +++ b/drivers/scsi/mac_scsi.c @@ -286,13 +286,14 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, while (!NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, BASR_DRQ | BASR_PHASE_MATCH, BASR_DRQ | BASR_PHASE_MATCH, 0)) { - int bytes; + int bytes, chunk_bytes; if (macintosh_config->ident == MAC_MODEL_IIFX) write_ctrl_reg(hostdata, CTRL_HANDSHAKE_MODE | CTRL_INTERRUPTS_ENABLE); - bytes = mac_pdma_recv(s, d, min(hostdata->pdma_residual, 512)); + chunk_bytes = min(hostdata->pdma_residual, 512); + bytes = mac_pdma_recv(s, d, chunk_bytes); if (bytes > 0) { d += bytes; @@ -302,23 +303,23 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, if (hostdata->pdma_residual == 0) goto out; - if (NCR5380_poll_politely2(hostdata, STATUS_REG, SR_REQ, SR_REQ, - BUS_AND_STATUS_REG, BASR_ACK, - BASR_ACK, 0) < 0) - scmd_printk(KERN_DEBUG, hostdata->connected, - "%s: !REQ and !ACK\n", __func__); if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) goto out; if (bytes == 0) udelay(MAC_PDMA_DELAY); - if (bytes >= 0) + if (bytes > 0) continue; - dsprintk(NDEBUG_PSEUDO_DMA, hostdata->host, - "%s: bus error (%d/%d)\n", __func__, d - dst, len); NCR5380_dprint(NDEBUG_PSEUDO_DMA, hostdata->host); + dsprintk(NDEBUG_PSEUDO_DMA, hostdata->host, + "%s: bus error [%d/%d] (%d/%d)\n", + __func__, d - dst, len, bytes, chunk_bytes); + + if (bytes == 0) + continue; + result = -1; goto out; } @@ -345,13 +346,14 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, while (!NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, BASR_DRQ | BASR_PHASE_MATCH, BASR_DRQ | BASR_PHASE_MATCH, 0)) { - int bytes; + int bytes, chunk_bytes; if (macintosh_config->ident == MAC_MODEL_IIFX) write_ctrl_reg(hostdata, CTRL_HANDSHAKE_MODE | CTRL_INTERRUPTS_ENABLE); - bytes = mac_pdma_send(s, d, min(hostdata->pdma_residual, 512)); + chunk_bytes = min(hostdata->pdma_residual, 512); + bytes = mac_pdma_send(s, d, chunk_bytes); if (bytes > 0) { s += bytes; @@ -370,23 +372,23 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, goto out; } - if (NCR5380_poll_politely2(hostdata, STATUS_REG, SR_REQ, SR_REQ, - BUS_AND_STATUS_REG, BASR_ACK, - BASR_ACK, 0) < 0) - scmd_printk(KERN_DEBUG, hostdata->connected, - "%s: !REQ and !ACK\n", __func__); if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) goto out; if (bytes == 0) udelay(MAC_PDMA_DELAY); - if (bytes >= 0) + if (bytes > 0) continue; - dsprintk(NDEBUG_PSEUDO_DMA, hostdata->host, - "%s: bus error (%d/%d)\n", __func__, s - src, len); NCR5380_dprint(NDEBUG_PSEUDO_DMA, hostdata->host); + dsprintk(NDEBUG_PSEUDO_DMA, hostdata->host, + "%s: bus error [%d/%d] (%d/%d)\n", + __func__, s - src, len, bytes, chunk_bytes); + + if (bytes == 0) + continue; + result = -1; goto out; } From 5545c3165cbc98615fe65a44f41167cbb557e410 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 041/123] scsi: mac_scsi: Refactor polling loop Before the error handling can be revised, some preparation is needed. Refactor the polling loop with a new function, macscsi_wait_for_drq(). This function will gain more call sites in the next patch. Cc: stable@vger.kernel.org # 5.15+ Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/6a5ffabb4290c0d138c6d285fda8fa3902e926f0.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mac_scsi.c | 80 +++++++++++++++++++++-------------------- 1 file changed, 42 insertions(+), 38 deletions(-) diff --git a/drivers/scsi/mac_scsi.c b/drivers/scsi/mac_scsi.c index e67b038a3577..99a2008f8752 100644 --- a/drivers/scsi/mac_scsi.c +++ b/drivers/scsi/mac_scsi.c @@ -208,8 +208,6 @@ __setup("mac5380=", mac_scsi_setup); ".previous \n" \ : "+a" (addr), "+r" (n), "+r" (result) : "a" (io)) -#define MAC_PDMA_DELAY 32 - static inline int mac_pdma_recv(void __iomem *io, unsigned char *start, int n) { unsigned char *addr = start; @@ -274,6 +272,36 @@ static inline void write_ctrl_reg(struct NCR5380_hostdata *hostdata, u32 value) out_be32(hostdata->io + (CTRL_REG << 4), value); } +static inline int macscsi_wait_for_drq(struct NCR5380_hostdata *hostdata) +{ + unsigned int n = 1; /* effectively multiplies NCR5380_REG_POLL_TIME */ + unsigned char basr; + +again: + basr = NCR5380_read(BUS_AND_STATUS_REG); + + if (!(basr & BASR_PHASE_MATCH)) + return 1; + + if (basr & BASR_IRQ) + return -1; + + if (basr & BASR_DRQ) + return 0; + + if (n-- == 0) { + NCR5380_dprint(NDEBUG_PSEUDO_DMA, hostdata->host); + dsprintk(NDEBUG_PSEUDO_DMA, hostdata->host, + "%s: DRQ timeout\n", __func__); + return -1; + } + + NCR5380_poll_politely2(hostdata, + BUS_AND_STATUS_REG, BASR_DRQ, BASR_DRQ, + BUS_AND_STATUS_REG, BASR_PHASE_MATCH, 0, 0); + goto again; +} + static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, unsigned char *dst, int len) { @@ -283,9 +311,7 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, hostdata->pdma_residual = len; - while (!NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, - BASR_DRQ | BASR_PHASE_MATCH, - BASR_DRQ | BASR_PHASE_MATCH, 0)) { + while (macscsi_wait_for_drq(hostdata) == 0) { int bytes, chunk_bytes; if (macintosh_config->ident == MAC_MODEL_IIFX) @@ -295,19 +321,16 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, chunk_bytes = min(hostdata->pdma_residual, 512); bytes = mac_pdma_recv(s, d, chunk_bytes); + if (macintosh_config->ident == MAC_MODEL_IIFX) + write_ctrl_reg(hostdata, CTRL_INTERRUPTS_ENABLE); + if (bytes > 0) { d += bytes; hostdata->pdma_residual -= bytes; } if (hostdata->pdma_residual == 0) - goto out; - - if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) - goto out; - - if (bytes == 0) - udelay(MAC_PDMA_DELAY); + break; if (bytes > 0) continue; @@ -321,16 +344,9 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, continue; result = -1; - goto out; + break; } - scmd_printk(KERN_ERR, hostdata->connected, - "%s: phase mismatch or !DRQ\n", __func__); - NCR5380_dprint(NDEBUG_PSEUDO_DMA, hostdata->host); - result = -1; -out: - if (macintosh_config->ident == MAC_MODEL_IIFX) - write_ctrl_reg(hostdata, CTRL_INTERRUPTS_ENABLE); return result; } @@ -343,9 +359,7 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, hostdata->pdma_residual = len; - while (!NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, - BASR_DRQ | BASR_PHASE_MATCH, - BASR_DRQ | BASR_PHASE_MATCH, 0)) { + while (macscsi_wait_for_drq(hostdata) == 0) { int bytes, chunk_bytes; if (macintosh_config->ident == MAC_MODEL_IIFX) @@ -355,6 +369,9 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, chunk_bytes = min(hostdata->pdma_residual, 512); bytes = mac_pdma_send(s, d, chunk_bytes); + if (macintosh_config->ident == MAC_MODEL_IIFX) + write_ctrl_reg(hostdata, CTRL_INTERRUPTS_ENABLE); + if (bytes > 0) { s += bytes; hostdata->pdma_residual -= bytes; @@ -369,15 +386,9 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, "%s: Last Byte Sent timeout\n", __func__); result = -1; } - goto out; + break; } - if (!(NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) - goto out; - - if (bytes == 0) - udelay(MAC_PDMA_DELAY); - if (bytes > 0) continue; @@ -390,16 +401,9 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, continue; result = -1; - goto out; + break; } - scmd_printk(KERN_ERR, hostdata->connected, - "%s: phase mismatch or !DRQ\n", __func__); - NCR5380_dprint(NDEBUG_PSEUDO_DMA, hostdata->host); - result = -1; -out: - if (macintosh_config->ident == MAC_MODEL_IIFX) - write_ctrl_reg(hostdata, CTRL_INTERRUPTS_ENABLE); return result; } From 5551bc30e4a69ad86d0d008e2f56cd59b6583476 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 042/123] scsi: mac_scsi: Disallow bus errors during PDMA send SD cards can produce write latency spikes on the order of a hundred milliseconds. If the target firmware does not hide that latency during DATA IN and OUT phases it can cause the PDMA circuitry to raise a processor bus fault which in turn leads to an unreliable byte count and a DMA overrun. The Last Byte Sent flag is used to detect the overrun but this mechanism is unreliable on some systems. Instead, set a DID_ERROR result whenever there is a bus fault during a PDMA send, unless the cause was a phase mismatch. Cc: stable@vger.kernel.org # 5.15+ Reported-and-tested-by: Stan Johnson Fixes: 7c1f3e3447a1 ("scsi: mac_scsi: Treat Last Byte Sent time-out as failure") Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/cc38df687ace2c4ffc375a683b2502fc476b600d.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mac_scsi.c | 44 ++++++++++++++++++----------------------- 1 file changed, 19 insertions(+), 25 deletions(-) diff --git a/drivers/scsi/mac_scsi.c b/drivers/scsi/mac_scsi.c index 99a2008f8752..3958f7dc679f 100644 --- a/drivers/scsi/mac_scsi.c +++ b/drivers/scsi/mac_scsi.c @@ -102,11 +102,15 @@ __setup("mac5380=", mac_scsi_setup); * Linux SCSI drivers lack knowledge of the timing behaviour of SCSI targets * so bus errors are unavoidable. * - * If a MOVE.B instruction faults, we assume that zero bytes were transferred - * and simply retry. That assumption probably depends on target behaviour but - * seems to hold up okay. The NOP provides synchronization: without it the - * fault can sometimes occur after the program counter has moved past the - * offending instruction. Post-increment addressing can't be used. + * If a MOVE.B instruction faults during a receive operation, we assume the + * target sent nothing and try again. That assumption probably depends on + * target firmware but it seems to hold up okay. If a fault happens during a + * send operation, the target may or may not have seen /ACK and got the byte. + * It's uncertain so the whole SCSI command gets retried. + * + * The NOP is needed for synchronization because the fault address in the + * exception stack frame may or may not be the instruction that actually + * caused the bus error. Post-increment addressing can't be used. */ #define MOVE_BYTE(operands) \ @@ -243,22 +247,21 @@ static inline int mac_pdma_send(unsigned char *start, void __iomem *io, int n) if (n >= 1) { MOVE_BYTE("%0@,%3@"); if (result) - goto out; + return -1; } if (n >= 1 && ((unsigned long)addr & 1)) { MOVE_BYTE("%0@,%3@"); if (result) - goto out; + return -2; } while (n >= 32) MOVE_16_WORDS("%0@+,%3@"); while (n >= 2) MOVE_WORD("%0@+,%3@"); if (result) - return start - addr; /* Negated to indicate uncertain length */ + return start - addr - 1; /* Negated to indicate uncertain length */ if (n == 1) MOVE_BYTE("%0@,%3@"); -out: return addr - start; } @@ -307,7 +310,6 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, { u8 __iomem *s = hostdata->pdma_io + (INPUT_DATA_REG << 4); unsigned char *d = dst; - int result = 0; hostdata->pdma_residual = len; @@ -343,11 +345,12 @@ static inline int macscsi_pread(struct NCR5380_hostdata *hostdata, if (bytes == 0) continue; - result = -1; + if (macscsi_wait_for_drq(hostdata) <= 0) + set_host_byte(hostdata->connected, DID_ERROR); break; } - return result; + return 0; } static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, @@ -355,7 +358,6 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, { unsigned char *s = src; u8 __iomem *d = hostdata->pdma_io + (OUTPUT_DATA_REG << 4); - int result = 0; hostdata->pdma_residual = len; @@ -377,17 +379,8 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, hostdata->pdma_residual -= bytes; } - if (hostdata->pdma_residual == 0) { - if (NCR5380_poll_politely(hostdata, TARGET_COMMAND_REG, - TCR_LAST_BYTE_SENT, - TCR_LAST_BYTE_SENT, - 0) < 0) { - scmd_printk(KERN_ERR, hostdata->connected, - "%s: Last Byte Sent timeout\n", __func__); - result = -1; - } + if (hostdata->pdma_residual == 0) break; - } if (bytes > 0) continue; @@ -400,11 +393,12 @@ static inline int macscsi_pwrite(struct NCR5380_hostdata *hostdata, if (bytes == 0) continue; - result = -1; + if (macscsi_wait_for_drq(hostdata) <= 0) + set_host_byte(hostdata->connected, DID_ERROR); break; } - return result; + return 0; } static int macscsi_dma_xfer_len(struct NCR5380_hostdata *hostdata, From 5768718da9417331803fc4bc090544c2a93b88dc Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 043/123] scsi: NCR5380: Check for phase match during PDMA fixup It's not an error for a target to change the bus phase during a transfer. Unfortunately, the FLAG_DMA_FIXUP workaround does not allow for that -- a phase change produces a DRQ timeout error and the device borken flag will be set. Check the phase match bit during FLAG_DMA_FIXUP processing. Don't forget to decrement the command residual. While we are here, change shost_printk() into scmd_printk() for better consistency with other DMA error messages. Tested-by: Stan Johnson Fixes: 55181be8ced1 ("ncr5380: Replace redundant flags with FLAG_NO_DMA_FIXUP") Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/99dc7d1f4c825621b5b120963a69f6cd3e9ca659.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 78 +++++++++++++++++++++--------------------- 1 file changed, 39 insertions(+), 39 deletions(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index cea3a79d538e..00e245173320 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -1485,6 +1485,7 @@ static int NCR5380_transfer_dma(struct Scsi_Host *instance, unsigned char **data) { struct NCR5380_hostdata *hostdata = shost_priv(instance); + struct NCR5380_cmd *ncmd = NCR5380_to_ncmd(hostdata->connected); int c = *count; unsigned char p = *phase; unsigned char *d = *data; @@ -1496,7 +1497,7 @@ static int NCR5380_transfer_dma(struct Scsi_Host *instance, return -1; } - NCR5380_to_ncmd(hostdata->connected)->phase = p; + ncmd->phase = p; if (p & SR_IO) { if (hostdata->read_overruns) @@ -1608,45 +1609,44 @@ static int NCR5380_transfer_dma(struct Scsi_Host *instance, * request. */ - if (hostdata->flags & FLAG_DMA_FIXUP) { - if (p & SR_IO) { - /* - * The workaround was to transfer fewer bytes than we - * intended to with the pseudo-DMA read function, wait for - * the chip to latch the last byte, read it, and then disable - * pseudo-DMA mode. - * - * After REQ is asserted, the NCR5380 asserts DRQ and ACK. - * REQ is deasserted when ACK is asserted, and not reasserted - * until ACK goes false. Since the NCR5380 won't lower ACK - * until DACK is asserted, which won't happen unless we twiddle - * the DMA port or we take the NCR5380 out of DMA mode, we - * can guarantee that we won't handshake another extra - * byte. - */ - - if (NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, - BASR_DRQ, BASR_DRQ, 0) < 0) { - result = -1; - shost_printk(KERN_ERR, instance, "PDMA read: DRQ timeout\n"); - } - if (NCR5380_poll_politely(hostdata, STATUS_REG, - SR_REQ, 0, 0) < 0) { - result = -1; - shost_printk(KERN_ERR, instance, "PDMA read: !REQ timeout\n"); - } - d[*count - 1] = NCR5380_read(INPUT_DATA_REG); - } else { - /* - * Wait for the last byte to be sent. If REQ is being asserted for - * the byte we're interested, we'll ACK it and it will go false. - */ - if (NCR5380_poll_politely2(hostdata, - BUS_AND_STATUS_REG, BASR_DRQ, BASR_DRQ, - BUS_AND_STATUS_REG, BASR_PHASE_MATCH, 0, 0) < 0) { - result = -1; - shost_printk(KERN_ERR, instance, "PDMA write: DRQ and phase timeout\n"); + if ((hostdata->flags & FLAG_DMA_FIXUP) && + (NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) { + /* + * The workaround was to transfer fewer bytes than we + * intended to with the pseudo-DMA receive function, wait for + * the chip to latch the last byte, read it, and then disable + * DMA mode. + * + * After REQ is asserted, the NCR5380 asserts DRQ and ACK. + * REQ is deasserted when ACK is asserted, and not reasserted + * until ACK goes false. Since the NCR5380 won't lower ACK + * until DACK is asserted, which won't happen unless we twiddle + * the DMA port or we take the NCR5380 out of DMA mode, we + * can guarantee that we won't handshake another extra + * byte. + * + * If sending, wait for the last byte to be sent. If REQ is + * being asserted for the byte we're interested, we'll ACK it + * and it will go false. + */ + if (!NCR5380_poll_politely(hostdata, BUS_AND_STATUS_REG, + BASR_DRQ, BASR_DRQ, 0)) { + if ((p & SR_IO) && + (NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) { + if (!NCR5380_poll_politely(hostdata, STATUS_REG, + SR_REQ, 0, 0)) { + d[c] = NCR5380_read(INPUT_DATA_REG); + --ncmd->this_residual; + } else { + result = -1; + scmd_printk(KERN_ERR, hostdata->connected, + "PDMA fixup: !REQ timeout\n"); + } } + } else if (NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH) { + result = -1; + scmd_printk(KERN_ERR, hostdata->connected, + "PDMA fixup: DRQ timeout\n"); } } From 2ac6d29716cd7a3a013b34fad6ba6e88767e28c9 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 044/123] scsi: mac_scsi: Enable scatter/gather by default Now that FLAG_DMA_FIXUP has itself been fixed up, it can be used to enable scatter/gather. Increase the default value for sg_tablesize to SG_ALL for those systems which are compatible with FLAG_DMA_FIXUP. Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/f155ba5ce93055cbc6ac6d4026673f40f826edb8.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mac_scsi.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/drivers/scsi/mac_scsi.c b/drivers/scsi/mac_scsi.c index 3958f7dc679f..f225bb20aa22 100644 --- a/drivers/scsi/mac_scsi.c +++ b/drivers/scsi/mac_scsi.c @@ -432,7 +432,7 @@ static struct scsi_host_template mac_scsi_template = { .eh_host_reset_handler = macscsi_host_reset, .can_queue = 16, .this_id = 7, - .sg_tablesize = 1, + .sg_tablesize = SG_ALL, .cmd_per_lun = 2, .dma_boundary = PAGE_SIZE - 1, .cmd_size = sizeof(struct NCR5380_cmd), @@ -470,6 +470,9 @@ static int __init mac_scsi_probe(struct platform_device *pdev) if (setup_hostid >= 0) mac_scsi_template.this_id = setup_hostid & 7; + if (macintosh_config->ident == MAC_MODEL_IIFX) + mac_scsi_template.sg_tablesize = 1; + instance = scsi_host_alloc(&mac_scsi_template, sizeof(struct NCR5380_hostdata)); if (!instance) @@ -491,6 +494,9 @@ static int __init mac_scsi_probe(struct platform_device *pdev) host_flags |= setup_toshiba_delay > 0 ? FLAG_TOSHIBA_DELAY : 0; + if (instance->sg_tablesize > 1) + host_flags |= FLAG_DMA_FIXUP; + error = NCR5380_init(instance, host_flags | FLAG_LATE_DMA_SETUP); if (error) goto fail_init; From 1c71065df2df693d208dd32758171c1dece66341 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 045/123] scsi: NCR5380: Initialize buffer for MSG IN and STATUS transfers Following an incomplete transfer in MSG IN phase, the driver would not notice the problem and would make use of invalid data. Initialize 'tmp' appropriately and bail out if no message was received. For STATUS phase, preserve the existing status code unless a new value was transferred. Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/52e02a8812ae1a2d810d7f9f7fd800c3ccc320c4.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index 00e245173320..4fcb73b727aa 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -1807,8 +1807,11 @@ static void NCR5380_information_transfer(struct Scsi_Host *instance) return; case PHASE_MSGIN: len = 1; + tmp = 0xff; data = &tmp; NCR5380_transfer_pio(instance, &phase, &len, &data, 0); + if (tmp == 0xff) + break; ncmd->message = tmp; switch (tmp) { @@ -1996,6 +1999,7 @@ static void NCR5380_information_transfer(struct Scsi_Host *instance) break; case PHASE_STATIN: len = 1; + tmp = ncmd->status; data = &tmp; NCR5380_transfer_pio(instance, &phase, &len, &data, 0); ncmd->status = tmp; From 086c4802cf99ab7275cd4ec6651f696f2cc61bfa Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 046/123] scsi: NCR5380: Handle BSY signal loss during information transfer phases Improve robustness by checking for a lost BSY signal during the information transfer loop. The status register is being polled anyway, so a BSY check costs nothing. BSY signal loss could be caused by a target error or a kicked plug etc. A bus reset is another possibility but that is already handled and hostdata->connected would be NULL. Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/d253dddaf4d9bc17b8ee02ea2b731d92f25b16f1.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index 4fcb73b727aa..8a9df2ab9569 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -1244,8 +1244,6 @@ out: * is in same phase. * * Also, *phase, *count, *data are modified in place. - * - * XXX Note : handling for bus free may be useful. */ /* @@ -1277,8 +1275,8 @@ static int NCR5380_transfer_pio(struct Scsi_Host *instance, * valid */ - if (NCR5380_poll_politely(hostdata, STATUS_REG, SR_REQ, SR_REQ, - HZ * can_sleep) < 0) + if (NCR5380_poll_politely(hostdata, STATUS_REG, SR_REQ | SR_BSY, + SR_REQ | SR_BSY, HZ * can_sleep) < 0) break; dsprintk(NDEBUG_HANDSHAKE, instance, "REQ asserted\n"); @@ -1666,9 +1664,6 @@ static int NCR5380_transfer_dma(struct Scsi_Host *instance, * Side effects : SCSI things happen, the disconnected queue will be * modified if a command disconnects, *instance->connected will * change. - * - * XXX Note : we need to watch for bus free or a reset condition here - * to recover from an unexpected bus free condition. */ static void NCR5380_information_transfer(struct Scsi_Host *instance) @@ -2009,9 +2004,20 @@ static void NCR5380_information_transfer(struct Scsi_Host *instance) NCR5380_dprint(NDEBUG_ANY, instance); } /* switch(phase) */ } else { + int err; + spin_unlock_irq(&hostdata->lock); - NCR5380_poll_politely(hostdata, STATUS_REG, SR_REQ, SR_REQ, HZ); + err = NCR5380_poll_politely(hostdata, STATUS_REG, + SR_REQ, SR_REQ, HZ); spin_lock_irq(&hostdata->lock); + + if (err < 0 && hostdata->connected && + !(NCR5380_read(STATUS_REG) & SR_BSY)) { + scmd_printk(KERN_ERR, hostdata->connected, + "BSY signal lost\n"); + do_reset(instance); + bus_reset_cleanup(instance); + } } } } From 476f8c82e2187369be982ef1b3739307d67c10d3 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 047/123] scsi: NCR5380: Drop redundant member from struct NCR5380_cmd The 'message' member is stored but never loaded so just remove it. Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/4dc903a95a814d0c9b09656f3651a1bd798fcbbb.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 2 -- drivers/scsi/NCR5380.h | 1 - 2 files changed, 3 deletions(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index 8a9df2ab9569..a47a825e7220 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -157,7 +157,6 @@ static inline void initialize_SCp(struct scsi_cmnd *cmd) } ncmd->status = 0; - ncmd->message = 0; } static inline void advance_sg_buffer(struct NCR5380_cmd *ncmd) @@ -1807,7 +1806,6 @@ static void NCR5380_information_transfer(struct Scsi_Host *instance) NCR5380_transfer_pio(instance, &phase, &len, &data, 0); if (tmp == 0xff) break; - ncmd->message = tmp; switch (tmp) { case ABORT: diff --git a/drivers/scsi/NCR5380.h b/drivers/scsi/NCR5380.h index 8dc2be4212dc..84db14b036e4 100644 --- a/drivers/scsi/NCR5380.h +++ b/drivers/scsi/NCR5380.h @@ -231,7 +231,6 @@ struct NCR5380_cmd { int this_residual; struct scatterlist *buffer; int status; - int message; int phase; struct list_head list; }; From 8663cadefd15bafee457cb23bb561704383fb1ce Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 048/123] scsi: NCR5380: Remove redundant result calculation from NCR5380_transfer_pio() NCR5380_transfer_pio() returns an ambiguous value which is ignored by callers. Make it void and remove the redundant calculation. Adopt kernel-doc format for the updated description. Tested-by: Stan Johnson Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/c07a52d0d7610b4b9969d6dd4fc9a62458fe15de.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 34 +++++++++++----------------------- drivers/scsi/NCR5380.h | 5 +++-- 2 files changed, 14 insertions(+), 25 deletions(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index a47a825e7220..931b2581a33d 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -1227,22 +1227,15 @@ out: return ret; } -/* - * Function : int NCR5380_transfer_pio (struct Scsi_Host *instance, - * unsigned char *phase, int *count, unsigned char **data) +/** + * NCR5380_transfer_pio() - transfers data in given phase using polled I/O + * @instance: instance of driver + * @phase: pointer to what phase is expected + * @count: pointer to number of bytes to transfer + * @data: pointer to data pointer + * @can_sleep: 1 or 0 when sleeping is permitted or not, respectively * - * Purpose : transfers data in given phase using polled I/O - * - * Inputs : instance - instance of driver, *phase - pointer to - * what phase is expected, *count - pointer to number of - * bytes to transfer, **data - pointer to data pointer, - * can_sleep - 1 or 0 when sleeping is permitted or not, respectively. - * - * Returns : -1 when different phase is entered without transferring - * maximum number of bytes, 0 if all bytes are transferred or exit - * is in same phase. - * - * Also, *phase, *count, *data are modified in place. + * Returns: void. *phase, *count, *data are modified in place. */ /* @@ -1251,9 +1244,9 @@ out: * counts, we will always do a pseudo DMA or DMA transfer. */ -static int NCR5380_transfer_pio(struct Scsi_Host *instance, - unsigned char *phase, int *count, - unsigned char **data, unsigned int can_sleep) +static void NCR5380_transfer_pio(struct Scsi_Host *instance, + unsigned char *phase, int *count, + unsigned char **data, unsigned int can_sleep) { struct NCR5380_hostdata *hostdata = shost_priv(instance); unsigned char p = *phase, tmp; @@ -1358,11 +1351,6 @@ static int NCR5380_transfer_pio(struct Scsi_Host *instance, *phase = tmp & PHASE_MASK; else *phase = PHASE_UNKNOWN; - - if (!c || (*phase == p)) - return 0; - else - return -1; } /** diff --git a/drivers/scsi/NCR5380.h b/drivers/scsi/NCR5380.h index 84db14b036e4..64a1c6ce5e1b 100644 --- a/drivers/scsi/NCR5380.h +++ b/drivers/scsi/NCR5380.h @@ -285,8 +285,9 @@ static const char *NCR5380_info(struct Scsi_Host *instance); static void NCR5380_reselect(struct Scsi_Host *instance); static bool NCR5380_select(struct Scsi_Host *, struct scsi_cmnd *); static int NCR5380_transfer_dma(struct Scsi_Host *instance, unsigned char *phase, int *count, unsigned char **data); -static int NCR5380_transfer_pio(struct Scsi_Host *instance, unsigned char *phase, int *count, unsigned char **data, - unsigned int can_sleep); +static void NCR5380_transfer_pio(struct Scsi_Host *instance, + unsigned char *phase, int *count, + unsigned char **data, unsigned int can_sleep); static int NCR5380_poll_politely2(struct NCR5380_hostdata *, unsigned int, u8, u8, unsigned int, u8, u8, unsigned long); From c331df3d4a8db8a8d285b31895923cefabe77ecc Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 049/123] scsi: NCR5380: Remove obsolete comment This comment should have been removed in commit e7734ef14ead ("scsi: NCR5380: Remove context check") when the irqs_disabled() conditional was removed. Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/c54aff198b5a60be8ecfd50df0a9a77850730501.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index 931b2581a33d..94501773506b 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -198,7 +198,6 @@ static inline void set_resid_from_SCp(struct scsi_cmnd *cmd) * Polls the chip in a reasonably efficient manner waiting for an * event to occur. After a short quick poll we begin to yield the CPU * (if possible). In irq contexts the time-out is arbitrarily limited. - * Callers may hold locks as long as they are held in irq mode. * * Returns 0 if either or both event(s) occurred otherwise -ETIMEDOUT. */ From a8ebca904f8e0e02afcff961f342734d36b69c69 Mon Sep 17 00:00:00 2001 From: Finn Thain Date: Wed, 7 Aug 2024 13:36:28 +1000 Subject: [PATCH 050/123] scsi: NCR5380: Clean up indentation Tidy up a few indentation annoyances. No functional change. Signed-off-by: Finn Thain Link: https://lore.kernel.org/r/8541ea096fde9f8716b79e4f0707aed916a8c58d.1723001788.git.fthain@linux-m68k.org Signed-off-by: Martin K. Petersen --- drivers/scsi/NCR5380.c | 92 +++++++++++++++++++++------------------- drivers/scsi/NCR5380.h | 14 +++--- drivers/scsi/sun3_scsi.c | 2 +- 3 files changed, 56 insertions(+), 52 deletions(-) diff --git a/drivers/scsi/NCR5380.c b/drivers/scsi/NCR5380.c index 94501773506b..0e10502660de 100644 --- a/drivers/scsi/NCR5380.c +++ b/drivers/scsi/NCR5380.c @@ -1318,17 +1318,19 @@ static void NCR5380_transfer_pio(struct Scsi_Host *instance, dsprintk(NDEBUG_HANDSHAKE, instance, "REQ negated, handshake complete\n"); -/* - * We have several special cases to consider during REQ/ACK handshaking : - * 1. We were in MSGOUT phase, and we are on the last byte of the - * message. ATN must be dropped as ACK is dropped. - * - * 2. We are in a MSGIN phase, and we are on the last byte of the - * message. We must exit with ACK asserted, so that the calling - * code may raise ATN before dropping ACK to reject the message. - * - * 3. ACK and ATN are clear and the target may proceed as normal. - */ + /* + * We have several special cases to consider during REQ/ACK + * handshaking: + * + * 1. We were in MSGOUT phase, and we are on the last byte of + * the message. ATN must be dropped as ACK is dropped. + * + * 2. We are in MSGIN phase, and we are on the last byte of the + * message. We must exit with ACK asserted, so that the calling + * code may raise ATN before dropping ACK to reject the message. + * + * 3. ACK and ATN are clear & the target may proceed as normal. + */ if (!(p == PHASE_MSGIN && c == 1)) { if (p == PHASE_MSGOUT && c > 1) NCR5380_write(INITIATOR_COMMAND_REG, ICR_BASE | ICR_ASSERT_ATN); @@ -1559,39 +1561,41 @@ static int NCR5380_transfer_dma(struct Scsi_Host *instance, /* The result is zero iff pseudo DMA send/receive was completed. */ hostdata->dma_len = c; -/* - * A note regarding the DMA errata workarounds for early NMOS silicon. - * - * For DMA sends, we want to wait until the last byte has been - * transferred out over the bus before we turn off DMA mode. Alas, there - * seems to be no terribly good way of doing this on a 5380 under all - * conditions. For non-scatter-gather operations, we can wait until REQ - * and ACK both go false, or until a phase mismatch occurs. Gather-sends - * are nastier, since the device will be expecting more data than we - * are prepared to send it, and REQ will remain asserted. On a 53C8[01] we - * could test Last Byte Sent to assure transfer (I imagine this is precisely - * why this signal was added to the newer chips) but on the older 538[01] - * this signal does not exist. The workaround for this lack is a watchdog; - * we bail out of the wait-loop after a modest amount of wait-time if - * the usual exit conditions are not met. Not a terribly clean or - * correct solution :-% - * - * DMA receive is equally tricky due to a nasty characteristic of the NCR5380. - * If the chip is in DMA receive mode, it will respond to a target's - * REQ by latching the SCSI data into the INPUT DATA register and asserting - * ACK, even if it has _already_ been notified by the DMA controller that - * the current DMA transfer has completed! If the NCR5380 is then taken - * out of DMA mode, this already-acknowledged byte is lost. This is - * not a problem for "one DMA transfer per READ command", because - * the situation will never arise... either all of the data is DMA'ed - * properly, or the target switches to MESSAGE IN phase to signal a - * disconnection (either operation bringing the DMA to a clean halt). - * However, in order to handle scatter-receive, we must work around the - * problem. The chosen fix is to DMA fewer bytes, then check for the - * condition before taking the NCR5380 out of DMA mode. One or two extra - * bytes are transferred via PIO as necessary to fill out the original - * request. - */ + /* + * A note regarding the DMA errata workarounds for early NMOS silicon. + * + * For DMA sends, we want to wait until the last byte has been + * transferred out over the bus before we turn off DMA mode. Alas, there + * seems to be no terribly good way of doing this on a 5380 under all + * conditions. For non-scatter-gather operations, we can wait until REQ + * and ACK both go false, or until a phase mismatch occurs. Gather-sends + * are nastier, since the device will be expecting more data than we + * are prepared to send it, and REQ will remain asserted. On a 53C8[01] + * we could test Last Byte Sent to assure transfer (I imagine this is + * precisely why this signal was added to the newer chips) but on the + * older 538[01] this signal does not exist. The workaround for this + * lack is a watchdog; we bail out of the wait-loop after a modest + * amount of wait-time if the usual exit conditions are not met. + * Not a terribly clean or correct solution :-% + * + * DMA receive is equally tricky due to a nasty characteristic of the + * NCR5380. If the chip is in DMA receive mode, it will respond to a + * target's REQ by latching the SCSI data into the INPUT DATA register + * and asserting ACK, even if it has _already_ been notified by the + * DMA controller that the current DMA transfer has completed! If the + * NCR5380 is then taken out of DMA mode, this already-acknowledged + * byte is lost. + * + * This is not a problem for "one DMA transfer per READ + * command", because the situation will never arise... either all of + * the data is DMA'ed properly, or the target switches to MESSAGE IN + * phase to signal a disconnection (either operation bringing the DMA + * to a clean halt). However, in order to handle scatter-receive, we + * must work around the problem. The chosen fix is to DMA fewer bytes, + * then check for the condition before taking the NCR5380 out of DMA + * mode. One or two extra bytes are transferred via PIO as necessary + * to fill out the original request. + */ if ((hostdata->flags & FLAG_DMA_FIXUP) && (NCR5380_read(BUS_AND_STATUS_REG) & BASR_PHASE_MATCH)) { diff --git a/drivers/scsi/NCR5380.h b/drivers/scsi/NCR5380.h index 64a1c6ce5e1b..d402d4bffcb2 100644 --- a/drivers/scsi/NCR5380.h +++ b/drivers/scsi/NCR5380.h @@ -3,10 +3,10 @@ * NCR 5380 defines * * Copyright 1993, Drew Eckhardt - * Visionary Computing - * (Unix consulting and custom programming) - * drew@colorado.edu - * +1 (303) 666-5836 + * Visionary Computing + * (Unix consulting and custom programming) + * drew@colorado.edu + * +1 (303) 666-5836 * * For more information, please consult * @@ -78,7 +78,7 @@ #define ICR_DIFF_ENABLE 0x20 /* wo Set to enable diff. drivers */ #define ICR_ASSERT_ACK 0x10 /* rw ini Set to assert ACK */ #define ICR_ASSERT_BSY 0x08 /* rw Set to assert BSY */ -#define ICR_ASSERT_SEL 0x04 /* rw Set to assert SEL */ +#define ICR_ASSERT_SEL 0x04 /* rw Set to assert SEL */ #define ICR_ASSERT_ATN 0x02 /* rw Set to assert ATN */ #define ICR_ASSERT_DATA 0x01 /* rw SCSI_DATA_REG is asserted */ @@ -135,7 +135,7 @@ #define BASR_IRQ 0x10 /* ro mirror of IRQ pin */ #define BASR_PHASE_MATCH 0x08 /* ro Set when MSG CD IO match TCR */ #define BASR_BUSY_ERROR 0x04 /* ro Unexpected change to inactive state */ -#define BASR_ATN 0x02 /* ro BUS status */ +#define BASR_ATN 0x02 /* ro BUS status */ #define BASR_ACK 0x01 /* ro BUS status */ /* Write any value to this register to start a DMA send */ @@ -170,7 +170,7 @@ #define CSR_BASE CSR_53C80_INTR /* Note : PHASE_* macros are based on the values of the STATUS register */ -#define PHASE_MASK (SR_MSG | SR_CD | SR_IO) +#define PHASE_MASK (SR_MSG | SR_CD | SR_IO) #define PHASE_DATAOUT 0 #define PHASE_DATAIN SR_IO diff --git a/drivers/scsi/sun3_scsi.c b/drivers/scsi/sun3_scsi.c index f51702893306..fffc0fa52594 100644 --- a/drivers/scsi/sun3_scsi.c +++ b/drivers/scsi/sun3_scsi.c @@ -304,7 +304,7 @@ static int sun3scsi_dma_setup(struct NCR5380_hostdata *hostdata, sun3_udc_write(UDC_INT_ENABLE, UDC_CSR); #endif - return count; + return count; } From b9d104465a6c2f1bb27ae247bd90b4ab079a1699 Mon Sep 17 00:00:00 2001 From: Avri Altman Date: Sun, 11 Aug 2024 17:37:56 +0300 Subject: [PATCH 051/123] scsi: ufs: Prepare to add HCI capabilities sysfs Prepare so we'll be able to read various other HCI registers. While at it, fix the HCPID & HCMID register names to stand for what they really are. Also replace the pm_runtime_{get/put}_sync() calls in auto_hibern8_show to ufshcd_rpm_{get/put}_sync() as any host controller register reads should. Reviewed-by: Keoseong Park Reviewed-by: Bart Van Assche Reviewed-by: Bean Huo Signed-off-by: Avri Altman Link: https://lore.kernel.org/r/20240811143757.2538212-2-avri.altman@wdc.com Signed-off-by: Martin K. Petersen --- drivers/ufs/core/ufs-sysfs.c | 38 +++++++++++++++++++++--------------- include/ufs/ufshci.h | 5 +++-- 2 files changed, 25 insertions(+), 18 deletions(-) diff --git a/drivers/ufs/core/ufs-sysfs.c b/drivers/ufs/core/ufs-sysfs.c index e80a32421a8c..dec7746c98e0 100644 --- a/drivers/ufs/core/ufs-sysfs.c +++ b/drivers/ufs/core/ufs-sysfs.c @@ -198,6 +198,24 @@ static u32 ufshcd_us_to_ahit(unsigned int timer) FIELD_PREP(UFSHCI_AHIBERN8_SCALE_MASK, scale); } +static int ufshcd_read_hci_reg(struct ufs_hba *hba, u32 *val, unsigned int reg) +{ + down(&hba->host_sem); + if (!ufshcd_is_user_access_allowed(hba)) { + up(&hba->host_sem); + return -EBUSY; + } + + ufshcd_rpm_get_sync(hba); + ufshcd_hold(hba); + *val = ufshcd_readl(hba, reg); + ufshcd_release(hba); + ufshcd_rpm_put_sync(hba); + + up(&hba->host_sem); + return 0; +} + static ssize_t auto_hibern8_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -208,23 +226,11 @@ static ssize_t auto_hibern8_show(struct device *dev, if (!ufshcd_is_auto_hibern8_supported(hba)) return -EOPNOTSUPP; - down(&hba->host_sem); - if (!ufshcd_is_user_access_allowed(hba)) { - ret = -EBUSY; - goto out; - } + ret = ufshcd_read_hci_reg(hba, &ahit, REG_AUTO_HIBERNATE_IDLE_TIMER); + if (ret) + return ret; - pm_runtime_get_sync(hba->dev); - ufshcd_hold(hba); - ahit = ufshcd_readl(hba, REG_AUTO_HIBERNATE_IDLE_TIMER); - ufshcd_release(hba); - pm_runtime_put_sync(hba->dev); - - ret = sysfs_emit(buf, "%d\n", ufshcd_ahit_to_us(ahit)); - -out: - up(&hba->host_sem); - return ret; + return sysfs_emit(buf, "%d\n", ufshcd_ahit_to_us(ahit)); } static ssize_t auto_hibern8_store(struct device *dev, diff --git a/include/ufs/ufshci.h b/include/ufs/ufshci.h index 38fe97971a65..194e3655902e 100644 --- a/include/ufs/ufshci.h +++ b/include/ufs/ufshci.h @@ -25,8 +25,9 @@ enum { REG_CONTROLLER_CAPABILITIES = 0x00, REG_MCQCAP = 0x04, REG_UFS_VERSION = 0x08, - REG_CONTROLLER_DEV_ID = 0x10, - REG_CONTROLLER_PROD_ID = 0x14, + REG_EXT_CONTROLLER_CAPABILITIES = 0x0C, + REG_CONTROLLER_PID = 0x10, + REG_CONTROLLER_MID = 0x14, REG_AUTO_HIBERNATE_IDLE_TIMER = 0x18, REG_INTERRUPT_STATUS = 0x20, REG_INTERRUPT_ENABLE = 0x24, From f51d748195773c7780deca815dc787c281d77eb5 Mon Sep 17 00:00:00 2001 From: Avri Altman Date: Sun, 11 Aug 2024 17:37:57 +0300 Subject: [PATCH 052/123] scsi: ufs: Add HCI capabilities sysfs group The standard register map of UFSHCI is comprised of several groups. The first group (starting from offset 0x00), is the host capabilities group. It contains some interesting information that otherwise is not available, e.g. the UFS version of the platform etc. Reviewed-by: Keoseong Park Reviewed-by: Bart Van Assche Reviewed-by: Bean Huo Signed-off-by: Avri Altman Link: https://lore.kernel.org/r/20240811143757.2538212-3-avri.altman@wdc.com Signed-off-by: Martin K. Petersen --- Documentation/ABI/testing/sysfs-driver-ufs | 27 +++++++++++ drivers/ufs/core/ufs-sysfs.c | 53 ++++++++++++++++++++++ 2 files changed, 80 insertions(+) diff --git a/Documentation/ABI/testing/sysfs-driver-ufs b/Documentation/ABI/testing/sysfs-driver-ufs index fe943ce76c60..5fa6655aee84 100644 --- a/Documentation/ABI/testing/sysfs-driver-ufs +++ b/Documentation/ABI/testing/sysfs-driver-ufs @@ -1532,3 +1532,30 @@ Contact: Bean Huo Description: rtc_update_ms indicates how often the host should synchronize or update the UFS RTC. If set to 0, this will disable UFS RTC periodic update. + +What: /sys/devices/platform/.../ufshci_capabilities/version +Date: August 2024 +Contact: Avri Altman +Description: + Host Capabilities register group: UFS version register. + Symbol - VER. This file shows the UFSHCD version. + Example: Version 3.12 would be represented as 0000_0312h. + The file is read only. + +What: /sys/devices/platform/.../ufshci_capabilities/product_id +Date: August 2024 +Contact: Avri Altman +Description: + Host Capabilities register group: product ID register. + Symbol - HCPID. This file shows the UFSHCD product id. + The content of this register is vendor specific. + The file is read only. + +What: /sys/devices/platform/.../ufshci_capabilities/man_id +Date: August 2024 +Contact: Avri Altman +Description: + Host Capabilities register group: manufacturer ID register. + Symbol - HCMID. This file shows the UFSHCD manufacturer id. + The Manufacturer ID is defined by JEDEC in JEDEC-JEP106. + The file is read only. diff --git a/drivers/ufs/core/ufs-sysfs.c b/drivers/ufs/core/ufs-sysfs.c index dec7746c98e0..fe313800aed0 100644 --- a/drivers/ufs/core/ufs-sysfs.c +++ b/drivers/ufs/core/ufs-sysfs.c @@ -525,6 +525,58 @@ static const struct attribute_group ufs_sysfs_capabilities_group = { .attrs = ufs_sysfs_capabilities_attrs, }; +static ssize_t version_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + struct ufs_hba *hba = dev_get_drvdata(dev); + + return sysfs_emit(buf, "0x%x\n", hba->ufs_version); +} + +static ssize_t product_id_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int ret; + u32 val; + struct ufs_hba *hba = dev_get_drvdata(dev); + + ret = ufshcd_read_hci_reg(hba, &val, REG_CONTROLLER_PID); + if (ret) + return ret; + + return sysfs_emit(buf, "0x%x\n", val); +} + +static ssize_t man_id_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + int ret; + u32 val; + struct ufs_hba *hba = dev_get_drvdata(dev); + + ret = ufshcd_read_hci_reg(hba, &val, REG_CONTROLLER_MID); + if (ret) + return ret; + + return sysfs_emit(buf, "0x%x\n", val); +} + +static DEVICE_ATTR_RO(version); +static DEVICE_ATTR_RO(product_id); +static DEVICE_ATTR_RO(man_id); + +static struct attribute *ufs_sysfs_ufshci_cap_attrs[] = { + &dev_attr_version.attr, + &dev_attr_product_id.attr, + &dev_attr_man_id.attr, + NULL +}; + +static const struct attribute_group ufs_sysfs_ufshci_group = { + .name = "ufshci_capabilities", + .attrs = ufs_sysfs_ufshci_cap_attrs, +}; + static ssize_t monitor_enable_show(struct device *dev, struct device_attribute *attr, char *buf) { @@ -1508,6 +1560,7 @@ static const struct attribute_group ufs_sysfs_attributes_group = { static const struct attribute_group *ufs_sysfs_groups[] = { &ufs_sysfs_default_group, &ufs_sysfs_capabilities_group, + &ufs_sysfs_ufshci_group, &ufs_sysfs_monitor_group, &ufs_sysfs_power_info_group, &ufs_sysfs_device_descriptor_group, From 3ba963597d19d88eb06b50af8e8757abbdc9035b Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 15 Aug 2024 14:24:36 +0300 Subject: [PATCH 053/123] scsi: ufs: ufshcd-pltfrm: Signedness bug in ufshcd_parse_clock_info() The "sz" variable needs to be a signed type for the error handling to work as intended. Fortunately, there is some sanity checking on "sz" on the next line, so negative values would be caught and it doesn't really affect runtime. Fixes: eab0dce11dd9 ("scsi: ufs: ufshcd-pltfrm: Use of_property_count_u32_elems() to get property length") Signed-off-by: Dan Carpenter Link: https://lore.kernel.org/r/404a4727-89c6-410b-9ece-301fa399d4db@stanley.mountain Reviewed-by: Manivannan Sadhasivam Signed-off-by: Martin K. Petersen --- drivers/ufs/host/ufshcd-pltfrm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/ufs/host/ufshcd-pltfrm.c b/drivers/ufs/host/ufshcd-pltfrm.c index 0c9b303ccfa0..1f4f30d6cb42 100644 --- a/drivers/ufs/host/ufshcd-pltfrm.c +++ b/drivers/ufs/host/ufshcd-pltfrm.c @@ -31,7 +31,7 @@ static int ufshcd_parse_clock_info(struct ufs_hba *hba) const char *name; u32 *clkfreq = NULL; struct ufs_clk_info *clki; - size_t sz = 0; + ssize_t sz = 0; if (!np) goto out; From fa3c109a6d302b56437a6412c5f3044c3e12de03 Mon Sep 17 00:00:00 2001 From: Yosry Ahmed Date: Sat, 10 Aug 2024 00:59:11 +0000 Subject: [PATCH 054/123] dma-mapping: use bit masking to check VM_DMA_COHERENT In dma_common_find_pages(), area->flags are compared directly with VM_DMA_COHERENT. This works because VM_DMA_COHERENT is the only set flag. During development of a new feature (ASI [1]), a new VM flag is introduced, and that flag can be injected into VM_DMA_COHERENT mappings (among others). The presence of that flag caused dma_common_find_pages() to return NULL for VM_DMA_COHERENT addresses, leading to a lot of problems ending in crashing during boot. It took a bit of time to figure this problem out. It was a mistake to inject a VM flag to begin with, but it took a significant amount of debugging to figure out the problem. Most users of area->flags use bitmasking rather than equivalency to check for flags. Update dma_common_find_pages() and dma_common_free_remap() to do the same, which would have avoided the boot crashing. Instead, add a warning in dma_common_find_pages() if any extra VM flags are set to catch such problems more easily during development. No functional change intended. [1]https://lore.kernel.org/lkml/20240712-asi-rfc-24-v1-0-144b319a40d8@google.com/ Signed-off-by: Yosry Ahmed Signed-off-by: Christoph Hellwig --- kernel/dma/remap.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/kernel/dma/remap.c b/kernel/dma/remap.c index 27596f3b4aef..9e2afad1c615 100644 --- a/kernel/dma/remap.c +++ b/kernel/dma/remap.c @@ -10,8 +10,10 @@ struct page **dma_common_find_pages(void *cpu_addr) { struct vm_struct *area = find_vm_area(cpu_addr); - if (!area || area->flags != VM_DMA_COHERENT) + if (!area || !(area->flags & VM_DMA_COHERENT)) return NULL; + WARN(area->flags != VM_DMA_COHERENT, + "unexpected flags in area: %p\n", cpu_addr); return area->pages; } @@ -61,7 +63,7 @@ void dma_common_free_remap(void *cpu_addr, size_t size) { struct vm_struct *area = find_vm_area(cpu_addr); - if (!area || area->flags != VM_DMA_COHERENT) { + if (!area || !(area->flags & VM_DMA_COHERENT)) { WARN(1, "trying to free invalid coherent area: %p\n", cpu_addr); return; } From ba0fb44aed47693cc2482427f63ba6cd19051327 Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Sun, 11 Aug 2024 10:09:35 +0300 Subject: [PATCH 055/123] dma-mapping: replace zone_dma_bits by zone_dma_limit The hardware DMA limit might not be power of 2. When RAM range starts above 0, say 4GB, DMA limit of 30 bits should end at 5GB. A single high bit can not encode this limit. Use a plain address for the DMA zone limit instead. Since the DMA zone can now potentially span beyond 4GB physical limit of DMA32, make sure to use DMA zone for GFP_DMA32 allocations in that case. Signed-off-by: Catalin Marinas Co-developed-by: Baruch Siach Signed-off-by: Baruch Siach Reviewed-by: Catalin Marinas Reviewed-by: Petr Tesarik Signed-off-by: Christoph Hellwig --- arch/arm64/mm/init.c | 30 +++++++++++++++--------------- arch/powerpc/mm/mem.c | 5 ++++- arch/s390/mm/init.c | 2 +- include/linux/dma-direct.h | 2 +- kernel/dma/direct.c | 6 +++--- kernel/dma/pool.c | 4 ++-- kernel/dma/swiotlb.c | 6 +++--- 7 files changed, 29 insertions(+), 26 deletions(-) diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index 9b5ab6818f7f..c45e2152ca9e 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -115,35 +115,35 @@ static void __init arch_reserve_crashkernel(void) } /* - * Return the maximum physical address for a zone accessible by the given bits - * limit. If DRAM starts above 32-bit, expand the zone to the maximum + * Return the maximum physical address for a zone given its limit. + * If DRAM starts above 32-bit, expand the zone to the maximum * available memory, otherwise cap it at 32-bit. */ -static phys_addr_t __init max_zone_phys(unsigned int zone_bits) +static phys_addr_t __init max_zone_phys(phys_addr_t zone_limit) { - phys_addr_t zone_mask = DMA_BIT_MASK(zone_bits); phys_addr_t phys_start = memblock_start_of_DRAM(); if (phys_start > U32_MAX) - zone_mask = PHYS_ADDR_MAX; - else if (phys_start > zone_mask) - zone_mask = U32_MAX; + zone_limit = PHYS_ADDR_MAX; + else if (phys_start > zone_limit) + zone_limit = U32_MAX; - return min(zone_mask, memblock_end_of_DRAM() - 1) + 1; + return min(zone_limit, memblock_end_of_DRAM() - 1) + 1; } static void __init zone_sizes_init(void) { unsigned long max_zone_pfns[MAX_NR_ZONES] = {0}; - unsigned int __maybe_unused acpi_zone_dma_bits; - unsigned int __maybe_unused dt_zone_dma_bits; - phys_addr_t __maybe_unused dma32_phys_limit = max_zone_phys(32); + phys_addr_t __maybe_unused acpi_zone_dma_limit; + phys_addr_t __maybe_unused dt_zone_dma_limit; + phys_addr_t __maybe_unused dma32_phys_limit = + max_zone_phys(DMA_BIT_MASK(32)); #ifdef CONFIG_ZONE_DMA - acpi_zone_dma_bits = fls64(acpi_iort_dma_get_max_cpu_address()); - dt_zone_dma_bits = fls64(of_dma_get_max_cpu_address(NULL)); - zone_dma_bits = min3(32U, dt_zone_dma_bits, acpi_zone_dma_bits); - arm64_dma_phys_limit = max_zone_phys(zone_dma_bits); + acpi_zone_dma_limit = acpi_iort_dma_get_max_cpu_address(); + dt_zone_dma_limit = of_dma_get_max_cpu_address(NULL); + zone_dma_limit = min(dt_zone_dma_limit, acpi_zone_dma_limit); + arm64_dma_phys_limit = max_zone_phys(zone_dma_limit); max_zone_pfns[ZONE_DMA] = PFN_DOWN(arm64_dma_phys_limit); #endif #ifdef CONFIG_ZONE_DMA32 diff --git a/arch/powerpc/mm/mem.c b/arch/powerpc/mm/mem.c index da21cb018984..7e217aa4a274 100644 --- a/arch/powerpc/mm/mem.c +++ b/arch/powerpc/mm/mem.c @@ -216,7 +216,7 @@ static int __init mark_nonram_nosave(void) * everything else. GFP_DMA32 page allocations automatically fall back to * ZONE_DMA. * - * By using 31-bit unconditionally, we can exploit zone_dma_bits to inform the + * By using 31-bit unconditionally, we can exploit zone_dma_limit to inform the * generic DMA mapping code. 32-bit only devices (if not handled by an IOMMU * anyway) will take a first dip into ZONE_NORMAL and get otherwise served by * ZONE_DMA. @@ -230,6 +230,7 @@ void __init paging_init(void) { unsigned long long total_ram = memblock_phys_mem_size(); phys_addr_t top_of_ram = memblock_end_of_DRAM(); + int zone_dma_bits; #ifdef CONFIG_HIGHMEM unsigned long v = __fix_to_virt(FIX_KMAP_END); @@ -256,6 +257,8 @@ void __init paging_init(void) else zone_dma_bits = 31; + zone_dma_limit = DMA_BIT_MASK(zone_dma_bits); + #ifdef CONFIG_ZONE_DMA max_zone_pfns[ZONE_DMA] = min(max_low_pfn, 1UL << (zone_dma_bits - PAGE_SHIFT)); diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c index e3d258f9e726..688abc65c79e 100644 --- a/arch/s390/mm/init.c +++ b/arch/s390/mm/init.c @@ -97,7 +97,7 @@ void __init paging_init(void) vmem_map_init(); sparse_init(); - zone_dma_bits = 31; + zone_dma_limit = DMA_BIT_MASK(31); memset(max_zone_pfns, 0, sizeof(max_zone_pfns)); max_zone_pfns[ZONE_DMA] = virt_to_pfn(MAX_DMA_ADDRESS); max_zone_pfns[ZONE_NORMAL] = max_low_pfn; diff --git a/include/linux/dma-direct.h b/include/linux/dma-direct.h index edbe13d00776..d7e30d4f7503 100644 --- a/include/linux/dma-direct.h +++ b/include/linux/dma-direct.h @@ -12,7 +12,7 @@ #include #include -extern unsigned int zone_dma_bits; +extern u64 zone_dma_limit; /* * Record the mapping of CPU physical to DMA addresses for a given region. diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c index 4480a3cd92e0..f2ba074a6a54 100644 --- a/kernel/dma/direct.c +++ b/kernel/dma/direct.c @@ -20,7 +20,7 @@ * it for entirely different regions. In that case the arch code needs to * override the variable below for dma-direct to work properly. */ -unsigned int zone_dma_bits __ro_after_init = 24; +u64 zone_dma_limit __ro_after_init = DMA_BIT_MASK(24); static inline dma_addr_t phys_to_dma_direct(struct device *dev, phys_addr_t phys) @@ -59,7 +59,7 @@ static gfp_t dma_direct_optimal_gfp_mask(struct device *dev, u64 *phys_limit) * zones. */ *phys_limit = dma_to_phys(dev, dma_limit); - if (*phys_limit <= DMA_BIT_MASK(zone_dma_bits)) + if (*phys_limit <= zone_dma_limit) return GFP_DMA; if (*phys_limit <= DMA_BIT_MASK(32)) return GFP_DMA32; @@ -580,7 +580,7 @@ int dma_direct_supported(struct device *dev, u64 mask) * part of the check. */ if (IS_ENABLED(CONFIG_ZONE_DMA)) - min_mask = min_t(u64, min_mask, DMA_BIT_MASK(zone_dma_bits)); + min_mask = min_t(u64, min_mask, zone_dma_limit); return mask >= phys_to_dma_unencrypted(dev, min_mask); } diff --git a/kernel/dma/pool.c b/kernel/dma/pool.c index d10613eb0f63..7b04f7575796 100644 --- a/kernel/dma/pool.c +++ b/kernel/dma/pool.c @@ -70,9 +70,9 @@ static bool cma_in_zone(gfp_t gfp) /* CMA can't cross zone boundaries, see cma_activate_area() */ end = cma_get_base(cma) + size - 1; if (IS_ENABLED(CONFIG_ZONE_DMA) && (gfp & GFP_DMA)) - return end <= DMA_BIT_MASK(zone_dma_bits); + return end <= zone_dma_limit; if (IS_ENABLED(CONFIG_ZONE_DMA32) && (gfp & GFP_DMA32)) - return end <= DMA_BIT_MASK(32); + return end <= max(DMA_BIT_MASK(32), zone_dma_limit); return true; } diff --git a/kernel/dma/swiotlb.c b/kernel/dma/swiotlb.c index df68d29740a0..abcf3fa63a56 100644 --- a/kernel/dma/swiotlb.c +++ b/kernel/dma/swiotlb.c @@ -450,9 +450,9 @@ int swiotlb_init_late(size_t size, gfp_t gfp_mask, if (!remap) io_tlb_default_mem.can_grow = true; if (IS_ENABLED(CONFIG_ZONE_DMA) && (gfp_mask & __GFP_DMA)) - io_tlb_default_mem.phys_limit = DMA_BIT_MASK(zone_dma_bits); + io_tlb_default_mem.phys_limit = zone_dma_limit; else if (IS_ENABLED(CONFIG_ZONE_DMA32) && (gfp_mask & __GFP_DMA32)) - io_tlb_default_mem.phys_limit = DMA_BIT_MASK(32); + io_tlb_default_mem.phys_limit = max(DMA_BIT_MASK(32), zone_dma_limit); else io_tlb_default_mem.phys_limit = virt_to_phys(high_memory - 1); #endif @@ -629,7 +629,7 @@ static struct page *swiotlb_alloc_tlb(struct device *dev, size_t bytes, } gfp &= ~GFP_ZONEMASK; - if (phys_limit <= DMA_BIT_MASK(zone_dma_bits)) + if (phys_limit <= zone_dma_limit) gfp |= __GFP_DMA; else if (phys_limit <= DMA_BIT_MASK(32)) gfp |= __GFP_DMA32; From 3be9b846896d1d8b66040ece6e216cd58d03242e Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Sun, 11 Aug 2024 10:09:36 +0300 Subject: [PATCH 056/123] arm64: support DMA zone above 4GB Commit 791ab8b2e3db ("arm64: Ignore any DMA offsets in the max_zone_phys() calculation") made arm64 DMA/DMA32 zones span the entire RAM when RAM starts above 32-bits. This breaks hardware with DMA area that start above 32-bits. But the commit log says that "we haven't noticed any such hardware". It turns out that such hardware does exist. One such platform has RAM starting at 32GB with an internal bus that has the following DMA limits: #address-cells = <2>; #size-cells = <2>; dma-ranges = <0x00 0xc0000000 0x08 0x00000000 0x00 0x40000000>; That is, devices under this bus see 1GB of DMA range between 3GB-4GB in their address space. This range is mapped to CPU memory at 32GB-33GB. With current code DMA allocations for devices under this bus are not limited to DMA area, leading to run-time allocation failure. This commit reinstates DMA zone at the bottom of RAM. The result is DMA zone that properly reflects the hardware constraints as follows: [ 0.000000] Zone ranges: [ 0.000000] DMA [mem 0x0000000800000000-0x000000083fffffff] [ 0.000000] DMA32 empty [ 0.000000] Normal [mem 0x0000000840000000-0x0000000bffffffff] Signed-off-by: Catalin Marinas [baruch: split off the original patch] Signed-off-by: Baruch Siach Reviewed-by: Petr Tesarik Signed-off-by: Christoph Hellwig --- arch/arm64/mm/init.c | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index c45e2152ca9e..bfb10969cbf0 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -114,20 +114,8 @@ static void __init arch_reserve_crashkernel(void) low_size, high); } -/* - * Return the maximum physical address for a zone given its limit. - * If DRAM starts above 32-bit, expand the zone to the maximum - * available memory, otherwise cap it at 32-bit. - */ static phys_addr_t __init max_zone_phys(phys_addr_t zone_limit) { - phys_addr_t phys_start = memblock_start_of_DRAM(); - - if (phys_start > U32_MAX) - zone_limit = PHYS_ADDR_MAX; - else if (phys_start > zone_limit) - zone_limit = U32_MAX; - return min(zone_limit, memblock_end_of_DRAM() - 1) + 1; } From f69e342eec008e1bab772d3963c3dd9979293e13 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 24 Jul 2024 21:04:48 +0300 Subject: [PATCH 057/123] dma-mapping: call ->unmap_page and ->unmap_sg unconditionally Almost all instances of the dma_map_ops ->map_page()/map_sg() methods implement ->unmap_page()/unmap_sg() too. The once instance which doesn't dma_dummy_ops which is used to fail the DMA mapping and thus there won't be any calls to ->unmap_page()/unmap_sg(). Remove the checks for ->unmap_page()/unmap_sg() and call them directly to create an interface that is symmetrical to ->map_page()/map_sg(). Signed-off-by: Leon Romanovsky Signed-off-by: Leon Romanovsky Reviewed-by: Robin Murphy Signed-off-by: Christoph Hellwig --- kernel/dma/dummy.c | 21 +++++++++++++++++++++ kernel/dma/mapping.c | 4 ++-- 2 files changed, 23 insertions(+), 2 deletions(-) diff --git a/kernel/dma/dummy.c b/kernel/dma/dummy.c index b492d59ac77e..92de80e5b057 100644 --- a/kernel/dma/dummy.c +++ b/kernel/dma/dummy.c @@ -17,6 +17,15 @@ static dma_addr_t dma_dummy_map_page(struct device *dev, struct page *page, { return DMA_MAPPING_ERROR; } +static void dma_dummy_unmap_page(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir, unsigned long attrs) +{ + /* + * Dummy ops doesn't support map_page, so unmap_page should never be + * called. + */ + WARN_ON_ONCE(true); +} static int dma_dummy_map_sg(struct device *dev, struct scatterlist *sgl, int nelems, enum dma_data_direction dir, @@ -25,6 +34,16 @@ static int dma_dummy_map_sg(struct device *dev, struct scatterlist *sgl, return -EINVAL; } +static void dma_dummy_unmap_sg(struct device *dev, struct scatterlist *sgl, + int nelems, enum dma_data_direction dir, + unsigned long attrs) +{ + /* + * Dummy ops doesn't support map_sg, so unmap_sg should never be called. + */ + WARN_ON_ONCE(true); +} + static int dma_dummy_supported(struct device *hwdev, u64 mask) { return 0; @@ -33,6 +52,8 @@ static int dma_dummy_supported(struct device *hwdev, u64 mask) const struct dma_map_ops dma_dummy_ops = { .mmap = dma_dummy_mmap, .map_page = dma_dummy_map_page, + .unmap_page = dma_dummy_unmap_page, .map_sg = dma_dummy_map_sg, + .unmap_sg = dma_dummy_unmap_sg, .dma_supported = dma_dummy_supported, }; diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index b1c18058d55f..71416b156bb5 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -177,7 +177,7 @@ void dma_unmap_page_attrs(struct device *dev, dma_addr_t addr, size_t size, if (dma_map_direct(dev, ops) || arch_dma_unmap_page_direct(dev, addr + size)) dma_direct_unmap_page(dev, addr, size, dir, attrs); - else if (ops->unmap_page) + else ops->unmap_page(dev, addr, size, dir, attrs); debug_dma_unmap_page(dev, addr, size, dir); } @@ -291,7 +291,7 @@ void dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sg, if (dma_map_direct(dev, ops) || arch_dma_unmap_sg_direct(dev, sg, nents)) dma_direct_unmap_sg(dev, sg, nents, dir, attrs); - else if (ops->unmap_sg) + else ops->unmap_sg(dev, sg, nents, dir, attrs); } EXPORT_SYMBOL(dma_unmap_sg_attrs); From b5c58b2fdc427e7958412ecb2de2804a1f7c1572 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 24 Jul 2024 21:04:49 +0300 Subject: [PATCH 058/123] dma-mapping: direct calls for dma-iommu Directly call into dma-iommu just like we have been doing for dma-direct for a while. This avoids the indirect call overhead for IOMMU ops and removes the need to have DMA ops entirely for many common configurations. Signed-off-by: Leon Romanovsky Signed-off-by: Leon Romanovsky Acked-by: Greg Kroah-Hartman Acked-by: Robin Murphy Signed-off-by: Christoph Hellwig --- MAINTAINERS | 1 + drivers/iommu/Kconfig | 2 +- drivers/iommu/dma-iommu.c | 104 +++++++++---------------- drivers/iommu/intel/Kconfig | 1 - include/linux/device.h | 5 ++ include/linux/dma-map-ops.h | 13 ---- include/linux/iommu-dma.h | 147 ++++++++++++++++++++++++++++++++++++ kernel/dma/Kconfig | 4 + kernel/dma/Makefile | 2 +- kernel/dma/mapping.c | 83 +++++++++++++++++--- 10 files changed, 269 insertions(+), 93 deletions(-) create mode 100644 include/linux/iommu-dma.h diff --git a/MAINTAINERS b/MAINTAINERS index f328373463b0..6e653684efe9 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -11722,6 +11722,7 @@ T: git git://git.kernel.org/pub/scm/linux/kernel/git/iommu/linux.git F: drivers/iommu/dma-iommu.c F: drivers/iommu/dma-iommu.h F: drivers/iommu/iova.c +F: include/linux/iommu-dma.h F: include/linux/iova.h IOMMU SUBSYSTEM diff --git a/drivers/iommu/Kconfig b/drivers/iommu/Kconfig index a82f10054aec..61a6889f4dee 100644 --- a/drivers/iommu/Kconfig +++ b/drivers/iommu/Kconfig @@ -151,7 +151,7 @@ config OF_IOMMU # IOMMU-agnostic DMA-mapping layer config IOMMU_DMA def_bool ARM64 || X86 || S390 - select DMA_OPS + select DMA_OPS_HELPERS select IOMMU_API select IOMMU_IOVA select IRQ_MSI_IOMMU diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c index 7b1dfa0665df..3672d619bcb6 100644 --- a/drivers/iommu/dma-iommu.c +++ b/drivers/iommu/dma-iommu.c @@ -17,6 +17,7 @@ #include #include #include +#include #include #include #include @@ -1037,9 +1038,8 @@ out_unmap: return NULL; } -static struct sg_table *iommu_dma_alloc_noncontiguous(struct device *dev, - size_t size, enum dma_data_direction dir, gfp_t gfp, - unsigned long attrs) +struct sg_table *iommu_dma_alloc_noncontiguous(struct device *dev, size_t size, + enum dma_data_direction dir, gfp_t gfp, unsigned long attrs) { struct dma_sgt_handle *sh; @@ -1055,7 +1055,7 @@ static struct sg_table *iommu_dma_alloc_noncontiguous(struct device *dev, return &sh->sgt; } -static void iommu_dma_free_noncontiguous(struct device *dev, size_t size, +void iommu_dma_free_noncontiguous(struct device *dev, size_t size, struct sg_table *sgt, enum dma_data_direction dir) { struct dma_sgt_handle *sh = sgt_handle(sgt); @@ -1066,8 +1066,8 @@ static void iommu_dma_free_noncontiguous(struct device *dev, size_t size, kfree(sh); } -static void iommu_dma_sync_single_for_cpu(struct device *dev, - dma_addr_t dma_handle, size_t size, enum dma_data_direction dir) +void iommu_dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir) { phys_addr_t phys; @@ -1081,8 +1081,8 @@ static void iommu_dma_sync_single_for_cpu(struct device *dev, swiotlb_sync_single_for_cpu(dev, phys, size, dir); } -static void iommu_dma_sync_single_for_device(struct device *dev, - dma_addr_t dma_handle, size_t size, enum dma_data_direction dir) +void iommu_dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir) { phys_addr_t phys; @@ -1096,9 +1096,8 @@ static void iommu_dma_sync_single_for_device(struct device *dev, arch_sync_dma_for_device(phys, size, dir); } -static void iommu_dma_sync_sg_for_cpu(struct device *dev, - struct scatterlist *sgl, int nelems, - enum dma_data_direction dir) +void iommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sgl, + int nelems, enum dma_data_direction dir) { struct scatterlist *sg; int i; @@ -1112,9 +1111,8 @@ static void iommu_dma_sync_sg_for_cpu(struct device *dev, arch_sync_dma_for_cpu(sg_phys(sg), sg->length, dir); } -static void iommu_dma_sync_sg_for_device(struct device *dev, - struct scatterlist *sgl, int nelems, - enum dma_data_direction dir) +void iommu_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sgl, + int nelems, enum dma_data_direction dir) { struct scatterlist *sg; int i; @@ -1129,9 +1127,9 @@ static void iommu_dma_sync_sg_for_device(struct device *dev, arch_sync_dma_for_device(sg_phys(sg), sg->length, dir); } -static dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, - unsigned long offset, size_t size, enum dma_data_direction dir, - unsigned long attrs) +dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, + unsigned long offset, size_t size, enum dma_data_direction dir, + unsigned long attrs) { phys_addr_t phys = page_to_phys(page) + offset; bool coherent = dev_is_dma_coherent(dev); @@ -1189,7 +1187,7 @@ static dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, return iova; } -static void iommu_dma_unmap_page(struct device *dev, dma_addr_t dma_handle, +void iommu_dma_unmap_page(struct device *dev, dma_addr_t dma_handle, size_t size, enum dma_data_direction dir, unsigned long attrs) { struct iommu_domain *domain = iommu_get_dma_domain(dev); @@ -1342,8 +1340,8 @@ out_unmap: * impedance-matching, to be able to hand off a suitably-aligned list, * but still preserve the original offsets and sizes for the caller. */ -static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, - int nents, enum dma_data_direction dir, unsigned long attrs) +int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir, unsigned long attrs) { struct iommu_domain *domain = iommu_get_dma_domain(dev); struct iommu_dma_cookie *cookie = domain->iova_cookie; @@ -1462,8 +1460,8 @@ out: return ret; } -static void iommu_dma_unmap_sg(struct device *dev, struct scatterlist *sg, - int nents, enum dma_data_direction dir, unsigned long attrs) +void iommu_dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir, unsigned long attrs) { dma_addr_t end = 0, start; struct scatterlist *tmp; @@ -1512,7 +1510,7 @@ static void iommu_dma_unmap_sg(struct device *dev, struct scatterlist *sg, __iommu_dma_unmap(dev, start, end - start); } -static dma_addr_t iommu_dma_map_resource(struct device *dev, phys_addr_t phys, +dma_addr_t iommu_dma_map_resource(struct device *dev, phys_addr_t phys, size_t size, enum dma_data_direction dir, unsigned long attrs) { return __iommu_dma_map(dev, phys, size, @@ -1520,7 +1518,7 @@ static dma_addr_t iommu_dma_map_resource(struct device *dev, phys_addr_t phys, dma_get_mask(dev)); } -static void iommu_dma_unmap_resource(struct device *dev, dma_addr_t handle, +void iommu_dma_unmap_resource(struct device *dev, dma_addr_t handle, size_t size, enum dma_data_direction dir, unsigned long attrs) { __iommu_dma_unmap(dev, handle, size); @@ -1557,7 +1555,7 @@ static void __iommu_dma_free(struct device *dev, size_t size, void *cpu_addr) dma_free_contiguous(dev, page, alloc_size); } -static void iommu_dma_free(struct device *dev, size_t size, void *cpu_addr, +void iommu_dma_free(struct device *dev, size_t size, void *cpu_addr, dma_addr_t handle, unsigned long attrs) { __iommu_dma_unmap(dev, handle, size); @@ -1601,8 +1599,8 @@ out_free_pages: return NULL; } -static void *iommu_dma_alloc(struct device *dev, size_t size, - dma_addr_t *handle, gfp_t gfp, unsigned long attrs) +void *iommu_dma_alloc(struct device *dev, size_t size, dma_addr_t *handle, + gfp_t gfp, unsigned long attrs) { bool coherent = dev_is_dma_coherent(dev); int ioprot = dma_info_to_prot(DMA_BIDIRECTIONAL, coherent, attrs); @@ -1635,7 +1633,7 @@ static void *iommu_dma_alloc(struct device *dev, size_t size, return cpu_addr; } -static int iommu_dma_mmap(struct device *dev, struct vm_area_struct *vma, +int iommu_dma_mmap(struct device *dev, struct vm_area_struct *vma, void *cpu_addr, dma_addr_t dma_addr, size_t size, unsigned long attrs) { @@ -1666,7 +1664,7 @@ static int iommu_dma_mmap(struct device *dev, struct vm_area_struct *vma, vma->vm_page_prot); } -static int iommu_dma_get_sgtable(struct device *dev, struct sg_table *sgt, +int iommu_dma_get_sgtable(struct device *dev, struct sg_table *sgt, void *cpu_addr, dma_addr_t dma_addr, size_t size, unsigned long attrs) { @@ -1693,19 +1691,19 @@ static int iommu_dma_get_sgtable(struct device *dev, struct sg_table *sgt, return ret; } -static unsigned long iommu_dma_get_merge_boundary(struct device *dev) +unsigned long iommu_dma_get_merge_boundary(struct device *dev) { struct iommu_domain *domain = iommu_get_dma_domain(dev); return (1UL << __ffs(domain->pgsize_bitmap)) - 1; } -static size_t iommu_dma_opt_mapping_size(void) +size_t iommu_dma_opt_mapping_size(void) { return iova_rcache_range(); } -static size_t iommu_dma_max_mapping_size(struct device *dev) +size_t iommu_dma_max_mapping_size(struct device *dev) { if (dev_is_untrusted(dev)) return swiotlb_max_mapping_size(dev); @@ -1713,32 +1711,6 @@ static size_t iommu_dma_max_mapping_size(struct device *dev) return SIZE_MAX; } -static const struct dma_map_ops iommu_dma_ops = { - .flags = DMA_F_PCI_P2PDMA_SUPPORTED | - DMA_F_CAN_SKIP_SYNC, - .alloc = iommu_dma_alloc, - .free = iommu_dma_free, - .alloc_pages_op = dma_common_alloc_pages, - .free_pages = dma_common_free_pages, - .alloc_noncontiguous = iommu_dma_alloc_noncontiguous, - .free_noncontiguous = iommu_dma_free_noncontiguous, - .mmap = iommu_dma_mmap, - .get_sgtable = iommu_dma_get_sgtable, - .map_page = iommu_dma_map_page, - .unmap_page = iommu_dma_unmap_page, - .map_sg = iommu_dma_map_sg, - .unmap_sg = iommu_dma_unmap_sg, - .sync_single_for_cpu = iommu_dma_sync_single_for_cpu, - .sync_single_for_device = iommu_dma_sync_single_for_device, - .sync_sg_for_cpu = iommu_dma_sync_sg_for_cpu, - .sync_sg_for_device = iommu_dma_sync_sg_for_device, - .map_resource = iommu_dma_map_resource, - .unmap_resource = iommu_dma_unmap_resource, - .get_merge_boundary = iommu_dma_get_merge_boundary, - .opt_mapping_size = iommu_dma_opt_mapping_size, - .max_mapping_size = iommu_dma_max_mapping_size, -}; - void iommu_setup_dma_ops(struct device *dev) { struct iommu_domain *domain = iommu_get_domain_for_dev(dev); @@ -1746,19 +1718,15 @@ void iommu_setup_dma_ops(struct device *dev) if (dev_is_pci(dev)) dev->iommu->pci_32bit_workaround = !iommu_dma_forcedac; - if (iommu_is_dma_domain(domain)) { - if (iommu_dma_init_domain(domain, dev)) - goto out_err; - dev->dma_ops = &iommu_dma_ops; - } else if (dev->dma_ops == &iommu_dma_ops) { - /* Clean up if we've switched *from* a DMA domain */ - dev->dma_ops = NULL; - } + dev->dma_iommu = iommu_is_dma_domain(domain); + if (dev->dma_iommu && iommu_dma_init_domain(domain, dev)) + goto out_err; return; out_err: - pr_warn("Failed to set up IOMMU for device %s; retaining platform DMA ops\n", - dev_name(dev)); + pr_warn("Failed to set up IOMMU for device %s; retaining platform DMA ops\n", + dev_name(dev)); + dev->dma_iommu = false; } static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev, diff --git a/drivers/iommu/intel/Kconfig b/drivers/iommu/intel/Kconfig index f52fb39c968e..88fd32a9323c 100644 --- a/drivers/iommu/intel/Kconfig +++ b/drivers/iommu/intel/Kconfig @@ -12,7 +12,6 @@ config DMAR_DEBUG config INTEL_IOMMU bool "Support for Intel IOMMU using DMA Remapping Devices" depends on PCI_MSI && ACPI && X86 - select DMA_OPS select IOMMU_API select IOMMU_IOVA select IOMMUFD_DRIVER if IOMMUFD diff --git a/include/linux/device.h b/include/linux/device.h index 34eb20f5966f..1c5280d28bc3 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -707,6 +707,8 @@ struct device_physical_location { * for dma allocations. This flag is managed by the dma ops * instance from ->dma_supported. * @dma_skip_sync: DMA sync operations can be skipped for coherent buffers. + * @dma_iommu: Device is using default IOMMU implementation for DMA and + * doesn't rely on dma_ops structure. * * At the lowest level, every device in a Linux system is represented by an * instance of struct device. The device structure contains the information @@ -822,6 +824,9 @@ struct device { #ifdef CONFIG_DMA_NEED_SYNC bool dma_skip_sync:1; #endif +#ifdef CONFIG_IOMMU_DMA + bool dma_iommu:1; +#endif }; /** diff --git a/include/linux/dma-map-ops.h b/include/linux/dma-map-ops.h index 02a1c825896b..077b15c93bb8 100644 --- a/include/linux/dma-map-ops.h +++ b/include/linux/dma-map-ops.h @@ -13,20 +13,7 @@ struct cma; struct iommu_ops; -/* - * Values for struct dma_map_ops.flags: - * - * DMA_F_PCI_P2PDMA_SUPPORTED: Indicates the dma_map_ops implementation can - * handle PCI P2PDMA pages in the map_sg/unmap_sg operation. - * DMA_F_CAN_SKIP_SYNC: DMA sync operations can be skipped if the device is - * coherent and it's not an SWIOTLB buffer. - */ -#define DMA_F_PCI_P2PDMA_SUPPORTED (1 << 0) -#define DMA_F_CAN_SKIP_SYNC (1 << 1) - struct dma_map_ops { - unsigned int flags; - void *(*alloc)(struct device *dev, size_t size, dma_addr_t *dma_handle, gfp_t gfp, unsigned long attrs); diff --git a/include/linux/iommu-dma.h b/include/linux/iommu-dma.h new file mode 100644 index 000000000000..d30a58bf00fd --- /dev/null +++ b/include/linux/iommu-dma.h @@ -0,0 +1,147 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved + * + * DMA operations that map physical memory through IOMMU. + */ +#ifndef _LINUX_IOMMU_DMA_H +#define _LINUX_IOMMU_DMA_H + +#include + +#ifdef CONFIG_IOMMU_DMA +dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, + unsigned long offset, size_t size, enum dma_data_direction dir, + unsigned long attrs); +void iommu_dma_unmap_page(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir, unsigned long attrs); +int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir, unsigned long attrs); +void iommu_dma_unmap_sg(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir, unsigned long attrs); +void *iommu_dma_alloc(struct device *dev, size_t size, dma_addr_t *handle, + gfp_t gfp, unsigned long attrs); +int iommu_dma_mmap(struct device *dev, struct vm_area_struct *vma, + void *cpu_addr, dma_addr_t dma_addr, size_t size, + unsigned long attrs); +int iommu_dma_get_sgtable(struct device *dev, struct sg_table *sgt, + void *cpu_addr, dma_addr_t dma_addr, size_t size, + unsigned long attrs); +unsigned long iommu_dma_get_merge_boundary(struct device *dev); +size_t iommu_dma_opt_mapping_size(void); +size_t iommu_dma_max_mapping_size(struct device *dev); +void iommu_dma_free(struct device *dev, size_t size, void *cpu_addr, + dma_addr_t handle, unsigned long attrs); +dma_addr_t iommu_dma_map_resource(struct device *dev, phys_addr_t phys, + size_t size, enum dma_data_direction dir, unsigned long attrs); +void iommu_dma_unmap_resource(struct device *dev, dma_addr_t handle, + size_t size, enum dma_data_direction dir, unsigned long attrs); +struct sg_table *iommu_dma_alloc_noncontiguous(struct device *dev, size_t size, + enum dma_data_direction dir, gfp_t gfp, unsigned long attrs); +void iommu_dma_free_noncontiguous(struct device *dev, size_t size, + struct sg_table *sgt, enum dma_data_direction dir); +void iommu_dma_sync_single_for_cpu(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir); +void iommu_dma_sync_single_for_device(struct device *dev, dma_addr_t dma_handle, + size_t size, enum dma_data_direction dir); +void iommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sgl, + int nelems, enum dma_data_direction dir); +void iommu_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sgl, + int nelems, enum dma_data_direction dir); +#else +static inline dma_addr_t iommu_dma_map_page(struct device *dev, + struct page *page, unsigned long offset, size_t size, + enum dma_data_direction dir, unsigned long attrs) +{ + return DMA_MAPPING_ERROR; +} +static inline void iommu_dma_unmap_page(struct device *dev, + dma_addr_t dma_handle, size_t size, enum dma_data_direction dir, + unsigned long attrs) +{ +} +static inline int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg, + int nents, enum dma_data_direction dir, unsigned long attrs) +{ + return -EINVAL; +} +static inline void iommu_dma_unmap_sg(struct device *dev, + struct scatterlist *sg, int nents, enum dma_data_direction dir, + unsigned long attrs) +{ +} +static inline void *iommu_dma_alloc(struct device *dev, size_t size, + dma_addr_t *handle, gfp_t gfp, unsigned long attrs) +{ + return NULL; +} +static inline int iommu_dma_mmap(struct device *dev, struct vm_area_struct *vma, + void *cpu_addr, dma_addr_t dma_addr, size_t size, + unsigned long attrs) +{ + return -EINVAL; +} +static inline int iommu_dma_get_sgtable(struct device *dev, + struct sg_table *sgt, void *cpu_addr, dma_addr_t dma_addr, + size_t size, unsigned long attrs) +{ + return -EINVAL; +} +static inline unsigned long iommu_dma_get_merge_boundary(struct device *dev) +{ + return 0; +} +static inline size_t iommu_dma_opt_mapping_size(void) +{ + return 0; +} +static inline size_t iommu_dma_max_mapping_size(struct device *dev) +{ + return 0; +} +static inline void iommu_dma_free(struct device *dev, size_t size, + void *cpu_addr, dma_addr_t handle, unsigned long attrs) +{ +} +static inline dma_addr_t iommu_dma_map_resource(struct device *dev, + phys_addr_t phys, size_t size, enum dma_data_direction dir, + unsigned long attrs) +{ + return DMA_MAPPING_ERROR; +} +static inline void iommu_dma_unmap_resource(struct device *dev, + dma_addr_t handle, size_t size, enum dma_data_direction dir, + unsigned long attrs) +{ +} +static inline struct sg_table * +iommu_dma_alloc_noncontiguous(struct device *dev, size_t size, + enum dma_data_direction dir, gfp_t gfp, unsigned long attrs) +{ + return NULL; +} +static inline void iommu_dma_free_noncontiguous(struct device *dev, size_t size, + struct sg_table *sgt, enum dma_data_direction dir) +{ +} +static inline void iommu_dma_sync_single_for_cpu(struct device *dev, + dma_addr_t dma_handle, size_t size, + enum dma_data_direction dir) +{ +} +static inline void iommu_dma_sync_single_for_device(struct device *dev, + dma_addr_t dma_handle, size_t size, enum dma_data_direction dir) +{ +} +static inline void iommu_dma_sync_sg_for_cpu(struct device *dev, + struct scatterlist *sgl, int nelems, + enum dma_data_direction dir) +{ +} +static inline void iommu_dma_sync_sg_for_device(struct device *dev, + struct scatterlist *sgl, int nelems, + enum dma_data_direction dir) +{ +} +#endif /* CONFIG_IOMMU_DMA */ +#endif /* _LINUX_IOMMU_DMA_H */ diff --git a/kernel/dma/Kconfig b/kernel/dma/Kconfig index c06e56be0ca1..21bae1700836 100644 --- a/kernel/dma/Kconfig +++ b/kernel/dma/Kconfig @@ -8,8 +8,12 @@ config HAS_DMA depends on !NO_DMA default y +config DMA_OPS_HELPERS + bool + config DMA_OPS depends on HAS_DMA + select DMA_OPS_HELPERS bool # diff --git a/kernel/dma/Makefile b/kernel/dma/Makefile index 21926e46ef4f..2e6e933cf7f3 100644 --- a/kernel/dma/Makefile +++ b/kernel/dma/Makefile @@ -1,7 +1,7 @@ # SPDX-License-Identifier: GPL-2.0 obj-$(CONFIG_HAS_DMA) += mapping.o direct.o -obj-$(CONFIG_DMA_OPS) += ops_helpers.o +obj-$(CONFIG_DMA_OPS_HELPERS) += ops_helpers.o obj-$(CONFIG_DMA_OPS) += dummy.o obj-$(CONFIG_DMA_CMA) += contiguous.o obj-$(CONFIG_DMA_DECLARE_COHERENT) += coherent.o diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index 71416b156bb5..b50ae3d198a6 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -10,6 +10,7 @@ #include #include #include +#include #include #include #include @@ -113,11 +114,27 @@ void *dmam_alloc_attrs(struct device *dev, size_t size, dma_addr_t *dma_handle, } EXPORT_SYMBOL(dmam_alloc_attrs); +#ifdef CONFIG_IOMMU_DMA +static bool use_dma_iommu(struct device *dev) +{ + return dev->dma_iommu; +} +#else +static bool use_dma_iommu(struct device *dev) +{ + return false; +} +#endif + static bool dma_go_direct(struct device *dev, dma_addr_t mask, const struct dma_map_ops *ops) { + if (use_dma_iommu(dev)) + return false; + if (likely(!ops)) return true; + #ifdef CONFIG_DMA_OPS_BYPASS if (dev->dma_ops_bypass) return min_not_zero(mask, dev->bus_dma_limit) >= @@ -159,6 +176,8 @@ dma_addr_t dma_map_page_attrs(struct device *dev, struct page *page, if (dma_map_direct(dev, ops) || arch_dma_map_page_direct(dev, page_to_phys(page) + offset + size)) addr = dma_direct_map_page(dev, page, offset, size, dir, attrs); + else if (use_dma_iommu(dev)) + addr = iommu_dma_map_page(dev, page, offset, size, dir, attrs); else addr = ops->map_page(dev, page, offset, size, dir, attrs); kmsan_handle_dma(page, offset, size, dir); @@ -177,6 +196,8 @@ void dma_unmap_page_attrs(struct device *dev, dma_addr_t addr, size_t size, if (dma_map_direct(dev, ops) || arch_dma_unmap_page_direct(dev, addr + size)) dma_direct_unmap_page(dev, addr, size, dir, attrs); + else if (use_dma_iommu(dev)) + iommu_dma_unmap_page(dev, addr, size, dir, attrs); else ops->unmap_page(dev, addr, size, dir, attrs); debug_dma_unmap_page(dev, addr, size, dir); @@ -197,6 +218,8 @@ static int __dma_map_sg_attrs(struct device *dev, struct scatterlist *sg, if (dma_map_direct(dev, ops) || arch_dma_map_sg_direct(dev, sg, nents)) ents = dma_direct_map_sg(dev, sg, nents, dir, attrs); + else if (use_dma_iommu(dev)) + ents = iommu_dma_map_sg(dev, sg, nents, dir, attrs); else ents = ops->map_sg(dev, sg, nents, dir, attrs); @@ -291,7 +314,9 @@ void dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sg, if (dma_map_direct(dev, ops) || arch_dma_unmap_sg_direct(dev, sg, nents)) dma_direct_unmap_sg(dev, sg, nents, dir, attrs); - else + else if (use_dma_iommu(dev)) + iommu_dma_unmap_sg(dev, sg, nents, dir, attrs); + else if (ops->unmap_sg) ops->unmap_sg(dev, sg, nents, dir, attrs); } EXPORT_SYMBOL(dma_unmap_sg_attrs); @@ -309,6 +334,8 @@ dma_addr_t dma_map_resource(struct device *dev, phys_addr_t phys_addr, if (dma_map_direct(dev, ops)) addr = dma_direct_map_resource(dev, phys_addr, size, dir, attrs); + else if (use_dma_iommu(dev)) + addr = iommu_dma_map_resource(dev, phys_addr, size, dir, attrs); else if (ops->map_resource) addr = ops->map_resource(dev, phys_addr, size, dir, attrs); @@ -323,7 +350,11 @@ void dma_unmap_resource(struct device *dev, dma_addr_t addr, size_t size, const struct dma_map_ops *ops = get_dma_ops(dev); BUG_ON(!valid_dma_direction(dir)); - if (!dma_map_direct(dev, ops) && ops->unmap_resource) + if (dma_map_direct(dev, ops)) + ; /* nothing to do: uncached and no swiotlb */ + else if (use_dma_iommu(dev)) + iommu_dma_unmap_resource(dev, addr, size, dir, attrs); + else if (ops->unmap_resource) ops->unmap_resource(dev, addr, size, dir, attrs); debug_dma_unmap_resource(dev, addr, size, dir); } @@ -338,6 +369,8 @@ void __dma_sync_single_for_cpu(struct device *dev, dma_addr_t addr, size_t size, BUG_ON(!valid_dma_direction(dir)); if (dma_map_direct(dev, ops)) dma_direct_sync_single_for_cpu(dev, addr, size, dir); + else if (use_dma_iommu(dev)) + iommu_dma_sync_single_for_cpu(dev, addr, size, dir); else if (ops->sync_single_for_cpu) ops->sync_single_for_cpu(dev, addr, size, dir); debug_dma_sync_single_for_cpu(dev, addr, size, dir); @@ -352,6 +385,8 @@ void __dma_sync_single_for_device(struct device *dev, dma_addr_t addr, BUG_ON(!valid_dma_direction(dir)); if (dma_map_direct(dev, ops)) dma_direct_sync_single_for_device(dev, addr, size, dir); + else if (use_dma_iommu(dev)) + iommu_dma_sync_single_for_device(dev, addr, size, dir); else if (ops->sync_single_for_device) ops->sync_single_for_device(dev, addr, size, dir); debug_dma_sync_single_for_device(dev, addr, size, dir); @@ -366,6 +401,8 @@ void __dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, BUG_ON(!valid_dma_direction(dir)); if (dma_map_direct(dev, ops)) dma_direct_sync_sg_for_cpu(dev, sg, nelems, dir); + else if (use_dma_iommu(dev)) + iommu_dma_sync_sg_for_cpu(dev, sg, nelems, dir); else if (ops->sync_sg_for_cpu) ops->sync_sg_for_cpu(dev, sg, nelems, dir); debug_dma_sync_sg_for_cpu(dev, sg, nelems, dir); @@ -380,6 +417,8 @@ void __dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, BUG_ON(!valid_dma_direction(dir)); if (dma_map_direct(dev, ops)) dma_direct_sync_sg_for_device(dev, sg, nelems, dir); + else if (use_dma_iommu(dev)) + iommu_dma_sync_sg_for_device(dev, sg, nelems, dir); else if (ops->sync_sg_for_device) ops->sync_sg_for_device(dev, sg, nelems, dir); debug_dma_sync_sg_for_device(dev, sg, nelems, dir); @@ -405,7 +444,7 @@ static void dma_setup_need_sync(struct device *dev) { const struct dma_map_ops *ops = get_dma_ops(dev); - if (dma_map_direct(dev, ops) || (ops->flags & DMA_F_CAN_SKIP_SYNC)) + if (dma_map_direct(dev, ops) || use_dma_iommu(dev)) /* * dma_skip_sync will be reset to %false on first SWIOTLB buffer * mapping, if any. During the device initialization, it's @@ -446,6 +485,9 @@ int dma_get_sgtable_attrs(struct device *dev, struct sg_table *sgt, if (dma_alloc_direct(dev, ops)) return dma_direct_get_sgtable(dev, sgt, cpu_addr, dma_addr, size, attrs); + if (use_dma_iommu(dev)) + return iommu_dma_get_sgtable(dev, sgt, cpu_addr, dma_addr, + size, attrs); if (!ops->get_sgtable) return -ENXIO; return ops->get_sgtable(dev, sgt, cpu_addr, dma_addr, size, attrs); @@ -482,6 +524,8 @@ bool dma_can_mmap(struct device *dev) if (dma_alloc_direct(dev, ops)) return dma_direct_can_mmap(dev); + if (use_dma_iommu(dev)) + return true; return ops->mmap != NULL; } EXPORT_SYMBOL_GPL(dma_can_mmap); @@ -508,6 +552,9 @@ int dma_mmap_attrs(struct device *dev, struct vm_area_struct *vma, if (dma_alloc_direct(dev, ops)) return dma_direct_mmap(dev, vma, cpu_addr, dma_addr, size, attrs); + if (use_dma_iommu(dev)) + return iommu_dma_mmap(dev, vma, cpu_addr, dma_addr, size, + attrs); if (!ops->mmap) return -ENXIO; return ops->mmap(dev, vma, cpu_addr, dma_addr, size, attrs); @@ -559,6 +606,8 @@ void *dma_alloc_attrs(struct device *dev, size_t size, dma_addr_t *dma_handle, if (dma_alloc_direct(dev, ops)) cpu_addr = dma_direct_alloc(dev, size, dma_handle, flag, attrs); + else if (use_dma_iommu(dev)) + cpu_addr = iommu_dma_alloc(dev, size, dma_handle, flag, attrs); else if (ops->alloc) cpu_addr = ops->alloc(dev, size, dma_handle, flag, attrs); else @@ -591,6 +640,8 @@ void dma_free_attrs(struct device *dev, size_t size, void *cpu_addr, debug_dma_free_coherent(dev, size, cpu_addr, dma_handle); if (dma_alloc_direct(dev, ops)) dma_direct_free(dev, size, cpu_addr, dma_handle, attrs); + else if (use_dma_iommu(dev)) + iommu_dma_free(dev, size, cpu_addr, dma_handle, attrs); else if (ops->free) ops->free(dev, size, cpu_addr, dma_handle, attrs); } @@ -611,6 +662,8 @@ static struct page *__dma_alloc_pages(struct device *dev, size_t size, size = PAGE_ALIGN(size); if (dma_alloc_direct(dev, ops)) return dma_direct_alloc_pages(dev, size, dma_handle, dir, gfp); + if (use_dma_iommu(dev)) + return dma_common_alloc_pages(dev, size, dma_handle, dir, gfp); if (!ops->alloc_pages_op) return NULL; return ops->alloc_pages_op(dev, size, dma_handle, dir, gfp); @@ -635,6 +688,8 @@ static void __dma_free_pages(struct device *dev, size_t size, struct page *page, size = PAGE_ALIGN(size); if (dma_alloc_direct(dev, ops)) dma_direct_free_pages(dev, size, page, dma_handle, dir); + else if (use_dma_iommu(dev)) + dma_common_free_pages(dev, size, page, dma_handle, dir); else if (ops->free_pages) ops->free_pages(dev, size, page, dma_handle, dir); } @@ -697,6 +752,8 @@ struct sg_table *dma_alloc_noncontiguous(struct device *dev, size_t size, if (ops && ops->alloc_noncontiguous) sgt = ops->alloc_noncontiguous(dev, size, dir, gfp, attrs); + else if (use_dma_iommu(dev)) + sgt = iommu_dma_alloc_noncontiguous(dev, size, dir, gfp, attrs); else sgt = alloc_single_sgt(dev, size, dir, gfp); @@ -725,6 +782,8 @@ void dma_free_noncontiguous(struct device *dev, size_t size, debug_dma_unmap_sg(dev, sgt->sgl, sgt->orig_nents, dir); if (ops && ops->free_noncontiguous) ops->free_noncontiguous(dev, size, sgt, dir); + else if (use_dma_iommu(dev)) + iommu_dma_free_noncontiguous(dev, size, sgt, dir); else free_single_sgt(dev, size, sgt, dir); } @@ -772,6 +831,8 @@ static int dma_supported(struct device *dev, u64 mask) { const struct dma_map_ops *ops = get_dma_ops(dev); + if (WARN_ON(ops && use_dma_iommu(dev))) + return false; /* * ->dma_supported sets the bypass flag, so we must always call * into the method here unless the device is truly direct mapped. @@ -787,17 +848,14 @@ bool dma_pci_p2pdma_supported(struct device *dev) { const struct dma_map_ops *ops = get_dma_ops(dev); - /* if ops is not set, dma direct will be used which supports P2PDMA */ - if (!ops) - return true; - /* * Note: dma_ops_bypass is not checked here because P2PDMA should * not be used with dma mapping ops that do not have support even * if the specific device is bypassing them. */ - return ops->flags & DMA_F_PCI_P2PDMA_SUPPORTED; + /* if ops is not set, dma direct and default IOMMU support P2PDMA */ + return !ops; } EXPORT_SYMBOL_GPL(dma_pci_p2pdma_supported); @@ -865,6 +923,8 @@ size_t dma_max_mapping_size(struct device *dev) if (dma_map_direct(dev, ops)) size = dma_direct_max_mapping_size(dev); + else if (use_dma_iommu(dev)) + size = iommu_dma_max_mapping_size(dev); else if (ops && ops->max_mapping_size) size = ops->max_mapping_size(dev); @@ -877,7 +937,9 @@ size_t dma_opt_mapping_size(struct device *dev) const struct dma_map_ops *ops = get_dma_ops(dev); size_t size = SIZE_MAX; - if (ops && ops->opt_mapping_size) + if (use_dma_iommu(dev)) + size = iommu_dma_opt_mapping_size(); + else if (ops && ops->opt_mapping_size) size = ops->opt_mapping_size(); return min(dma_max_mapping_size(dev), size); @@ -888,6 +950,9 @@ unsigned long dma_get_merge_boundary(struct device *dev) { const struct dma_map_ops *ops = get_dma_ops(dev); + if (use_dma_iommu(dev)) + return iommu_dma_get_merge_boundary(dev); + if (!ops || !ops->get_merge_boundary) return 0; /* can't merge */ From eabc10e60dac08b9f4f05872d785532d0856e09f Mon Sep 17 00:00:00 2001 From: GiSeong Ji Date: Thu, 22 Aug 2024 18:18:06 +0900 Subject: [PATCH 059/123] security: smack: Fix indentation in smack_netfilter.c Aligned parameters in the function declaration of smack_ip_output to adhere to the Linux kernel coding style guidelines. The parameters of the smack_ip_output function were previously misaligned, with the second and third parameters not aligned under the first parameter. This change corrects the indentation, improving code readability and maintaining consistency with the rest of the codebase. Signed-off-by: GiSeong Ji Signed-off-by: Casey Schaufler --- security/smack/smack_netfilter.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/security/smack/smack_netfilter.c b/security/smack/smack_netfilter.c index b945c1d3a743..709b1fcff514 100644 --- a/security/smack/smack_netfilter.c +++ b/security/smack/smack_netfilter.c @@ -19,8 +19,8 @@ #include "smack.h" static unsigned int smack_ip_output(void *priv, - struct sk_buff *skb, - const struct nf_hook_state *state) + struct sk_buff *skb, + const struct nf_hook_state *state) { struct sock *sk = skb_to_full_sk(skb); struct socket_smack *ssp; From 89835a58f5f54d52537709f2513fb91024e2d069 Mon Sep 17 00:00:00 2001 From: Avri Altman Date: Wed, 21 Aug 2024 08:54:11 +0300 Subject: [PATCH 060/123] scsi: ufs: Move UFS trace events to private header UFS trace events are called exclusively from the UFS core drivers. Make those events private to the core driver. The MAINTAINERS file does not need updating as the maintainership remains the same and the relevant directory is already covered. Reviewed-by: Bart Van Assche Signed-off-by: Avri Altman Link: https://lore.kernel.org/r/20240821055411.3128159-1-avri.altman@wdc.com Acked-by: Bean Huo Signed-off-by: Martin K. Petersen --- include/trace/events/ufs.h => drivers/ufs/core/ufs_trace.h | 6 ++++++ drivers/ufs/core/ufshcd.c | 2 +- include/ufs/ufs.h | 4 ++-- 3 files changed, 9 insertions(+), 3 deletions(-) rename include/trace/events/ufs.h => drivers/ufs/core/ufs_trace.h (98%) diff --git a/include/trace/events/ufs.h b/drivers/ufs/core/ufs_trace.h similarity index 98% rename from include/trace/events/ufs.h rename to drivers/ufs/core/ufs_trace.h index c4e209fbdfbb..84deca2b841d 100644 --- a/include/trace/events/ufs.h +++ b/drivers/ufs/core/ufs_trace.h @@ -9,6 +9,7 @@ #if !defined(_TRACE_UFS_H) || defined(TRACE_HEADER_MULTI_READ) #define _TRACE_UFS_H +#include #include #define str_opcode(opcode) \ @@ -395,5 +396,10 @@ TRACE_EVENT(ufshcd_exception_event, #endif /* if !defined(_TRACE_UFS_H) || defined(TRACE_HEADER_MULTI_READ) */ +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH ../../drivers/ufs/core +#undef TRACE_INCLUDE_FILE +#define TRACE_INCLUDE_FILE ufs_trace + /* This part must be outside protection */ #include diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index 0dd26059f5d7..db30d0c4d91e 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -39,7 +39,7 @@ #include #define CREATE_TRACE_POINTS -#include +#include "ufs_trace.h" #define UFSHCD_ENABLE_INTRS (UTP_TRANSFER_REQ_COMPL |\ UTP_TASK_REQ_COMPL |\ diff --git a/include/ufs/ufs.h b/include/ufs/ufs.h index 853e95957c31..e594abe5d05f 100644 --- a/include/ufs/ufs.h +++ b/include/ufs/ufs.h @@ -597,7 +597,7 @@ struct ufs_dev_info { }; /* - * This enum is used in string mapping in include/trace/events/ufs.h. + * This enum is used in string mapping in ufs_trace.h. */ enum ufs_trace_str_t { UFS_CMD_SEND, UFS_CMD_COMP, UFS_DEV_COMP, @@ -607,7 +607,7 @@ enum ufs_trace_str_t { /* * Transaction Specific Fields (TSF) type in the UPIU package, this enum is - * used in include/trace/events/ufs.h for UFS command trace. + * used in ufs_trace.h for UFS command trace. */ enum ufs_trace_tsf_t { UFS_TSF_CDB, UFS_TSF_OSF, UFS_TSF_TM_INPUT, UFS_TSF_TM_OUTPUT From 2e4b02fad094976763af08fec2c620f4f8edd9ae Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Thu, 15 Aug 2024 14:29:05 +0300 Subject: [PATCH 061/123] scsi: elx: libefc: Fix potential use after free in efc_nport_vport_del() The kref_put() function will call nport->release if the refcount drops to zero. The nport->release release function is _efc_nport_free() which frees "nport". But then we dereference "nport" on the next line which is a use after free. Re-order these lines to avoid the use after free. Fixes: fcd427303eb9 ("scsi: elx: libefc: SLI and FC PORT state machine interfaces") Signed-off-by: Dan Carpenter Link: https://lore.kernel.org/r/b666ab26-6581-4213-9a3d-32a9147f0399@stanley.mountain Reviewed-by: Daniel Wagner Signed-off-by: Martin K. Petersen --- drivers/scsi/elx/libefc/efc_nport.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/elx/libefc/efc_nport.c b/drivers/scsi/elx/libefc/efc_nport.c index 2e83a667901f..1a7437f4328e 100644 --- a/drivers/scsi/elx/libefc/efc_nport.c +++ b/drivers/scsi/elx/libefc/efc_nport.c @@ -705,9 +705,9 @@ efc_nport_vport_del(struct efc *efc, struct efc_domain *domain, spin_lock_irqsave(&efc->lock, flags); list_for_each_entry(nport, &domain->nport_list, list_entry) { if (nport->wwpn == wwpn && nport->wwnn == wwnn) { - kref_put(&nport->ref, nport->release); /* Shutdown this NPORT */ efc_sm_post_event(&nport->sm, EFC_EVT_SHUTDOWN, NULL); + kref_put(&nport->ref, nport->release); break; } } From 3c9265ed191d34349889ee426ed4a2394ba4a77b Mon Sep 17 00:00:00 2001 From: Yue Haibing Date: Sat, 10 Aug 2024 17:34:37 +0800 Subject: [PATCH 062/123] scsi: target: Remove unused declarations Commit 13247018d68f ("scsi: target: iscsi: Fix hang in the iSCSI login code") removed iscsi_handle_login_thread_timeout() but left declaration. Commit 3e1c81a95f0d ("iscsi-target: Refactor RX PDU logic + export request PDU handling") left iscsi_target_get_initial_payload() declaration. Commit d703ce2f7f4d ("iscsi/iser-target: Convert to command priv_size usage") remove iscsit_alloc_cmd() but left declaration. And finally, a few other declarations were never implenmented since introduction in commit e48354ce078c ("iscsi-target: Add iSCSI fabric support for target v4.1"). Signed-off-by: Yue Haibing Link: https://lore.kernel.org/r/20240810093437.2586476-1-yuehaibing@huawei.com Signed-off-by: Martin K. Petersen --- drivers/target/iscsi/iscsi_target.h | 2 -- drivers/target/iscsi/iscsi_target_login.h | 1 - drivers/target/iscsi/iscsi_target_nego.h | 2 -- drivers/target/iscsi/iscsi_target_tpg.h | 5 ----- drivers/target/iscsi/iscsi_target_util.h | 5 ----- 5 files changed, 15 deletions(-) diff --git a/drivers/target/iscsi/iscsi_target.h b/drivers/target/iscsi/iscsi_target.h index 0c997a08adec..873411e95ed2 100644 --- a/drivers/target/iscsi/iscsi_target.h +++ b/drivers/target/iscsi/iscsi_target.h @@ -15,7 +15,6 @@ struct kref; struct sockaddr_storage; extern struct iscsi_tiqn *iscsit_get_tiqn_for_login(unsigned char *); -extern struct iscsi_tiqn *iscsit_get_tiqn(unsigned char *, int); extern void iscsit_put_tiqn_for_login(struct iscsi_tiqn *); extern struct iscsi_tiqn *iscsit_add_tiqn(unsigned char *); extern void iscsit_del_tiqn(struct iscsi_tiqn *); @@ -35,7 +34,6 @@ extern void iscsit_set_unsolicited_dataout(struct iscsit_cmd *); extern int iscsit_logout_closesession(struct iscsit_cmd *, struct iscsit_conn *); extern int iscsit_logout_closeconnection(struct iscsit_cmd *, struct iscsit_conn *); extern int iscsit_logout_removeconnforrecovery(struct iscsit_cmd *, struct iscsit_conn *); -extern int iscsit_send_async_msg(struct iscsit_conn *, u16, u8, u8); extern int iscsit_build_r2ts_for_cmd(struct iscsit_conn *, struct iscsit_cmd *, bool recovery); extern void iscsit_thread_get_cpumask(struct iscsit_conn *); extern int iscsi_target_tx_thread(void *); diff --git a/drivers/target/iscsi/iscsi_target_login.h b/drivers/target/iscsi/iscsi_target_login.h index 3ca2f232b387..e8760735486b 100644 --- a/drivers/target/iscsi/iscsi_target_login.h +++ b/drivers/target/iscsi/iscsi_target_login.h @@ -24,6 +24,5 @@ extern int iscsit_start_kthreads(struct iscsit_conn *); extern void iscsi_post_login_handler(struct iscsi_np *, struct iscsit_conn *, u8); extern void iscsi_target_login_sess_out(struct iscsit_conn *, bool, bool); extern int iscsi_target_login_thread(void *); -extern void iscsi_handle_login_thread_timeout(struct timer_list *t); #endif /*** ISCSI_TARGET_LOGIN_H ***/ diff --git a/drivers/target/iscsi/iscsi_target_nego.h b/drivers/target/iscsi/iscsi_target_nego.h index 41c3db3ddeaa..e60a46d34835 100644 --- a/drivers/target/iscsi/iscsi_target_nego.h +++ b/drivers/target/iscsi/iscsi_target_nego.h @@ -15,8 +15,6 @@ extern int extract_param(const char *, const char *, unsigned int, char *, unsigned char *); extern int iscsi_target_check_login_request(struct iscsit_conn *, struct iscsi_login *); -extern int iscsi_target_get_initial_payload(struct iscsit_conn *, - struct iscsi_login *); extern int iscsi_target_locate_portal(struct iscsi_np *, struct iscsit_conn *, struct iscsi_login *); extern int iscsi_target_start_negotiation( diff --git a/drivers/target/iscsi/iscsi_target_tpg.h b/drivers/target/iscsi/iscsi_target_tpg.h index 71d067f62177..d44d09f2dde9 100644 --- a/drivers/target/iscsi/iscsi_target_tpg.h +++ b/drivers/target/iscsi/iscsi_target_tpg.h @@ -24,12 +24,7 @@ extern int iscsit_tpg_del_portal_group(struct iscsi_tiqn *, struct iscsi_portal_ int); extern int iscsit_tpg_enable_portal_group(struct iscsi_portal_group *); extern int iscsit_tpg_disable_portal_group(struct iscsi_portal_group *, int); -extern struct iscsi_node_acl *iscsit_tpg_add_initiator_node_acl( - struct iscsi_portal_group *, const char *, u32); -extern void iscsit_tpg_del_initiator_node_acl(struct iscsi_portal_group *, - struct se_node_acl *); extern struct iscsi_node_attrib *iscsit_tpg_get_node_attrib(struct iscsit_session *); -extern void iscsit_tpg_del_external_nps(struct iscsi_tpg_np *); extern struct iscsi_tpg_np *iscsit_tpg_locate_child_np(struct iscsi_tpg_np *, int); extern struct iscsi_tpg_np *iscsit_tpg_add_network_portal(struct iscsi_portal_group *, struct sockaddr_storage *, struct iscsi_tpg_np *, diff --git a/drivers/target/iscsi/iscsi_target_util.h b/drivers/target/iscsi/iscsi_target_util.h index 24b8e577575a..336da4fb0a77 100644 --- a/drivers/target/iscsi/iscsi_target_util.h +++ b/drivers/target/iscsi/iscsi_target_util.h @@ -17,7 +17,6 @@ extern struct iscsi_r2t *iscsit_get_r2t_for_eos(struct iscsit_cmd *, u32, u32); extern struct iscsi_r2t *iscsit_get_r2t_from_list(struct iscsit_cmd *); extern void iscsit_free_r2t(struct iscsi_r2t *, struct iscsit_cmd *); extern void iscsit_free_r2ts_from_list(struct iscsit_cmd *); -extern struct iscsit_cmd *iscsit_alloc_cmd(struct iscsit_conn *, gfp_t); extern struct iscsit_cmd *iscsit_allocate_cmd(struct iscsit_conn *, int); extern struct iscsi_seq *iscsit_get_seq_holder_for_datain(struct iscsit_cmd *, u32); extern struct iscsi_seq *iscsit_get_seq_holder_for_r2t(struct iscsit_cmd *); @@ -34,7 +33,6 @@ extern void iscsit_add_cmd_to_immediate_queue(struct iscsit_cmd *, struct iscsit extern struct iscsi_queue_req *iscsit_get_cmd_from_immediate_queue(struct iscsit_conn *); extern int iscsit_add_cmd_to_response_queue(struct iscsit_cmd *, struct iscsit_conn *, u8); extern struct iscsi_queue_req *iscsit_get_cmd_from_response_queue(struct iscsit_conn *); -extern void iscsit_remove_cmd_from_tx_queues(struct iscsit_cmd *, struct iscsit_conn *); extern bool iscsit_conn_all_queues_empty(struct iscsit_conn *); extern void iscsit_free_queue_reqs_for_conn(struct iscsit_conn *); extern void iscsit_release_cmd(struct iscsit_cmd *); @@ -64,9 +62,6 @@ extern int iscsit_send_tx_data(struct iscsit_cmd *, struct iscsit_conn *, int); extern int iscsit_fe_sendpage_sg(struct iscsit_cmd *, struct iscsit_conn *); extern int iscsit_tx_login_rsp(struct iscsit_conn *, u8, u8); extern void iscsit_print_session_params(struct iscsit_session *); -extern int iscsit_print_dev_to_proc(char *, char **, off_t, int); -extern int iscsit_print_sessions_to_proc(char *, char **, off_t, int); -extern int iscsit_print_tpg_to_proc(char *, char **, off_t, int); extern int rx_data(struct iscsit_conn *, struct kvec *, int, int); extern int tx_data(struct iscsit_conn *, struct kvec *, int, int); extern void iscsit_collect_login_stats(struct iscsit_conn *, u8, u8); From b97c0741c7dccedec60524b596c4fa9d6a136523 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:05 -0700 Subject: [PATCH 063/123] scsi: Expand all create*_workqueue() invocations The workqueue maintainer wants to remove the create*_workqueue() macros because these macros always set the WQ_MEM_RECLAIM flag and because these only support literal workqueue names. Hence this patch that replaces the create*_workqueue() invocations with the definition of this macro. The WQ_MEM_RECLAIM flag has been retained because I think that flag is necessary for workqueues created by storage drivers. This patch has been generated by running spatch and git clang-format. spatch has been invoked as follows: spatch --in-place --sp-file expand-create-workqueue.spatch $(git grep -lEw 'create_(freezable_|singlethread_|)workqueue' */scsi */ufs) The contents of the expand-create-workqueue.spatch file is as follows: @@ expression name; @@ -create_workqueue(name) +alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, name) @@ expression name; @@ -create_freezable_workqueue(name) +alloc_workqueue("%s", WQ_FREEZABLE | WQ_UNBOUND | WQ_MEM_RECLAIM, 1, name) @@ expression name; @@ -create_singlethread_workqueue(name) +alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, name) Reviewed-by: Peter Wang Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-2-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/bfa/bfad_im.c | 3 ++- drivers/scsi/bnx2fc/bnx2fc_fcoe.c | 4 ++-- drivers/scsi/device_handler/scsi_dh_rdac.c | 3 ++- drivers/scsi/elx/efct/efct_lio.c | 3 ++- drivers/scsi/esas2r/esas2r_init.c | 3 ++- drivers/scsi/fcoe/fcoe_sysfs.c | 8 ++++---- drivers/scsi/fnic/fnic_main.c | 6 ++++-- drivers/scsi/hisi_sas/hisi_sas_main.c | 3 ++- drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 2 +- drivers/scsi/libfc/fc_exch.c | 3 ++- drivers/scsi/libfc/fc_rport.c | 3 ++- drivers/scsi/libsas/sas_init.c | 4 ++-- drivers/scsi/megaraid/megaraid_sas_fusion.c | 4 ++-- drivers/scsi/mpi3mr/mpi3mr_fw.c | 4 ++-- drivers/scsi/mpt3sas/mpt3sas_base.c | 4 ++-- drivers/scsi/myrb.c | 3 ++- drivers/scsi/myrs.c | 3 ++- drivers/scsi/qedf/qedf_main.c | 13 +++++++------ drivers/scsi/qedi/qedi_main.c | 6 ++++-- drivers/scsi/qla2xxx/qla_os.c | 6 ++++-- drivers/scsi/qla4xxx/ql4_os.c | 2 +- drivers/scsi/snic/snic_main.c | 6 ++++-- drivers/scsi/stex.c | 3 ++- drivers/scsi/vmw_pvscsi.c | 3 ++- drivers/ufs/core/ufshcd.c | 5 +++-- 25 files changed, 64 insertions(+), 43 deletions(-) diff --git a/drivers/scsi/bfa/bfad_im.c b/drivers/scsi/bfa/bfad_im.c index a9d3d8562d3c..a1d015356063 100644 --- a/drivers/scsi/bfa/bfad_im.c +++ b/drivers/scsi/bfa/bfad_im.c @@ -768,7 +768,8 @@ bfad_thread_workq(struct bfad_s *bfad) bfa_trc(bfad, 0); snprintf(im->drv_workq_name, KOBJ_NAME_LEN, "bfad_wq_%d", bfad->inst_no); - im->drv_workq = create_singlethread_workqueue(im->drv_workq_name); + im->drv_workq = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + im->drv_workq_name); if (!im->drv_workq) return BFA_STATUS_FAILED; diff --git a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c index 1078c20c5ef6..f49783b89d04 100644 --- a/drivers/scsi/bnx2fc/bnx2fc_fcoe.c +++ b/drivers/scsi/bnx2fc/bnx2fc_fcoe.c @@ -2363,8 +2363,8 @@ static int _bnx2fc_create(struct net_device *netdev, interface->vlan_id = vlan_id; interface->tm_timeout = BNX2FC_TM_TIMEOUT; - interface->timer_work_queue = - create_singlethread_workqueue("bnx2fc_timer_wq"); + interface->timer_work_queue = alloc_ordered_workqueue( + "%s", WQ_MEM_RECLAIM, "bnx2fc_timer_wq"); if (!interface->timer_work_queue) { printk(KERN_ERR PFX "ulp_init could not create timer_wq\n"); rc = -EINVAL; diff --git a/drivers/scsi/device_handler/scsi_dh_rdac.c b/drivers/scsi/device_handler/scsi_dh_rdac.c index f8a09e3eba58..6e1b252cea0e 100644 --- a/drivers/scsi/device_handler/scsi_dh_rdac.c +++ b/drivers/scsi/device_handler/scsi_dh_rdac.c @@ -822,7 +822,8 @@ static int __init rdac_init(void) /* * Create workqueue to handle mode selects for rdac */ - kmpath_rdacd = create_singlethread_workqueue("kmpath_rdacd"); + kmpath_rdacd = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, "kmpath_rdacd"); if (!kmpath_rdacd) { scsi_unregister_device_handler(&rdac_dh); printk(KERN_ERR "kmpath_rdacd creation failed.\n"); diff --git a/drivers/scsi/elx/efct/efct_lio.c b/drivers/scsi/elx/efct/efct_lio.c index 6a6ec32c46bd..9ac69356b13e 100644 --- a/drivers/scsi/elx/efct/efct_lio.c +++ b/drivers/scsi/elx/efct/efct_lio.c @@ -1114,7 +1114,8 @@ int efct_scsi_tgt_new_device(struct efct *efct) atomic_set(&efct->tgt_efct.watermark_hit, 0); atomic_set(&efct->tgt_efct.initiator_count, 0); - lio_wq = create_singlethread_workqueue("efct_lio_worker"); + lio_wq = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + "efct_lio_worker"); if (!lio_wq) { efc_log_err(efct, "workqueue create failed\n"); return -EIO; diff --git a/drivers/scsi/esas2r/esas2r_init.c b/drivers/scsi/esas2r/esas2r_init.c index c1a5ab662dc8..ff1fa3160c61 100644 --- a/drivers/scsi/esas2r/esas2r_init.c +++ b/drivers/scsi/esas2r/esas2r_init.c @@ -313,7 +313,8 @@ int esas2r_init_adapter(struct Scsi_Host *host, struct pci_dev *pcid, esas2r_fw_event_off(a); snprintf(a->fw_event_q_name, ESAS2R_KOBJ_NAME_LEN, "esas2r/%d", a->index); - a->fw_event_q = create_singlethread_workqueue(a->fw_event_q_name); + a->fw_event_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + a->fw_event_q_name); init_waitqueue_head(&a->buffered_ioctl_waiter); init_waitqueue_head(&a->nvram_waiter); diff --git a/drivers/scsi/fcoe/fcoe_sysfs.c b/drivers/scsi/fcoe/fcoe_sysfs.c index 7d3b904af9e8..06357bbf6b2c 100644 --- a/drivers/scsi/fcoe/fcoe_sysfs.c +++ b/drivers/scsi/fcoe/fcoe_sysfs.c @@ -799,16 +799,16 @@ struct fcoe_ctlr_device *fcoe_ctlr_device_add(struct device *parent, snprintf(ctlr->work_q_name, sizeof(ctlr->work_q_name), "ctlr_wq_%d", ctlr->id); - ctlr->work_q = create_singlethread_workqueue( - ctlr->work_q_name); + ctlr->work_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + ctlr->work_q_name); if (!ctlr->work_q) goto out_del; snprintf(ctlr->devloss_work_q_name, sizeof(ctlr->devloss_work_q_name), "ctlr_dl_wq_%d", ctlr->id); - ctlr->devloss_work_q = create_singlethread_workqueue( - ctlr->devloss_work_q_name); + ctlr->devloss_work_q = alloc_ordered_workqueue( + "%s", WQ_MEM_RECLAIM, ctlr->devloss_work_q_name); if (!ctlr->devloss_work_q) goto out_del_q; diff --git a/drivers/scsi/fnic/fnic_main.c b/drivers/scsi/fnic/fnic_main.c index 29eead383eb9..0044717d4486 100644 --- a/drivers/scsi/fnic/fnic_main.c +++ b/drivers/scsi/fnic/fnic_main.c @@ -1161,14 +1161,16 @@ static int __init fnic_init_module(void) goto err_create_fnic_ioreq_slab; } - fnic_event_queue = create_singlethread_workqueue("fnic_event_wq"); + fnic_event_queue = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, "fnic_event_wq"); if (!fnic_event_queue) { printk(KERN_ERR PFX "fnic work queue create failed\n"); err = -ENOMEM; goto err_create_fnic_workq; } - fnic_fip_queue = create_singlethread_workqueue("fnic_fip_q"); + fnic_fip_queue = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, "fnic_fip_q"); if (!fnic_fip_queue) { printk(KERN_ERR PFX "fnic FIP work queue create failed\n"); err = -ENOMEM; diff --git a/drivers/scsi/hisi_sas/hisi_sas_main.c b/drivers/scsi/hisi_sas/hisi_sas_main.c index ec1a3e7ee94d..6219807ce3b9 100644 --- a/drivers/scsi/hisi_sas/hisi_sas_main.c +++ b/drivers/scsi/hisi_sas/hisi_sas_main.c @@ -2302,7 +2302,8 @@ int hisi_sas_alloc(struct hisi_hba *hisi_hba) hisi_hba->last_slot_index = 0; - hisi_hba->wq = create_singlethread_workqueue(dev_name(dev)); + hisi_hba->wq = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, dev_name(dev)); if (!hisi_hba->wq) { dev_err(dev, "sas_alloc: failed to create workqueue\n"); goto err_out; diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c index 2fca17cf8b51..639f72f28911 100644 --- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c +++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c @@ -3537,7 +3537,7 @@ static int ibmvscsis_probe(struct vio_dev *vdev, init_completion(&vscsi->unconfig); snprintf(wq_name, 24, "ibmvscsis%s", dev_name(&vdev->dev)); - vscsi->work_q = create_workqueue(wq_name); + vscsi->work_q = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, wq_name); if (!vscsi->work_q) { rc = -ENOMEM; dev_err(&vscsi->dev, "create_workqueue failed\n"); diff --git a/drivers/scsi/libfc/fc_exch.c b/drivers/scsi/libfc/fc_exch.c index 1d91c457527f..f84a7e6ae379 100644 --- a/drivers/scsi/libfc/fc_exch.c +++ b/drivers/scsi/libfc/fc_exch.c @@ -2693,7 +2693,8 @@ int fc_setup_exch_mgr(void) fc_cpu_order = ilog2(roundup_pow_of_two(nr_cpu_ids)); fc_cpu_mask = (1 << fc_cpu_order) - 1; - fc_exch_workqueue = create_singlethread_workqueue("fc_exch_workqueue"); + fc_exch_workqueue = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + "fc_exch_workqueue"); if (!fc_exch_workqueue) goto err; return 0; diff --git a/drivers/scsi/libfc/fc_rport.c b/drivers/scsi/libfc/fc_rport.c index 33da3c1085f0..308cb4872f96 100644 --- a/drivers/scsi/libfc/fc_rport.c +++ b/drivers/scsi/libfc/fc_rport.c @@ -2263,7 +2263,8 @@ struct fc4_prov fc_rport_t0_prov = { */ int fc_setup_rport(void) { - rport_event_queue = create_singlethread_workqueue("fc_rport_eq"); + rport_event_queue = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, "fc_rport_eq"); if (!rport_event_queue) return -ENOMEM; return 0; diff --git a/drivers/scsi/libsas/sas_init.c b/drivers/scsi/libsas/sas_init.c index 9c8cc723170d..8566bb1208a0 100644 --- a/drivers/scsi/libsas/sas_init.c +++ b/drivers/scsi/libsas/sas_init.c @@ -122,12 +122,12 @@ int sas_register_ha(struct sas_ha_struct *sas_ha) error = -ENOMEM; snprintf(name, sizeof(name), "%s_event_q", dev_name(sas_ha->dev)); - sas_ha->event_q = create_singlethread_workqueue(name); + sas_ha->event_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, name); if (!sas_ha->event_q) goto Undo_ports; snprintf(name, sizeof(name), "%s_disco_q", dev_name(sas_ha->dev)); - sas_ha->disco_q = create_singlethread_workqueue(name); + sas_ha->disco_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, name); if (!sas_ha->disco_q) goto Undo_event_q; diff --git a/drivers/scsi/megaraid/megaraid_sas_fusion.c b/drivers/scsi/megaraid/megaraid_sas_fusion.c index 6c1fb8149553..1eec23da28e2 100644 --- a/drivers/scsi/megaraid/megaraid_sas_fusion.c +++ b/drivers/scsi/megaraid/megaraid_sas_fusion.c @@ -1988,8 +1988,8 @@ megasas_fusion_start_watchdog(struct megasas_instance *instance) sizeof(instance->fault_handler_work_q_name), "poll_megasas%d_status", instance->host->host_no); - instance->fw_fault_work_q = - create_singlethread_workqueue(instance->fault_handler_work_q_name); + instance->fw_fault_work_q = alloc_ordered_workqueue( + "%s", WQ_MEM_RECLAIM, instance->fault_handler_work_q_name); if (!instance->fw_fault_work_q) { dev_err(&instance->pdev->dev, "Failed from %s %d\n", __func__, __LINE__); diff --git a/drivers/scsi/mpi3mr/mpi3mr_fw.c b/drivers/scsi/mpi3mr/mpi3mr_fw.c index c196dc14ad20..7c739468dca5 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_fw.c +++ b/drivers/scsi/mpi3mr/mpi3mr_fw.c @@ -2742,8 +2742,8 @@ void mpi3mr_start_watchdog(struct mpi3mr_ioc *mrioc) snprintf(mrioc->watchdog_work_q_name, sizeof(mrioc->watchdog_work_q_name), "watchdog_%s%d", mrioc->name, mrioc->id); - mrioc->watchdog_work_q = - create_singlethread_workqueue(mrioc->watchdog_work_q_name); + mrioc->watchdog_work_q = alloc_ordered_workqueue( + "%s", WQ_MEM_RECLAIM, mrioc->watchdog_work_q_name); if (!mrioc->watchdog_work_q) { ioc_err(mrioc, "%s: failed (line=%d)\n", __func__, __LINE__); return; diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c b/drivers/scsi/mpt3sas/mpt3sas_base.c index b2bcf4a27ddc..2d3eeda5a6a0 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_base.c +++ b/drivers/scsi/mpt3sas/mpt3sas_base.c @@ -846,8 +846,8 @@ mpt3sas_base_start_watchdog(struct MPT3SAS_ADAPTER *ioc) snprintf(ioc->fault_reset_work_q_name, sizeof(ioc->fault_reset_work_q_name), "poll_%s%d_status", ioc->driver_name, ioc->id); - ioc->fault_reset_work_q = - create_singlethread_workqueue(ioc->fault_reset_work_q_name); + ioc->fault_reset_work_q = alloc_ordered_workqueue( + "%s", WQ_MEM_RECLAIM, ioc->fault_reset_work_q_name); if (!ioc->fault_reset_work_q) { ioc_err(ioc, "%s: failed (line=%d)\n", __func__, __LINE__); return; diff --git a/drivers/scsi/myrb.c b/drivers/scsi/myrb.c index f684eb5e0489..140dc0e9cead 100644 --- a/drivers/scsi/myrb.c +++ b/drivers/scsi/myrb.c @@ -114,7 +114,8 @@ static bool myrb_create_mempools(struct pci_dev *pdev, struct myrb_hba *cb) snprintf(cb->work_q_name, sizeof(cb->work_q_name), "myrb_wq_%d", cb->host->host_no); - cb->work_q = create_singlethread_workqueue(cb->work_q_name); + cb->work_q = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, cb->work_q_name); if (!cb->work_q) { dma_pool_destroy(cb->dcdb_pool); cb->dcdb_pool = NULL; diff --git a/drivers/scsi/myrs.c b/drivers/scsi/myrs.c index e824be9d9bbb..8a8f26633cda 100644 --- a/drivers/scsi/myrs.c +++ b/drivers/scsi/myrs.c @@ -2208,7 +2208,8 @@ static bool myrs_create_mempools(struct pci_dev *pdev, struct myrs_hba *cs) snprintf(cs->work_q_name, sizeof(cs->work_q_name), "myrs_wq_%d", shost->host_no); - cs->work_q = create_singlethread_workqueue(cs->work_q_name); + cs->work_q = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, cs->work_q_name); if (!cs->work_q) { dma_pool_destroy(cs->dcdb_pool); cs->dcdb_pool = NULL; diff --git a/drivers/scsi/qedf/qedf_main.c b/drivers/scsi/qedf/qedf_main.c index 4813087e58a1..119afcaf6e13 100644 --- a/drivers/scsi/qedf/qedf_main.c +++ b/drivers/scsi/qedf/qedf_main.c @@ -3374,7 +3374,8 @@ retry_probe: sprintf(host_buf, "qedf_%u_link", qedf->lport->host->host_no); - qedf->link_update_wq = create_workqueue(host_buf); + qedf->link_update_wq = + alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); INIT_DELAYED_WORK(&qedf->link_update, qedf_handle_link_update); INIT_DELAYED_WORK(&qedf->link_recovery, qedf_link_recovery); INIT_DELAYED_WORK(&qedf->grcdump_work, qedf_wq_grcdump); @@ -3585,8 +3586,7 @@ retry_probe: /* Start LL2 processing thread */ snprintf(host_buf, 20, "qedf_%d_ll2", host->host_no); - qedf->ll2_recv_wq = - create_workqueue(host_buf); + qedf->ll2_recv_wq = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); if (!qedf->ll2_recv_wq) { QEDF_ERR(&(qedf->dbg_ctx), "Failed to LL2 workqueue.\n"); rc = -ENOMEM; @@ -3629,7 +3629,7 @@ retry_probe: sprintf(host_buf, "qedf_%u_timer", qedf->lport->host->host_no); qedf->timer_work_queue = - create_workqueue(host_buf); + alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); if (!qedf->timer_work_queue) { QEDF_ERR(&(qedf->dbg_ctx), "Failed to start timer " "workqueue.\n"); @@ -3641,7 +3641,8 @@ retry_probe: if (mode != QEDF_MODE_RECOVERY) { sprintf(host_buf, "qedf_%u_dpc", qedf->lport->host->host_no); - qedf->dpc_wq = create_workqueue(host_buf); + qedf->dpc_wq = + alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); } INIT_DELAYED_WORK(&qedf->recovery_work, qedf_recovery_handler); @@ -4182,7 +4183,7 @@ static int __init qedf_init(void) goto err3; } - qedf_io_wq = create_workqueue("qedf_io_wq"); + qedf_io_wq = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, "qedf_io_wq"); if (!qedf_io_wq) { QEDF_ERR(NULL, "Could not create qedf_io_wq.\n"); goto err4; diff --git a/drivers/scsi/qedi/qedi_main.c b/drivers/scsi/qedi/qedi_main.c index cd0180b1f5b9..319c1da549f7 100644 --- a/drivers/scsi/qedi/qedi_main.c +++ b/drivers/scsi/qedi/qedi_main.c @@ -2767,7 +2767,8 @@ retry_probe: } sprintf(host_buf, "host_%d", qedi->shost->host_no); - qedi->tmf_thread = create_singlethread_workqueue(host_buf); + qedi->tmf_thread = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, host_buf); if (!qedi->tmf_thread) { QEDI_ERR(&qedi->dbg_ctx, "Unable to start tmf thread!\n"); @@ -2776,7 +2777,8 @@ retry_probe: } sprintf(host_buf, "qedi_ofld%d", qedi->shost->host_no); - qedi->offload_thread = create_workqueue(host_buf); + qedi->offload_thread = + alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); if (!qedi->offload_thread) { QEDI_ERR(&qedi->dbg_ctx, "Unable to start offload thread!\n"); diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c index bc3b2aea3f8b..7f980e6141c2 100644 --- a/drivers/scsi/qla2xxx/qla_os.c +++ b/drivers/scsi/qla2xxx/qla_os.c @@ -3501,11 +3501,13 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct pci_device_id *id) if (IS_QLA8031(ha) || IS_MCTP_CAPABLE(ha)) { sprintf(wq_name, "qla2xxx_%lu_dpc_lp_wq", base_vha->host_no); - ha->dpc_lp_wq = create_singlethread_workqueue(wq_name); + ha->dpc_lp_wq = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, wq_name); INIT_WORK(&ha->idc_aen, qla83xx_service_idc_aen); sprintf(wq_name, "qla2xxx_%lu_dpc_hp_wq", base_vha->host_no); - ha->dpc_hp_wq = create_singlethread_workqueue(wq_name); + ha->dpc_hp_wq = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, wq_name); INIT_WORK(&ha->nic_core_reset, qla83xx_nic_core_reset_work); INIT_WORK(&ha->idc_state_handler, qla83xx_idc_state_handler_work); diff --git a/drivers/scsi/qla4xxx/ql4_os.c b/drivers/scsi/qla4xxx/ql4_os.c index 17cccd14765f..d91f54a6e752 100644 --- a/drivers/scsi/qla4xxx/ql4_os.c +++ b/drivers/scsi/qla4xxx/ql4_os.c @@ -8806,7 +8806,7 @@ skip_retry_init: DEBUG2(printk("scsi: %s: Starting kernel thread for " "qla4xxx_dpc\n", __func__)); sprintf(buf, "qla4xxx_%lu_dpc", ha->host_no); - ha->dpc_thread = create_singlethread_workqueue(buf); + ha->dpc_thread = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, buf); if (!ha->dpc_thread) { ql4_printk(KERN_WARNING, ha, "Unable to start DPC thread!\n"); ret = -ENODEV; diff --git a/drivers/scsi/snic/snic_main.c b/drivers/scsi/snic/snic_main.c index cc824dcfe7da..2bd01eb57869 100644 --- a/drivers/scsi/snic/snic_main.c +++ b/drivers/scsi/snic/snic_main.c @@ -302,7 +302,8 @@ snic_add_host(struct Scsi_Host *shost, struct pci_dev *pdev) SNIC_BUG_ON(shost->work_q != NULL); snprintf(shost->work_q_name, sizeof(shost->work_q_name), "scsi_wq_%d", shost->host_no); - shost->work_q = create_singlethread_workqueue(shost->work_q_name); + shost->work_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, + shost->work_q_name); if (!shost->work_q) { SNIC_HOST_ERR(shost, "Failed to Create ScsiHost wq.\n"); @@ -884,7 +885,8 @@ snic_global_data_init(void) snic_glob->req_cache[SNIC_REQ_TM_CACHE] = cachep; /* snic_event queue */ - snic_glob->event_q = create_singlethread_workqueue("snic_event_wq"); + snic_glob->event_q = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, "snic_event_wq"); if (!snic_glob->event_q) { SNIC_ERR("snic event queue create failed\n"); ret = -ENOMEM; diff --git a/drivers/scsi/stex.c b/drivers/scsi/stex.c index 8ffb75be99bc..fbee7db4a835 100644 --- a/drivers/scsi/stex.c +++ b/drivers/scsi/stex.c @@ -1797,7 +1797,8 @@ static int stex_probe(struct pci_dev *pdev, const struct pci_device_id *id) snprintf(hba->work_q_name, sizeof(hba->work_q_name), "stex_wq_%d", host->host_no); - hba->work_q = create_singlethread_workqueue(hba->work_q_name); + hba->work_q = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, hba->work_q_name); if (!hba->work_q) { printk(KERN_ERR DRV_NAME "(%s): create workqueue failed\n", pci_name(pdev)); diff --git a/drivers/scsi/vmw_pvscsi.c b/drivers/scsi/vmw_pvscsi.c index c4fea077265e..32242d86cf5b 100644 --- a/drivers/scsi/vmw_pvscsi.c +++ b/drivers/scsi/vmw_pvscsi.c @@ -1137,7 +1137,8 @@ static int pvscsi_setup_msg_workqueue(struct pvscsi_adapter *adapter) snprintf(name, sizeof(name), "vmw_pvscsi_wq_%u", adapter->host->host_no); - adapter->workqueue = create_singlethread_workqueue(name); + adapter->workqueue = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, name); if (!adapter->workqueue) { printk(KERN_ERR "vmw_pvscsi: failed to create work queue\n"); return 0; diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index dc757ba47522..930b15d9356b 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -1800,7 +1800,8 @@ static void ufshcd_init_clk_scaling(struct ufs_hba *hba) snprintf(wq_name, sizeof(wq_name), "ufs_clkscaling_%d", hba->host->host_no); - hba->clk_scaling.workq = create_singlethread_workqueue(wq_name); + hba->clk_scaling.workq = + alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, wq_name); hba->clk_scaling.is_initialized = true; } @@ -10444,7 +10445,7 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq) /* Initialize work queues */ snprintf(eh_wq_name, sizeof(eh_wq_name), "ufs_eh_wq_%d", hba->host->host_no); - hba->eh_wq = create_singlethread_workqueue(eh_wq_name); + hba->eh_wq = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, eh_wq_name); if (!hba->eh_wq) { dev_err(hba->dev, "%s: failed to create eh workqueue\n", __func__); From dec523975b85e725e2bf786a0b7dcee76b77a64e Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:06 -0700 Subject: [PATCH 064/123] scsi: mptfusion: Simplify the alloc*_workqueue() invocations Let alloc*_workqueue() format the workqueue names instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-3-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/message/fusion/mptbase.c | 10 +++------- drivers/message/fusion/mptbase.h | 3 --- drivers/message/fusion/mptfc.c | 7 ++----- 3 files changed, 5 insertions(+), 15 deletions(-) diff --git a/drivers/message/fusion/mptbase.c b/drivers/message/fusion/mptbase.c index 4bf669c55649..738bc4e60a18 100644 --- a/drivers/message/fusion/mptbase.c +++ b/drivers/message/fusion/mptbase.c @@ -1856,10 +1856,8 @@ mpt_attach(struct pci_dev *pdev, const struct pci_device_id *id) /* Initialize workqueue */ INIT_DELAYED_WORK(&ioc->fault_reset_work, mpt_fault_reset_work); - snprintf(ioc->reset_work_q_name, MPT_KOBJ_NAME_LEN, - "mpt_poll_%d", ioc->id); - ioc->reset_work_q = alloc_workqueue(ioc->reset_work_q_name, - WQ_MEM_RECLAIM, 0); + ioc->reset_work_q = + alloc_workqueue("mpt_poll_%d", WQ_MEM_RECLAIM, 0, ioc->id); if (!ioc->reset_work_q) { printk(MYIOC_s_ERR_FMT "Insufficient memory to add adapter!\n", ioc->name); @@ -1986,9 +1984,7 @@ mpt_attach(struct pci_dev *pdev, const struct pci_device_id *id) INIT_LIST_HEAD(&ioc->fw_event_list); spin_lock_init(&ioc->fw_event_lock); - snprintf(ioc->fw_event_q_name, MPT_KOBJ_NAME_LEN, "mpt/%d", ioc->id); - ioc->fw_event_q = alloc_workqueue(ioc->fw_event_q_name, - WQ_MEM_RECLAIM, 0); + ioc->fw_event_q = alloc_workqueue("mpt/%d", WQ_MEM_RECLAIM, 0, ioc->id); if (!ioc->fw_event_q) { printk(MYIOC_s_ERR_FMT "Insufficient memory to add adapter!\n", ioc->name); diff --git a/drivers/message/fusion/mptbase.h b/drivers/message/fusion/mptbase.h index 8031173c3655..b406fd676da0 100644 --- a/drivers/message/fusion/mptbase.h +++ b/drivers/message/fusion/mptbase.h @@ -729,7 +729,6 @@ typedef struct _MPT_ADAPTER struct list_head fw_event_list; spinlock_t fw_event_lock; u8 fw_events_off; /* if '1', then ignore events */ - char fw_event_q_name[MPT_KOBJ_NAME_LEN]; struct mutex sas_discovery_mutex; u8 sas_discovery_runtime; @@ -764,7 +763,6 @@ typedef struct _MPT_ADAPTER u8 fc_link_speed[2]; spinlock_t fc_rescan_work_lock; struct work_struct fc_rescan_work; - char fc_rescan_work_q_name[MPT_KOBJ_NAME_LEN]; struct workqueue_struct *fc_rescan_work_q; /* driver forced bus resets count */ @@ -778,7 +776,6 @@ typedef struct _MPT_ADAPTER spinlock_t scsi_lookup_lock; u64 dma_mask; u32 broadcast_aen_busy; - char reset_work_q_name[MPT_KOBJ_NAME_LEN]; struct workqueue_struct *reset_work_q; struct delayed_work fault_reset_work; diff --git a/drivers/message/fusion/mptfc.c b/drivers/message/fusion/mptfc.c index a3c17c4fe69c..91242f26defb 100644 --- a/drivers/message/fusion/mptfc.c +++ b/drivers/message/fusion/mptfc.c @@ -1349,11 +1349,8 @@ mptfc_probe(struct pci_dev *pdev, const struct pci_device_id *id) /* initialize workqueue */ - snprintf(ioc->fc_rescan_work_q_name, sizeof(ioc->fc_rescan_work_q_name), - "mptfc_wq_%d", sh->host_no); - ioc->fc_rescan_work_q = - alloc_ordered_workqueue(ioc->fc_rescan_work_q_name, - WQ_MEM_RECLAIM); + ioc->fc_rescan_work_q = alloc_ordered_workqueue( + "mptfc_wq_%d", WQ_MEM_RECLAIM, sh->host_no); if (!ioc->fc_rescan_work_q) { error = -ENOMEM; goto out_mptfc_host; From 66088e7b918275a4520f5d7939e7cf9d495e3b83 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:07 -0700 Subject: [PATCH 065/123] scsi: be2iscsi: Simplify an alloc_workqueue() invocation Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-4-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/be2iscsi/be_main.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/scsi/be2iscsi/be_main.c b/drivers/scsi/be2iscsi/be_main.c index 06acb5ff609e..76a1e373386e 100644 --- a/drivers/scsi/be2iscsi/be_main.c +++ b/drivers/scsi/be2iscsi/be_main.c @@ -5528,7 +5528,6 @@ static int beiscsi_dev_probe(struct pci_dev *pcidev, struct beiscsi_hba *phba = NULL; struct be_eq_obj *pbe_eq; unsigned int s_handle; - char wq_name[20]; int ret, i; ret = beiscsi_enable_pci(pcidev); @@ -5634,9 +5633,8 @@ static int beiscsi_dev_probe(struct pci_dev *pcidev, phba->ctrl.mcc_alloc_index = phba->ctrl.mcc_free_index = 0; - snprintf(wq_name, sizeof(wq_name), "beiscsi_%02x_wq", - phba->shost->host_no); - phba->wq = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, wq_name); + phba->wq = alloc_workqueue("beiscsi_%02x_wq", WQ_MEM_RECLAIM, 1, + phba->shost->host_no); if (!phba->wq) { beiscsi_log(phba, KERN_ERR, BEISCSI_LOG_INIT, "BM_%d : beiscsi_dev_probe-" From 70fbb7c11507576011511dbf4d00940ebb3a09fb Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:08 -0700 Subject: [PATCH 066/123] scsi: bfa: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-5-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/bfa/bfad_im.c | 6 ++---- drivers/scsi/bfa/bfad_im.h | 1 - 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/bfa/bfad_im.c b/drivers/scsi/bfa/bfad_im.c index a1d015356063..66fb701401de 100644 --- a/drivers/scsi/bfa/bfad_im.c +++ b/drivers/scsi/bfa/bfad_im.c @@ -766,10 +766,8 @@ bfad_thread_workq(struct bfad_s *bfad) struct bfad_im_s *im = bfad->im; bfa_trc(bfad, 0); - snprintf(im->drv_workq_name, KOBJ_NAME_LEN, "bfad_wq_%d", - bfad->inst_no); - im->drv_workq = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, - im->drv_workq_name); + im->drv_workq = alloc_ordered_workqueue("bfad_wq_%d", WQ_MEM_RECLAIM, + bfad->inst_no); if (!im->drv_workq) return BFA_STATUS_FAILED; diff --git a/drivers/scsi/bfa/bfad_im.h b/drivers/scsi/bfa/bfad_im.h index 4353feedf76a..0884af04bd1f 100644 --- a/drivers/scsi/bfa/bfad_im.h +++ b/drivers/scsi/bfa/bfad_im.h @@ -134,7 +134,6 @@ struct bfad_fcp_binding { struct bfad_im_s { struct bfad_s *bfad; struct workqueue_struct *drv_workq; - char drv_workq_name[KOBJ_NAME_LEN]; struct work_struct aen_im_notify_work; }; From 4cb1b41a5ee4c513897441bda42d56673b0c6edb Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:09 -0700 Subject: [PATCH 067/123] scsi: esas2r: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-6-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/esas2r/esas2r.h | 1 - drivers/scsi/esas2r/esas2r_init.c | 6 ++---- 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/esas2r/esas2r.h b/drivers/scsi/esas2r/esas2r.h index ed63f7a9ea54..8a133254c4f6 100644 --- a/drivers/scsi/esas2r/esas2r.h +++ b/drivers/scsi/esas2r/esas2r.h @@ -929,7 +929,6 @@ struct esas2r_adapter { struct list_head fw_event_list; spinlock_t fw_event_lock; u8 fw_events_off; /* if '1', then ignore events */ - char fw_event_q_name[ESAS2R_KOBJ_NAME_LEN]; /* * intr_mode stores the interrupt mode currently being used by this * adapter. it is based on the interrupt_mode module parameter, but diff --git a/drivers/scsi/esas2r/esas2r_init.c b/drivers/scsi/esas2r/esas2r_init.c index ff1fa3160c61..0cea5f3d1a08 100644 --- a/drivers/scsi/esas2r/esas2r_init.c +++ b/drivers/scsi/esas2r/esas2r_init.c @@ -311,10 +311,8 @@ int esas2r_init_adapter(struct Scsi_Host *host, struct pci_dev *pcid, sema_init(&a->nvram_semaphore, 1); esas2r_fw_event_off(a); - snprintf(a->fw_event_q_name, ESAS2R_KOBJ_NAME_LEN, "esas2r/%d", - a->index); - a->fw_event_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, - a->fw_event_q_name); + a->fw_event_q = + alloc_ordered_workqueue("esas2r/%d", WQ_MEM_RECLAIM, a->index); init_waitqueue_head(&a->buffered_ioctl_waiter); init_waitqueue_head(&a->nvram_waiter); From d77381c2f62a557f630a64280ff09675128be363 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:10 -0700 Subject: [PATCH 068/123] scsi: fcoe: Simplify alloc_ordered_workqueue() invocations Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Reviewed-by: Hannes Reinecke Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-7-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/fcoe/fcoe_sysfs.c | 18 +++++------------- include/scsi/fcoe_sysfs.h | 2 -- 2 files changed, 5 insertions(+), 15 deletions(-) diff --git a/drivers/scsi/fcoe/fcoe_sysfs.c b/drivers/scsi/fcoe/fcoe_sysfs.c index 06357bbf6b2c..0609ca6b9353 100644 --- a/drivers/scsi/fcoe/fcoe_sysfs.c +++ b/drivers/scsi/fcoe/fcoe_sysfs.c @@ -45,12 +45,8 @@ MODULE_PARM_DESC(fcf_dev_loss_tmo, */ #define fcoe_ctlr_id(x) \ ((x)->id) -#define fcoe_ctlr_work_q_name(x) \ - ((x)->work_q_name) #define fcoe_ctlr_work_q(x) \ ((x)->work_q) -#define fcoe_ctlr_devloss_work_q_name(x) \ - ((x)->devloss_work_q_name) #define fcoe_ctlr_devloss_work_q(x) \ ((x)->devloss_work_q) #define fcoe_ctlr_mode(x) \ @@ -797,18 +793,14 @@ struct fcoe_ctlr_device *fcoe_ctlr_device_add(struct device *parent, ctlr->fcf_dev_loss_tmo = fcoe_fcf_dev_loss_tmo; - snprintf(ctlr->work_q_name, sizeof(ctlr->work_q_name), - "ctlr_wq_%d", ctlr->id); - ctlr->work_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, - ctlr->work_q_name); + ctlr->work_q = alloc_ordered_workqueue("ctlr_wq_%d", WQ_MEM_RECLAIM, + ctlr->id); if (!ctlr->work_q) goto out_del; - snprintf(ctlr->devloss_work_q_name, - sizeof(ctlr->devloss_work_q_name), - "ctlr_dl_wq_%d", ctlr->id); - ctlr->devloss_work_q = alloc_ordered_workqueue( - "%s", WQ_MEM_RECLAIM, ctlr->devloss_work_q_name); + ctlr->devloss_work_q = alloc_ordered_workqueue("ctlr_dl_wq_%d", + WQ_MEM_RECLAIM, + ctlr->id); if (!ctlr->devloss_work_q) goto out_del_q; diff --git a/include/scsi/fcoe_sysfs.h b/include/scsi/fcoe_sysfs.h index 4b1216de3f22..2b28a05e492b 100644 --- a/include/scsi/fcoe_sysfs.h +++ b/include/scsi/fcoe_sysfs.h @@ -50,9 +50,7 @@ struct fcoe_ctlr_device { struct fcoe_sysfs_function_template *f; struct list_head fcfs; - char work_q_name[20]; struct workqueue_struct *work_q; - char devloss_work_q_name[20]; struct workqueue_struct *devloss_work_q; struct mutex lock; From 5615cfb3cbadf7c7159c5de18e6f02ca9c6630fc Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:11 -0700 Subject: [PATCH 069/123] scsi: ibmvscsi_tgt: Simplify an alloc_workqueue() invocation Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-8-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c index 639f72f28911..16d085d56e9d 100644 --- a/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c +++ b/drivers/scsi/ibmvscsi_tgt/ibmvscsi_tgt.c @@ -3425,7 +3425,6 @@ static int ibmvscsis_probe(struct vio_dev *vdev, struct scsi_info *vscsi; int rc = 0; long hrc = 0; - char wq_name[24]; vscsi = kzalloc(sizeof(*vscsi), GFP_KERNEL); if (!vscsi) { @@ -3536,8 +3535,8 @@ static int ibmvscsis_probe(struct vio_dev *vdev, init_completion(&vscsi->wait_idle); init_completion(&vscsi->unconfig); - snprintf(wq_name, 24, "ibmvscsis%s", dev_name(&vdev->dev)); - vscsi->work_q = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, wq_name); + vscsi->work_q = alloc_workqueue("ibmvscsis%s", WQ_MEM_RECLAIM, 1, + dev_name(&vdev->dev)); if (!vscsi->work_q) { rc = -ENOMEM; dev_err(&vscsi->dev, "create_workqueue failed\n"); From 1bd289620e4240ebf882e5e213ea0af200501be9 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:12 -0700 Subject: [PATCH 070/123] scsi: mpi3mr: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-9-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mpi3mr/mpi3mr.h | 2 -- drivers/scsi/mpi3mr/mpi3mr_os.c | 4 +--- 2 files changed, 1 insertion(+), 5 deletions(-) diff --git a/drivers/scsi/mpi3mr/mpi3mr.h b/drivers/scsi/mpi3mr/mpi3mr.h index dc2cdd5f0311..c1c97ed1eb38 100644 --- a/drivers/scsi/mpi3mr/mpi3mr.h +++ b/drivers/scsi/mpi3mr/mpi3mr.h @@ -1059,7 +1059,6 @@ struct scmd_priv { * @sbq_lock: Sense buffer queue lock * @sbq_host_index: Sense buffer queuehost index * @event_masks: Event mask bitmap - * @fwevt_worker_name: Firmware event worker thread name * @fwevt_worker_thread: Firmware event worker thread * @fwevt_lock: Firmware event lock * @fwevt_list: Firmware event list @@ -1240,7 +1239,6 @@ struct mpi3mr_ioc { u32 sbq_host_index; u32 event_masks[MPI3_EVENT_NOTIFY_EVENTMASK_WORDS]; - char fwevt_worker_name[MPI3MR_NAME_LENGTH]; struct workqueue_struct *fwevt_worker_thread; spinlock_t fwevt_lock; struct list_head fwevt_list; diff --git a/drivers/scsi/mpi3mr/mpi3mr_os.c b/drivers/scsi/mpi3mr/mpi3mr_os.c index 69b14918de59..e18c44b53b99 100644 --- a/drivers/scsi/mpi3mr/mpi3mr_os.c +++ b/drivers/scsi/mpi3mr/mpi3mr_os.c @@ -5305,10 +5305,8 @@ mpi3mr_probe(struct pci_dev *pdev, const struct pci_device_id *id) else scsi_host_set_guard(shost, SHOST_DIX_GUARD_CRC); - snprintf(mrioc->fwevt_worker_name, sizeof(mrioc->fwevt_worker_name), - "%s%d_fwevt_wrkr", mrioc->driver_name, mrioc->id); mrioc->fwevt_worker_thread = alloc_ordered_workqueue( - mrioc->fwevt_worker_name, 0); + "%s%d_fwevt_wrkr", 0, mrioc->driver_name, mrioc->id); if (!mrioc->fwevt_worker_thread) { ioc_err(mrioc, "failure at %s:%d/%s()!\n", __FILE__, __LINE__, __func__); From b3b359ac726763ead9cee1c3674f5c50d26651c4 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:13 -0700 Subject: [PATCH 071/123] scsi: mpt3sas: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-10-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/mpt3sas/mpt3sas_base.h | 4 +--- drivers/scsi/mpt3sas/mpt3sas_scsih.c | 4 +--- 2 files changed, 2 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h b/drivers/scsi/mpt3sas/mpt3sas_base.h index fe1e96fda284..eceb5eeb4651 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_base.h +++ b/drivers/scsi/mpt3sas/mpt3sas_base.h @@ -1162,8 +1162,7 @@ typedef void (*MPT3SAS_FLUSH_RUNNING_CMDS)(struct MPT3SAS_ADAPTER *ioc); * @fault_reset_work_q_name: fw fault work queue * @fault_reset_work_q: "" * @fault_reset_work: "" - * @firmware_event_name: fw event work queue - * @firmware_event_thread: "" + * @firmware_event_thread: fw event work queue * @fw_event_lock: * @fw_event_list: list of fw events * @current_evet: current processing firmware event @@ -1351,7 +1350,6 @@ struct MPT3SAS_ADAPTER { struct delayed_work fault_reset_work; /* fw event handler */ - char firmware_event_name[20]; struct workqueue_struct *firmware_event_thread; spinlock_t fw_event_lock; struct list_head fw_event_list; diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c index 97c2472cd434..728cced42b0e 100644 --- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c +++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c @@ -12301,10 +12301,8 @@ _scsih_probe(struct pci_dev *pdev, const struct pci_device_id *id) scsi_host_set_guard(shost, SHOST_DIX_GUARD_CRC); /* event thread */ - snprintf(ioc->firmware_event_name, sizeof(ioc->firmware_event_name), - "fw_event_%s%d", ioc->driver_name, ioc->id); ioc->firmware_event_thread = alloc_ordered_workqueue( - ioc->firmware_event_name, 0); + "fw_event_%s%d", 0, ioc->driver_name, ioc->id); if (!ioc->firmware_event_thread) { ioc_err(ioc, "failure at %s:%d/%s()!\n", __FILE__, __LINE__, __func__); From c57a617fd53fae6fcfa7d1737adfee209b8051ca Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:14 -0700 Subject: [PATCH 072/123] scsi: myrb: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Reviewed-by: Hannes Reinecke Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-11-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/myrb.c | 6 ++---- drivers/scsi/myrb.h | 1 - 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/myrb.c b/drivers/scsi/myrb.c index 140dc0e9cead..bfc2b835e612 100644 --- a/drivers/scsi/myrb.c +++ b/drivers/scsi/myrb.c @@ -112,10 +112,8 @@ static bool myrb_create_mempools(struct pci_dev *pdev, struct myrb_hba *cb) return false; } - snprintf(cb->work_q_name, sizeof(cb->work_q_name), - "myrb_wq_%d", cb->host->host_no); - cb->work_q = - alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, cb->work_q_name); + cb->work_q = alloc_ordered_workqueue("myrb_wq_%d", WQ_MEM_RECLAIM, + cb->host->host_no); if (!cb->work_q) { dma_pool_destroy(cb->dcdb_pool); cb->dcdb_pool = NULL; diff --git a/drivers/scsi/myrb.h b/drivers/scsi/myrb.h index fb8eacfceee8..78dc4136fb10 100644 --- a/drivers/scsi/myrb.h +++ b/drivers/scsi/myrb.h @@ -712,7 +712,6 @@ struct myrb_hba { struct Scsi_Host *host; struct workqueue_struct *work_q; - char work_q_name[20]; struct delayed_work monitor_work; unsigned long primary_monitor_time; unsigned long secondary_monitor_time; From f30679166255148f7781b0726d8e1b5e22dd5b48 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:15 -0700 Subject: [PATCH 073/123] scsi: myrs: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Reviewed-by: Hannes Reinecke Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-12-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/myrs.c | 6 ++---- drivers/scsi/myrs.h | 1 - 2 files changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/myrs.c b/drivers/scsi/myrs.c index 8a8f26633cda..3392feb15cb4 100644 --- a/drivers/scsi/myrs.c +++ b/drivers/scsi/myrs.c @@ -2206,10 +2206,8 @@ static bool myrs_create_mempools(struct pci_dev *pdev, struct myrs_hba *cs) return false; } - snprintf(cs->work_q_name, sizeof(cs->work_q_name), - "myrs_wq_%d", shost->host_no); - cs->work_q = - alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, cs->work_q_name); + cs->work_q = alloc_ordered_workqueue("myrs_wq_%d", WQ_MEM_RECLAIM, + shost->host_no); if (!cs->work_q) { dma_pool_destroy(cs->dcdb_pool); cs->dcdb_pool = NULL; diff --git a/drivers/scsi/myrs.h b/drivers/scsi/myrs.h index 9f6696d0ddd5..e1d6b123de7b 100644 --- a/drivers/scsi/myrs.h +++ b/drivers/scsi/myrs.h @@ -904,7 +904,6 @@ struct myrs_hba { bool disable_enc_msg; struct workqueue_struct *work_q; - char work_q_name[20]; struct delayed_work monitor_work; unsigned long primary_monitor_time; unsigned long secondary_monitor_time; From 8bbe60bbd43daa215106d263538e0822c4bba172 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:16 -0700 Subject: [PATCH 074/123] scsi: qedf: Simplify alloc_workqueue() invocations Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-13-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/qedf/qedf_main.c | 15 ++++++--------- 1 file changed, 6 insertions(+), 9 deletions(-) diff --git a/drivers/scsi/qedf/qedf_main.c b/drivers/scsi/qedf/qedf_main.c index 119afcaf6e13..cf13148ba281 100644 --- a/drivers/scsi/qedf/qedf_main.c +++ b/drivers/scsi/qedf/qedf_main.c @@ -3372,10 +3372,8 @@ retry_probe: QEDF_INFO(&(qedf->dbg_ctx), QEDF_LOG_INFO, "qedf->io_mempool=%p.\n", qedf->io_mempool); - sprintf(host_buf, "qedf_%u_link", - qedf->lport->host->host_no); - qedf->link_update_wq = - alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); + qedf->link_update_wq = alloc_workqueue("qedf_%u_link", WQ_MEM_RECLAIM, + 1, qedf->lport->host->host_no); INIT_DELAYED_WORK(&qedf->link_update, qedf_handle_link_update); INIT_DELAYED_WORK(&qedf->link_recovery, qedf_link_recovery); INIT_DELAYED_WORK(&qedf->grcdump_work, qedf_wq_grcdump); @@ -3585,8 +3583,8 @@ retry_probe: ether_addr_copy(params.ll2_mac_address, qedf->mac); /* Start LL2 processing thread */ - snprintf(host_buf, 20, "qedf_%d_ll2", host->host_no); - qedf->ll2_recv_wq = alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); + qedf->ll2_recv_wq = alloc_workqueue("qedf_%d_ll2", WQ_MEM_RECLAIM, 1, + host->host_no); if (!qedf->ll2_recv_wq) { QEDF_ERR(&(qedf->dbg_ctx), "Failed to LL2 workqueue.\n"); rc = -ENOMEM; @@ -3627,9 +3625,8 @@ retry_probe: } } - sprintf(host_buf, "qedf_%u_timer", qedf->lport->host->host_no); - qedf->timer_work_queue = - alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); + qedf->timer_work_queue = alloc_workqueue("qedf_%u_timer", + WQ_MEM_RECLAIM, 1, qedf->lport->host->host_no); if (!qedf->timer_work_queue) { QEDF_ERR(&(qedf->dbg_ctx), "Failed to start timer " "workqueue.\n"); From 19d7cda1c6304e032c1f7cd0ff3e6b5dc29f76dd Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:17 -0700 Subject: [PATCH 075/123] scsi: qedi: Simplify an alloc_workqueue() invocation Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-14-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/qedi/qedi_main.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/qedi/qedi_main.c b/drivers/scsi/qedi/qedi_main.c index 319c1da549f7..c5aec26019d6 100644 --- a/drivers/scsi/qedi/qedi_main.c +++ b/drivers/scsi/qedi/qedi_main.c @@ -2776,9 +2776,9 @@ retry_probe: goto free_cid_que; } - sprintf(host_buf, "qedi_ofld%d", qedi->shost->host_no); - qedi->offload_thread = - alloc_workqueue("%s", WQ_MEM_RECLAIM, 1, host_buf); + qedi->offload_thread = alloc_workqueue("qedi_ofld%d", + WQ_MEM_RECLAIM, + 1, qedi->shost->host_no); if (!qedi->offload_thread) { QEDI_ERR(&qedi->dbg_ctx, "Unable to start offload thread!\n"); From 6411307b63286eff7c5c72a8ad64796ae8549101 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:18 -0700 Subject: [PATCH 076/123] scsi: snic: Simplify alloc_workqueue() invocations Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Not setting shost->work_q_name is safe because there is no code that reads the value set by the removed code. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-15-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/snic/snic_main.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/scsi/snic/snic_main.c b/drivers/scsi/snic/snic_main.c index 2bd01eb57869..5ca8bc89dfa7 100644 --- a/drivers/scsi/snic/snic_main.c +++ b/drivers/scsi/snic/snic_main.c @@ -300,10 +300,8 @@ snic_add_host(struct Scsi_Host *shost, struct pci_dev *pdev) } SNIC_BUG_ON(shost->work_q != NULL); - snprintf(shost->work_q_name, sizeof(shost->work_q_name), "scsi_wq_%d", - shost->host_no); - shost->work_q = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, - shost->work_q_name); + shost->work_q = alloc_ordered_workqueue("scsi_wq_%d", WQ_MEM_RECLAIM, + shost->host_no); if (!shost->work_q) { SNIC_HOST_ERR(shost, "Failed to Create ScsiHost wq.\n"); From 06d53789761cad6bbbc3c00e8c96a631a41f4bf3 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:19 -0700 Subject: [PATCH 077/123] scsi: scsi_transport_fc: Simplify alloc_workqueue() invocations Let alloc_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-16-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/scsi_transport_fc.c | 11 +++-------- include/scsi/scsi_transport_fc.h | 6 ------ 2 files changed, 3 insertions(+), 14 deletions(-) diff --git a/drivers/scsi/scsi_transport_fc.c b/drivers/scsi/scsi_transport_fc.c index 7d088b8da075..62ea7e44460e 100644 --- a/drivers/scsi/scsi_transport_fc.c +++ b/drivers/scsi/scsi_transport_fc.c @@ -441,18 +441,13 @@ static int fc_host_setup(struct transport_container *tc, struct device *dev, fc_host->next_vport_number = 0; fc_host->npiv_vports_inuse = 0; - snprintf(fc_host->work_q_name, sizeof(fc_host->work_q_name), - "fc_wq_%d", shost->host_no); - fc_host->work_q = alloc_workqueue("%s", 0, 0, fc_host->work_q_name); + fc_host->work_q = alloc_workqueue("fc_wq_%d", 0, 0, shost->host_no); if (!fc_host->work_q) return -ENOMEM; fc_host->dev_loss_tmo = fc_dev_loss_tmo; - snprintf(fc_host->devloss_work_q_name, - sizeof(fc_host->devloss_work_q_name), - "fc_dl_%d", shost->host_no); - fc_host->devloss_work_q = alloc_workqueue("%s", 0, 0, - fc_host->devloss_work_q_name); + fc_host->devloss_work_q = alloc_workqueue("fc_dl_%d", 0, 0, + shost->host_no); if (!fc_host->devloss_work_q) { destroy_workqueue(fc_host->work_q); fc_host->work_q = NULL; diff --git a/include/scsi/scsi_transport_fc.h b/include/scsi/scsi_transport_fc.h index 4b884b8013e0..8e6c60090c62 100644 --- a/include/scsi/scsi_transport_fc.h +++ b/include/scsi/scsi_transport_fc.h @@ -575,9 +575,7 @@ struct fc_host_attrs { u16 npiv_vports_inuse; /* work queues for rport state manipulation */ - char work_q_name[20]; struct workqueue_struct *work_q; - char devloss_work_q_name[20]; struct workqueue_struct *devloss_work_q; /* bsg support */ @@ -654,12 +652,8 @@ struct fc_host_attrs { (((struct fc_host_attrs *)(x)->shost_data)->next_vport_number) #define fc_host_npiv_vports_inuse(x) \ (((struct fc_host_attrs *)(x)->shost_data)->npiv_vports_inuse) -#define fc_host_work_q_name(x) \ - (((struct fc_host_attrs *)(x)->shost_data)->work_q_name) #define fc_host_work_q(x) \ (((struct fc_host_attrs *)(x)->shost_data)->work_q) -#define fc_host_devloss_work_q_name(x) \ - (((struct fc_host_attrs *)(x)->shost_data)->devloss_work_q_name) #define fc_host_devloss_work_q(x) \ (((struct fc_host_attrs *)(x)->shost_data)->devloss_work_q) #define fc_host_dev_loss_tmo(x) \ From 0ef9b0186dae7039e95e9367ed1d9402a763511b Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:20 -0700 Subject: [PATCH 078/123] scsi: stex: Simplify an alloc_ordered_workqueue() invocation Let alloc_ordered_workqueue() format the workqueue name instead of calling snprintf() explicitly. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-17-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/stex.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/drivers/scsi/stex.c b/drivers/scsi/stex.c index fbee7db4a835..0e81125df8c7 100644 --- a/drivers/scsi/stex.c +++ b/drivers/scsi/stex.c @@ -334,7 +334,6 @@ struct st_hba { struct st_ccb *wait_ccb; __le32 *scratch; - char work_q_name[20]; struct workqueue_struct *work_q; struct work_struct reset_work; wait_queue_head_t reset_waitq; @@ -1795,10 +1794,8 @@ static int stex_probe(struct pci_dev *pdev, const struct pci_device_id *id) hba->pdev = pdev; init_waitqueue_head(&hba->reset_waitq); - snprintf(hba->work_q_name, sizeof(hba->work_q_name), - "stex_wq_%d", host->host_no); - hba->work_q = - alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, hba->work_q_name); + hba->work_q = alloc_ordered_workqueue("stex_wq_%d", WQ_MEM_RECLAIM, + host->host_no); if (!hba->work_q) { printk(KERN_ERR DRV_NAME "(%s): create workqueue failed\n", pci_name(pdev)); From 1aa992cbc272beceba40ae39b33b2848c228c0c9 Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:21 -0700 Subject: [PATCH 079/123] scsi: ufs: Simplify alloc*_workqueue() invocation Let alloc*_workqueue() format the workqueue name instead of calling snprintf() explicitly. Reviewed-by: Manivannan Sadhasivam Reviewed-by: Peter Wang Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-18-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/ufs/core/ufshcd.c | 23 +++++++---------------- 1 file changed, 7 insertions(+), 16 deletions(-) diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index 930b15d9356b..d89b2f14e764 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -1785,8 +1785,6 @@ static void ufshcd_remove_clk_scaling_sysfs(struct ufs_hba *hba) static void ufshcd_init_clk_scaling(struct ufs_hba *hba) { - char wq_name[sizeof("ufs_clkscaling_00")]; - if (!ufshcd_is_clkscaling_supported(hba)) return; @@ -1798,10 +1796,8 @@ static void ufshcd_init_clk_scaling(struct ufs_hba *hba) INIT_WORK(&hba->clk_scaling.resume_work, ufshcd_clk_scaling_resume_work); - snprintf(wq_name, sizeof(wq_name), "ufs_clkscaling_%d", - hba->host->host_no); - hba->clk_scaling.workq = - alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, wq_name); + hba->clk_scaling.workq = alloc_ordered_workqueue( + "ufs_clkscaling_%d", WQ_MEM_RECLAIM, hba->host->host_no); hba->clk_scaling.is_initialized = true; } @@ -2125,8 +2121,6 @@ static void ufshcd_remove_clk_gating_sysfs(struct ufs_hba *hba) static void ufshcd_init_clk_gating(struct ufs_hba *hba) { - char wq_name[sizeof("ufs_clk_gating_00")]; - if (!ufshcd_is_clkgating_allowed(hba)) return; @@ -2136,10 +2130,9 @@ static void ufshcd_init_clk_gating(struct ufs_hba *hba) INIT_DELAYED_WORK(&hba->clk_gating.gate_work, ufshcd_gate_work); INIT_WORK(&hba->clk_gating.ungate_work, ufshcd_ungate_work); - snprintf(wq_name, ARRAY_SIZE(wq_name), "ufs_clk_gating_%d", - hba->host->host_no); - hba->clk_gating.clk_gating_workq = alloc_ordered_workqueue(wq_name, - WQ_MEM_RECLAIM | WQ_HIGHPRI); + hba->clk_gating.clk_gating_workq = alloc_ordered_workqueue( + "ufs_clk_gating_%d", WQ_MEM_RECLAIM | WQ_HIGHPRI, + hba->host->host_no); ufshcd_init_clk_gating_sysfs(hba); @@ -10376,7 +10369,6 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq) int err; struct Scsi_Host *host = hba->host; struct device *dev = hba->dev; - char eh_wq_name[sizeof("ufs_eh_wq_00")]; /* * dev_set_drvdata() must be called before any callbacks are registered @@ -10443,9 +10435,8 @@ int ufshcd_init(struct ufs_hba *hba, void __iomem *mmio_base, unsigned int irq) hba->max_pwr_info.is_valid = false; /* Initialize work queues */ - snprintf(eh_wq_name, sizeof(eh_wq_name), "ufs_eh_wq_%d", - hba->host->host_no); - hba->eh_wq = alloc_ordered_workqueue("%s", WQ_MEM_RECLAIM, eh_wq_name); + hba->eh_wq = alloc_ordered_workqueue("ufs_eh_wq_%d", WQ_MEM_RECLAIM, + hba->host->host_no); if (!hba->eh_wq) { dev_err(hba->dev, "%s: failed to create eh workqueue\n", __func__); From ba52850cb6b4db5f4ec4636c73d2ad85d0e9adba Mon Sep 17 00:00:00 2001 From: Bart Van Assche Date: Thu, 22 Aug 2024 12:59:22 -0700 Subject: [PATCH 080/123] scsi: core: Simplify an alloc_workqueue() invocation Let alloc_workqueue() format the workqueue name. Remove the work_q_name[] member from struct Scsi_Host because it is no longer used by any SCSI driver nor by the SCSI core. Signed-off-by: Bart Van Assche Link: https://lore.kernel.org/r/20240822195944.654691-19-bvanassche@acm.org Signed-off-by: Martin K. Petersen --- drivers/scsi/hosts.c | 9 ++++----- include/scsi/scsi_host.h | 1 - 2 files changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/hosts.c b/drivers/scsi/hosts.c index 7f987335b44c..e021f1106bea 100644 --- a/drivers/scsi/hosts.c +++ b/drivers/scsi/hosts.c @@ -292,11 +292,10 @@ int scsi_add_host_with_dma(struct Scsi_Host *shost, struct device *dev, } if (shost->transportt->create_work_queue) { - snprintf(shost->work_q_name, sizeof(shost->work_q_name), - "scsi_wq_%d", shost->host_no); - shost->work_q = alloc_workqueue("%s", - WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM | WQ_UNBOUND, - 1, shost->work_q_name); + shost->work_q = alloc_workqueue( + "scsi_wq_%d", + WQ_SYSFS | __WQ_LEGACY | WQ_MEM_RECLAIM | WQ_UNBOUND, 1, + shost->host_no); if (!shost->work_q) { error = -EINVAL; diff --git a/include/scsi/scsi_host.h b/include/scsi/scsi_host.h index 19a1c5c48935..2b4ab0369ffb 100644 --- a/include/scsi/scsi_host.h +++ b/include/scsi/scsi_host.h @@ -677,7 +677,6 @@ struct Scsi_Host { /* * Optional work queue to be utilized by the transport */ - char work_q_name[20]; struct workqueue_struct *work_q; /* From adedd0f46c923f8d63aeb42d504c82431febed31 Mon Sep 17 00:00:00 2001 From: Yue Haibing Date: Sat, 24 Aug 2024 16:47:24 +0800 Subject: [PATCH 081/123] scsi: bnx2i: Remove unused declarations Commit cf4e6363859d ("[SCSI] bnx2i: Add bnx2i iSCSI driver.") declared but never implemented these. Signed-off-by: Yue Haibing Link: https://lore.kernel.org/r/20240824084724.3647307-1-yuehaibing@huawei.com Signed-off-by: Martin K. Petersen --- drivers/scsi/bnx2i/bnx2i.h | 11 ----------- 1 file changed, 11 deletions(-) diff --git a/drivers/scsi/bnx2i/bnx2i.h b/drivers/scsi/bnx2i/bnx2i.h index df7d04afce05..7030efee5c46 100644 --- a/drivers/scsi/bnx2i/bnx2i.h +++ b/drivers/scsi/bnx2i/bnx2i.h @@ -815,11 +815,6 @@ extern struct bnx2i_hba *get_adapter_list_head(void); struct bnx2i_conn *bnx2i_get_conn_from_id(struct bnx2i_hba *hba, u16 iscsi_cid); -int bnx2i_alloc_ep_pool(void); -void bnx2i_release_ep_pool(void); -struct bnx2i_endpoint *bnx2i_ep_ofld_list_next(struct bnx2i_hba *hba); -struct bnx2i_endpoint *bnx2i_ep_destroy_list_next(struct bnx2i_hba *hba); - struct bnx2i_hba *bnx2i_find_hba_for_cnic(struct cnic_dev *cnic); struct bnx2i_hba *bnx2i_alloc_hba(struct cnic_dev *cnic); @@ -869,12 +864,6 @@ extern int bnx2i_arm_cq_event_coalescing(struct bnx2i_endpoint *ep, u8 action); extern int bnx2i_hw_ep_disconnect(struct bnx2i_endpoint *bnx2i_ep); -/* Debug related function prototypes */ -extern void bnx2i_print_pend_cmd_queue(struct bnx2i_conn *conn); -extern void bnx2i_print_active_cmd_queue(struct bnx2i_conn *conn); -extern void bnx2i_print_xmit_pdu_queue(struct bnx2i_conn *conn); -extern void bnx2i_print_recv_state(struct bnx2i_conn *conn); - extern int bnx2i_percpu_io_thread(void *arg); extern int bnx2i_process_scsi_cmd_resp(struct iscsi_session *session, struct bnx2i_conn *bnx2i_conn, From 3e3ac9c39e1b575e970ecab90504b6cb090f6b05 Mon Sep 17 00:00:00 2001 From: Gaosheng Cui Date: Mon, 26 Aug 2024 11:20:05 +0800 Subject: [PATCH 082/123] scsi: core: Remove obsoleted declaration for scsi_driverbyte_string() scsi_driverbyte_string() has been unused since commit 54c29086195f ("scsi: core: Drop the now obsolete driver_byte definitions"). Remove it. Signed-off-by: Gaosheng Cui Link: https://lore.kernel.org/r/20240826032005.4007834-1-cuigaosheng1@huawei.com Reviewed-by: Bart Van Assche Signed-off-by: Martin K. Petersen --- include/scsi/scsi_dbg.h | 7 ------- 1 file changed, 7 deletions(-) diff --git a/include/scsi/scsi_dbg.h b/include/scsi/scsi_dbg.h index 7b196d234626..bd29cdb513a5 100644 --- a/include/scsi/scsi_dbg.h +++ b/include/scsi/scsi_dbg.h @@ -24,7 +24,6 @@ extern const char *scsi_extd_sense_format(unsigned char, unsigned char, const char **); extern const char *scsi_mlreturn_string(int); extern const char *scsi_hostbyte_string(int); -extern const char *scsi_driverbyte_string(int); #else static inline bool scsi_opcode_sa_name(int cmd, int sa, @@ -76,12 +75,6 @@ scsi_hostbyte_string(int result) return NULL; } -static inline const char * -scsi_driverbyte_string(int result) -{ - return NULL; -} - #endif #endif /* _SCSI_SCSI_DBG_H */ From 0a5167e21126c4b1d35aaab64ad04ea056ec8aab Mon Sep 17 00:00:00 2001 From: "Bao D. Nguyen" Date: Tue, 27 Aug 2024 16:14:13 -0700 Subject: [PATCH 083/123] scsi: ufs: core: Remove ufshcd_urgent_bkops() ufshcd_urgent_bkops() is a wrapper function. It only calls ufshcd_bkops_ctrl(). Remove it to simplify the ufs core driver. Replace any references to ufshcd_urgent_bkops() with ufshcd_bkops_ctrl(). In addition, remove the second parameter in the ufshcd_bkops_ctrl() because the information can be retrieved from the first parameter. Signed-off-by: Bao D. Nguyen Link: https://lore.kernel.org/r/0c7f2c8d68408e39c28e3e81addce09cc0ee3969.1724800328.git.quic_nguyenb@quicinc.com Acked-by: Avri Altman Reviewed-by: Manivannan Sadhasivam Signed-off-by: Martin K. Petersen --- drivers/ufs/core/ufshcd.c | 32 +++++++------------------------- 1 file changed, 7 insertions(+), 25 deletions(-) diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c index e13b9ac145f6..8ea5a82503a9 100644 --- a/drivers/ufs/core/ufshcd.c +++ b/drivers/ufs/core/ufshcd.c @@ -5870,12 +5870,11 @@ static inline int ufshcd_get_bkops_status(struct ufs_hba *hba, u32 *status) /** * ufshcd_bkops_ctrl - control the auto bkops based on current bkops status * @hba: per-adapter instance - * @status: bkops_status value * * Read the bkops_status from the UFS device and Enable fBackgroundOpsEn * flag in the device to permit background operations if the device - * bkops_status is greater than or equal to "status" argument passed to - * this function, disable otherwise. + * bkops_status is greater than or equal to the "hba->urgent_bkops_lvl", + * disable otherwise. * * Return: 0 for success, non-zero in case of failure. * @@ -5883,11 +5882,11 @@ static inline int ufshcd_get_bkops_status(struct ufs_hba *hba, u32 *status) * to know whether auto bkops is enabled or disabled after this function * returns control to it. */ -static int ufshcd_bkops_ctrl(struct ufs_hba *hba, - enum bkops_status status) +static int ufshcd_bkops_ctrl(struct ufs_hba *hba) { - int err; + enum bkops_status status = hba->urgent_bkops_lvl; u32 curr_status = 0; + int err; err = ufshcd_get_bkops_status(hba, &curr_status); if (err) { @@ -5909,23 +5908,6 @@ out: return err; } -/** - * ufshcd_urgent_bkops - handle urgent bkops exception event - * @hba: per-adapter instance - * - * Enable fBackgroundOpsEn flag in the device to permit background - * operations. - * - * If BKOPs is enabled, this function returns 0, 1 if the bkops in not enabled - * and negative error value for any other failure. - * - * Return: 0 upon success; < 0 upon failure. - */ -static int ufshcd_urgent_bkops(struct ufs_hba *hba) -{ - return ufshcd_bkops_ctrl(hba, hba->urgent_bkops_lvl); -} - static inline int ufshcd_get_ee_status(struct ufs_hba *hba, u32 *status) { return ufshcd_query_attr_retry(hba, UPIU_QUERY_OPCODE_READ_ATTR, @@ -9682,7 +9664,7 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op) * allow background operations if bkops status shows * that performance might be impacted. */ - ret = ufshcd_urgent_bkops(hba); + ret = ufshcd_bkops_ctrl(hba); if (ret) { /* * If return err in suspend flow, IO will hang. @@ -9871,7 +9853,7 @@ static int __ufshcd_wl_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op) * If BKOPs operations are urgently needed at this moment then * keep auto-bkops enabled or else disable it. */ - ufshcd_urgent_bkops(hba); + ufshcd_bkops_ctrl(hba); if (hba->ee_usr_mask) ufshcd_write_ee_control(hba); From e59f43fb647733cd58191e6cfec4122f9bdd0aab Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Mon, 19 Aug 2024 21:11:43 +0200 Subject: [PATCH 084/123] scsi: qla2xxx: Remove the unused 'del_list_entry' field in struct fc_port The 'del_list_entry' field in "struct fc_port" is unused. The field was introduced in commit 2d70c103fd2a ("[SCSI] qla2xxx: Add LLD target-mode infrastructure for >= 24xx series") in 2012-05 and the last user was removed in commit 726b85487067 ("qla2xxx: Add framework for async fabric discovery") in 2017-02. Remove this unused field. Signed-off-by: Christophe JAILLET Link: https://lore.kernel.org/r/69155321ab26c1f4d473d5bb6cd44b59b9b6a020.1724094686.git.christophe.jaillet@wanadoo.fr Signed-off-by: Martin K. Petersen --- drivers/scsi/qla2xxx/qla_def.h | 1 - 1 file changed, 1 deletion(-) diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h index 7cf998e3cc68..15066c112817 100644 --- a/drivers/scsi/qla2xxx/qla_def.h +++ b/drivers/scsi/qla2xxx/qla_def.h @@ -2621,7 +2621,6 @@ typedef struct fc_port { struct kref sess_kref; struct qla_tgt *tgt; unsigned long expires; - struct list_head del_list_entry; struct work_struct free_work; struct work_struct reg_work; uint64_t jiffies_at_registration; From d5a4b0d64242574bc12f7864809e03aa2863b83b Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Fri, 23 Aug 2024 09:59:05 +0200 Subject: [PATCH 085/123] scsi: bnx2fc: Remove some unused fields in struct bnx2fc_rport Some fields are unused in struct bnx2fc_rport. Remove them in order to save 96 bytes on a x86_64. Signed-off-by: Christophe JAILLET Link: https://lore.kernel.org/r/42e20b159f3bbb12da7796463a521ca051bd5274.1724399924.git.christophe.jaillet@wanadoo.fr Signed-off-by: Martin K. Petersen --- drivers/scsi/bnx2fc/bnx2fc.h | 6 ------ 1 file changed, 6 deletions(-) diff --git a/drivers/scsi/bnx2fc/bnx2fc.h b/drivers/scsi/bnx2fc/bnx2fc.h index 7e74f77da14f..6d47a4d8eed6 100644 --- a/drivers/scsi/bnx2fc/bnx2fc.h +++ b/drivers/scsi/bnx2fc/bnx2fc.h @@ -358,18 +358,12 @@ struct bnx2fc_rport { dma_addr_t lcq_dma; u32 lcq_mem_size; - void *ofld_req[4]; - dma_addr_t ofld_req_dma[4]; - void *enbl_req; - dma_addr_t enbl_req_dma; - spinlock_t tgt_lock; spinlock_t cq_lock; atomic_t num_active_ios; u32 flush_in_prog; unsigned long timestamp; unsigned long retry_delay_timestamp; - struct list_head free_task_list; struct bnx2fc_cmd *pending_queue[BNX2FC_SQ_WQES_MAX+1]; struct list_head active_cmd_queue; struct list_head els_queue; From 058311b72f54890de824b063feb603942269b732 Mon Sep 17 00:00:00 2001 From: Murthy Bhat Date: Tue, 27 Aug 2024 13:54:55 -0500 Subject: [PATCH 086/123] scsi: smartpqi: Add fw log to kdump Add controller logs to kdump. Driver allocates DMA memory and communicates this address to FW. In the event of system crash, host driver notifies the firmware about the crash and firmware posts all the necessary logs in the pre-allocated host buffer for firmware debugging. Once firmware notifies the completion of the log uploading to the host memory and host continues with the OS crash dump saving. This is a "feature" driven capability and is backward compatible with existing controller FW. Rename some prefixes for OFA (Online-Firmware Activation ofa_*) buffers to host_memory_*. So, not a lot of actual functional changes to smartpqi_init.c, mainly determining the memory size allocation. Added a function to notify the controller to copy debug data into host memory before continuing kdump. Most of the functional changes are in smartpqi_sis.c where the actual handshaking is done. Reviewed-by: Scott Benesh Reviewed-by: Mike McGowen Signed-off-by: Murthy Bhat Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-2-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi.h | 32 ++- drivers/scsi/smartpqi/smartpqi_init.c | 396 +++++++++++++++----------- drivers/scsi/smartpqi/smartpqi_sis.c | 60 ++++ drivers/scsi/smartpqi/smartpqi_sis.h | 3 + 4 files changed, 306 insertions(+), 185 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi.h b/drivers/scsi/smartpqi/smartpqi.h index 023fbce04e7a..f493006bee9d 100644 --- a/drivers/scsi/smartpqi/smartpqi.h +++ b/drivers/scsi/smartpqi/smartpqi.h @@ -505,7 +505,7 @@ struct pqi_vendor_general_request { __le64 buffer_address; __le32 buffer_length; u8 reserved[40]; - } ofa_memory_allocation; + } host_memory_allocation; } data; }; @@ -517,21 +517,30 @@ struct pqi_vendor_general_response { u8 reserved[2]; }; -#define PQI_VENDOR_GENERAL_CONFIG_TABLE_UPDATE 0 -#define PQI_VENDOR_GENERAL_HOST_MEMORY_UPDATE 1 +#define PQI_VENDOR_GENERAL_CONFIG_TABLE_UPDATE 0 +#define PQI_VENDOR_GENERAL_OFA_MEMORY_UPDATE 1 +#define PQI_VENDOR_GENERAL_CTRL_LOG_MEMORY_UPDATE 2 #define PQI_OFA_VERSION 1 #define PQI_OFA_SIGNATURE "OFA_QRM" -#define PQI_OFA_MAX_SG_DESCRIPTORS 64 +#define PQI_CTRL_LOG_VERSION 1 +#define PQI_CTRL_LOG_SIGNATURE "FW_DATA" +#define PQI_HOST_MAX_SG_DESCRIPTORS 64 -struct pqi_ofa_memory { - __le64 signature; /* "OFA_QRM" */ +struct pqi_host_memory { + __le64 signature; /* "OFA_QRM", "FW_DATA", etc. */ __le16 version; /* version of this struct (1 = 1st version) */ u8 reserved[62]; __le32 bytes_allocated; /* total allocated memory in bytes */ __le16 num_memory_descriptors; u8 reserved1[2]; - struct pqi_sg_descriptor sg_descriptor[PQI_OFA_MAX_SG_DESCRIPTORS]; + struct pqi_sg_descriptor sg_descriptor[PQI_HOST_MAX_SG_DESCRIPTORS]; +}; + +struct pqi_host_memory_descriptor { + struct pqi_host_memory *host_memory; + dma_addr_t host_memory_dma_handle; + void **host_chunk_virt_address; }; struct pqi_aio_error_info { @@ -867,7 +876,8 @@ struct pqi_config_table_firmware_features { #define PQI_FIRMWARE_FEATURE_FW_TRIAGE 17 #define PQI_FIRMWARE_FEATURE_RPL_EXTENDED_FORMAT_4_5 18 #define PQI_FIRMWARE_FEATURE_MULTI_LUN_DEVICE_SUPPORT 21 -#define PQI_FIRMWARE_FEATURE_MAXIMUM 21 +#define PQI_FIRMWARE_FEATURE_CTRL_LOGGING 22 +#define PQI_FIRMWARE_FEATURE_MAXIMUM 22 struct pqi_config_table_debug { struct pqi_config_table_section_header header; @@ -1357,6 +1367,7 @@ struct pqi_ctrl_info { u8 firmware_triage_supported : 1; u8 rpl_extended_format_4_5_supported : 1; u8 multi_lun_device_supported : 1; + u8 ctrl_logging_supported : 1; u8 enable_r1_writes : 1; u8 enable_r5_writes : 1; u8 enable_r6_writes : 1; @@ -1398,13 +1409,12 @@ struct pqi_ctrl_info { wait_queue_head_t block_requests_wait; struct mutex ofa_mutex; - struct pqi_ofa_memory *pqi_ofa_mem_virt_addr; - dma_addr_t pqi_ofa_mem_dma_handle; - void **pqi_ofa_chunk_virt_addr; struct work_struct ofa_memory_alloc_work; struct work_struct ofa_quiesce_work; u32 ofa_bytes_requested; u16 ofa_cancel_reason; + struct pqi_host_memory_descriptor ofa_memory; + struct pqi_host_memory_descriptor ctrl_log_memory; enum pqi_ctrl_removal_state ctrl_removal_state; }; diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index b5396d722d52..54f7fe843445 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -92,9 +92,9 @@ static int pqi_aio_submit_r56_write_io(struct pqi_ctrl_info *ctrl_info, static void pqi_ofa_ctrl_quiesce(struct pqi_ctrl_info *ctrl_info); static void pqi_ofa_ctrl_unquiesce(struct pqi_ctrl_info *ctrl_info); static int pqi_ofa_ctrl_restart(struct pqi_ctrl_info *ctrl_info, unsigned int delay_secs); -static void pqi_ofa_setup_host_buffer(struct pqi_ctrl_info *ctrl_info); -static void pqi_ofa_free_host_buffer(struct pqi_ctrl_info *ctrl_info); -static int pqi_ofa_host_memory_update(struct pqi_ctrl_info *ctrl_info); +static void pqi_host_setup_buffer(struct pqi_ctrl_info *ctrl_info, struct pqi_host_memory_descriptor *host_memory_descriptor, u32 total_size, u32 min_size); +static void pqi_host_free_buffer(struct pqi_ctrl_info *ctrl_info, struct pqi_host_memory_descriptor *host_memory_descriptor); +static int pqi_host_memory_update(struct pqi_ctrl_info *ctrl_info, struct pqi_host_memory_descriptor *host_memory_descriptor, u16 function_code); static int pqi_device_wait_for_pending_io(struct pqi_ctrl_info *ctrl_info, struct pqi_scsi_dev *device, u8 lun, unsigned long timeout_msecs); static void pqi_fail_all_outstanding_requests(struct pqi_ctrl_info *ctrl_info); @@ -3634,7 +3634,7 @@ static void pqi_process_soft_reset(struct pqi_ctrl_info *ctrl_info) ctrl_info->pqi_mode_enabled = false; pqi_save_ctrl_mode(ctrl_info, SIS_MODE); rc = pqi_ofa_ctrl_restart(ctrl_info, delay_secs); - pqi_ofa_free_host_buffer(ctrl_info); + pqi_host_free_buffer(ctrl_info, &ctrl_info->ofa_memory); pqi_ctrl_ofa_done(ctrl_info); dev_info(&ctrl_info->pci_dev->dev, "Online Firmware Activation: %s\n", @@ -3645,7 +3645,7 @@ static void pqi_process_soft_reset(struct pqi_ctrl_info *ctrl_info) "Online Firmware Activation ABORTED\n"); if (ctrl_info->soft_reset_handshake_supported) pqi_clear_soft_reset_status(ctrl_info); - pqi_ofa_free_host_buffer(ctrl_info); + pqi_host_free_buffer(ctrl_info, &ctrl_info->ofa_memory); pqi_ctrl_ofa_done(ctrl_info); pqi_ofa_ctrl_unquiesce(ctrl_info); break; @@ -3655,7 +3655,7 @@ static void pqi_process_soft_reset(struct pqi_ctrl_info *ctrl_info) dev_err(&ctrl_info->pci_dev->dev, "unexpected Online Firmware Activation reset status: 0x%x\n", reset_status); - pqi_ofa_free_host_buffer(ctrl_info); + pqi_host_free_buffer(ctrl_info, &ctrl_info->ofa_memory); pqi_ctrl_ofa_done(ctrl_info); pqi_ofa_ctrl_unquiesce(ctrl_info); pqi_take_ctrl_offline(ctrl_info, PQI_OFA_RESPONSE_TIMEOUT); @@ -3670,8 +3670,8 @@ static void pqi_ofa_memory_alloc_worker(struct work_struct *work) ctrl_info = container_of(work, struct pqi_ctrl_info, ofa_memory_alloc_work); pqi_ctrl_ofa_start(ctrl_info); - pqi_ofa_setup_host_buffer(ctrl_info); - pqi_ofa_host_memory_update(ctrl_info); + pqi_host_setup_buffer(ctrl_info, &ctrl_info->ofa_memory, ctrl_info->ofa_bytes_requested, ctrl_info->ofa_bytes_requested); + pqi_host_memory_update(ctrl_info, &ctrl_info->ofa_memory, PQI_VENDOR_GENERAL_OFA_MEMORY_UPDATE); } static void pqi_ofa_quiesce_worker(struct work_struct *work) @@ -3711,7 +3711,7 @@ static bool pqi_ofa_process_event(struct pqi_ctrl_info *ctrl_info, dev_info(&ctrl_info->pci_dev->dev, "received Online Firmware Activation cancel request: reason: %u\n", ctrl_info->ofa_cancel_reason); - pqi_ofa_free_host_buffer(ctrl_info); + pqi_host_free_buffer(ctrl_info, &ctrl_info->ofa_memory); pqi_ctrl_ofa_done(ctrl_info); break; default: @@ -7883,6 +7883,9 @@ static void pqi_ctrl_update_feature_flags(struct pqi_ctrl_info *ctrl_info, case PQI_FIRMWARE_FEATURE_MULTI_LUN_DEVICE_SUPPORT: ctrl_info->multi_lun_device_supported = firmware_feature->enabled; break; + case PQI_FIRMWARE_FEATURE_CTRL_LOGGING: + ctrl_info->ctrl_logging_supported = firmware_feature->enabled; + break; } pqi_firmware_feature_status(ctrl_info, firmware_feature); @@ -7988,6 +7991,11 @@ static struct pqi_firmware_feature pqi_firmware_features[] = { .feature_bit = PQI_FIRMWARE_FEATURE_MULTI_LUN_DEVICE_SUPPORT, .feature_status = pqi_ctrl_update_feature_flags, }, + { + .feature_name = "Controller Data Logging", + .feature_bit = PQI_FIRMWARE_FEATURE_CTRL_LOGGING, + .feature_status = pqi_ctrl_update_feature_flags, + }, }; static void pqi_process_firmware_features( @@ -8090,6 +8098,7 @@ static void pqi_ctrl_reset_config(struct pqi_ctrl_info *ctrl_info) ctrl_info->firmware_triage_supported = false; ctrl_info->rpl_extended_format_4_5_supported = false; ctrl_info->multi_lun_device_supported = false; + ctrl_info->ctrl_logging_supported = false; } static int pqi_process_config_table(struct pqi_ctrl_info *ctrl_info) @@ -8230,6 +8239,9 @@ static void pqi_perform_lockup_action(void) } } +#define PQI_CTRL_LOG_TOTAL_SIZE (4 * 1024 * 1024) +#define PQI_CTRL_LOG_MIN_SIZE (PQI_CTRL_LOG_TOTAL_SIZE / PQI_HOST_MAX_SG_DESCRIPTORS) + static int pqi_ctrl_init(struct pqi_ctrl_info *ctrl_info) { int rc; @@ -8241,6 +8253,12 @@ static int pqi_ctrl_init(struct pqi_ctrl_info *ctrl_info) if (rc) return rc; } + if (sis_is_ctrl_logging_supported(ctrl_info)) { + sis_notify_kdump(ctrl_info); + rc = sis_wait_for_ctrl_logging_completion(ctrl_info); + if (rc) + return rc; + } sis_soft_reset(ctrl_info); ssleep(PQI_POST_RESET_DELAY_SECS); } else { @@ -8422,6 +8440,11 @@ static int pqi_ctrl_init(struct pqi_ctrl_info *ctrl_info) if (rc) return rc; + if (ctrl_info->ctrl_logging_supported && !reset_devices) { + pqi_host_setup_buffer(ctrl_info, &ctrl_info->ctrl_log_memory, PQI_CTRL_LOG_TOTAL_SIZE, PQI_CTRL_LOG_MIN_SIZE); + pqi_host_memory_update(ctrl_info, &ctrl_info->ctrl_log_memory, PQI_VENDOR_GENERAL_CTRL_LOG_MEMORY_UPDATE); + } + rc = pqi_get_ctrl_product_details(ctrl_info); if (rc) { dev_err(&ctrl_info->pci_dev->dev, @@ -8606,8 +8629,22 @@ static int pqi_ctrl_init_resume(struct pqi_ctrl_info *ctrl_info) return rc; } - if (pqi_ofa_in_progress(ctrl_info)) + if (pqi_ofa_in_progress(ctrl_info)) { pqi_ctrl_unblock_scan(ctrl_info); + if (ctrl_info->ctrl_logging_supported) { + if (!ctrl_info->ctrl_log_memory.host_memory) + pqi_host_setup_buffer(ctrl_info, + &ctrl_info->ctrl_log_memory, + PQI_CTRL_LOG_TOTAL_SIZE, + PQI_CTRL_LOG_MIN_SIZE); + pqi_host_memory_update(ctrl_info, + &ctrl_info->ctrl_log_memory, PQI_VENDOR_GENERAL_CTRL_LOG_MEMORY_UPDATE); + } else { + if (ctrl_info->ctrl_log_memory.host_memory) + pqi_host_free_buffer(ctrl_info, + &ctrl_info->ctrl_log_memory); + } + } pqi_scan_scsi_devices(ctrl_info); @@ -8797,6 +8834,7 @@ static void pqi_remove_ctrl(struct pqi_ctrl_info *ctrl_info) pqi_fail_all_outstanding_requests(ctrl_info); ctrl_info->pqi_mode_enabled = false; } + pqi_host_free_buffer(ctrl_info, &ctrl_info->ctrl_log_memory); pqi_unregister_scsi(ctrl_info); if (ctrl_info->pqi_mode_enabled) pqi_revert_to_sis_mode(ctrl_info); @@ -8822,170 +8860,6 @@ static void pqi_ofa_ctrl_unquiesce(struct pqi_ctrl_info *ctrl_info) pqi_ctrl_unblock_scan(ctrl_info); } -static int pqi_ofa_alloc_mem(struct pqi_ctrl_info *ctrl_info, u32 total_size, u32 chunk_size) -{ - int i; - u32 sg_count; - struct device *dev; - struct pqi_ofa_memory *ofap; - struct pqi_sg_descriptor *mem_descriptor; - dma_addr_t dma_handle; - - ofap = ctrl_info->pqi_ofa_mem_virt_addr; - - sg_count = DIV_ROUND_UP(total_size, chunk_size); - if (sg_count == 0 || sg_count > PQI_OFA_MAX_SG_DESCRIPTORS) - goto out; - - ctrl_info->pqi_ofa_chunk_virt_addr = kmalloc_array(sg_count, sizeof(void *), GFP_KERNEL); - if (!ctrl_info->pqi_ofa_chunk_virt_addr) - goto out; - - dev = &ctrl_info->pci_dev->dev; - - for (i = 0; i < sg_count; i++) { - ctrl_info->pqi_ofa_chunk_virt_addr[i] = - dma_alloc_coherent(dev, chunk_size, &dma_handle, GFP_KERNEL); - if (!ctrl_info->pqi_ofa_chunk_virt_addr[i]) - goto out_free_chunks; - mem_descriptor = &ofap->sg_descriptor[i]; - put_unaligned_le64((u64)dma_handle, &mem_descriptor->address); - put_unaligned_le32(chunk_size, &mem_descriptor->length); - } - - put_unaligned_le32(CISS_SG_LAST, &mem_descriptor->flags); - put_unaligned_le16(sg_count, &ofap->num_memory_descriptors); - put_unaligned_le32(sg_count * chunk_size, &ofap->bytes_allocated); - - return 0; - -out_free_chunks: - while (--i >= 0) { - mem_descriptor = &ofap->sg_descriptor[i]; - dma_free_coherent(dev, chunk_size, - ctrl_info->pqi_ofa_chunk_virt_addr[i], - get_unaligned_le64(&mem_descriptor->address)); - } - kfree(ctrl_info->pqi_ofa_chunk_virt_addr); - -out: - return -ENOMEM; -} - -static int pqi_ofa_alloc_host_buffer(struct pqi_ctrl_info *ctrl_info) -{ - u32 total_size; - u32 chunk_size; - u32 min_chunk_size; - - if (ctrl_info->ofa_bytes_requested == 0) - return 0; - - total_size = PAGE_ALIGN(ctrl_info->ofa_bytes_requested); - min_chunk_size = DIV_ROUND_UP(total_size, PQI_OFA_MAX_SG_DESCRIPTORS); - min_chunk_size = PAGE_ALIGN(min_chunk_size); - - for (chunk_size = total_size; chunk_size >= min_chunk_size;) { - if (pqi_ofa_alloc_mem(ctrl_info, total_size, chunk_size) == 0) - return 0; - chunk_size /= 2; - chunk_size = PAGE_ALIGN(chunk_size); - } - - return -ENOMEM; -} - -static void pqi_ofa_setup_host_buffer(struct pqi_ctrl_info *ctrl_info) -{ - struct device *dev; - struct pqi_ofa_memory *ofap; - - dev = &ctrl_info->pci_dev->dev; - - ofap = dma_alloc_coherent(dev, sizeof(*ofap), - &ctrl_info->pqi_ofa_mem_dma_handle, GFP_KERNEL); - if (!ofap) - return; - - ctrl_info->pqi_ofa_mem_virt_addr = ofap; - - if (pqi_ofa_alloc_host_buffer(ctrl_info) < 0) { - dev_err(dev, - "failed to allocate host buffer for Online Firmware Activation\n"); - dma_free_coherent(dev, sizeof(*ofap), ofap, ctrl_info->pqi_ofa_mem_dma_handle); - ctrl_info->pqi_ofa_mem_virt_addr = NULL; - return; - } - - put_unaligned_le16(PQI_OFA_VERSION, &ofap->version); - memcpy(&ofap->signature, PQI_OFA_SIGNATURE, sizeof(ofap->signature)); -} - -static void pqi_ofa_free_host_buffer(struct pqi_ctrl_info *ctrl_info) -{ - unsigned int i; - struct device *dev; - struct pqi_ofa_memory *ofap; - struct pqi_sg_descriptor *mem_descriptor; - unsigned int num_memory_descriptors; - - ofap = ctrl_info->pqi_ofa_mem_virt_addr; - if (!ofap) - return; - - dev = &ctrl_info->pci_dev->dev; - - if (get_unaligned_le32(&ofap->bytes_allocated) == 0) - goto out; - - mem_descriptor = ofap->sg_descriptor; - num_memory_descriptors = - get_unaligned_le16(&ofap->num_memory_descriptors); - - for (i = 0; i < num_memory_descriptors; i++) { - dma_free_coherent(dev, - get_unaligned_le32(&mem_descriptor[i].length), - ctrl_info->pqi_ofa_chunk_virt_addr[i], - get_unaligned_le64(&mem_descriptor[i].address)); - } - kfree(ctrl_info->pqi_ofa_chunk_virt_addr); - -out: - dma_free_coherent(dev, sizeof(*ofap), ofap, - ctrl_info->pqi_ofa_mem_dma_handle); - ctrl_info->pqi_ofa_mem_virt_addr = NULL; -} - -static int pqi_ofa_host_memory_update(struct pqi_ctrl_info *ctrl_info) -{ - u32 buffer_length; - struct pqi_vendor_general_request request; - struct pqi_ofa_memory *ofap; - - memset(&request, 0, sizeof(request)); - - request.header.iu_type = PQI_REQUEST_IU_VENDOR_GENERAL; - put_unaligned_le16(sizeof(request) - PQI_REQUEST_HEADER_LENGTH, - &request.header.iu_length); - put_unaligned_le16(PQI_VENDOR_GENERAL_HOST_MEMORY_UPDATE, - &request.function_code); - - ofap = ctrl_info->pqi_ofa_mem_virt_addr; - - if (ofap) { - buffer_length = offsetof(struct pqi_ofa_memory, sg_descriptor) + - get_unaligned_le16(&ofap->num_memory_descriptors) * - sizeof(struct pqi_sg_descriptor); - - put_unaligned_le64((u64)ctrl_info->pqi_ofa_mem_dma_handle, - &request.data.ofa_memory_allocation.buffer_address); - put_unaligned_le32(buffer_length, - &request.data.ofa_memory_allocation.buffer_length); - } - - return pqi_submit_raid_request_synchronous(ctrl_info, &request.header, 0, NULL); -} - static int pqi_ofa_ctrl_restart(struct pqi_ctrl_info *ctrl_info, unsigned int delay_secs) { ssleep(delay_secs); @@ -8993,6 +8867,180 @@ static int pqi_ofa_ctrl_restart(struct pqi_ctrl_info *ctrl_info, unsigned int de return pqi_ctrl_init_resume(ctrl_info); } +static int pqi_host_alloc_mem(struct pqi_ctrl_info *ctrl_info, + struct pqi_host_memory_descriptor *host_memory_descriptor, + u32 total_size, u32 chunk_size) +{ + int i; + u32 sg_count; + struct device *dev; + struct pqi_host_memory *host_memory; + struct pqi_sg_descriptor *mem_descriptor; + dma_addr_t dma_handle; + + sg_count = DIV_ROUND_UP(total_size, chunk_size); + if (sg_count == 0 || sg_count > PQI_HOST_MAX_SG_DESCRIPTORS) + goto out; + + host_memory_descriptor->host_chunk_virt_address = kmalloc(sg_count * sizeof(void *), GFP_KERNEL); + if (!host_memory_descriptor->host_chunk_virt_address) + goto out; + + dev = &ctrl_info->pci_dev->dev; + host_memory = host_memory_descriptor->host_memory; + + for (i = 0; i < sg_count; i++) { + host_memory_descriptor->host_chunk_virt_address[i] = dma_alloc_coherent(dev, chunk_size, &dma_handle, GFP_KERNEL); + if (!host_memory_descriptor->host_chunk_virt_address[i]) + goto out_free_chunks; + mem_descriptor = &host_memory->sg_descriptor[i]; + put_unaligned_le64((u64)dma_handle, &mem_descriptor->address); + put_unaligned_le32(chunk_size, &mem_descriptor->length); + } + + put_unaligned_le32(CISS_SG_LAST, &mem_descriptor->flags); + put_unaligned_le16(sg_count, &host_memory->num_memory_descriptors); + put_unaligned_le32(sg_count * chunk_size, &host_memory->bytes_allocated); + + return 0; + +out_free_chunks: + while (--i >= 0) { + mem_descriptor = &host_memory->sg_descriptor[i]; + dma_free_coherent(dev, chunk_size, + host_memory_descriptor->host_chunk_virt_address[i], + get_unaligned_le64(&mem_descriptor->address)); + } + kfree(host_memory_descriptor->host_chunk_virt_address); +out: + return -ENOMEM; +} + +static int pqi_host_alloc_buffer(struct pqi_ctrl_info *ctrl_info, + struct pqi_host_memory_descriptor *host_memory_descriptor, + u32 total_required_size, u32 min_required_size) +{ + u32 chunk_size; + u32 min_chunk_size; + + if (total_required_size == 0 || min_required_size == 0) + return 0; + + total_required_size = PAGE_ALIGN(total_required_size); + min_required_size = PAGE_ALIGN(min_required_size); + min_chunk_size = DIV_ROUND_UP(total_required_size, PQI_HOST_MAX_SG_DESCRIPTORS); + min_chunk_size = PAGE_ALIGN(min_chunk_size); + + while (total_required_size >= min_required_size) { + for (chunk_size = total_required_size; chunk_size >= min_chunk_size;) { + if (pqi_host_alloc_mem(ctrl_info, + host_memory_descriptor, total_required_size, + chunk_size) == 0) + return 0; + chunk_size /= 2; + chunk_size = PAGE_ALIGN(chunk_size); + } + total_required_size /= 2; + total_required_size = PAGE_ALIGN(total_required_size); + } + + return -ENOMEM; +} + +static void pqi_host_setup_buffer(struct pqi_ctrl_info *ctrl_info, + struct pqi_host_memory_descriptor *host_memory_descriptor, + u32 total_size, u32 min_size) +{ + struct device *dev; + struct pqi_host_memory *host_memory; + + dev = &ctrl_info->pci_dev->dev; + + host_memory = dma_alloc_coherent(dev, sizeof(*host_memory), + &host_memory_descriptor->host_memory_dma_handle, GFP_KERNEL); + if (!host_memory) + return; + + host_memory_descriptor->host_memory = host_memory; + + if (pqi_host_alloc_buffer(ctrl_info, host_memory_descriptor, + total_size, min_size) < 0) { + dev_err(dev, "failed to allocate firmware usable host buffer\n"); + dma_free_coherent(dev, sizeof(*host_memory), host_memory, + host_memory_descriptor->host_memory_dma_handle); + host_memory_descriptor->host_memory = NULL; + return; + } +} + +static void pqi_host_free_buffer(struct pqi_ctrl_info *ctrl_info, + struct pqi_host_memory_descriptor *host_memory_descriptor) +{ + unsigned int i; + struct device *dev; + struct pqi_host_memory *host_memory; + struct pqi_sg_descriptor *mem_descriptor; + unsigned int num_memory_descriptors; + + host_memory = host_memory_descriptor->host_memory; + if (!host_memory) + return; + + dev = &ctrl_info->pci_dev->dev; + + if (get_unaligned_le32(&host_memory->bytes_allocated) == 0) + goto out; + + mem_descriptor = host_memory->sg_descriptor; + num_memory_descriptors = get_unaligned_le16(&host_memory->num_memory_descriptors); + + for (i = 0; i < num_memory_descriptors; i++) { + dma_free_coherent(dev, + get_unaligned_le32(&mem_descriptor[i].length), + host_memory_descriptor->host_chunk_virt_address[i], + get_unaligned_le64(&mem_descriptor[i].address)); + } + kfree(host_memory_descriptor->host_chunk_virt_address); + +out: + dma_free_coherent(dev, sizeof(*host_memory), host_memory, + host_memory_descriptor->host_memory_dma_handle); + host_memory_descriptor->host_memory = NULL; +} + +static int pqi_host_memory_update(struct pqi_ctrl_info *ctrl_info, + struct pqi_host_memory_descriptor *host_memory_descriptor, + u16 function_code) +{ + u32 buffer_length; + struct pqi_vendor_general_request request; + struct pqi_host_memory *host_memory; + + memset(&request, 0, sizeof(request)); + + request.header.iu_type = PQI_REQUEST_IU_VENDOR_GENERAL; + put_unaligned_le16(sizeof(request) - PQI_REQUEST_HEADER_LENGTH, &request.header.iu_length); + put_unaligned_le16(function_code, &request.function_code); + + host_memory = host_memory_descriptor->host_memory; + + if (host_memory) { + buffer_length = offsetof(struct pqi_host_memory, sg_descriptor) + get_unaligned_le16(&host_memory->num_memory_descriptors) * sizeof(struct pqi_sg_descriptor); + put_unaligned_le64((u64)host_memory_descriptor->host_memory_dma_handle, &request.data.host_memory_allocation.buffer_address); + put_unaligned_le32(buffer_length, &request.data.host_memory_allocation.buffer_length); + + if (function_code == PQI_VENDOR_GENERAL_OFA_MEMORY_UPDATE) { + put_unaligned_le16(PQI_OFA_VERSION, &host_memory->version); + memcpy(&host_memory->signature, PQI_OFA_SIGNATURE, sizeof(host_memory->signature)); + } else if (function_code == PQI_VENDOR_GENERAL_CTRL_LOG_MEMORY_UPDATE) { + put_unaligned_le16(PQI_CTRL_LOG_VERSION, &host_memory->version); + memcpy(&host_memory->signature, PQI_CTRL_LOG_SIGNATURE, sizeof(host_memory->signature)); + } + } + + return pqi_submit_raid_request_synchronous(ctrl_info, &request.header, 0, NULL); +} + static struct pqi_raid_error_info pqi_ctrl_offline_raid_error_info = { .data_out_result = PQI_DATA_IN_OUT_HARDWARE_ERROR, .status = SAM_STAT_CHECK_CONDITION, diff --git a/drivers/scsi/smartpqi/smartpqi_sis.c b/drivers/scsi/smartpqi/smartpqi_sis.c index 673437c7152b..ca1df36b83f7 100644 --- a/drivers/scsi/smartpqi/smartpqi_sis.c +++ b/drivers/scsi/smartpqi/smartpqi_sis.c @@ -29,6 +29,7 @@ #define SIS_ENABLE_INTX 0x80 #define SIS_SOFT_RESET 0x100 #define SIS_CMD_READY 0x200 +#define SIS_NOTIFY_KDUMP 0x400 #define SIS_TRIGGER_SHUTDOWN 0x800000 #define SIS_PQI_RESET_QUIESCE 0x1000000 @@ -52,6 +53,8 @@ #define SIS_BASE_STRUCT_ALIGNMENT 16 #define SIS_CTRL_KERNEL_FW_TRIAGE 0x3 +#define SIS_CTRL_KERNEL_CTRL_LOGGING 0x4 +#define SIS_CTRL_KERNEL_CTRL_LOGGING_STATUS 0x18 #define SIS_CTRL_KERNEL_UP 0x80 #define SIS_CTRL_KERNEL_PANIC 0x100 #define SIS_CTRL_READY_TIMEOUT_SECS 180 @@ -65,6 +68,13 @@ enum sis_fw_triage_status { FW_TRIAGE_COMPLETED }; +enum sis_ctrl_logging_status { + CTRL_LOGGING_NOT_STARTED = 0, + CTRL_LOGGING_STARTED, + CTRL_LOGGING_COND_INVALID, + CTRL_LOGGING_COMPLETED +}; + #pragma pack(1) /* for use with SIS_CMD_INIT_BASE_STRUCT_ADDRESS command */ @@ -442,6 +452,21 @@ static inline enum sis_fw_triage_status SIS_CTRL_KERNEL_FW_TRIAGE)); } +bool sis_is_ctrl_logging_supported(struct pqi_ctrl_info *ctrl_info) +{ + return readl(&ctrl_info->registers->sis_firmware_status) & SIS_CTRL_KERNEL_CTRL_LOGGING; +} + +void sis_notify_kdump(struct pqi_ctrl_info *ctrl_info) +{ + sis_set_doorbell_bit(ctrl_info, SIS_NOTIFY_KDUMP); +} + +static inline enum sis_ctrl_logging_status sis_read_ctrl_logging_status(struct pqi_ctrl_info *ctrl_info) +{ + return ((enum sis_ctrl_logging_status)((readl(&ctrl_info->registers->sis_firmware_status) & SIS_CTRL_KERNEL_CTRL_LOGGING_STATUS) >> 3)); +} + void sis_soft_reset(struct pqi_ctrl_info *ctrl_info) { writel(SIS_SOFT_RESET, @@ -484,6 +509,41 @@ int sis_wait_for_fw_triage_completion(struct pqi_ctrl_info *ctrl_info) return rc; } +#define SIS_CTRL_LOGGING_STATUS_TIMEOUT_SECS 180 +#define SIS_CTRL_LOGGING_STATUS_POLL_INTERVAL_SECS 1 + +int sis_wait_for_ctrl_logging_completion(struct pqi_ctrl_info *ctrl_info) +{ + int rc; + enum sis_ctrl_logging_status status; + unsigned long timeout; + + timeout = (SIS_CTRL_LOGGING_STATUS_TIMEOUT_SECS * HZ) + jiffies; + while (1) { + status = sis_read_ctrl_logging_status(ctrl_info); + if (status == CTRL_LOGGING_COND_INVALID) { + dev_err(&ctrl_info->pci_dev->dev, + "controller data logging condition invalid\n"); + rc = -EINVAL; + break; + } else if (status == CTRL_LOGGING_COMPLETED) { + rc = 0; + break; + } + + if (time_after(jiffies, timeout)) { + dev_err(&ctrl_info->pci_dev->dev, + "timed out waiting for controller data logging status\n"); + rc = -ETIMEDOUT; + break; + } + + ssleep(SIS_CTRL_LOGGING_STATUS_POLL_INTERVAL_SECS); + } + + return rc; +} + void sis_verify_structures(void) { BUILD_BUG_ON(offsetof(struct sis_base_struct, diff --git a/drivers/scsi/smartpqi/smartpqi_sis.h b/drivers/scsi/smartpqi/smartpqi_sis.h index 0c97626d87d4..7e0eac3d07de 100644 --- a/drivers/scsi/smartpqi/smartpqi_sis.h +++ b/drivers/scsi/smartpqi/smartpqi_sis.h @@ -31,6 +31,9 @@ u32 sis_read_driver_scratch(struct pqi_ctrl_info *ctrl_info); void sis_soft_reset(struct pqi_ctrl_info *ctrl_info); u32 sis_get_product_id(struct pqi_ctrl_info *ctrl_info); int sis_wait_for_fw_triage_completion(struct pqi_ctrl_info *ctrl_info); +bool sis_is_ctrl_logging_supported(struct pqi_ctrl_info *ctrl_info); +void sis_notify_kdump(struct pqi_ctrl_info *ctrl_info); +int sis_wait_for_ctrl_logging_completion(struct pqi_ctrl_info *ctrl_info); extern unsigned int sis_ctrl_ready_timeout_secs; From 4c76114932d1d6fad2e72823e7898a3c960cf2a7 Mon Sep 17 00:00:00 2001 From: Mahesh Rajashekhara Date: Tue, 27 Aug 2024 13:54:56 -0500 Subject: [PATCH 087/123] scsi: smartpqi: correct stream detection Correct stream detection by initializing the structure pqi_scsi_dev_raid_map_data to 0s. When the OS issues SCSI READ commands, the driver erroneously considers them as SCSI WRITES. If they are identified as sequential IOs, the driver then submits those requests via the RAID path instead of the AIO path. The 'is_write' flag might be set for SCSI READ commands also. The driver may interpret SCSI READ commands as SCSI WRITE commands, resulting in IOs being submitted through the RAID path. Note: This does not cause data corruption. Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Signed-off-by: Mahesh Rajashekhara Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-3-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 54f7fe843445..8e2e71ab49ae 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -5942,7 +5942,7 @@ static bool pqi_is_parity_write_stream(struct pqi_ctrl_info *ctrl_info, int rc; struct pqi_scsi_dev *device; struct pqi_stream_data *pqi_stream_data; - struct pqi_scsi_dev_raid_map_data rmd; + struct pqi_scsi_dev_raid_map_data rmd = { 0 }; if (!ctrl_info->enable_stream_detection) return false; From 283dcc1b142ebd60786f8f5e3fbbd53a51035739 Mon Sep 17 00:00:00 2001 From: Mahesh Rajashekhara Date: Tue, 27 Aug 2024 13:54:57 -0500 Subject: [PATCH 088/123] scsi: smartpqi: add counter for parity write stream requests Add sysfs entry to check for write stream requests. Move existing raid_bypass_cnt into a structure named pqi_raid_io_stats and add member write_stream_cnt. These two counters are related because write stream detection is only checked if an I/O request is eligible for bypass (AIO). Example usage: lsscsi [15:1:0:0] disk Adaptec LOGICAL VOLUME 0129 /dev/sdae cat /sys/block/sdae/device/ssd_smart_path_enabled 1 ^ | +---- NOTE: here bypass has been enabled on device sdae To read the counter for parity write stream requests: cat /sys/block/sdae/device/write_stream_cnt 0x60cd507 Reviewed-by: Scott Benesh Reviewed-by: Mike McGowen Signed-off-by: Mahesh Rajashekhara Co-developed-by: Kevin Barnett Signed-off-by: Kevin Barnett Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-4-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi.h | 7 ++- drivers/scsi/smartpqi/smartpqi_init.c | 62 +++++++++++++++++++++------ 2 files changed, 55 insertions(+), 14 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi.h b/drivers/scsi/smartpqi/smartpqi.h index f493006bee9d..fae6db20a6e9 100644 --- a/drivers/scsi/smartpqi/smartpqi.h +++ b/drivers/scsi/smartpqi/smartpqi.h @@ -1106,6 +1106,11 @@ struct pqi_tmf_work { u8 scsi_opcode; }; +struct pqi_raid_io_stats { + u64 raid_bypass_cnt; + u64 write_stream_cnt; +}; + struct pqi_scsi_dev { int devtype; /* as reported by INQUIRY command */ u8 device_type; /* as reported by */ @@ -1168,7 +1173,7 @@ struct pqi_scsi_dev { struct pqi_stream_data stream_data[NUM_STREAMS_PER_LUN]; atomic_t scsi_cmds_outstanding[PQI_MAX_LUNS_PER_DEVICE]; - u64 __percpu *raid_bypass_cnt; + struct pqi_raid_io_stats __percpu *raid_io_stats; struct pqi_tmf_work tmf_work[PQI_MAX_LUNS_PER_DEVICE]; }; diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 8e2e71ab49ae..6a941735c982 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -1508,8 +1508,8 @@ static int pqi_get_raid_map(struct pqi_ctrl_info *ctrl_info, if (rc) goto error; - device->raid_bypass_cnt = alloc_percpu(u64); - if (!device->raid_bypass_cnt) { + device->raid_io_stats = alloc_percpu(struct pqi_raid_io_stats); + if (!device->raid_io_stats) { rc = -ENOMEM; goto error; } @@ -2105,9 +2105,9 @@ static void pqi_scsi_update_device(struct pqi_ctrl_info *ctrl_info, /* To prevent this from being freed later. */ new_device->raid_map = NULL; } - if (new_device->raid_bypass_enabled && existing_device->raid_bypass_cnt == NULL) { - existing_device->raid_bypass_cnt = new_device->raid_bypass_cnt; - new_device->raid_bypass_cnt = NULL; + if (new_device->raid_bypass_enabled && existing_device->raid_io_stats == NULL) { + existing_device->raid_io_stats = new_device->raid_io_stats; + new_device->raid_io_stats = NULL; } existing_device->raid_bypass_configured = new_device->raid_bypass_configured; existing_device->raid_bypass_enabled = new_device->raid_bypass_enabled; @@ -2131,7 +2131,7 @@ static void pqi_scsi_update_device(struct pqi_ctrl_info *ctrl_info, static inline void pqi_free_device(struct pqi_scsi_dev *device) { if (device) { - free_percpu(device->raid_bypass_cnt); + free_percpu(device->raid_io_stats); kfree(device->raid_map); kfree(device); } @@ -5984,6 +5984,7 @@ static bool pqi_is_parity_write_stream(struct pqi_ctrl_info *ctrl_info, pqi_stream_data->next_lba = rmd.first_block + rmd.block_cnt; pqi_stream_data->last_accessed = jiffies; + per_cpu_ptr(device->raid_io_stats, smp_processor_id())->write_stream_cnt++; return true; } @@ -6016,7 +6017,6 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm u16 hw_queue; struct pqi_queue_group *queue_group; bool raid_bypassed; - u64 *raid_bypass_cnt; u8 lun; scmd->host_scribble = PQI_NO_COMPLETION; @@ -6063,8 +6063,7 @@ static int pqi_scsi_queue_command(struct Scsi_Host *shost, struct scsi_cmnd *scm rc = pqi_raid_bypass_submit_scsi_cmd(ctrl_info, device, scmd, queue_group); if (rc == 0 || rc == SCSI_MLQUEUE_HOST_BUSY) { raid_bypassed = true; - raid_bypass_cnt = per_cpu_ptr(device->raid_bypass_cnt, smp_processor_id()); - (*raid_bypass_cnt)++; + per_cpu_ptr(device->raid_io_stats, smp_processor_id())->raid_bypass_cnt++; } } if (!raid_bypassed) @@ -7363,7 +7362,6 @@ static ssize_t pqi_raid_bypass_cnt_show(struct device *dev, unsigned long flags; u64 raid_bypass_cnt; int cpu; - u64 *per_cpu_bypass_cnt_ptr; sdev = to_scsi_device(dev); ctrl_info = shost_to_hba(sdev->host); @@ -7381,10 +7379,9 @@ static ssize_t pqi_raid_bypass_cnt_show(struct device *dev, raid_bypass_cnt = 0; - if (device->raid_bypass_cnt) { + if (device->raid_io_stats) { for_each_online_cpu(cpu) { - per_cpu_bypass_cnt_ptr = per_cpu_ptr(device->raid_bypass_cnt, cpu); - raid_bypass_cnt += *per_cpu_bypass_cnt_ptr; + raid_bypass_cnt += per_cpu_ptr(device->raid_io_stats, cpu)->raid_bypass_cnt; } } @@ -7472,6 +7469,43 @@ static ssize_t pqi_numa_node_show(struct device *dev, return scnprintf(buffer, PAGE_SIZE, "%d\n", ctrl_info->numa_node); } +static ssize_t pqi_write_stream_cnt_show(struct device *dev, + struct device_attribute *attr, char *buffer) +{ + struct pqi_ctrl_info *ctrl_info; + struct scsi_device *sdev; + struct pqi_scsi_dev *device; + unsigned long flags; + u64 write_stream_cnt; + int cpu; + + sdev = to_scsi_device(dev); + ctrl_info = shost_to_hba(sdev->host); + + if (pqi_ctrl_offline(ctrl_info)) + return -ENODEV; + + spin_lock_irqsave(&ctrl_info->scsi_device_list_lock, flags); + + device = sdev->hostdata; + if (!device) { + spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); + return -ENODEV; + } + + write_stream_cnt = 0; + + if (device->raid_io_stats) { + for_each_online_cpu(cpu) { + write_stream_cnt += per_cpu_ptr(device->raid_io_stats, cpu)->write_stream_cnt; + } + } + + spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); + + return scnprintf(buffer, PAGE_SIZE, "0x%llx\n", write_stream_cnt); +} + static DEVICE_ATTR(lunid, 0444, pqi_lunid_show, NULL); static DEVICE_ATTR(unique_id, 0444, pqi_unique_id_show, NULL); static DEVICE_ATTR(path_info, 0444, pqi_path_info_show, NULL); @@ -7482,6 +7516,7 @@ static DEVICE_ATTR(raid_bypass_cnt, 0444, pqi_raid_bypass_cnt_show, NULL); static DEVICE_ATTR(sas_ncq_prio_enable, 0644, pqi_sas_ncq_prio_enable_show, pqi_sas_ncq_prio_enable_store); static DEVICE_ATTR(numa_node, 0444, pqi_numa_node_show, NULL); +static DEVICE_ATTR(write_stream_cnt, 0444, pqi_write_stream_cnt_show, NULL); static struct attribute *pqi_sdev_attrs[] = { &dev_attr_lunid.attr, @@ -7493,6 +7528,7 @@ static struct attribute *pqi_sdev_attrs[] = { &dev_attr_raid_bypass_cnt.attr, &dev_attr_sas_ncq_prio_enable.attr, &dev_attr_numa_node.attr, + &dev_attr_write_stream_cnt.attr, NULL }; From dbc39b84540f746cc814e69b21e53e6d3e12329a Mon Sep 17 00:00:00 2001 From: David Strahan Date: Tue, 27 Aug 2024 13:54:58 -0500 Subject: [PATCH 089/123] scsi: smartpqi: add new controller PCI IDs All PCI ID entries in Hex. Add new cisco pci ids: VID / DID / SVID / SDID ---- ---- ---- ---- 9005 028f 1137 02fe 9005 028f 1137 02ff 9005 028f 1137 0300 Add new h3c pci ids: VID / DID / SVID / SDID ---- ---- ---- ---- 9005 028f 193d 0462 9005 028f 193d 8462 Add new ieit pci ids: VID / DID / SVID / SDID ---- ---- ---- ---- 9005 028f 1ff9 00a3 Reviewed-by: Scott Benesh Reviewed-by: Mike McGowen Signed-off-by: David Strahan Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-5-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 6a941735c982..46bef2cf95c4 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -9548,6 +9548,10 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x152d, 0x8a37) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x193d, 0x0462) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x193d, 0x1104) @@ -9588,6 +9592,10 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x193d, 0x8461) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x193d, 0x8462) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x193d, 0xc460) @@ -10296,6 +10304,18 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1137, 0x02fa) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1137, 0x02fe) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1137, 0x02ff) + }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1137, 0x0300) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1ff9, 0x0045) @@ -10472,6 +10492,10 @@ static const struct pci_device_id pqi_pci_id_table[] = { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, 0x1f51, 0x1045) }, + { + PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, + 0x1ff9, 0x00a3) + }, { PCI_DEVICE_SUB(PCI_VENDOR_ID_ADAPTEC2, 0x028f, PCI_ANY_ID, PCI_ANY_ID) From 4e0a51716dae782822b5df6c3b29cc0915a9f802 Mon Sep 17 00:00:00 2001 From: Murthy Bhat Date: Tue, 27 Aug 2024 13:54:59 -0500 Subject: [PATCH 090/123] scsi: smartpqi: fix rare system hang during LUN reset Correct a rare case where in a LUN reset occurs on a device and I/O requests for other devices persist in the driver's internal request queue. Part of a LUN reset involves waiting for our internal request queue to empty before proceeding. The internal request queue contains requests not yet sent down to the controller. We were clearing the requests queued for the LUN undergoing a reset, but not all of the queued requests. Causing a hang. For all requests in our internal request queue: Complete requests with DID_RESET for queued requests for the device undergoing a reset. Complete requests with DID_REQUEUE for all other queued requests. Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Signed-off-by: Murthy Bhat Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-6-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 46bef2cf95c4..d1d117d5d08d 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -6200,14 +6200,12 @@ static void pqi_fail_io_queued_for_device(struct pqi_ctrl_info *ctrl_info, continue; scsi_device = scmd->device->hostdata; - if (scsi_device != device) - continue; - - if ((u8)scmd->device->lun != lun) - continue; list_del(&io_request->request_list_entry); - set_host_byte(scmd, DID_RESET); + if (scsi_device == device && (u8)scmd->device->lun == lun) + set_host_byte(scmd, DID_RESET); + else + set_host_byte(scmd, DID_REQUEUE); pqi_free_io_request(io_request); scsi_dma_unmap(scmd); pqi_scsi_done(scmd); From 07dde72ff173742137546cb4e5e9264c8e1ba2ba Mon Sep 17 00:00:00 2001 From: Don Brace Date: Tue, 27 Aug 2024 13:55:00 -0500 Subject: [PATCH 091/123] scsi: smartpqi: fix volume size updates Correct logical volume size changes by moving the check for a volume rescan outside of the check for a queue depth change. Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-7-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 22 ++++++++++++++-------- 1 file changed, 14 insertions(+), 8 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index d1d117d5d08d..0dab30753f0a 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -2303,17 +2303,23 @@ static void pqi_update_device_list(struct pqi_ctrl_info *ctrl_info, * queue depth, device size. */ list_for_each_entry(device, &ctrl_info->scsi_device_list, scsi_device_list_entry) { + /* + * Check for queue depth change. + */ if (device->sdev && device->queue_depth != device->advertised_queue_depth) { device->advertised_queue_depth = device->queue_depth; scsi_change_queue_depth(device->sdev, device->advertised_queue_depth); - spin_lock_irqsave(&ctrl_info->scsi_device_list_lock, flags); - if (pqi_volume_rescan_needed(device)) { - device->rescan = false; - spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); - scsi_rescan_device(device->sdev); - } else { - spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); - } + } + spin_lock_irqsave(&ctrl_info->scsi_device_list_lock, flags); + /* + * Check for changes in the device, such as size. + */ + if (pqi_volume_rescan_needed(device)) { + device->rescan = false; + spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); + scsi_rescan_device(device->sdev); + } else { + spin_unlock_irqrestore(&ctrl_info->scsi_device_list_lock, flags); } } From bda1c931e2993cbd43515dccec8bdd200a39cbf0 Mon Sep 17 00:00:00 2001 From: Don Brace Date: Tue, 27 Aug 2024 13:55:01 -0500 Subject: [PATCH 092/123] scsi: smartpqi: update driver version to 2.1.30-031 Update driver version to 2.1.30-031. Reviewed-by: David Strahan Reviewed-by: Scott Benesh Reviewed-by: Scott Teel Reviewed-by: Mike McGowen Reviewed-by: Gerry Morong Signed-off-by: Don Brace Link: https://lore.kernel.org/r/20240827185501.692804-8-don.brace@microchip.com Signed-off-by: Martin K. Petersen --- drivers/scsi/smartpqi/smartpqi_init.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/scsi/smartpqi/smartpqi_init.c b/drivers/scsi/smartpqi/smartpqi_init.c index 0dab30753f0a..7fd5a8c813dc 100644 --- a/drivers/scsi/smartpqi/smartpqi_init.c +++ b/drivers/scsi/smartpqi/smartpqi_init.c @@ -33,11 +33,11 @@ #define BUILD_TIMESTAMP #endif -#define DRIVER_VERSION "2.1.28-025" +#define DRIVER_VERSION "2.1.30-031" #define DRIVER_MAJOR 2 #define DRIVER_MINOR 1 -#define DRIVER_RELEASE 28 -#define DRIVER_REVISION 25 +#define DRIVER_RELEASE 30 +#define DRIVER_REVISION 31 #define DRIVER_NAME "Microchip SmartPQI Driver (v" \ DRIVER_VERSION BUILD_TIMESTAMP ")" From 833bd284a45448f04ab9fcd4fdf7ef594d458eed Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Wed, 28 Aug 2024 12:51:24 +0300 Subject: [PATCH 093/123] arm64: mm: fix DMA zone when dma-ranges is missing Some platforms, like Rockchip RK3568 based Odroid M1, do not provide DMA limits information in device-tree dma-ranges property. Still some device drivers set DMA limit that relies on DMA zone at low 4GB memory area. Until commit ba0fb44aed47 ("dma-mapping: replace zone_dma_bits by zone_dma_limit"), zone_sizes_init() restricted DMA zone to low 32-bit. Restore DMA zone 32-bit limit when the platform provides no DMA bus limit information. Fixes: ba0fb44aed47 ("dma-mapping: replace zone_dma_bits by zone_dma_limit") Reported-by: Marek Szyprowski Link: https://lore.kernel.org/r/53d988b1-bdce-422a-ae4e-158f305ad703@samsung.com Suggested-by: Robin Murphy Signed-off-by: Baruch Siach Reviewed-by: Catalin Marinas Tested-by: Marek Szyprowski Signed-off-by: Christoph Hellwig --- arch/arm64/mm/init.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index bfb10969cbf0..5710129b8033 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -116,6 +116,9 @@ static void __init arch_reserve_crashkernel(void) static phys_addr_t __init max_zone_phys(phys_addr_t zone_limit) { + if (zone_limit == PHYS_ADDR_MAX) + zone_limit = U32_MAX; + return min(zone_limit, memblock_end_of_DRAM() - 1) + 1; } From b5a73bf4d1de95e620bf5f592557b81f71c76f0e Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 22 Aug 2024 06:56:31 +0200 Subject: [PATCH 094/123] scsi: check that busses support the DMA API before setting dma parameters We'll start throwing warnings soon when dma_set_seg_boundary and dma_set_max_seg_size are called on devices for buses that don't fully support the DMA API. Prepare for that by making the calls in the SCSI midlayer conditional. Signed-off-by: Christoph Hellwig Reviewed-by: Robin Murphy Reviewed-by: Martin K. Petersen --- drivers/scsi/scsi_lib.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index 3958a6d14bf4..7f0394c44920 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -1988,8 +1988,15 @@ void scsi_init_limits(struct Scsi_Host *shost, struct queue_limits *lim) if (shost->no_highmem) lim->features |= BLK_FEAT_BOUNCE_HIGH; - dma_set_seg_boundary(dev, shost->dma_boundary); - dma_set_max_seg_size(dev, shost->max_segment_size); + /* + * Propagate the DMA formation properties to the dma-mapping layer as + * a courtesy service to the LLDDs. This needs to check that the buses + * actually support the DMA API first, though. + */ + if (dev->dma_parms) { + dma_set_seg_boundary(dev, shost->dma_boundary); + dma_set_max_seg_size(dev, shost->max_segment_size); + } } EXPORT_SYMBOL_GPL(scsi_init_limits); From c42a01264ba1497eb3193c08ff3c2656d98250a6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 19 Jul 2024 06:10:24 +0200 Subject: [PATCH 095/123] dma-mapping: don't return errors from dma_set_min_align_mask A NULL dev->dma_parms indicates either a bus that is not DMA capable or grave bug in the implementation of the bus code. There isn't much the driver can do in terms of error handling for either case, so just warn and continue as DMA operations will fail anyway. Signed-off-by: Christoph Hellwig Reviewed-by: Robin Murphy Reviewed-by: Martin K. Petersen --- include/linux/dma-mapping.h | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h index f693aafe221f..cfd6bafec3f9 100644 --- a/include/linux/dma-mapping.h +++ b/include/linux/dma-mapping.h @@ -575,13 +575,12 @@ static inline unsigned int dma_get_min_align_mask(struct device *dev) return 0; } -static inline int dma_set_min_align_mask(struct device *dev, +static inline void dma_set_min_align_mask(struct device *dev, unsigned int min_align_mask) { if (WARN_ON_ONCE(!dev->dma_parms)) - return -EIO; + return; dev->dma_parms->min_align_mask = min_align_mask; - return 0; } #ifndef dma_get_cache_alignment From 560a861ab4174b42240157ab5cebe36b8c7bc418 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 19 Jul 2024 06:06:28 +0200 Subject: [PATCH 096/123] dma-mapping: don't return errors from dma_set_seg_boundary A NULL dev->dma_parms indicates either a bus that is not DMA capable or grave bug in the implementation of the bus code. There isn't much the driver can do in terms of error handling for either case, so just warn and continue as DMA operations will fail anyway. Signed-off-by: Christoph Hellwig Reviewed-by: Robin Murphy Reviewed-by: Martin K. Petersen --- include/linux/dma-mapping.h | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-) diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h index cfd6bafec3f9..6bd1333dbacb 100644 --- a/include/linux/dma-mapping.h +++ b/include/linux/dma-mapping.h @@ -559,13 +559,11 @@ static inline unsigned long dma_get_seg_boundary_nr_pages(struct device *dev, return (dma_get_seg_boundary(dev) >> page_shift) + 1; } -static inline int dma_set_seg_boundary(struct device *dev, unsigned long mask) +static inline void dma_set_seg_boundary(struct device *dev, unsigned long mask) { - if (dev->dma_parms) { - dev->dma_parms->segment_boundary_mask = mask; - return 0; - } - return -EIO; + if (WARN_ON_ONCE(!dev->dma_parms)) + return; + dev->dma_parms->segment_boundary_mask = mask; } static inline unsigned int dma_get_min_align_mask(struct device *dev) From 334304ac2baca7f3e821c47cf5129d90e7a6b1e6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Fri, 19 Jul 2024 06:07:38 +0200 Subject: [PATCH 097/123] dma-mapping: don't return errors from dma_set_max_seg_size A NULL dev->dma_parms indicates either a bus that is not DMA capable or grave bug in the implementation of the bus code. There isn't much the driver can do in terms of error handling for either case, so just warn and continue as DMA operations will fail anyway. Signed-off-by: Christoph Hellwig Reviewed-by: Robin Murphy Reviewed-by: Martin K. Petersen Acked-by: Ulf Hansson # For MMC --- drivers/accel/qaic/qaic_drv.c | 4 +--- drivers/dma/idma64.c | 4 +--- drivers/dma/pl330.c | 5 +---- drivers/dma/qcom/bam_dma.c | 6 +----- drivers/dma/sh/rcar-dmac.c | 4 +--- drivers/dma/ste_dma40.c | 6 +----- drivers/gpu/drm/mediatek/mtk_drm_drv.c | 6 +----- drivers/media/common/videobuf2/videobuf2-dma-contig.c | 3 +-- drivers/media/pci/intel/ipu6/ipu6.c | 4 +--- drivers/mmc/host/mmci_stm32_sdmmc.c | 3 ++- drivers/net/ethernet/microsoft/mana/gdma_main.c | 6 +----- drivers/scsi/lpfc/lpfc_init.c | 7 +------ include/linux/dma-mapping.h | 10 ++++------ 13 files changed, 17 insertions(+), 51 deletions(-) diff --git a/drivers/accel/qaic/qaic_drv.c b/drivers/accel/qaic/qaic_drv.c index 580b29ed1902..bf10156c334e 100644 --- a/drivers/accel/qaic/qaic_drv.c +++ b/drivers/accel/qaic/qaic_drv.c @@ -447,9 +447,7 @@ static int init_pci(struct qaic_device *qdev, struct pci_dev *pdev) ret = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64)); if (ret) return ret; - ret = dma_set_max_seg_size(&pdev->dev, UINT_MAX); - if (ret) - return ret; + dma_set_max_seg_size(&pdev->dev, UINT_MAX); qdev->bar_0 = devm_ioremap_resource(&pdev->dev, &pdev->resource[0]); if (IS_ERR(qdev->bar_0)) diff --git a/drivers/dma/idma64.c b/drivers/dma/idma64.c index e3505e56784b..1398814d8fbb 100644 --- a/drivers/dma/idma64.c +++ b/drivers/dma/idma64.c @@ -598,9 +598,7 @@ static int idma64_probe(struct idma64_chip *chip) idma64->dma.dev = chip->sysdev; - ret = dma_set_max_seg_size(idma64->dma.dev, IDMA64C_CTLH_BLOCK_TS_MASK); - if (ret) - return ret; + dma_set_max_seg_size(idma64->dma.dev, IDMA64C_CTLH_BLOCK_TS_MASK); ret = dma_async_device_register(&idma64->dma); if (ret) diff --git a/drivers/dma/pl330.c b/drivers/dma/pl330.c index 60c4de8dac1d..82a9fe88ad54 100644 --- a/drivers/dma/pl330.c +++ b/drivers/dma/pl330.c @@ -3163,10 +3163,7 @@ pl330_probe(struct amba_device *adev, const struct amba_id *id) * This is the limit for transfers with a buswidth of 1, larger * buswidths will have larger limits. */ - ret = dma_set_max_seg_size(&adev->dev, 1900800); - if (ret) - dev_err(&adev->dev, "unable to set the seg size\n"); - + dma_set_max_seg_size(&adev->dev, 1900800); init_pl330_debugfs(pl330); dev_info(&adev->dev, diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c index 5e7d332731e0..368ffaa40037 100644 --- a/drivers/dma/qcom/bam_dma.c +++ b/drivers/dma/qcom/bam_dma.c @@ -1325,11 +1325,7 @@ static int bam_dma_probe(struct platform_device *pdev) /* set max dma segment size */ bdev->common.dev = bdev->dev; - ret = dma_set_max_seg_size(bdev->common.dev, BAM_FIFO_SIZE); - if (ret) { - dev_err(bdev->dev, "cannot set maximum segment size\n"); - goto err_bam_channel_exit; - } + dma_set_max_seg_size(bdev->common.dev, BAM_FIFO_SIZE); platform_set_drvdata(pdev, bdev); diff --git a/drivers/dma/sh/rcar-dmac.c b/drivers/dma/sh/rcar-dmac.c index 40482cb73d79..1094a2f82164 100644 --- a/drivers/dma/sh/rcar-dmac.c +++ b/drivers/dma/sh/rcar-dmac.c @@ -1868,9 +1868,7 @@ static int rcar_dmac_probe(struct platform_device *pdev) dmac->dev = &pdev->dev; platform_set_drvdata(pdev, dmac); - ret = dma_set_max_seg_size(dmac->dev, RCAR_DMATCR_MASK); - if (ret) - return ret; + dma_set_max_seg_size(dmac->dev, RCAR_DMATCR_MASK); ret = dma_set_mask_and_coherent(dmac->dev, DMA_BIT_MASK(40)); if (ret) diff --git a/drivers/dma/ste_dma40.c b/drivers/dma/ste_dma40.c index 2c489299148e..d52e1685aed5 100644 --- a/drivers/dma/ste_dma40.c +++ b/drivers/dma/ste_dma40.c @@ -3632,11 +3632,7 @@ static int __init d40_probe(struct platform_device *pdev) if (ret) goto destroy_cache; - ret = dma_set_max_seg_size(base->dev, STEDMA40_MAX_SEG_SIZE); - if (ret) { - d40_err(dev, "Failed to set dma max seg size\n"); - goto destroy_cache; - } + dma_set_max_seg_size(base->dev, STEDMA40_MAX_SEG_SIZE); d40_hw_init(base); diff --git a/drivers/gpu/drm/mediatek/mtk_drm_drv.c b/drivers/gpu/drm/mediatek/mtk_drm_drv.c index 77b50c56c124..3e807195a0d0 100644 --- a/drivers/gpu/drm/mediatek/mtk_drm_drv.c +++ b/drivers/gpu/drm/mediatek/mtk_drm_drv.c @@ -559,11 +559,7 @@ static int mtk_drm_kms_init(struct drm_device *drm) * Configure the DMA segment size to make sure we get contiguous IOVA * when importing PRIME buffers. */ - ret = dma_set_max_seg_size(dma_dev, UINT_MAX); - if (ret) { - dev_err(dma_dev, "Failed to set DMA segment size\n"); - goto err_component_unbind; - } + dma_set_max_seg_size(dma_dev, UINT_MAX); ret = drm_vblank_init(drm, MAX_CRTC); if (ret < 0) diff --git a/drivers/media/common/videobuf2/videobuf2-dma-contig.c b/drivers/media/common/videobuf2/videobuf2-dma-contig.c index 3d4fd4ef5310..bb0b7fa67b53 100644 --- a/drivers/media/common/videobuf2/videobuf2-dma-contig.c +++ b/drivers/media/common/videobuf2/videobuf2-dma-contig.c @@ -854,8 +854,7 @@ int vb2_dma_contig_set_max_seg_size(struct device *dev, unsigned int size) return -ENODEV; } if (dma_get_max_seg_size(dev) < size) - return dma_set_max_seg_size(dev, size); - + dma_set_max_seg_size(dev, size); return 0; } EXPORT_SYMBOL_GPL(vb2_dma_contig_set_max_seg_size); diff --git a/drivers/media/pci/intel/ipu6/ipu6.c b/drivers/media/pci/intel/ipu6/ipu6.c index bbd646378ab3..83e70c692d95 100644 --- a/drivers/media/pci/intel/ipu6/ipu6.c +++ b/drivers/media/pci/intel/ipu6/ipu6.c @@ -576,9 +576,7 @@ static int ipu6_pci_probe(struct pci_dev *pdev, const struct pci_device_id *id) if (ret) return dev_err_probe(dev, ret, "Failed to set DMA mask\n"); - ret = dma_set_max_seg_size(dev, UINT_MAX); - if (ret) - return dev_err_probe(dev, ret, "Failed to set max_seg_size\n"); + dma_set_max_seg_size(dev, UINT_MAX); ret = ipu6_pci_config_setup(pdev, isp->hw_ver); if (ret) diff --git a/drivers/mmc/host/mmci_stm32_sdmmc.c b/drivers/mmc/host/mmci_stm32_sdmmc.c index f5da7f9baa52..9dc51859c2e5 100644 --- a/drivers/mmc/host/mmci_stm32_sdmmc.c +++ b/drivers/mmc/host/mmci_stm32_sdmmc.c @@ -213,7 +213,8 @@ static int sdmmc_idma_setup(struct mmci_host *host) host->mmc->max_seg_size = host->mmc->max_req_size; } - return dma_set_max_seg_size(dev, host->mmc->max_seg_size); + dma_set_max_seg_size(dev, host->mmc->max_seg_size); + return 0; } static int sdmmc_idma_start(struct mmci_host *host, unsigned int *datactrl) diff --git a/drivers/net/ethernet/microsoft/mana/gdma_main.c b/drivers/net/ethernet/microsoft/mana/gdma_main.c index ddb8f68d80a2..ca4ed58f1206 100644 --- a/drivers/net/ethernet/microsoft/mana/gdma_main.c +++ b/drivers/net/ethernet/microsoft/mana/gdma_main.c @@ -1496,11 +1496,7 @@ static int mana_gd_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (err) goto release_region; - err = dma_set_max_seg_size(&pdev->dev, UINT_MAX); - if (err) { - dev_err(&pdev->dev, "Failed to set dma device segment size\n"); - goto release_region; - } + dma_set_max_seg_size(&pdev->dev, UINT_MAX); err = -ENOMEM; gc = vzalloc(sizeof(*gc)); diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c index e1dfa96c2a55..50620918becd 100644 --- a/drivers/scsi/lpfc/lpfc_init.c +++ b/drivers/scsi/lpfc/lpfc_init.c @@ -13861,12 +13861,7 @@ fcponly: if (sli4_params->sge_supp_len > LPFC_MAX_SGE_SIZE) sli4_params->sge_supp_len = LPFC_MAX_SGE_SIZE; - rc = dma_set_max_seg_size(&phba->pcidev->dev, sli4_params->sge_supp_len); - if (unlikely(rc)) { - lpfc_printf_log(phba, KERN_INFO, LOG_INIT, - "6400 Can't set dma maximum segment size\n"); - return rc; - } + dma_set_max_seg_size(&phba->pcidev->dev, sli4_params->sge_supp_len); /* * Check whether the adapter supports an embedded copy of the diff --git a/include/linux/dma-mapping.h b/include/linux/dma-mapping.h index 6bd1333dbacb..1524da363734 100644 --- a/include/linux/dma-mapping.h +++ b/include/linux/dma-mapping.h @@ -524,13 +524,11 @@ static inline unsigned int dma_get_max_seg_size(struct device *dev) return SZ_64K; } -static inline int dma_set_max_seg_size(struct device *dev, unsigned int size) +static inline void dma_set_max_seg_size(struct device *dev, unsigned int size) { - if (dev->dma_parms) { - dev->dma_parms->max_segment_size = size; - return 0; - } - return -EIO; + if (WARN_ON_ONCE(!dev->dma_parms)) + return; + dev->dma_parms->max_segment_size = size; } static inline unsigned long dma_get_seg_boundary(struct device *dev) From 122c234ef4e182440f8a60490e9344bfc4e2b5e9 Mon Sep 17 00:00:00 2001 From: Baruch Siach Date: Sun, 1 Sep 2024 09:12:34 +0300 Subject: [PATCH 098/123] arm64: mm: keep low RAM dma zone Commit ba0fb44aed47 ("dma-mapping: replace zone_dma_bits by zone_dma_limit") optimistically assumed that device-tree dma-ranges property describes the system DMA limits. That assumption ignores DMA limits of individual devices that are not encoded in device tree. Commit 833bd284a45 ("arm64: mm: fix DMA zone when dma-ranges is missing") fixed part of the problem for platforms that do not provide dma-ranges at all. However platforms like SM8550-HDK provide DMA bus limit, but have devices with stronger DMA limits. of_dma_get_max_cpu_address() does not take device limitations into account. These platforms implicitly rely on DMA zone in low 32-bit RAM area. Until we find a better way to figure out the optimal DMA zone range, restore the low RAM DMA zone we had before commit ba0fb44aed47. Fixes: ba0fb44aed47 ("dma-mapping: replace zone_dma_bits by zone_dma_limit") Closes: https://lore.kernel.org/r/1a0c7282-63e0-4add-8e38-3abe3e0a8e2f@linaro.org Reported-by: Neil Armstrong Reviewed-by: Catalin Marinas Suggested-by: Robin Murphy Signed-off-by: Baruch Siach Tested-by: Neil Armstrong # on SM8550-HDK Tested-by: Marek Szyprowski Signed-off-by: Christoph Hellwig --- arch/arm64/mm/init.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/arch/arm64/mm/init.c b/arch/arm64/mm/init.c index 5710129b8033..eb76ac6d87c0 100644 --- a/arch/arm64/mm/init.c +++ b/arch/arm64/mm/init.c @@ -116,8 +116,14 @@ static void __init arch_reserve_crashkernel(void) static phys_addr_t __init max_zone_phys(phys_addr_t zone_limit) { - if (zone_limit == PHYS_ADDR_MAX) - zone_limit = U32_MAX; + /** + * Information we get from firmware (e.g. DT dma-ranges) describe DMA + * bus constraints. Devices using DMA might have their own limitations. + * Some of them rely on DMA zone in low 32-bit memory. Keep low RAM + * DMA zone on platforms that have RAM there. + */ + if (memblock_start_of_DRAM() < U32_MAX) + zone_limit = min(zone_limit, U32_MAX); return min(zone_limit, memblock_end_of_DRAM() - 1) + 1; } From 2749749afa071f8a0e405605de9da615e771a7ce Mon Sep 17 00:00:00 2001 From: Jiawei Ye Date: Mon, 2 Sep 2024 08:47:26 +0000 Subject: [PATCH 099/123] smackfs: Use rcu_assign_pointer() to ensure safe assignment in smk_set_cipso In the `smk_set_cipso` function, the `skp->smk_netlabel.attr.mls.cat` field is directly assigned to a new value without using the appropriate RCU pointer assignment functions. According to RCU usage rules, this is illegal and can lead to unpredictable behavior, including data inconsistencies and impossible-to-diagnose memory corruption issues. This possible bug was identified using a static analysis tool developed by myself, specifically designed to detect RCU-related issues. To address this, the assignment is now done using rcu_assign_pointer(), which ensures that the pointer assignment is done safely, with the necessary memory barriers and synchronization. This change prevents potential RCU dereference issues by ensuring that the `cat` field is safely updated while still adhering to RCU's requirements. Fixes: 0817534ff9ea ("smackfs: Fix use-after-free in netlbl_catmap_walk()") Signed-off-by: Jiawei Ye Signed-off-by: Casey Schaufler --- security/smack/smackfs.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/security/smack/smackfs.c b/security/smack/smackfs.c index e22aad7604e8..5dd1e164f9b1 100644 --- a/security/smack/smackfs.c +++ b/security/smack/smackfs.c @@ -932,7 +932,7 @@ static ssize_t smk_set_cipso(struct file *file, const char __user *buf, } if (rc >= 0) { old_cat = skp->smk_netlabel.attr.mls.cat; - skp->smk_netlabel.attr.mls.cat = ncats.attr.mls.cat; + rcu_assign_pointer(skp->smk_netlabel.attr.mls.cat, ncats.attr.mls.cat); skp->smk_netlabel.attr.mls.lvl = ncats.attr.mls.lvl; synchronize_rcu(); netlbl_catmap_free(old_cat); From 92193b3569ade4968011007356c8606def0657a1 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 28 Aug 2024 09:00:08 +0300 Subject: [PATCH 100/123] vdpa_sim: don't select DMA_OPS vdpa_sim has been fixed to not override the dma_map_ops in commit 6c3d329e6486 ("vdpa_sim: get rid of DMA ops"), so don't select the symbol and don't depend on HAS_DMA. Signed-off-by: Christoph Hellwig Acked-by: Michael S. Tsirkin --- drivers/vdpa/Kconfig | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/vdpa/Kconfig b/drivers/vdpa/Kconfig index 5265d09fc1c4..b08de3b77061 100644 --- a/drivers/vdpa/Kconfig +++ b/drivers/vdpa/Kconfig @@ -11,8 +11,7 @@ if VDPA config VDPA_SIM tristate "vDPA device simulator core" - depends on RUNTIME_TESTING_MENU && HAS_DMA - select DMA_OPS + depends on RUNTIME_TESTING_MENU select VHOST_RING select IOMMU_IOVA help From de6c85bf918ea52d5c680f0d130b37ee2ff152d6 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 28 Aug 2024 09:02:47 +0300 Subject: [PATCH 101/123] dma-mapping: clearly mark DMA ops as an architecture feature DMA ops are a helper for architectures and not for drivers to override the DMA implementation. Unfortunately driver authors keep ignoring this. Make the fact more clear by renaming the symbol to ARCH_HAS_DMA_OPS and having the two drivers overriding their dma_ops depend on that. These drivers should probably be marked broken, but we can give them a bit of a grace period for that. Signed-off-by: Christoph Hellwig Reviewed-by: Thomas Gleixner Acked-by: Sakari Ailus # for IPU6 Acked-by: Robin Murphy --- arch/Kconfig | 9 +++++++++ arch/alpha/Kconfig | 2 +- arch/arm/Kconfig | 2 +- arch/arm64/Kconfig | 1 + arch/mips/Kconfig | 2 +- arch/parisc/Kconfig | 2 +- arch/powerpc/Kconfig | 2 +- arch/s390/Kconfig | 2 +- arch/sparc/Kconfig | 2 +- arch/x86/Kconfig | 2 +- drivers/macintosh/macio_asic.c | 4 ++-- drivers/media/pci/intel/ipu6/Kconfig | 7 ++++++- drivers/vdpa/Kconfig | 7 ++++++- drivers/xen/Kconfig | 4 ++-- include/linux/device.h | 2 +- include/linux/dma-map-ops.h | 6 +++--- kernel/dma/Kconfig | 9 ++------- kernel/dma/Makefile | 2 +- 18 files changed, 41 insertions(+), 26 deletions(-) diff --git a/arch/Kconfig b/arch/Kconfig index 975dd22a2dbd..61c4ec048787 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -17,6 +17,15 @@ config CPU_MITIGATIONS def_bool y endif +# +# Selected by architectures that need custom DMA operations for e.g. legacy +# IOMMUs not handled by dma-iommu. Drivers must never select this symbol. +# +config ARCH_HAS_DMA_OPS + depends on HAS_DMA + select DMA_OPS_HELPERS + bool + menu "General architecture-dependent options" config ARCH_HAS_SUBPAGE_FAULTS diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig index 50ff06d5b799..109a4cddcd13 100644 --- a/arch/alpha/Kconfig +++ b/arch/alpha/Kconfig @@ -4,12 +4,12 @@ config ALPHA default y select ARCH_32BIT_USTAT_F_TINODE select ARCH_HAS_CURRENT_STACK_POINTER + select ARCH_HAS_DMA_OPS if PCI select ARCH_MIGHT_HAVE_PC_PARPORT select ARCH_MIGHT_HAVE_PC_SERIO select ARCH_NO_PREEMPT select ARCH_NO_SG_CHAIN select ARCH_USE_CMPXCHG_LOCKREF - select DMA_OPS if PCI select FORCE_PCI select PCI_DOMAINS if PCI select PCI_SYSCALL if PCI diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 54b2bb817a7f..f5f7995a2f8f 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -10,6 +10,7 @@ config ARM select ARCH_HAS_CURRENT_STACK_POINTER select ARCH_HAS_DEBUG_VIRTUAL if MMU select ARCH_HAS_DMA_ALLOC if MMU + select ARCH_HAS_DMA_OPS select ARCH_HAS_DMA_WRITE_COMBINE if !ARM_DMA_MEM_BUFFERABLE select ARCH_HAS_ELF_RANDOMIZE select ARCH_HAS_FORTIFY_SOURCE @@ -54,7 +55,6 @@ config ARM select DCACHE_WORD_ACCESS if HAVE_EFFICIENT_UNALIGNED_ACCESS select DMA_DECLARE_COHERENT select DMA_GLOBAL_POOL if !MMU - select DMA_OPS select DMA_NONCOHERENT_MMAP if MMU select EDAC_SUPPORT select EDAC_ATOMIC_SCRUB diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index a2f8ff354ca6..40940cbde435 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -24,6 +24,7 @@ config ARM64 select ARCH_HAS_CURRENT_STACK_POINTER select ARCH_HAS_DEBUG_VIRTUAL select ARCH_HAS_DEBUG_VM_PGTABLE + select ARCH_HAS_DMA_OPS if XEN select ARCH_HAS_DMA_PREP_COHERENT select ARCH_HAS_ACPI_TABLE_UPGRADE if ACPI select ARCH_HAS_FAST_MULTIPLIER diff --git a/arch/mips/Kconfig b/arch/mips/Kconfig index 60077e576935..023ad33a7e94 100644 --- a/arch/mips/Kconfig +++ b/arch/mips/Kconfig @@ -8,6 +8,7 @@ config MIPS select ARCH_HAS_CPU_FINALIZE_INIT select ARCH_HAS_CURRENT_STACK_POINTER if !CC_IS_CLANG || CLANG_VERSION >= 140000 select ARCH_HAS_DEBUG_VIRTUAL if !64BIT + select ARCH_HAS_DMA_OPS if MACH_JAZZ select ARCH_HAS_FORTIFY_SOURCE select ARCH_HAS_KCOV select ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE if !EVA @@ -393,7 +394,6 @@ config MACH_JAZZ select ARC_PROMLIB select ARCH_MIGHT_HAVE_PC_PARPORT select ARCH_MIGHT_HAVE_PC_SERIO - select DMA_OPS select FW_ARC select FW_ARC32 select ARCH_MAY_HAVE_PC_FDC diff --git a/arch/parisc/Kconfig b/arch/parisc/Kconfig index b0a2ac3ba916..859835a0692c 100644 --- a/arch/parisc/Kconfig +++ b/arch/parisc/Kconfig @@ -10,6 +10,7 @@ config PARISC select ARCH_WANT_FRAME_POINTERS select ARCH_HAS_CPU_CACHE_ALIASING select ARCH_HAS_DMA_ALLOC if PA11 + select ARCH_HAS_DMA_OPS select ARCH_HAS_ELF_RANDOMIZE select ARCH_HAS_STRICT_KERNEL_RWX select ARCH_HAS_STRICT_MODULE_RWX @@ -23,7 +24,6 @@ config PARISC select ARCH_HAS_CACHE_LINE_SIZE select ARCH_HAS_DEBUG_VM_PGTABLE select HAVE_RELIABLE_STACKTRACE - select DMA_OPS select RTC_CLASS select RTC_DRV_GENERIC select INIT_ALL_POSSIBLE diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index d7b09b064a8a..f87fc0375a92 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -133,6 +133,7 @@ config PPC select ARCH_HAS_DEBUG_WX if STRICT_KERNEL_RWX select ARCH_HAS_DEVMEM_IS_ALLOWED select ARCH_HAS_DMA_MAP_DIRECT if PPC_PSERIES + select ARCH_HAS_DMA_OPS if PPC64 select ARCH_HAS_FORTIFY_SOURCE select ARCH_HAS_GCOV_PROFILE_ALL select ARCH_HAS_KCOV @@ -185,7 +186,6 @@ config PPC select CPUMASK_OFFSTACK if NR_CPUS >= 8192 select DCACHE_WORD_ACCESS if PPC64 && CPU_LITTLE_ENDIAN select DMA_OPS_BYPASS if PPC64 - select DMA_OPS if PPC64 select DYNAMIC_FTRACE if FUNCTION_TRACER select EDAC_ATOMIC_SCRUB select EDAC_SUPPORT diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig index a822f952f64a..f31f6b85c25c 100644 --- a/arch/s390/Kconfig +++ b/arch/s390/Kconfig @@ -70,6 +70,7 @@ config S390 select ARCH_HAS_DEBUG_VM_PGTABLE select ARCH_HAS_DEBUG_WX select ARCH_HAS_DEVMEM_IS_ALLOWED + select ARCH_HAS_DMA_OPS if PCI select ARCH_HAS_ELF_RANDOMIZE select ARCH_HAS_FORCE_DMA_UNENCRYPTED select ARCH_HAS_FORTIFY_SOURCE @@ -137,7 +138,6 @@ config S390 select BUILDTIME_TABLE_SORT select CLONE_BACKWARDS2 select DCACHE_WORD_ACCESS if !KMSAN - select DMA_OPS if PCI select DYNAMIC_FTRACE if FUNCTION_TRACER select FUNCTION_ALIGNMENT_8B if CC_IS_GCC select FUNCTION_ALIGNMENT_16B if !CC_IS_GCC diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig index 11bf9d312318..dcfdb7f1dae9 100644 --- a/arch/sparc/Kconfig +++ b/arch/sparc/Kconfig @@ -14,9 +14,9 @@ config SPARC bool default y select ARCH_HAS_CPU_CACHE_ALIASING + select ARCH_HAS_DMA_OPS select ARCH_MIGHT_HAVE_PC_PARPORT if SPARC64 && PCI select ARCH_MIGHT_HAVE_PC_SERIO - select DMA_OPS select OF select OF_PROMTREE select HAVE_ASM_MODVERSIONS diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 007bab9f2a0e..9e2e7d361019 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -79,6 +79,7 @@ config X86 select ARCH_HAS_DEBUG_VIRTUAL select ARCH_HAS_DEBUG_VM_PGTABLE if !X86_PAE select ARCH_HAS_DEVMEM_IS_ALLOWED + select ARCH_HAS_DMA_OPS if GART_IOMMU || XEN select ARCH_HAS_EARLY_DEBUG if KGDB select ARCH_HAS_ELF_RANDOMIZE select ARCH_HAS_FAST_MULTIPLIER @@ -943,7 +944,6 @@ config DMI config GART_IOMMU bool "Old AMD GART IOMMU support" - select DMA_OPS select IOMMU_HELPER select SWIOTLB depends on X86_64 && PCI && AMD_NB diff --git a/drivers/macintosh/macio_asic.c b/drivers/macintosh/macio_asic.c index 13626205530d..bede200e32e8 100644 --- a/drivers/macintosh/macio_asic.c +++ b/drivers/macintosh/macio_asic.c @@ -387,7 +387,7 @@ static struct macio_dev * macio_add_one_device(struct macio_chip *chip, dma_set_max_seg_size(&dev->ofdev.dev, 65536); dma_set_seg_boundary(&dev->ofdev.dev, 0xffffffff); -#if defined(CONFIG_PCI) && defined(CONFIG_DMA_OPS) +#if defined(CONFIG_PCI) && defined(CONFIG_ARCH_HAS_DMA_OPS) /* Set the DMA ops to the ones from the PCI device, this could be * fishy if we didn't know that on PowerMac it's always direct ops * or iommu ops that will work fine @@ -396,7 +396,7 @@ static struct macio_dev * macio_add_one_device(struct macio_chip *chip, */ dev->ofdev.dev.archdata = chip->lbus.pdev->dev.archdata; dev->ofdev.dev.dma_ops = chip->lbus.pdev->dev.dma_ops; -#endif /* CONFIG_PCI && CONFIG_DMA_OPS */ +#endif /* CONFIG_PCI && CONFIG_ARCH_HAS_DMA_OPS */ #ifdef DEBUG printk("preparing mdev @%p, ofdev @%p, dev @%p, kobj @%p\n", diff --git a/drivers/media/pci/intel/ipu6/Kconfig b/drivers/media/pci/intel/ipu6/Kconfig index 40e20f0aa5ae..49e4fb696573 100644 --- a/drivers/media/pci/intel/ipu6/Kconfig +++ b/drivers/media/pci/intel/ipu6/Kconfig @@ -4,8 +4,13 @@ config VIDEO_INTEL_IPU6 depends on VIDEO_DEV depends on X86 && X86_64 && HAS_DMA depends on IPU_BRIDGE || !IPU_BRIDGE + # + # This driver incorrectly tries to override the dma_ops. It should + # never have done that, but for now keep it working on architectures + # that use dma ops + # + depends on ARCH_HAS_DMA_OPS select AUXILIARY_BUS - select DMA_OPS select IOMMU_IOVA select VIDEO_V4L2_SUBDEV_API select MEDIA_CONTROLLER diff --git a/drivers/vdpa/Kconfig b/drivers/vdpa/Kconfig index b08de3b77061..559fb9d3271f 100644 --- a/drivers/vdpa/Kconfig +++ b/drivers/vdpa/Kconfig @@ -35,7 +35,12 @@ config VDPA_SIM_BLOCK config VDPA_USER tristate "VDUSE (vDPA Device in Userspace) support" depends on EVENTFD && MMU && HAS_DMA - select DMA_OPS + # + # This driver incorrectly tries to override the dma_ops. It should + # never have done that, but for now keep it working on architectures + # that use dma ops + # + depends on ARCH_HAS_DMA_OPS select VHOST_IOTLB select IOMMU_IOVA help diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig index d5989871dd5d..f7d6f47971fd 100644 --- a/drivers/xen/Kconfig +++ b/drivers/xen/Kconfig @@ -177,8 +177,8 @@ config XEN_GRANT_DMA_ALLOC config SWIOTLB_XEN def_bool y + depends on ARCH_HAS_DMA_OPS depends on XEN_PV || ARM || ARM64 - select DMA_OPS select SWIOTLB config XEN_PCI_STUB @@ -348,10 +348,10 @@ config XEN_GRANT_DMA_IOMMU config XEN_GRANT_DMA_OPS bool - select DMA_OPS config XEN_VIRTIO bool "Xen virtio support" + depends on ARCH_HAS_DMA_OPS depends on VIRTIO select XEN_GRANT_DMA_OPS select XEN_GRANT_DMA_IOMMU if OF diff --git a/include/linux/device.h b/include/linux/device.h index 1c5280d28bc3..b4bde8d22697 100644 --- a/include/linux/device.h +++ b/include/linux/device.h @@ -750,7 +750,7 @@ struct device { struct dev_pin_info *pins; #endif struct dev_msi_info msi; -#ifdef CONFIG_DMA_OPS +#ifdef CONFIG_ARCH_HAS_DMA_OPS const struct dma_map_ops *dma_ops; #endif u64 *dma_mask; /* dma mask (if dma'able device) */ diff --git a/include/linux/dma-map-ops.h b/include/linux/dma-map-ops.h index 077b15c93bb8..9668ddf3696e 100644 --- a/include/linux/dma-map-ops.h +++ b/include/linux/dma-map-ops.h @@ -75,7 +75,7 @@ struct dma_map_ops { unsigned long (*get_merge_boundary)(struct device *dev); }; -#ifdef CONFIG_DMA_OPS +#ifdef CONFIG_ARCH_HAS_DMA_OPS #include static inline const struct dma_map_ops *get_dma_ops(struct device *dev) @@ -90,7 +90,7 @@ static inline void set_dma_ops(struct device *dev, { dev->dma_ops = dma_ops; } -#else /* CONFIG_DMA_OPS */ +#else /* CONFIG_ARCH_HAS_DMA_OPS */ static inline const struct dma_map_ops *get_dma_ops(struct device *dev) { return NULL; @@ -99,7 +99,7 @@ static inline void set_dma_ops(struct device *dev, const struct dma_map_ops *dma_ops) { } -#endif /* CONFIG_DMA_OPS */ +#endif /* CONFIG_ARCH_HAS_DMA_OPS */ #ifdef CONFIG_DMA_CMA extern struct cma *dma_contiguous_default_area; diff --git a/kernel/dma/Kconfig b/kernel/dma/Kconfig index 21bae1700836..4c0dcd909121 100644 --- a/kernel/dma/Kconfig +++ b/kernel/dma/Kconfig @@ -11,11 +11,6 @@ config HAS_DMA config DMA_OPS_HELPERS bool -config DMA_OPS - depends on HAS_DMA - select DMA_OPS_HELPERS - bool - # # IOMMU drivers that can bypass the IOMMU code and optionally use the direct # mapping fast path should select this option and set the dma_ops_bypass @@ -113,8 +108,8 @@ config DMA_BOUNCE_UNALIGNED_KMALLOC config DMA_NEED_SYNC def_bool ARCH_HAS_SYNC_DMA_FOR_DEVICE || ARCH_HAS_SYNC_DMA_FOR_CPU || \ - ARCH_HAS_SYNC_DMA_FOR_CPU_ALL || DMA_API_DEBUG || DMA_OPS || \ - SWIOTLB + ARCH_HAS_SYNC_DMA_FOR_CPU_ALL || DMA_API_DEBUG || \ + ARCH_HAS_DMA_OPS || SWIOTLB config DMA_RESTRICTED_POOL bool "DMA Restricted Pool" diff --git a/kernel/dma/Makefile b/kernel/dma/Makefile index 2e6e933cf7f3..6977033444a3 100644 --- a/kernel/dma/Makefile +++ b/kernel/dma/Makefile @@ -2,7 +2,7 @@ obj-$(CONFIG_HAS_DMA) += mapping.o direct.o obj-$(CONFIG_DMA_OPS_HELPERS) += ops_helpers.o -obj-$(CONFIG_DMA_OPS) += dummy.o +obj-$(CONFIG_ARCH_HAS_DMA_OPS) += dummy.o obj-$(CONFIG_DMA_CMA) += contiguous.o obj-$(CONFIG_DMA_DECLARE_COHERENT) += coherent.o obj-$(CONFIG_DMA_API_DEBUG) += debug.o From f689a3ab7b8ece9e5787ff058b96b8630e4931ad Mon Sep 17 00:00:00 2001 From: Chen Yu Date: Sat, 31 Aug 2024 19:01:19 +0800 Subject: [PATCH 102/123] dma-direct: optimize page freeing when it is not addressable When the CMA allocation succeeds but isn't addressable, its buffer has already been released and the page is set to NULL. So later when the normal page allocation succeeds but isn't addressable, __free_pages() can be used to free that normal page rather than using dma_free_contiguous that does extra checks that are not needed. Signed-off-by: Chen Yu Signed-off-by: Christoph Hellwig --- kernel/dma/direct.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/kernel/dma/direct.c b/kernel/dma/direct.c index f2ba074a6a54..5b4e6d3bf7bc 100644 --- a/kernel/dma/direct.c +++ b/kernel/dma/direct.c @@ -140,7 +140,7 @@ again: if (!page) page = alloc_pages_node(node, gfp, get_order(size)); if (page && !dma_coherent_ok(dev, page_to_phys(page), size)) { - dma_free_contiguous(dev, page, size); + __free_pages(page, get_order(size)); page = NULL; if (IS_ENABLED(CONFIG_ZONE_DMA32) && From 19156263cb1f24128a9ba6ef7340be5cbacc3d22 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Thu, 5 Sep 2024 10:14:05 +0300 Subject: [PATCH 103/123] dma-mapping: use IOMMU DMA calls for common alloc/free page calls MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Common alloca and free pages routines are called when IOMMU DMA is used, and internally it calls to DMA ops structure which is not available for default IOMMU. This patch adds necessary if checks to call IOMMU DMA. It fixes the following crash: Unable to handle kernel NULL pointer dereference at virtual address 0000000000000040 Mem abort info: ESR = 0x0000000096000006 EC = 0x25: DABT (current EL), IL = 32 bits SET = 0, FnV = 0 EA = 0, S1PTW = 0 FSC = 0x06: level 2 translation fault Data abort info: ISV = 0, ISS = 0x00000006, ISS2 = 0x00000000 CM = 0, WnR = 0, TnD = 0, TagAccess = 0 GCS = 0, Overlay = 0, DirtyBit = 0, Xs = 0 user pgtable: 4k pages, 48-bit VAs, pgdp=00000000d20bb000 [0000000000000040] pgd=08000000d20c1003 , p4d=08000000d20c1003 , pud=08000000d20c2003, pmd=0000000000000000 Internal error: Oops: 0000000096000006 [#1] PREEMPT SMP Modules linked in: ipv6 hci_uart venus_core btqca v4l2_mem2mem btrtl qcom_spmi_adc5 sbs_battery btbcm qcom_vadc_common cros_ec_typec videobuf2_v4l2 leds_cros_ec cros_kbd_led_backlight cros_ec_chardev videodev elan_i2c videobuf2_common qcom_stats mc bluetooth coresight_stm stm_core ecdh_generic ecc pwrseq_core panel_edp icc_bwmon ath10k_snoc ath10k_core ath mac80211 phy_qcom_qmp_combo aux_bridge libarc4 coresight_replicator coresight_etm4x coresight_tmc coresight_funnel cfg80211 rfkill coresight qcom_wdt cbmem ramoops reed_solomon pwm_bl coreboot_table backlight crct10dif_ce CPU: 7 UID: 0 PID: 70 Comm: kworker/u32:4 Not tainted 6.11.0-rc6-next-20240903-00003-gdfc6015d0711 #660 Hardware name: Google Lazor Limozeen without Touchscreen (rev5 - rev8) (DT) Workqueue: events_unbound deferred_probe_work_func hub 2-1:1.0: 4 ports detected pstate: 80400009 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--) pc : dma_common_alloc_pages+0x54/0x1b4 lr : dma_common_alloc_pages+0x4c/0x1b4 sp : ffff8000807d3730 x29: ffff8000807d3730 x28: ffff02a7d312f880 x27: 0000000000000001 x26: 000000000000c000 x25: 0000000000000000 x24: 0000000000000001 x23: ffff02a7d23b6898 x22: 0000000000006cc0 x21: 000000000000c000 x20: ffff02a7858bf410 x19: fffffe0a60006000 x18: 0000000000000001 x17: 00000000000000d5 x16: 1fffe054f0bcc261 x15: 0000000000000001 x14: ffff02a7844dc680 x13: 0000000000100180 x12: dead000000000100 x11: dead000000000122 x10: 00000000001001ff x9 : ffff02a87f7b7b00 x8 : ffff02a87f7b7b00 x7 : ffff405977d6b000 x6 : ffff8000807d3310 x5 : ffff02a87f6b6398 x4 : 0000000000000001 x3 : ffff405977d6b000 x2 : ffff02a7844dc600 x1 : 0000000100000000 x0 : fffffe0a60006000 Call trace: dma_common_alloc_pages+0x54/0x1b4 __dma_alloc_pages+0x68/0x90 dma_alloc_pages+0x10/0x1c snd_dma_noncoherent_alloc+0x28/0x8c __snd_dma_alloc_pages+0x30/0x50 snd_dma_alloc_dir_pages+0x40/0x80 do_alloc_pages+0xb8/0x13c preallocate_pcm_pages+0x6c/0xf8 preallocate_pages+0x160/0x1a4 snd_pcm_set_managed_buffer_all+0x64/0xb0 lpass_platform_pcm_new+0xc0/0xe8 snd_soc_pcm_component_new+0x3c/0xc8 soc_new_pcm+0x4fc/0x668 snd_soc_bind_card+0xabc/0xbac snd_soc_register_card+0xf0/0x108 devm_snd_soc_register_card+0x4c/0xa4 sc7180_snd_platform_probe+0x180/0x224 platform_probe+0x68/0xc0 really_probe+0xbc/0x298 __driver_probe_device+0x78/0x12c driver_probe_device+0x3c/0x15c __device_attach_driver+0xb8/0x134 bus_for_each_drv+0x84/0xe0 __device_attach+0x9c/0x188 device_initial_probe+0x14/0x20 bus_probe_device+0xac/0xb0 deferred_probe_work_func+0x88/0xc0 process_one_work+0x14c/0x28c worker_thread+0x2cc/0x3d4 kthread+0x114/0x118 ret_from_fork+0x10/0x20 Code: f9411c19 940000c9 aa0003f3 b4000460 (f9402326) ---[ end trace 0000000000000000 ]--- Fixes: b5c58b2fdc42 ("dma-mapping: direct calls for dma-iommu") Closes: https://lore.kernel.org/all/10431dfd-ce04-4e0f-973b-c78477303c18@notapiano Reported-by: Nícolas F. R. A. Prado #KernelCI Signed-off-by: Leon Romanovsky Tested-by: Nícolas F. R. A. Prado Signed-off-by: Christoph Hellwig --- include/linux/iommu-dma.h | 8 ++++++++ kernel/dma/mapping.c | 12 ------------ kernel/dma/ops_helpers.c | 14 +++++++++++--- 3 files changed, 19 insertions(+), 15 deletions(-) diff --git a/include/linux/iommu-dma.h b/include/linux/iommu-dma.h index d30a58bf00fd..1bb55ca1ab79 100644 --- a/include/linux/iommu-dma.h +++ b/include/linux/iommu-dma.h @@ -10,6 +10,10 @@ #include #ifdef CONFIG_IOMMU_DMA +static inline bool use_dma_iommu(struct device *dev) +{ + return dev->dma_iommu; +} dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, unsigned long offset, size_t size, enum dma_data_direction dir, unsigned long attrs); @@ -49,6 +53,10 @@ void iommu_dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sgl, void iommu_dma_sync_sg_for_device(struct device *dev, struct scatterlist *sgl, int nelems, enum dma_data_direction dir); #else +static inline bool use_dma_iommu(struct device *dev) +{ + return false; +} static inline dma_addr_t iommu_dma_map_page(struct device *dev, struct page *page, unsigned long offset, size_t size, enum dma_data_direction dir, unsigned long attrs) diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index b50ae3d198a6..056f27962f69 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -114,18 +114,6 @@ void *dmam_alloc_attrs(struct device *dev, size_t size, dma_addr_t *dma_handle, } EXPORT_SYMBOL(dmam_alloc_attrs); -#ifdef CONFIG_IOMMU_DMA -static bool use_dma_iommu(struct device *dev) -{ - return dev->dma_iommu; -} -#else -static bool use_dma_iommu(struct device *dev) -{ - return false; -} -#endif - static bool dma_go_direct(struct device *dev, dma_addr_t mask, const struct dma_map_ops *ops) { diff --git a/kernel/dma/ops_helpers.c b/kernel/dma/ops_helpers.c index af4a6ef48ce0..9afd569eadb9 100644 --- a/kernel/dma/ops_helpers.c +++ b/kernel/dma/ops_helpers.c @@ -4,6 +4,7 @@ * the allocated memory contains normal pages in the direct kernel mapping. */ #include +#include static struct page *dma_common_vaddr_to_page(void *cpu_addr) { @@ -70,8 +71,12 @@ struct page *dma_common_alloc_pages(struct device *dev, size_t size, if (!page) return NULL; - *dma_handle = ops->map_page(dev, page, 0, size, dir, - DMA_ATTR_SKIP_CPU_SYNC); + if (use_dma_iommu(dev)) + *dma_handle = iommu_dma_map_page(dev, page, 0, size, dir, + DMA_ATTR_SKIP_CPU_SYNC); + else + *dma_handle = ops->map_page(dev, page, 0, size, dir, + DMA_ATTR_SKIP_CPU_SYNC); if (*dma_handle == DMA_MAPPING_ERROR) { dma_free_contiguous(dev, page, size); return NULL; @@ -86,7 +91,10 @@ void dma_common_free_pages(struct device *dev, size_t size, struct page *page, { const struct dma_map_ops *ops = get_dma_ops(dev); - if (ops->unmap_page) + if (use_dma_iommu(dev)) + iommu_dma_unmap_page(dev, dma_handle, size, dir, + DMA_ATTR_SKIP_CPU_SYNC); + else if (ops->unmap_page) ops->unmap_page(dev, dma_handle, size, dir, DMA_ATTR_SKIP_CPU_SYNC); dma_free_contiguous(dev, page, size); From 5784d9fcfd43bd853654bb80c87ef293b9e8e80a Mon Sep 17 00:00:00 2001 From: Julian Sun Date: Mon, 2 Sep 2024 11:08:44 +0800 Subject: [PATCH 104/123] ocfs2: fix null-ptr-deref when journal load failed. During the mounting process, if journal_reset() fails because of too short journal, then lead to jbd2_journal_load() fails with NULL j_sb_buffer. Subsequently, ocfs2_journal_shutdown() calls jbd2_journal_flush()->jbd2_cleanup_journal_tail()-> __jbd2_update_log_tail()->jbd2_journal_update_sb_log_tail() ->lock_buffer(journal->j_sb_buffer), resulting in a null-pointer dereference error. To resolve this issue, we should check the JBD2_LOADED flag to ensure the journal was properly loaded. Additionally, use journal instead of osb->journal directly to simplify the code. Link: https://syzkaller.appspot.com/bug?extid=05b9b39d8bdfe1a0861f Link: https://lkml.kernel.org/r/20240902030844.422725-1-sunjunchao2870@gmail.com Fixes: f6f50e28f0cb ("jbd2: Fail to load a journal if it is too short") Signed-off-by: Julian Sun Reported-by: syzbot+05b9b39d8bdfe1a0861f@syzkaller.appspotmail.com Suggested-by: Joseph Qi Reviewed-by: Joseph Qi Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/journal.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c index 530fba34f6d3..1bf188b6866a 100644 --- a/fs/ocfs2/journal.c +++ b/fs/ocfs2/journal.c @@ -1055,7 +1055,7 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb) if (!igrab(inode)) BUG(); - num_running_trans = atomic_read(&(osb->journal->j_num_trans)); + num_running_trans = atomic_read(&(journal->j_num_trans)); trace_ocfs2_journal_shutdown(num_running_trans); /* Do a commit_cache here. It will flush our journal, *and* @@ -1074,9 +1074,10 @@ void ocfs2_journal_shutdown(struct ocfs2_super *osb) osb->commit_task = NULL; } - BUG_ON(atomic_read(&(osb->journal->j_num_trans)) != 0); + BUG_ON(atomic_read(&(journal->j_num_trans)) != 0); - if (ocfs2_mount_local(osb)) { + if (ocfs2_mount_local(osb) && + (journal->j_journal->j_flags & JBD2_LOADED)) { jbd2_journal_lock_updates(journal->j_journal); status = jbd2_journal_flush(journal->j_journal, 0); jbd2_journal_unlock_updates(journal->j_journal); From c03a82b4a0c935774afa01fd6d128b444fd930a1 Mon Sep 17 00:00:00 2001 From: Lizhi Xu Date: Mon, 2 Sep 2024 10:36:35 +0800 Subject: [PATCH 105/123] ocfs2: remove unreasonable unlock in ocfs2_read_blocks Patch series "Misc fixes for ocfs2_read_blocks", v5. This series contains 2 fixes for ocfs2_read_blocks(). The first patch fix the issue reported by syzbot, which detects bad unlock balance in ocfs2_read_blocks(). The second patch fixes an issue reported by Heming Zhao when reviewing above fix. This patch (of 2): There was a lock release before exiting, so remove the unreasonable unlock. Link: https://lkml.kernel.org/r/20240902023636.1843422-1-joseph.qi@linux.alibaba.com Link: https://lkml.kernel.org/r/20240902023636.1843422-2-joseph.qi@linux.alibaba.com Fixes: cf76c78595ca ("ocfs2: don't put and assigning null to bh allocated outside") Signed-off-by: Lizhi Xu Signed-off-by: Joseph Qi Reviewed-by: Heming Zhao Reviewed-by: Joseph Qi Reported-by: syzbot+ab134185af9ef88dfed5@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=ab134185af9ef88dfed5 Tested-by: syzbot+ab134185af9ef88dfed5@syzkaller.appspotmail.com Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: [4.20+] Signed-off-by: Andrew Morton --- fs/ocfs2/buffer_head_io.c | 1 - 1 file changed, 1 deletion(-) diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index cdb9b9bdea1f..e62c7e1de4eb 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -235,7 +235,6 @@ int ocfs2_read_blocks(struct ocfs2_caching_info *ci, u64 block, int nr, if (bhs[i] == NULL) { bhs[i] = sb_getblk(sb, block++); if (bhs[i] == NULL) { - ocfs2_metadata_cache_io_unlock(ci); status = -ENOMEM; mlog_errno(status); /* Don't forget to put previous bh! */ From 33b525cef4cff49e216e4133cc48452e11c0391e Mon Sep 17 00:00:00 2001 From: Lizhi Xu Date: Mon, 2 Sep 2024 10:36:36 +0800 Subject: [PATCH 106/123] ocfs2: fix possible null-ptr-deref in ocfs2_set_buffer_uptodate When doing cleanup, if flags without OCFS2_BH_READAHEAD, it may trigger NULL pointer dereference in the following ocfs2_set_buffer_uptodate() if bh is NULL. Link: https://lkml.kernel.org/r/20240902023636.1843422-3-joseph.qi@linux.alibaba.com Fixes: cf76c78595ca ("ocfs2: don't put and assigning null to bh allocated outside") Signed-off-by: Lizhi Xu Signed-off-by: Joseph Qi Reviewed-by: Joseph Qi Reported-by: Heming Zhao Suggested-by: Heming Zhao Cc: [4.20+] Cc: Changwei Ge Cc: Gang He Cc: Joel Becker Cc: Jun Piao Cc: Junxiao Bi Cc: Mark Fasheh Signed-off-by: Andrew Morton --- fs/ocfs2/buffer_head_io.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/ocfs2/buffer_head_io.c b/fs/ocfs2/buffer_head_io.c index e62c7e1de4eb..8f714406528d 100644 --- a/fs/ocfs2/buffer_head_io.c +++ b/fs/ocfs2/buffer_head_io.c @@ -388,7 +388,8 @@ read_failure: /* Always set the buffer in the cache, even if it was * a forced read, or read-ahead which hasn't yet * completed. */ - ocfs2_set_buffer_uptodate(ci, bh); + if (bh) + ocfs2_set_buffer_uptodate(ci, bh); } ocfs2_metadata_cache_io_unlock(ci); From 35fccce29feb3706f649726d410122dd81b92c18 Mon Sep 17 00:00:00 2001 From: Joseph Qi Date: Wed, 4 Sep 2024 15:10:03 +0800 Subject: [PATCH 107/123] ocfs2: cancel dqi_sync_work before freeing oinfo ocfs2_global_read_info() will initialize and schedule dqi_sync_work at the end, if error occurs after successfully reading global quota, it will trigger the following warning with CONFIG_DEBUG_OBJECTS_* enabled: ODEBUG: free active (active state 0) object: 00000000d8b0ce28 object type: timer_list hint: qsync_work_fn+0x0/0x16c This reports that there is an active delayed work when freeing oinfo in error handling, so cancel dqi_sync_work first. BTW, return status instead of -1 when .read_file_info fails. Link: https://syzkaller.appspot.com/bug?extid=f7af59df5d6b25f0febd Link: https://lkml.kernel.org/r/20240904071004.2067695-1-joseph.qi@linux.alibaba.com Fixes: 171bf93ce11f ("ocfs2: Periodic quota syncing") Signed-off-by: Joseph Qi Reviewed-by: Heming Zhao Reported-by: syzbot+f7af59df5d6b25f0febd@syzkaller.appspotmail.com Tested-by: syzbot+f7af59df5d6b25f0febd@syzkaller.appspotmail.com Cc: Mark Fasheh Cc: Joel Becker Cc: Junxiao Bi Cc: Changwei Ge Cc: Gang He Cc: Jun Piao Cc: Signed-off-by: Andrew Morton --- fs/ocfs2/quota_local.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index 8ce462c64c51..73d3367c533b 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -692,7 +692,7 @@ static int ocfs2_local_read_info(struct super_block *sb, int type) int status; struct buffer_head *bh = NULL; struct ocfs2_quota_recovery *rec; - int locked = 0; + int locked = 0, global_read = 0; info->dqi_max_spc_limit = 0x7fffffffffffffffLL; info->dqi_max_ino_limit = 0x7fffffffffffffffLL; @@ -700,6 +700,7 @@ static int ocfs2_local_read_info(struct super_block *sb, int type) if (!oinfo) { mlog(ML_ERROR, "failed to allocate memory for ocfs2 quota" " info."); + status = -ENOMEM; goto out_err; } info->dqi_priv = oinfo; @@ -712,6 +713,7 @@ static int ocfs2_local_read_info(struct super_block *sb, int type) status = ocfs2_global_read_info(sb, type); if (status < 0) goto out_err; + global_read = 1; status = ocfs2_inode_lock(lqinode, &oinfo->dqi_lqi_bh, 1); if (status < 0) { @@ -782,10 +784,12 @@ out_err: if (locked) ocfs2_inode_unlock(lqinode, 1); ocfs2_release_local_quota_bitmaps(&oinfo->dqi_chunk); + if (global_read) + cancel_delayed_work_sync(&oinfo->dqi_sync_work); kfree(oinfo); } brelse(bh); - return -1; + return status; } /* Write local info to quota file */ From 0885ef4705607936fc36a38fd74356e1c465b023 Mon Sep 17 00:00:00 2001 From: Chris Li Date: Thu, 5 Sep 2024 01:08:17 -0700 Subject: [PATCH 108/123] mm: vmscan.c: fix OOM on swap stress test I found a regression on mm-unstable during my swap stress test, using tmpfs to compile linux. The test OOM very soon after the make spawns many cc processes. It bisects down to this change: 33dfe9204f29b415bbc0abb1a50642d1ba94f5e9 (mm/gup: clear the LRU flag of a page before adding to LRU batch) Yu Zhao propose the fix: "I think this is one of the potential side effects -- Huge mentioned earlier about isolate_lru_folios():" I test that with it the swap stress test no longer OOM. Link: https://lore.kernel.org/r/CAOUHufYi9h0kz5uW3LHHS3ZrVwEq-kKp8S6N-MZUmErNAXoXmw@mail.gmail.com/ Link: https://lkml.kernel.org/r/20240905-lru-flag-v2-1-8a2d9046c594@kernel.org Fixes: 33dfe9204f29 ("mm/gup: clear the LRU flag of a page before adding to LRU batch") Signed-off-by: Chris Li Suggested-by: Yu Zhao Suggested-by: Hugh Dickins Closes: https://lore.kernel.org/all/CAF8kJuNP5iTj2p07QgHSGOJsiUfYpJ2f4R1Q5-3BN9JiD9W_KA@mail.gmail.com/ Signed-off-by: Andrew Morton --- mm/vmscan.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/vmscan.c b/mm/vmscan.c index bd489c1af228..a8d61a8b6894 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -4300,7 +4300,7 @@ static bool sort_folio(struct lruvec *lruvec, struct folio *folio, struct scan_c } /* ineligible */ - if (zone > sc->reclaim_idx) { + if (!folio_test_lru(folio) || zone > sc->reclaim_idx) { gen = folio_inc_gen(lruvec, folio, false); list_move_tail(&folio->lru, &lrugen->folios[gen][type][zone]); return true; From fb497d6db7c19c797cbd694b52d1af87c4eebcc6 Mon Sep 17 00:00:00 2001 From: "Liam R. Howlett" Date: Wed, 4 Sep 2024 17:12:04 -0700 Subject: [PATCH 109/123] mm/damon/vaddr: protect vma traversal in __damon_va_thre_regions() with rcu read lock Traversing VMAs of a given maple tree should be protected by rcu read lock. However, __damon_va_three_regions() is not doing the protection. Hold the lock. Link: https://lkml.kernel.org/r/20240905001204.1481-1-sj@kernel.org Fixes: d0cf3dd47f0d ("damon: convert __damon_va_three_regions to use the VMA iterator") Signed-off-by: Liam R. Howlett Signed-off-by: SeongJae Park Reported-by: Guenter Roeck Closes: https://lore.kernel.org/b83651a0-5b24-4206-b860-cb54ffdf209b@roeck-us.net Tested-by: Guenter Roeck Cc: David Hildenbrand Cc: Matthew Wilcox Cc: Signed-off-by: Andrew Morton --- mm/damon/vaddr.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/mm/damon/vaddr.c b/mm/damon/vaddr.c index 58829baf8b5d..a0036dc78a3b 100644 --- a/mm/damon/vaddr.c +++ b/mm/damon/vaddr.c @@ -126,6 +126,7 @@ static int __damon_va_three_regions(struct mm_struct *mm, * If this is too slow, it can be optimised to examine the maple * tree gaps. */ + rcu_read_lock(); for_each_vma(vmi, vma) { unsigned long gap; @@ -146,6 +147,7 @@ static int __damon_va_three_regions(struct mm_struct *mm, next: prev = vma; } + rcu_read_unlock(); if (!sz_range(&second_gap) || !sz_range(&first_gap)) return -EINVAL; From 038eb433dc1474c4bc7d33188294e3d4778efdfd Mon Sep 17 00:00:00 2001 From: Sean Anderson Date: Fri, 6 Sep 2024 17:54:34 -0400 Subject: [PATCH 110/123] dma-mapping: add tracing for dma-mapping API calls When debugging drivers, it can often be useful to trace when memory gets (un)mapped for DMA (and can be accessed by the device). Add some tracepoints for this purpose. Use u64 instead of phys_addr_t and dma_addr_t (and similarly %llx instead of %pa) because libtraceevent can't handle typedefs in all cases. Signed-off-by: Sean Anderson Signed-off-by: Christoph Hellwig --- include/trace/events/dma.h | 341 +++++++++++++++++++++++++++++++++++++ kernel/dma/mapping.c | 24 ++- 2 files changed, 364 insertions(+), 1 deletion(-) create mode 100644 include/trace/events/dma.h diff --git a/include/trace/events/dma.h b/include/trace/events/dma.h new file mode 100644 index 000000000000..f57f05331d73 --- /dev/null +++ b/include/trace/events/dma.h @@ -0,0 +1,341 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM dma + +#if !defined(_TRACE_DMA_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_DMA_H + +#include +#include +#include +#include + +TRACE_DEFINE_ENUM(DMA_BIDIRECTIONAL); +TRACE_DEFINE_ENUM(DMA_TO_DEVICE); +TRACE_DEFINE_ENUM(DMA_FROM_DEVICE); +TRACE_DEFINE_ENUM(DMA_NONE); + +#define decode_dma_data_direction(dir) \ + __print_symbolic(dir, \ + { DMA_BIDIRECTIONAL, "BIDIRECTIONAL" }, \ + { DMA_TO_DEVICE, "TO_DEVICE" }, \ + { DMA_FROM_DEVICE, "FROM_DEVICE" }, \ + { DMA_NONE, "NONE" }) + +#define decode_dma_attrs(attrs) \ + __print_flags(attrs, "|", \ + { DMA_ATTR_WEAK_ORDERING, "WEAK_ORDERING" }, \ + { DMA_ATTR_WRITE_COMBINE, "WRITE_COMBINE" }, \ + { DMA_ATTR_NO_KERNEL_MAPPING, "NO_KERNEL_MAPPING" }, \ + { DMA_ATTR_SKIP_CPU_SYNC, "SKIP_CPU_SYNC" }, \ + { DMA_ATTR_FORCE_CONTIGUOUS, "FORCE_CONTIGUOUS" }, \ + { DMA_ATTR_ALLOC_SINGLE_PAGES, "ALLOC_SINGLE_PAGES" }, \ + { DMA_ATTR_NO_WARN, "NO_WARN" }, \ + { DMA_ATTR_PRIVILEGED, "PRIVILEGED" }) + +DECLARE_EVENT_CLASS(dma_map, + TP_PROTO(struct device *dev, phys_addr_t phys_addr, dma_addr_t dma_addr, + size_t size, enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, phys_addr, dma_addr, size, dir, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __field(u64, phys_addr) + __field(u64, dma_addr) + __field(size_t, size) + __field(enum dma_data_direction, dir) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + __assign_str(device); + __entry->phys_addr = phys_addr; + __entry->dma_addr = dma_addr; + __entry->size = size; + __entry->dir = dir; + __entry->attrs = attrs; + ), + + TP_printk("%s dir=%s dma_addr=%llx size=%zu phys_addr=%llx attrs=%s", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __entry->dma_addr, + __entry->size, + __entry->phys_addr, + decode_dma_attrs(__entry->attrs)) +); + +DEFINE_EVENT(dma_map, dma_map_page, + TP_PROTO(struct device *dev, phys_addr_t phys_addr, dma_addr_t dma_addr, + size_t size, enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, phys_addr, dma_addr, size, dir, attrs)); + +DEFINE_EVENT(dma_map, dma_map_resource, + TP_PROTO(struct device *dev, phys_addr_t phys_addr, dma_addr_t dma_addr, + size_t size, enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, phys_addr, dma_addr, size, dir, attrs)); + +DECLARE_EVENT_CLASS(dma_unmap, + TP_PROTO(struct device *dev, dma_addr_t addr, size_t size, + enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, addr, size, dir, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __field(u64, addr) + __field(size_t, size) + __field(enum dma_data_direction, dir) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + __assign_str(device); + __entry->addr = addr; + __entry->size = size; + __entry->dir = dir; + __entry->attrs = attrs; + ), + + TP_printk("%s dir=%s dma_addr=%llx size=%zu attrs=%s", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __entry->addr, + __entry->size, + decode_dma_attrs(__entry->attrs)) +); + +DEFINE_EVENT(dma_unmap, dma_unmap_page, + TP_PROTO(struct device *dev, dma_addr_t addr, size_t size, + enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, addr, size, dir, attrs)); + +DEFINE_EVENT(dma_unmap, dma_unmap_resource, + TP_PROTO(struct device *dev, dma_addr_t addr, size_t size, + enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, addr, size, dir, attrs)); + +TRACE_EVENT(dma_alloc, + TP_PROTO(struct device *dev, void *virt_addr, dma_addr_t dma_addr, + size_t size, gfp_t flags, unsigned long attrs), + TP_ARGS(dev, virt_addr, dma_addr, size, flags, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __field(u64, phys_addr) + __field(u64, dma_addr) + __field(size_t, size) + __field(gfp_t, flags) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + __assign_str(device); + __entry->phys_addr = virt_to_phys(virt_addr); + __entry->dma_addr = dma_addr; + __entry->size = size; + __entry->flags = flags; + __entry->attrs = attrs; + ), + + TP_printk("%s dma_addr=%llx size=%zu phys_addr=%llx flags=%s attrs=%s", + __get_str(device), + __entry->dma_addr, + __entry->size, + __entry->phys_addr, + show_gfp_flags(__entry->flags), + decode_dma_attrs(__entry->attrs)) +); + +TRACE_EVENT(dma_free, + TP_PROTO(struct device *dev, void *virt_addr, dma_addr_t dma_addr, + size_t size, unsigned long attrs), + TP_ARGS(dev, virt_addr, dma_addr, size, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __field(u64, phys_addr) + __field(u64, dma_addr) + __field(size_t, size) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + __assign_str(device); + __entry->phys_addr = virt_to_phys(virt_addr); + __entry->dma_addr = dma_addr; + __entry->size = size; + __entry->attrs = attrs; + ), + + TP_printk("%s dma_addr=%llx size=%zu phys_addr=%llx attrs=%s", + __get_str(device), + __entry->dma_addr, + __entry->size, + __entry->phys_addr, + decode_dma_attrs(__entry->attrs)) +); + +TRACE_EVENT(dma_map_sg, + TP_PROTO(struct device *dev, struct scatterlist *sg, int nents, + int ents, enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, sg, nents, ents, dir, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __dynamic_array(u64, phys_addrs, nents) + __dynamic_array(u64, dma_addrs, ents) + __dynamic_array(unsigned int, lengths, ents) + __field(enum dma_data_direction, dir) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + int i; + + __assign_str(device); + for (i = 0; i < nents; i++) + ((u64 *)__get_dynamic_array(phys_addrs))[i] = + sg_phys(sg + i); + for (i = 0; i < ents; i++) { + ((u64 *)__get_dynamic_array(dma_addrs))[i] = + sg_dma_address(sg + i); + ((unsigned int *)__get_dynamic_array(lengths))[i] = + sg_dma_len(sg + i); + } + __entry->dir = dir; + __entry->attrs = attrs; + ), + + TP_printk("%s dir=%s dma_addrs=%s sizes=%s phys_addrs=%s attrs=%s", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __print_array(__get_dynamic_array(dma_addrs), + __get_dynamic_array_len(dma_addrs) / + sizeof(u64), sizeof(u64)), + __print_array(__get_dynamic_array(lengths), + __get_dynamic_array_len(lengths) / + sizeof(unsigned int), sizeof(unsigned int)), + __print_array(__get_dynamic_array(phys_addrs), + __get_dynamic_array_len(phys_addrs) / + sizeof(u64), sizeof(u64)), + decode_dma_attrs(__entry->attrs)) +); + +TRACE_EVENT(dma_unmap_sg, + TP_PROTO(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir, unsigned long attrs), + TP_ARGS(dev, sg, nents, dir, attrs), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __dynamic_array(u64, addrs, nents) + __field(enum dma_data_direction, dir) + __field(unsigned long, attrs) + ), + + TP_fast_assign( + int i; + + __assign_str(device); + for (i = 0; i < nents; i++) + ((u64 *)__get_dynamic_array(addrs))[i] = + sg_phys(sg + i); + __entry->dir = dir; + __entry->attrs = attrs; + ), + + TP_printk("%s dir=%s phys_addrs=%s attrs=%s", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __print_array(__get_dynamic_array(addrs), + __get_dynamic_array_len(addrs) / + sizeof(u64), sizeof(u64)), + decode_dma_attrs(__entry->attrs)) +); + +DECLARE_EVENT_CLASS(dma_sync_single, + TP_PROTO(struct device *dev, dma_addr_t dma_addr, size_t size, + enum dma_data_direction dir), + TP_ARGS(dev, dma_addr, size, dir), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __field(u64, dma_addr) + __field(size_t, size) + __field(enum dma_data_direction, dir) + ), + + TP_fast_assign( + __assign_str(device); + __entry->dma_addr = dma_addr; + __entry->size = size; + __entry->dir = dir; + ), + + TP_printk("%s dir=%s dma_addr=%llx size=%zu", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __entry->dma_addr, + __entry->size) +); + +DEFINE_EVENT(dma_sync_single, dma_sync_single_for_cpu, + TP_PROTO(struct device *dev, dma_addr_t dma_addr, size_t size, + enum dma_data_direction dir), + TP_ARGS(dev, dma_addr, size, dir)); + +DEFINE_EVENT(dma_sync_single, dma_sync_single_for_device, + TP_PROTO(struct device *dev, dma_addr_t dma_addr, size_t size, + enum dma_data_direction dir), + TP_ARGS(dev, dma_addr, size, dir)); + +DECLARE_EVENT_CLASS(dma_sync_sg, + TP_PROTO(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir), + TP_ARGS(dev, sg, nents, dir), + + TP_STRUCT__entry( + __string(device, dev_name(dev)) + __dynamic_array(u64, dma_addrs, nents) + __dynamic_array(unsigned int, lengths, nents) + __field(enum dma_data_direction, dir) + ), + + TP_fast_assign( + int i; + + __assign_str(device); + for (i = 0; i < nents; i++) { + ((u64 *)__get_dynamic_array(dma_addrs))[i] = + sg_dma_address(sg + i); + ((unsigned int *)__get_dynamic_array(lengths))[i] = + sg_dma_len(sg + i); + } + __entry->dir = dir; + ), + + TP_printk("%s dir=%s dma_addrs=%s sizes=%s", + __get_str(device), + decode_dma_data_direction(__entry->dir), + __print_array(__get_dynamic_array(dma_addrs), + __get_dynamic_array_len(dma_addrs) / + sizeof(u64), sizeof(u64)), + __print_array(__get_dynamic_array(lengths), + __get_dynamic_array_len(lengths) / + sizeof(unsigned int), sizeof(unsigned int))) +); + +DEFINE_EVENT(dma_sync_sg, dma_sync_sg_for_cpu, + TP_PROTO(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir), + TP_ARGS(dev, sg, nents, dir)); + +DEFINE_EVENT(dma_sync_sg, dma_sync_sg_for_device, + TP_PROTO(struct device *dev, struct scatterlist *sg, int nents, + enum dma_data_direction dir), + TP_ARGS(dev, sg, nents, dir)); + +#endif /* _TRACE_DMA_H */ + +/* This part must be outside protection */ +#include diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index 056f27962f69..7550b5dc5e55 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -18,6 +18,9 @@ #include "debug.h" #include "direct.h" +#define CREATE_TRACE_POINTS +#include + #if defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_DEVICE) || \ defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU) || \ defined(CONFIG_ARCH_HAS_SYNC_DMA_FOR_CPU_ALL) @@ -169,6 +172,8 @@ dma_addr_t dma_map_page_attrs(struct device *dev, struct page *page, else addr = ops->map_page(dev, page, offset, size, dir, attrs); kmsan_handle_dma(page, offset, size, dir); + trace_dma_map_page(dev, page_to_phys(page) + offset, addr, size, dir, + attrs); debug_dma_map_page(dev, page, offset, size, dir, addr, attrs); return addr; @@ -188,6 +193,7 @@ void dma_unmap_page_attrs(struct device *dev, dma_addr_t addr, size_t size, iommu_dma_unmap_page(dev, addr, size, dir, attrs); else ops->unmap_page(dev, addr, size, dir, attrs); + trace_dma_unmap_page(dev, addr, size, dir, attrs); debug_dma_unmap_page(dev, addr, size, dir); } EXPORT_SYMBOL(dma_unmap_page_attrs); @@ -213,6 +219,7 @@ static int __dma_map_sg_attrs(struct device *dev, struct scatterlist *sg, if (ents > 0) { kmsan_handle_dma_sg(sg, nents, dir); + trace_dma_map_sg(dev, sg, nents, ents, dir, attrs); debug_dma_map_sg(dev, sg, nents, ents, dir, attrs); } else if (WARN_ON_ONCE(ents != -EINVAL && ents != -ENOMEM && ents != -EIO && ents != -EREMOTEIO)) { @@ -298,6 +305,7 @@ void dma_unmap_sg_attrs(struct device *dev, struct scatterlist *sg, const struct dma_map_ops *ops = get_dma_ops(dev); BUG_ON(!valid_dma_direction(dir)); + trace_dma_unmap_sg(dev, sg, nents, dir, attrs); debug_dma_unmap_sg(dev, sg, nents, dir); if (dma_map_direct(dev, ops) || arch_dma_unmap_sg_direct(dev, sg, nents)) @@ -327,6 +335,7 @@ dma_addr_t dma_map_resource(struct device *dev, phys_addr_t phys_addr, else if (ops->map_resource) addr = ops->map_resource(dev, phys_addr, size, dir, attrs); + trace_dma_map_resource(dev, phys_addr, addr, size, dir, attrs); debug_dma_map_resource(dev, phys_addr, size, dir, addr, attrs); return addr; } @@ -344,6 +353,7 @@ void dma_unmap_resource(struct device *dev, dma_addr_t addr, size_t size, iommu_dma_unmap_resource(dev, addr, size, dir, attrs); else if (ops->unmap_resource) ops->unmap_resource(dev, addr, size, dir, attrs); + trace_dma_unmap_resource(dev, addr, size, dir, attrs); debug_dma_unmap_resource(dev, addr, size, dir); } EXPORT_SYMBOL(dma_unmap_resource); @@ -361,6 +371,7 @@ void __dma_sync_single_for_cpu(struct device *dev, dma_addr_t addr, size_t size, iommu_dma_sync_single_for_cpu(dev, addr, size, dir); else if (ops->sync_single_for_cpu) ops->sync_single_for_cpu(dev, addr, size, dir); + trace_dma_sync_single_for_cpu(dev, addr, size, dir); debug_dma_sync_single_for_cpu(dev, addr, size, dir); } EXPORT_SYMBOL(__dma_sync_single_for_cpu); @@ -377,6 +388,7 @@ void __dma_sync_single_for_device(struct device *dev, dma_addr_t addr, iommu_dma_sync_single_for_device(dev, addr, size, dir); else if (ops->sync_single_for_device) ops->sync_single_for_device(dev, addr, size, dir); + trace_dma_sync_single_for_device(dev, addr, size, dir); debug_dma_sync_single_for_device(dev, addr, size, dir); } EXPORT_SYMBOL(__dma_sync_single_for_device); @@ -393,6 +405,7 @@ void __dma_sync_sg_for_cpu(struct device *dev, struct scatterlist *sg, iommu_dma_sync_sg_for_cpu(dev, sg, nelems, dir); else if (ops->sync_sg_for_cpu) ops->sync_sg_for_cpu(dev, sg, nelems, dir); + trace_dma_sync_sg_for_cpu(dev, sg, nelems, dir); debug_dma_sync_sg_for_cpu(dev, sg, nelems, dir); } EXPORT_SYMBOL(__dma_sync_sg_for_cpu); @@ -409,6 +422,7 @@ void __dma_sync_sg_for_device(struct device *dev, struct scatterlist *sg, iommu_dma_sync_sg_for_device(dev, sg, nelems, dir); else if (ops->sync_sg_for_device) ops->sync_sg_for_device(dev, sg, nelems, dir); + trace_dma_sync_sg_for_device(dev, sg, nelems, dir); debug_dma_sync_sg_for_device(dev, sg, nelems, dir); } EXPORT_SYMBOL(__dma_sync_sg_for_device); @@ -601,6 +615,7 @@ void *dma_alloc_attrs(struct device *dev, size_t size, dma_addr_t *dma_handle, else return NULL; + trace_dma_alloc(dev, cpu_addr, *dma_handle, size, flag, attrs); debug_dma_alloc_coherent(dev, size, *dma_handle, cpu_addr, attrs); return cpu_addr; } @@ -625,6 +640,7 @@ void dma_free_attrs(struct device *dev, size_t size, void *cpu_addr, if (!cpu_addr) return; + trace_dma_free(dev, cpu_addr, dma_handle, size, attrs); debug_dma_free_coherent(dev, size, cpu_addr, dma_handle); if (dma_alloc_direct(dev, ops)) dma_direct_free(dev, size, cpu_addr, dma_handle, attrs); @@ -662,8 +678,11 @@ struct page *dma_alloc_pages(struct device *dev, size_t size, { struct page *page = __dma_alloc_pages(dev, size, dma_handle, dir, gfp); - if (page) + if (page) { + trace_dma_map_page(dev, page_to_phys(page), *dma_handle, size, + dir, 0); debug_dma_map_page(dev, page, 0, size, dir, *dma_handle, 0); + } return page; } EXPORT_SYMBOL_GPL(dma_alloc_pages); @@ -685,6 +704,7 @@ static void __dma_free_pages(struct device *dev, size_t size, struct page *page, void dma_free_pages(struct device *dev, size_t size, struct page *page, dma_addr_t dma_handle, enum dma_data_direction dir) { + trace_dma_unmap_page(dev, dma_handle, size, dir, 0); debug_dma_unmap_page(dev, dma_handle, size, dir); __dma_free_pages(dev, size, page, dma_handle, dir); } @@ -747,6 +767,7 @@ struct sg_table *dma_alloc_noncontiguous(struct device *dev, size_t size, if (sgt) { sgt->nents = 1; + trace_dma_map_sg(dev, sgt->sgl, sgt->orig_nents, 1, dir, attrs); debug_dma_map_sg(dev, sgt->sgl, sgt->orig_nents, 1, dir, attrs); } return sgt; @@ -767,6 +788,7 @@ void dma_free_noncontiguous(struct device *dev, size_t size, { const struct dma_map_ops *ops = get_dma_ops(dev); + trace_dma_unmap_sg(dev, sgt->sgl, sgt->orig_nents, dir, 0); debug_dma_unmap_sg(dev, sgt->sgl, sgt->orig_nents, dir); if (ops && ops->free_noncontiguous) ops->free_noncontiguous(dev, size, sgt, dir); From 93701d3b84ac5f3ea07259d4ced405c53d757985 Mon Sep 17 00:00:00 2001 From: Tommy Huang Date: Wed, 11 Sep 2024 17:39:51 +0800 Subject: [PATCH 111/123] i2c: aspeed: Update the stop sw state when the bus recovery occurs When the i2c bus recovery occurs, driver will send i2c stop command in the scl low condition. In this case the sw state will still keep original situation. Under multi-master usage, i2c bus recovery will be called when i2c transfer timeout occurs. Update the stop command calling with aspeed_i2c_do_stop function to update master_state. Fixes: f327c686d3ba ("i2c: aspeed: added driver for Aspeed I2C") Cc: stable@vger.kernel.org # v4.13+ Signed-off-by: Tommy Huang Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-aspeed.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/drivers/i2c/busses/i2c-aspeed.c b/drivers/i2c/busses/i2c-aspeed.c index ce8c4846b7fa..2a03a221e2dd 100644 --- a/drivers/i2c/busses/i2c-aspeed.c +++ b/drivers/i2c/busses/i2c-aspeed.c @@ -170,6 +170,13 @@ struct aspeed_i2c_bus { static int aspeed_i2c_reset(struct aspeed_i2c_bus *bus); +/* precondition: bus.lock has been acquired. */ +static void aspeed_i2c_do_stop(struct aspeed_i2c_bus *bus) +{ + bus->master_state = ASPEED_I2C_MASTER_STOP; + writel(ASPEED_I2CD_M_STOP_CMD, bus->base + ASPEED_I2C_CMD_REG); +} + static int aspeed_i2c_recover_bus(struct aspeed_i2c_bus *bus) { unsigned long time_left, flags; @@ -187,7 +194,7 @@ static int aspeed_i2c_recover_bus(struct aspeed_i2c_bus *bus) command); reinit_completion(&bus->cmd_complete); - writel(ASPEED_I2CD_M_STOP_CMD, bus->base + ASPEED_I2C_CMD_REG); + aspeed_i2c_do_stop(bus); spin_unlock_irqrestore(&bus->lock, flags); time_left = wait_for_completion_timeout( @@ -390,13 +397,6 @@ static void aspeed_i2c_do_start(struct aspeed_i2c_bus *bus) writel(command, bus->base + ASPEED_I2C_CMD_REG); } -/* precondition: bus.lock has been acquired. */ -static void aspeed_i2c_do_stop(struct aspeed_i2c_bus *bus) -{ - bus->master_state = ASPEED_I2C_MASTER_STOP; - writel(ASPEED_I2CD_M_STOP_CMD, bus->base + ASPEED_I2C_CMD_REG); -} - /* precondition: bus.lock has been acquired. */ static void aspeed_i2c_next_msg_or_stop(struct aspeed_i2c_bus *bus) { From 521da1e9225450bd323db5fa5bca942b1dc485b7 Mon Sep 17 00:00:00 2001 From: Robert Hancock Date: Tue, 21 Nov 2023 18:11:16 +0000 Subject: [PATCH 112/123] i2c: xiic: Wait for TX empty to avoid missed TX NAKs Frequently an I2C write will be followed by a read, such as a register address write followed by a read of the register value. In this driver, when the TX FIFO half empty interrupt was raised and it was determined that there was enough space in the TX FIFO to send the following read command, it would do so without waiting for the TX FIFO to actually empty. Unfortunately it appears that in some cases this can result in a NAK that was raised by the target device on the write, such as due to an unsupported register address, being ignored and the subsequent read being done anyway. This can potentially put the I2C bus into an invalid state and/or result in invalid read data being processed. To avoid this, once a message has been fully written to the TX FIFO, wait for the TX FIFO empty interrupt before moving on to the next message, to ensure NAKs are handled properly. Fixes: e1d5b6598cdc ("i2c: Add support for Xilinx XPS IIC Bus Interface") Signed-off-by: Robert Hancock Cc: # v2.6.34+ Reviewed-by: Manikanta Guntupalli Acked-by: Michal Simek Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-xiic.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c index 19468565120e..bd6e107472b7 100644 --- a/drivers/i2c/busses/i2c-xiic.c +++ b/drivers/i2c/busses/i2c-xiic.c @@ -772,14 +772,17 @@ static irqreturn_t xiic_process(int irq, void *dev_id) goto out; } - xiic_fill_tx_fifo(i2c); - - /* current message sent and there is space in the fifo */ - if (!xiic_tx_space(i2c) && xiic_tx_fifo_space(i2c) >= 2) { + if (xiic_tx_space(i2c)) { + xiic_fill_tx_fifo(i2c); + } else { + /* current message fully written */ dev_dbg(i2c->adap.dev.parent, "%s end of message sent, nmsgs: %d\n", __func__, i2c->nmsgs); - if (i2c->nmsgs > 1) { + /* Don't move onto the next message until the TX FIFO empties, + * to ensure that a NAK is not missed. + */ + if (i2c->nmsgs > 1 && (pend & XIIC_INTR_TX_EMPTY_MASK)) { i2c->nmsgs--; i2c->tx_msg++; xfer_more = 1; @@ -790,11 +793,7 @@ static irqreturn_t xiic_process(int irq, void *dev_id) "%s Got TX IRQ but no more to do...\n", __func__); } - } else if (!xiic_tx_space(i2c) && (i2c->nmsgs == 1)) - /* current frame is sent and is last, - * make sure to disable tx half - */ - xiic_irq_dis(i2c, XIIC_INTR_TX_HALF_MASK); + } } if (pend & XIIC_INTR_BNB_MASK) { From 1d4a1adbed2582444aaf97671858b7d12915bd05 Mon Sep 17 00:00:00 2001 From: Robert Hancock Date: Wed, 11 Sep 2024 22:16:53 +0200 Subject: [PATCH 113/123] i2c: xiic: Try re-initialization on bus busy timeout In the event that the I2C bus was powered down when the I2C controller driver loads, or some spurious pulses occur on the I2C bus, it's possible that the controller detects a spurious I2C "start" condition. In this situation it may continue to report the bus is busy indefinitely and block the controller from working. The "single-master" DT flag can be specified to disable bus busy checks entirely, but this may not be safe to use in situations where other I2C masters may potentially exist. In the event that the controller reports "bus busy" for too long when starting a transaction, we can try reinitializing the controller to see if the busy condition clears. This allows recovering from this scenario. Fixes: e1d5b6598cdc ("i2c: Add support for Xilinx XPS IIC Bus Interface") Signed-off-by: Robert Hancock Cc: # v2.6.34+ Reviewed-by: Manikanta Guntupalli Acked-by: Michal Simek Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-xiic.c | 41 ++++++++++++++++++++++------------- 1 file changed, 26 insertions(+), 15 deletions(-) diff --git a/drivers/i2c/busses/i2c-xiic.c b/drivers/i2c/busses/i2c-xiic.c index bd6e107472b7..4c89aad02451 100644 --- a/drivers/i2c/busses/i2c-xiic.c +++ b/drivers/i2c/busses/i2c-xiic.c @@ -843,23 +843,11 @@ static int xiic_bus_busy(struct xiic_i2c *i2c) return (sr & XIIC_SR_BUS_BUSY_MASK) ? -EBUSY : 0; } -static int xiic_busy(struct xiic_i2c *i2c) +static int xiic_wait_not_busy(struct xiic_i2c *i2c) { int tries = 3; int err; - if (i2c->tx_msg || i2c->rx_msg) - return -EBUSY; - - /* In single master mode bus can only be busy, when in use by this - * driver. If the register indicates bus being busy for some reason we - * should ignore it, since bus will never be released and i2c will be - * stuck forever. - */ - if (i2c->singlemaster) { - return 0; - } - /* for instance if previous transfer was terminated due to TX error * it might be that the bus is on it's way to become available * give it at most 3 ms to wake @@ -1103,13 +1091,36 @@ static int xiic_start_xfer(struct xiic_i2c *i2c, struct i2c_msg *msgs, int num) mutex_lock(&i2c->lock); - ret = xiic_busy(i2c); - if (ret) { + if (i2c->tx_msg || i2c->rx_msg) { dev_err(i2c->adap.dev.parent, "cannot start a transfer while busy\n"); + ret = -EBUSY; goto out; } + /* In single master mode bus can only be busy, when in use by this + * driver. If the register indicates bus being busy for some reason we + * should ignore it, since bus will never be released and i2c will be + * stuck forever. + */ + if (!i2c->singlemaster) { + ret = xiic_wait_not_busy(i2c); + if (ret) { + /* If the bus is stuck in a busy state, such as due to spurious low + * pulses on the bus causing a false start condition to be detected, + * then try to recover by re-initializing the controller and check + * again if the bus is still busy. + */ + dev_warn(i2c->adap.dev.parent, "I2C bus busy timeout, reinitializing\n"); + ret = xiic_reinit(i2c); + if (ret) + goto out; + ret = xiic_wait_not_busy(i2c); + if (ret) + goto out; + } + } + i2c->tx_msg = msgs; i2c->rx_msg = NULL; i2c->nmsgs = num; From 1db4da55070d6a2754efeb3743f5312fc32f5961 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Wed, 11 Sep 2024 18:39:14 +0300 Subject: [PATCH 114/123] i2c: isch: Add missed 'else' In accordance with the existing comment and code analysis it is quite likely that there is a missed 'else' when adapter times out. Add it. Fixes: 5bc1200852c3 ("i2c: Add Intel SCH SMBus support") Signed-off-by: Andy Shevchenko Cc: # v2.6.27+ Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-isch.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/i2c/busses/i2c-isch.c b/drivers/i2c/busses/i2c-isch.c index 33dbc19d3848..f59158489ad9 100644 --- a/drivers/i2c/busses/i2c-isch.c +++ b/drivers/i2c/busses/i2c-isch.c @@ -99,8 +99,7 @@ static int sch_transaction(void) if (retries > MAX_RETRIES) { dev_err(&sch_adapter.dev, "SMBus Timeout!\n"); result = -ETIMEDOUT; - } - if (temp & 0x04) { + } else if (temp & 0x04) { result = -EIO; dev_dbg(&sch_adapter.dev, "Bus collision! SMBus may be " "locked until next hard reset. (sorry!)\n"); From e2c85d85a05f16af2223fcc0195ff50a7938b372 Mon Sep 17 00:00:00 2001 From: Jinjie Ruan Date: Thu, 12 Sep 2024 11:34:59 +0800 Subject: [PATCH 115/123] i2c: qcom-geni: Use IRQF_NO_AUTOEN flag in request_irq() disable_irq() after request_irq() still has a time gap in which interrupts can come. request_irq() with IRQF_NO_AUTOEN flag will disable IRQ auto-enable when request IRQ. Fixes: 37692de5d523 ("i2c: i2c-qcom-geni: Add bus driver for the Qualcomm GENI I2C controller") Signed-off-by: Jinjie Ruan Cc: # v4.19+ Acked-by: Mukesh Kumar Savaliya Reviewed-by: Vladimir Zapolskiy Signed-off-by: Andi Shyti --- drivers/i2c/busses/i2c-qcom-geni.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/i2c/busses/i2c-qcom-geni.c b/drivers/i2c/busses/i2c-qcom-geni.c index 06e836e3e877..4c9050a4d58e 100644 --- a/drivers/i2c/busses/i2c-qcom-geni.c +++ b/drivers/i2c/busses/i2c-qcom-geni.c @@ -818,15 +818,13 @@ static int geni_i2c_probe(struct platform_device *pdev) init_completion(&gi2c->done); spin_lock_init(&gi2c->lock); platform_set_drvdata(pdev, gi2c); - ret = devm_request_irq(dev, gi2c->irq, geni_i2c_irq, 0, + ret = devm_request_irq(dev, gi2c->irq, geni_i2c_irq, IRQF_NO_AUTOEN, dev_name(dev), gi2c); if (ret) { dev_err(dev, "Request_irq failed:%d: err:%d\n", gi2c->irq, ret); return ret; } - /* Disable the interrupt so that the system can enter low-power mode */ - disable_irq(gi2c->irq); i2c_set_adapdata(&gi2c->adap, gi2c); gi2c->adap.dev.parent = dev; gi2c->adap.dev.of_node = dev->of_node; From f45cfab28fcd5ac67a38750b6c68316b26d35ac8 Mon Sep 17 00:00:00 2001 From: Leon Romanovsky Date: Wed, 11 Sep 2024 13:15:00 +0300 Subject: [PATCH 116/123] dma-mapping: reliably inform about DMA support for IOMMU MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If the DMA IOMMU path is going to be used, the appropriate check should return that DMA is supported. Fixes: b5c58b2fdc42 ("dma-mapping: direct calls for dma-iommu") Closes: https://lore.kernel.org/all/181e06ff-35a3-434f-b505-672f430bd1cb@notapiano Reported-by: Nícolas F. R. A. Prado #KernelCI Signed-off-by: Leon Romanovsky Reviewed-by: Robin Murphy Tested-by: Nícolas F. R. A. Prado Signed-off-by: Christoph Hellwig --- kernel/dma/mapping.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index 7550b5dc5e55..022d670f8cad 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -843,6 +843,9 @@ static int dma_supported(struct device *dev, u64 mask) if (WARN_ON(ops && use_dma_iommu(dev))) return false; + + if (use_dma_iommu(dev)) + return true; /* * ->dma_supported sets the bypass flag, so we must always call * into the method here unless the device is truly direct mapped. From a5fb217f13f74b2af2ab366ffad522bae717f93c Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Thu, 12 Sep 2024 09:21:18 +0200 Subject: [PATCH 117/123] dma-mapping: reflow dma_supported dma_supported has become too much spaghetti for my taste. Reflow it to remove the duplicate use_dma_iommu condition and make the main path more obvious. Signed-off-by: Christoph Hellwig Reviewed-by: Leon Romanovsky --- kernel/dma/mapping.c | 25 ++++++++++++++----------- 1 file changed, 14 insertions(+), 11 deletions(-) diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index 022d670f8cad..b839683da0ba 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -841,20 +841,23 @@ static int dma_supported(struct device *dev, u64 mask) { const struct dma_map_ops *ops = get_dma_ops(dev); - if (WARN_ON(ops && use_dma_iommu(dev))) - return false; - - if (use_dma_iommu(dev)) + if (use_dma_iommu(dev)) { + if (WARN_ON(ops)) + return false; return true; + } + /* - * ->dma_supported sets the bypass flag, so we must always call - * into the method here unless the device is truly direct mapped. + * ->dma_supported sets and clears the bypass flag, so ignore it here + * and always call into the method if there is one. */ - if (!ops) - return dma_direct_supported(dev, mask); - if (!ops->dma_supported) - return 1; - return ops->dma_supported(dev, mask); + if (ops) { + if (!ops->dma_supported) + return true; + return ops->dma_supported(dev, mask); + } + + return dma_direct_supported(dev, mask); } bool dma_pci_p2pdma_supported(struct device *dev) From 6040f650c56862a4ac40b00c37ef6ab1ddfcebb5 Mon Sep 17 00:00:00 2001 From: Sergey Senozhatsky Date: Fri, 6 Sep 2024 12:45:44 +0900 Subject: [PATCH 118/123] zsmalloc: use unique zsmalloc caches names Each zsmalloc pool maintains several named kmem-caches for zs_handle-s and zspage-s. On a system with multiple zsmalloc pools and CONFIG_DEBUG_VM this triggers kmem_cache_sanity_check(): kmem_cache of name 'zspage' already exists WARNING: at mm/slab_common.c:108 do_kmem_cache_create_usercopy+0xb5/0x310 ... kmem_cache of name 'zs_handle' already exists WARNING: at mm/slab_common.c:108 do_kmem_cache_create_usercopy+0xb5/0x310 ... We provide zram device name when init its zsmalloc pool, so we can use that same name for zsmalloc caches and, hence, create unique names that can easily be linked to zram device that has created them. So instead of having this cat /proc/slabinfo slabinfo - version: 2.1 zspage 46 46 ... zs_handle 128 128 ... zspage 34270 34270 ... zs_handle 34816 34816 ... zspage 0 0 ... zs_handle 0 0 ... We now have this cat /proc/slabinfo slabinfo - version: 2.1 zspage-zram2 46 46 ... zs_handle-zram2 128 128 ... zspage-zram0 34270 34270 ... zs_handle-zram0 34816 34816 ... zspage-zram1 0 0 ... zs_handle-zram1 0 0 ... Link: https://lkml.kernel.org/r/20240906035103.2435557-1-senozhatsky@chromium.org Fixes: 2e40e163a25a ("zsmalloc: decouple handle and object") Signed-off-by: Sergey Senozhatsky Cc: Minchan Kim Signed-off-by: Andrew Morton --- mm/zsmalloc.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-) diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c index 2d3163e4da96..b572aa84823c 100644 --- a/mm/zsmalloc.c +++ b/mm/zsmalloc.c @@ -54,6 +54,7 @@ #include #include #include +#include #include #include #include @@ -293,17 +294,27 @@ static void SetZsPageMovable(struct zs_pool *pool, struct zspage *zspage) {} static int create_cache(struct zs_pool *pool) { - pool->handle_cachep = kmem_cache_create("zs_handle", ZS_HANDLE_SIZE, - 0, 0, NULL); - if (!pool->handle_cachep) - return 1; + char *name; - pool->zspage_cachep = kmem_cache_create("zspage", sizeof(struct zspage), - 0, 0, NULL); + name = kasprintf(GFP_KERNEL, "zs_handle-%s", pool->name); + if (!name) + return -ENOMEM; + pool->handle_cachep = kmem_cache_create(name, ZS_HANDLE_SIZE, + 0, 0, NULL); + kfree(name); + if (!pool->handle_cachep) + return -EINVAL; + + name = kasprintf(GFP_KERNEL, "zspage-%s", pool->name); + if (!name) + return -ENOMEM; + pool->zspage_cachep = kmem_cache_create(name, sizeof(struct zspage), + 0, 0, NULL); + kfree(name); if (!pool->zspage_cachep) { kmem_cache_destroy(pool->handle_cachep); pool->handle_cachep = NULL; - return 1; + return -EINVAL; } return 0; From b4afe4183ec77f230851ea139d91e5cf2644c68b Mon Sep 17 00:00:00 2001 From: Huang Ying Date: Fri, 6 Sep 2024 11:07:11 +0800 Subject: [PATCH 119/123] resource: fix region_intersects() vs add_memory_driver_managed() On a system with CXL memory, the resource tree (/proc/iomem) related to CXL memory may look like something as follows. 490000000-50fffffff : CXL Window 0 490000000-50fffffff : region0 490000000-50fffffff : dax0.0 490000000-50fffffff : System RAM (kmem) Because drivers/dax/kmem.c calls add_memory_driver_managed() during onlining CXL memory, which makes "System RAM (kmem)" a descendant of "CXL Window X". This confuses region_intersects(), which expects all "System RAM" resources to be at the top level of iomem_resource. This can lead to bugs. For example, when the following command line is executed to write some memory in CXL memory range via /dev/mem, $ dd if=data of=/dev/mem bs=$((1 << 10)) seek=$((0x490000000 >> 10)) count=1 dd: error writing '/dev/mem': Bad address 1+0 records in 0+0 records out 0 bytes copied, 0.0283507 s, 0.0 kB/s the command fails as expected. However, the error code is wrong. It should be "Operation not permitted" instead of "Bad address". More seriously, the /dev/mem permission checking in devmem_is_allowed() passes incorrectly. Although the accessing is prevented later because ioremap() isn't allowed to map system RAM, it is a potential security issue. During command executing, the following warning is reported in the kernel log for calling ioremap() on system RAM. ioremap on RAM at 0x0000000490000000 - 0x0000000490000fff WARNING: CPU: 2 PID: 416 at arch/x86/mm/ioremap.c:216 __ioremap_caller.constprop.0+0x131/0x35d Call Trace: memremap+0xcb/0x184 xlate_dev_mem_ptr+0x25/0x2f write_mem+0x94/0xfb vfs_write+0x128/0x26d ksys_write+0xac/0xfe do_syscall_64+0x9a/0xfd entry_SYSCALL_64_after_hwframe+0x4b/0x53 The details of command execution process are as follows. In the above resource tree, "System RAM" is a descendant of "CXL Window 0" instead of a top level resource. So, region_intersects() will report no System RAM resources in the CXL memory region incorrectly, because it only checks the top level resources. Consequently, devmem_is_allowed() will return 1 (allow access via /dev/mem) for CXL memory region incorrectly. Fortunately, ioremap() doesn't allow to map System RAM and reject the access. So, region_intersects() needs to be fixed to work correctly with the resource tree with "System RAM" not at top level as above. To fix it, if we found a unmatched resource in the top level, we will continue to search matched resources in its descendant resources. So, we will not miss any matched resources in resource tree anymore. In the new implementation, an example resource tree |------------- "CXL Window 0" ------------| |-- "System RAM" --| will behave similar as the following fake resource tree for region_intersects(, IORESOURCE_SYSTEM_RAM, ), |-- "System RAM" --||-- "CXL Window 0a" --| Where "CXL Window 0a" is part of the original "CXL Window 0" that isn't covered by "System RAM". Link: https://lkml.kernel.org/r/20240906030713.204292-2-ying.huang@intel.com Fixes: c221c0b0308f ("device-dax: "Hotplug" persistent memory for use like normal RAM") Signed-off-by: "Huang, Ying" Cc: Dan Williams Cc: David Hildenbrand Cc: Davidlohr Bueso Cc: Jonathan Cameron Cc: Dave Jiang Cc: Alison Schofield Cc: Vishal Verma Cc: Ira Weiny Cc: Alistair Popple Cc: Andy Shevchenko Cc: Bjorn Helgaas Cc: Baoquan He Cc: Signed-off-by: Andrew Morton --- kernel/resource.c | 58 ++++++++++++++++++++++++++++++++++++++++------- 1 file changed, 50 insertions(+), 8 deletions(-) diff --git a/kernel/resource.c b/kernel/resource.c index 14777afb0a99..235dc77f8add 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -540,20 +540,62 @@ static int __region_intersects(struct resource *parent, resource_size_t start, size_t size, unsigned long flags, unsigned long desc) { - struct resource res; + resource_size_t ostart, oend; int type = 0; int other = 0; - struct resource *p; + struct resource *p, *dp; + bool is_type, covered; + struct resource res; res.start = start; res.end = start + size - 1; for (p = parent->child; p ; p = p->sibling) { - bool is_type = (((p->flags & flags) == flags) && - ((desc == IORES_DESC_NONE) || - (desc == p->desc))); - - if (resource_overlaps(p, &res)) - is_type ? type++ : other++; + if (!resource_overlaps(p, &res)) + continue; + is_type = (p->flags & flags) == flags && + (desc == IORES_DESC_NONE || desc == p->desc); + if (is_type) { + type++; + continue; + } + /* + * Continue to search in descendant resources as if the + * matched descendant resources cover some ranges of 'p'. + * + * |------------- "CXL Window 0" ------------| + * |-- "System RAM" --| + * + * will behave similar as the following fake resource + * tree when searching "System RAM". + * + * |-- "System RAM" --||-- "CXL Window 0a" --| + */ + covered = false; + ostart = max(res.start, p->start); + oend = min(res.end, p->end); + for_each_resource(p, dp, false) { + if (!resource_overlaps(dp, &res)) + continue; + is_type = (dp->flags & flags) == flags && + (desc == IORES_DESC_NONE || desc == dp->desc); + if (is_type) { + type++; + /* + * Range from 'ostart' to 'dp->start' + * isn't covered by matched resource. + */ + if (dp->start > ostart) + break; + if (dp->end >= oend) { + covered = true; + break; + } + /* Remove covered range */ + ostart = max(ostart, dp->end + 1); + } + } + if (!covered) + other++; } if (type == 0) From 2a058ab3286d6475b2082b90c2d2182d2fea4b39 Mon Sep 17 00:00:00 2001 From: "Vishal Moola (Oracle)" Date: Sat, 14 Sep 2024 12:41:18 -0700 Subject: [PATCH 120/123] mm: change vmf_anon_prepare() to __vmf_anon_prepare() Some callers of vmf_anon_prepare() may not want us to release the per-VMA lock ourselves. Rename vmf_anon_prepare() to __vmf_anon_prepare() and let the callers drop the lock when desired. Also, make vmf_anon_prepare() a wrapper that releases the per-VMA lock itself for any callers that don't care. This is in preparation to fix this bug reported by syzbot: https://lore.kernel.org/linux-mm/00000000000067c20b06219fbc26@google.com/ Link: https://lkml.kernel.org/r/20240914194243.245-1-vishal.moola@gmail.com Fixes: 9acad7ba3e25 ("hugetlb: use vmf_anon_prepare() instead of anon_vma_prepare()") Reported-by: syzbot+2dab93857ee95f2eeb08@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-mm/00000000000067c20b06219fbc26@google.com/ Signed-off-by: Vishal Moola (Oracle) Cc: Muchun Song Cc: Signed-off-by: Andrew Morton --- mm/internal.h | 11 ++++++++++- mm/memory.c | 8 +++----- 2 files changed, 13 insertions(+), 6 deletions(-) diff --git a/mm/internal.h b/mm/internal.h index b4d86436565b..a963f67d3452 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -310,7 +310,16 @@ static inline void wake_throttle_isolated(pg_data_t *pgdat) wake_up(wqh); } -vm_fault_t vmf_anon_prepare(struct vm_fault *vmf); +vm_fault_t __vmf_anon_prepare(struct vm_fault *vmf); +static inline vm_fault_t vmf_anon_prepare(struct vm_fault *vmf) +{ + vm_fault_t ret = __vmf_anon_prepare(vmf); + + if (unlikely(ret & VM_FAULT_RETRY)) + vma_end_read(vmf->vma); + return ret; +} + vm_fault_t do_swap_page(struct vm_fault *vmf); void folio_rotate_reclaimable(struct folio *folio); bool __folio_end_writeback(struct folio *folio); diff --git a/mm/memory.c b/mm/memory.c index 3c01d68065be..95c347ba6305 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -3259,7 +3259,7 @@ static inline vm_fault_t vmf_can_call_fault(const struct vm_fault *vmf) } /** - * vmf_anon_prepare - Prepare to handle an anonymous fault. + * __vmf_anon_prepare - Prepare to handle an anonymous fault. * @vmf: The vm_fault descriptor passed from the fault handler. * * When preparing to insert an anonymous page into a VMA from a @@ -3273,7 +3273,7 @@ static inline vm_fault_t vmf_can_call_fault(const struct vm_fault *vmf) * Return: 0 if fault handling can proceed. Any other value should be * returned to the caller. */ -vm_fault_t vmf_anon_prepare(struct vm_fault *vmf) +vm_fault_t __vmf_anon_prepare(struct vm_fault *vmf) { struct vm_area_struct *vma = vmf->vma; vm_fault_t ret = 0; @@ -3281,10 +3281,8 @@ vm_fault_t vmf_anon_prepare(struct vm_fault *vmf) if (likely(vma->anon_vma)) return 0; if (vmf->flags & FAULT_FLAG_VMA_LOCK) { - if (!mmap_read_trylock(vma->vm_mm)) { - vma_end_read(vma); + if (!mmap_read_trylock(vma->vm_mm)) return VM_FAULT_RETRY; - } } if (__anon_vma_prepare(vma)) ret = VM_FAULT_OOM; From 98b74bb4d7e96b4da5ef3126511febe55b76b807 Mon Sep 17 00:00:00 2001 From: "Vishal Moola (Oracle)" Date: Sat, 14 Sep 2024 12:41:19 -0700 Subject: [PATCH 121/123] mm/hugetlb.c: fix UAF of vma in hugetlb fault pathway Syzbot reports a UAF in hugetlb_fault(). This happens because vmf_anon_prepare() could drop the per-VMA lock and allow the current VMA to be freed before hugetlb_vma_unlock_read() is called. We can fix this by using a modified version of vmf_anon_prepare() that doesn't release the VMA lock on failure, and then release it ourselves after hugetlb_vma_unlock_read(). Link: https://lkml.kernel.org/r/20240914194243.245-2-vishal.moola@gmail.com Fixes: 9acad7ba3e25 ("hugetlb: use vmf_anon_prepare() instead of anon_vma_prepare()") Reported-by: syzbot+2dab93857ee95f2eeb08@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-mm/00000000000067c20b06219fbc26@google.com/ Signed-off-by: Vishal Moola (Oracle) Cc: Muchun Song Cc: Signed-off-by: Andrew Morton --- mm/hugetlb.c | 20 ++++++++++++++++++-- 1 file changed, 18 insertions(+), 2 deletions(-) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index aaf508be0a2b..9a3a6e2dee97 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -6048,7 +6048,7 @@ retry_avoidcopy: * When the original hugepage is shared one, it does not have * anon_vma prepared. */ - ret = vmf_anon_prepare(vmf); + ret = __vmf_anon_prepare(vmf); if (unlikely(ret)) goto out_release_all; @@ -6247,7 +6247,7 @@ static vm_fault_t hugetlb_no_page(struct address_space *mapping, } if (!(vma->vm_flags & VM_MAYSHARE)) { - ret = vmf_anon_prepare(vmf); + ret = __vmf_anon_prepare(vmf); if (unlikely(ret)) goto out; } @@ -6378,6 +6378,14 @@ static vm_fault_t hugetlb_no_page(struct address_space *mapping, folio_unlock(folio); out: hugetlb_vma_unlock_read(vma); + + /* + * We must check to release the per-VMA lock. __vmf_anon_prepare() is + * the only way ret can be set to VM_FAULT_RETRY. + */ + if (unlikely(ret & VM_FAULT_RETRY)) + vma_end_read(vma); + mutex_unlock(&hugetlb_fault_mutex_table[hash]); return ret; @@ -6599,6 +6607,14 @@ out_ptl: } out_mutex: hugetlb_vma_unlock_read(vma); + + /* + * We must check to release the per-VMA lock. __vmf_anon_prepare() in + * hugetlb_wp() is the only way ret can be set to VM_FAULT_RETRY. + */ + if (unlikely(ret & VM_FAULT_RETRY)) + vma_end_read(vma); + mutex_unlock(&hugetlb_fault_mutex_table[hash]); /* * Generally it's safe to hold refcount during waiting page lock. But From 2a1b8648d9be9f37f808a36c0f74adb8c53d06e6 Mon Sep 17 00:00:00 2001 From: Miaohe Lin Date: Sat, 14 Sep 2024 09:53:06 +0800 Subject: [PATCH 122/123] mm/huge_memory: ensure huge_zero_folio won't have large_rmappable flag set Ensure huge_zero_folio won't have large_rmappable flag set. So it can be reported as thp,zero correctly through stable_page_flags(). Link: https://lkml.kernel.org/r/20240914015306.3656791-1-linmiaohe@huawei.com Fixes: 5691753d73a2 ("mm: convert huge_zero_page to huge_zero_folio") Signed-off-by: Miaohe Lin Cc: David Hildenbrand Cc: Matthew Wilcox (Oracle) Cc: Signed-off-by: Andrew Morton --- mm/huge_memory.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 67c86a5d64a6..99b146d16a18 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -220,6 +220,8 @@ retry: count_vm_event(THP_ZERO_PAGE_ALLOC_FAILED); return false; } + /* Ensure zero folio won't have large_rmappable flag set. */ + folio_clear_large_rmappable(zero_folio); preempt_disable(); if (cmpxchg(&huge_zero_folio, NULL, zero_folio)) { preempt_enable(); From 22af8caff7d1ca22a1ff1a554180e53f7a6555af Mon Sep 17 00:00:00 2001 From: Lorenzo Stoakes Date: Fri, 13 Sep 2024 15:06:28 +0100 Subject: [PATCH 123/123] mm/madvise: process_madvise() drop capability check if same mm In commit 96cfe2c0fd23 ("mm/madvise: replace ptrace attach requirement for process_madvise") process_madvise() was updated to require the caller to possess the CAP_SYS_NICE capability to perform the operation, in addition to a check against PTRACE_MODE_READ performed by mm_access(). The mm_access() function explicitly checks to see if the address space of the process being referenced is the current one, in which case no check is performed. We, however, do not do this when checking the CAP_SYS_NICE capability. This means that we insist on the caller possessing this capability in order to perform madvise() operations on its own address space, which seems nonsensical. Simply add a check to allow for an invocation of this function with pidfd set to the current process without elevation. Link: https://lkml.kernel.org/r/20240913140628.77047-1-lorenzo.stoakes@oracle.com Fixes: 96cfe2c0fd23 ("mm/madvise: replace ptrace attach requirement for process_madvise") Signed-off-by: Lorenzo Stoakes Reviewed-by: Liam R. Howlett Acked-by: Vlastimil Babka Acked-by: Shakeel Butt Acked-by: David Rientjes Cc: Kees Cook Cc: Minchan Kim Cc: Suren Baghdasaryan Signed-off-by: Andrew Morton --- mm/madvise.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/mm/madvise.c b/mm/madvise.c index 89089d84f8df..6e3a137b8e50 100644 --- a/mm/madvise.c +++ b/mm/madvise.c @@ -1527,7 +1527,7 @@ SYSCALL_DEFINE5(process_madvise, int, pidfd, const struct iovec __user *, vec, * Require CAP_SYS_NICE for influencing process performance. Note that * only non-destructive hints are currently supported. */ - if (!capable(CAP_SYS_NICE)) { + if (mm != current->mm && !capable(CAP_SYS_NICE)) { ret = -EPERM; goto release_mm; }