From 2cf66b9de406dadbe7598618aa4541261d7bf536 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 11 Jul 2024 09:46:32 +0800 Subject: [PATCH 001/107] f2fs: clean up data_blkaddr() and get_dnode_addr() Introudce a new help get_dnode_base() to wrap common code from get_dnode_addr() and data_blkaddr() for cleanup. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 48 ++++++++++++++++++------------------------------ 1 file changed, 18 insertions(+), 30 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ac19c61f0c3e..ee030a0b093f 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -2900,26 +2900,27 @@ static inline __le32 *blkaddr_in_node(struct f2fs_node *node) } static inline int f2fs_has_extra_attr(struct inode *inode); +static inline unsigned int get_dnode_base(struct inode *inode, + struct page *node_page) +{ + if (!IS_INODE(node_page)) + return 0; + + return inode ? get_extra_isize(inode) : + offset_in_addr(&F2FS_NODE(node_page)->i); +} + +static inline __le32 *get_dnode_addr(struct inode *inode, + struct page *node_page) +{ + return blkaddr_in_node(F2FS_NODE(node_page)) + + get_dnode_base(inode, node_page); +} + static inline block_t data_blkaddr(struct inode *inode, struct page *node_page, unsigned int offset) { - struct f2fs_node *raw_node; - __le32 *addr_array; - int base = 0; - bool is_inode = IS_INODE(node_page); - - raw_node = F2FS_NODE(node_page); - - if (is_inode) { - if (!inode) - /* from GC path only */ - base = offset_in_addr(&raw_node->i); - else if (f2fs_has_extra_attr(inode)) - base = get_extra_isize(inode); - } - - addr_array = blkaddr_in_node(raw_node); - return le32_to_cpu(addr_array[base + offset]); + return le32_to_cpu(*(get_dnode_addr(inode, node_page) + offset)); } static inline block_t f2fs_data_blkaddr(struct dnode_of_data *dn) @@ -3292,8 +3293,6 @@ static inline bool f2fs_is_cow_file(struct inode *inode) return is_inode_flag_set(inode, FI_COW_FILE); } -static inline __le32 *get_dnode_addr(struct inode *inode, - struct page *node_page); static inline void *inline_data_addr(struct inode *inode, struct page *page) { __le32 *addr = get_dnode_addr(inode, page); @@ -3432,17 +3431,6 @@ static inline int get_inline_xattr_addrs(struct inode *inode) return F2FS_I(inode)->i_inline_xattr_size; } -static inline __le32 *get_dnode_addr(struct inode *inode, - struct page *node_page) -{ - int base = 0; - - if (IS_INODE(node_page) && f2fs_has_extra_attr(inode)) - base = get_extra_isize(inode); - - return blkaddr_in_node(F2FS_NODE(node_page)) + base; -} - #define f2fs_get_inode_mode(i) \ ((is_inode_flag_set(i, FI_ACL_MODE)) ? \ (F2FS_I(i)->i_acl_mode) : ((i)->i_mode)) From 47f268f33dff4a5e31541a990dc09f116f80e61c Mon Sep 17 00:00:00 2001 From: Nikita Zhandarovich Date: Wed, 24 Jul 2024 10:05:44 -0700 Subject: [PATCH 002/107] f2fs: prevent possible int overflow in dir_block_index() The result of multiplication between values derived from functions dir_buckets() and bucket_blocks() *could* technically reach 2^30 * 2^2 = 2^32. While unlikely to happen, it is prudent to ensure that it will not lead to integer overflow. Thus, use mul_u32_u32() as it's more appropriate to mitigate the issue. Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE. Fixes: 3843154598a0 ("f2fs: introduce large directory support") Cc: stable@vger.kernel.org Signed-off-by: Nikita Zhandarovich Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/dir.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index cbd7a5e96a37..14900ca8a9ff 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -166,7 +166,8 @@ static unsigned long dir_block_index(unsigned int level, unsigned long bidx = 0; for (i = 0; i < level; i++) - bidx += dir_buckets(i, dir_level) * bucket_blocks(i); + bidx += mul_u32_u32(dir_buckets(i, dir_level), + bucket_blocks(i)); bidx += idx * bucket_blocks(level); return bidx; } From 1cade98cf6415897bf9342ee451cc5b40b58c638 Mon Sep 17 00:00:00 2001 From: Nikita Zhandarovich Date: Wed, 24 Jul 2024 10:28:38 -0700 Subject: [PATCH 003/107] f2fs: fix several potential integer overflows in file offsets When dealing with large extents and calculating file offsets by summing up according extent offsets and lengths of unsigned int type, one may encounter possible integer overflow if the values are big enough. Prevent this from happening by expanding one of the addends to (pgoff_t) type. Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE. Fixes: d323d005ac4a ("f2fs: support file defragment") Cc: stable@vger.kernel.org Signed-off-by: Nikita Zhandarovich Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/extent_cache.c | 4 ++-- fs/f2fs/file.c | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c index fd1fc06359ee..62ac440d9416 100644 --- a/fs/f2fs/extent_cache.c +++ b/fs/f2fs/extent_cache.c @@ -366,7 +366,7 @@ static unsigned int __free_extent_tree(struct f2fs_sb_info *sbi, static void __drop_largest_extent(struct extent_tree *et, pgoff_t fofs, unsigned int len) { - if (fofs < et->largest.fofs + et->largest.len && + if (fofs < (pgoff_t)et->largest.fofs + et->largest.len && fofs + len > et->largest.fofs) { et->largest.len = 0; et->largest_updated = true; @@ -456,7 +456,7 @@ static bool __lookup_extent_tree(struct inode *inode, pgoff_t pgofs, if (type == EX_READ && et->largest.fofs <= pgofs && - et->largest.fofs + et->largest.len > pgofs) { + (pgoff_t)et->largest.fofs + et->largest.len > pgofs) { *ei = et->largest; ret = true; stat_inc_largest_node_hit(sbi); diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 168f08507004..c598cfe5e0ed 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2710,7 +2710,7 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi, * block addresses are continuous. */ if (f2fs_lookup_read_extent_cache(inode, pg_start, &ei)) { - if (ei.fofs + ei.len >= pg_end) + if ((pgoff_t)ei.fofs + ei.len >= pg_end) goto out; } From 50438dbc483ca6a133d2bce9d5d6747bcee38371 Mon Sep 17 00:00:00 2001 From: Nikita Zhandarovich Date: Wed, 24 Jul 2024 10:51:58 -0700 Subject: [PATCH 004/107] f2fs: avoid potential int overflow in sanity_check_area_boundary() While calculating the end addresses of main area and segment 0, u32 may be not enough to hold the result without the danger of int overflow. Just in case, play it safe and cast one of the operands to a wider type (u64). Found by Linux Verification Center (linuxtesting.org) with static analysis tool SVACE. Fixes: fd694733d523 ("f2fs: cover large section in sanity check of super") Cc: stable@vger.kernel.org Signed-off-by: Nikita Zhandarovich Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 3959fd137cc9..4d8f38ca6fcd 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -3356,9 +3356,9 @@ static inline bool sanity_check_area_boundary(struct f2fs_sb_info *sbi, u32 segment_count = le32_to_cpu(raw_super->segment_count); u32 log_blocks_per_seg = le32_to_cpu(raw_super->log_blocks_per_seg); u64 main_end_blkaddr = main_blkaddr + - (segment_count_main << log_blocks_per_seg); + ((u64)segment_count_main << log_blocks_per_seg); u64 seg_end_blkaddr = segment0_blkaddr + - (segment_count << log_blocks_per_seg); + ((u64)segment_count << log_blocks_per_seg); if (segment0_blkaddr != cp_blkaddr) { f2fs_info(sbi, "Mismatch start address, segment0(%u) cp_blkaddr(%u)", From 8444ce524947daf441546b5b3a0c418706dade35 Mon Sep 17 00:00:00 2001 From: Liao Yuanhong Date: Mon, 15 Jul 2024 20:34:51 +0800 Subject: [PATCH 005/107] f2fs: add write priority option based on zone UFS Currently, we are using a mix of traditional UFS and zone UFS to support some functionalities that cannot be achieved on zone UFS alone. However, there are some issues with this approach. There exists a significant performance difference between traditional UFS and zone UFS. Under normal usage, we prioritize writes to zone UFS. However, in critical conditions (such as when the entire UFS is almost full), we cannot determine whether data will be written to traditional UFS or zone UFS. This can lead to significant performance fluctuations, which is not conducive to development and testing. To address this, we have added an option zlu_io_enable under sys with the following three modes: 1) zlu_io_enable == 0:Normal mode, prioritize writing to zone UFS; 2) zlu_io_enable == 1:Zone UFS only mode, only allow writing to zone UFS; 3) zlu_io_enable == 2:Traditional UFS priority mode, prioritize writing to traditional UFS. Signed-off-by: Liao Yuanhong Signed-off-by: Wu Bo Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 14 ++++++++++++++ fs/f2fs/f2fs.h | 8 ++++++++ fs/f2fs/segment.c | 25 ++++++++++++++++++++++++- fs/f2fs/super.c | 1 + fs/f2fs/sysfs.c | 11 +++++++++++ 5 files changed, 58 insertions(+), 1 deletion(-) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index cad6c3dc1f9c..3500920ab7ce 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -763,3 +763,17 @@ Date: November 2023 Contact: "Chao Yu" Description: It controls to enable/disable IO aware feature for background discard. By default, the value is 1 which indicates IO aware is on. + +What: /sys/fs/f2fs//blkzone_alloc_policy +Date: July 2024 +Contact: "Yuanhong Liao" +Description: The zone UFS we are currently using consists of two parts: + conventional zones and sequential zones. It can be used to control which part + to prioritize for writes, with a default value of 0. + + ======================== ========================================= + value description + blkzone_alloc_policy = 0 Prioritize writing to sequential zones + blkzone_alloc_policy = 1 Only allow writing to sequential zones + blkzone_alloc_policy = 2 Prioritize writing to conventional zones + ======================== ========================================= diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ee030a0b093f..ebcb059bd702 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -134,6 +134,12 @@ typedef u32 nid_t; #define COMPRESS_EXT_NUM 16 +enum blkzone_allocation_policy { + BLKZONE_ALLOC_PRIOR_SEQ, /* Prioritize writing to sequential zones */ + BLKZONE_ALLOC_ONLY_SEQ, /* Only allow writing to sequential zones */ + BLKZONE_ALLOC_PRIOR_CONV, /* Prioritize writing to conventional zones */ +}; + /* * An implementation of an rwsem that is explicitly unfair to readers. This * prevents priority inversion when a low-priority reader acquires the read lock @@ -1559,6 +1565,8 @@ struct f2fs_sb_info { #ifdef CONFIG_BLK_DEV_ZONED unsigned int blocks_per_blkz; /* F2FS blocks per zone */ unsigned int max_open_zones; /* max open zone resources of the zoned device */ + /* For adjust the priority writing position of data in zone UFS */ + unsigned int blkzone_alloc_policy; #endif /* for node-related operations */ diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 78c3198a6308..3aee71c9f3c6 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -2686,17 +2686,40 @@ static int get_new_segment(struct f2fs_sb_info *sbi, goto got_it; } +#ifdef CONFIG_BLK_DEV_ZONED /* * If we format f2fs on zoned storage, let's try to get pinned sections * from beginning of the storage, which should be a conventional one. */ if (f2fs_sb_has_blkzoned(sbi)) { - segno = pinning ? 0 : max(first_zoned_segno(sbi), *newseg); + /* Prioritize writing to conventional zones */ + if (sbi->blkzone_alloc_policy == BLKZONE_ALLOC_PRIOR_CONV || pinning) + segno = 0; + else + segno = max(first_zoned_segno(sbi), *newseg); hint = GET_SEC_FROM_SEG(sbi, segno); } +#endif find_other_zone: secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint); + +#ifdef CONFIG_BLK_DEV_ZONED + if (secno >= MAIN_SECS(sbi) && f2fs_sb_has_blkzoned(sbi)) { + /* Write only to sequential zones */ + if (sbi->blkzone_alloc_policy == BLKZONE_ALLOC_ONLY_SEQ) { + hint = GET_SEC_FROM_SEG(sbi, first_zoned_segno(sbi)); + secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint); + } else + secno = find_first_zero_bit(free_i->free_secmap, + MAIN_SECS(sbi)); + if (secno >= MAIN_SECS(sbi)) { + ret = -ENOSPC; + goto out_unlock; + } + } +#endif + if (secno >= MAIN_SECS(sbi)) { secno = find_first_zero_bit(free_i->free_secmap, MAIN_SECS(sbi)); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 4d8f38ca6fcd..aa23e4cfd89e 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -4219,6 +4219,7 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi) sbi->aligned_blksize = true; #ifdef CONFIG_BLK_DEV_ZONED sbi->max_open_zones = UINT_MAX; + sbi->blkzone_alloc_policy = BLKZONE_ALLOC_PRIOR_SEQ; #endif for (i = 0; i < max_devices; i++) { diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index fee7ee45ceaa..63ff2d1647eb 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -627,6 +627,15 @@ out: } #endif +#ifdef CONFIG_BLK_DEV_ZONED + if (!strcmp(a->attr.name, "blkzone_alloc_policy")) { + if (t < BLKZONE_ALLOC_PRIOR_SEQ || t > BLKZONE_ALLOC_PRIOR_CONV) + return -EINVAL; + sbi->blkzone_alloc_policy = t; + return count; + } +#endif + #ifdef CONFIG_F2FS_FS_COMPRESSION if (!strcmp(a->attr.name, "compr_written_block") || !strcmp(a->attr.name, "compr_saved_block")) { @@ -1033,6 +1042,7 @@ F2FS_SBI_GENERAL_RW_ATTR(warm_data_age_threshold); F2FS_SBI_GENERAL_RW_ATTR(last_age_weight); #ifdef CONFIG_BLK_DEV_ZONED F2FS_SBI_GENERAL_RO_ATTR(unusable_blocks_per_sec); +F2FS_SBI_GENERAL_RW_ATTR(blkzone_alloc_policy); #endif /* STAT_INFO ATTR */ @@ -1187,6 +1197,7 @@ static struct attribute *f2fs_attrs[] = { #endif #ifdef CONFIG_BLK_DEV_ZONED ATTR_LIST(unusable_blocks_per_sec), + ATTR_LIST(blkzone_alloc_policy), #endif #ifdef CONFIG_F2FS_FS_COMPRESSION ATTR_LIST(compr_written_block), From d1e1ff971d1aafeaedda69136e6974e3a8792cbd Mon Sep 17 00:00:00 2001 From: Julian Sun Date: Sun, 21 Jul 2024 22:39:12 -0400 Subject: [PATCH 006/107] f2fs: fix macro definition on_f2fs_build_free_nids The macro on_f2fs_build_free_nids accepts a parameter nmi, but it was not used, rather the variable nm_i was directly used, which may be a local variable inside a function that calls the macros. Signed-off-by: Julian Sun Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index b72ef96f7e33..58721a55f173 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -20,7 +20,7 @@ #include "iostat.h" #include -#define on_f2fs_build_free_nids(nmi) mutex_is_locked(&(nm_i)->build_lock) +#define on_f2fs_build_free_nids(nm_i) mutex_is_locked(&(nm_i)->build_lock) static struct kmem_cache *nat_entry_slab; static struct kmem_cache *free_nid_slab; From d72750e4a7528b83a07b65335c35b454a98593d4 Mon Sep 17 00:00:00 2001 From: Julian Sun Date: Sun, 21 Jul 2024 22:39:13 -0400 Subject: [PATCH 007/107] f2fs: fix macro definition stat_inc_cp_count The macro stat_inc_cp_count accepts a parameter si, but it was not used, rather the variable sbi was directly used, which may be a local variable inside a function that calls the macros. Signed-off-by: Julian Sun Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ebcb059bd702..0e181dde79c7 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3983,7 +3983,7 @@ static inline struct f2fs_stat_info *F2FS_STAT(struct f2fs_sb_info *sbi) #define stat_inc_cp_call_count(sbi, foreground) \ atomic_inc(&sbi->cp_call_count[(foreground)]) -#define stat_inc_cp_count(si) (F2FS_STAT(sbi)->cp_count++) +#define stat_inc_cp_count(sbi) (F2FS_STAT(sbi)->cp_count++) #define stat_io_skip_bggc_count(sbi) ((sbi)->io_skip_bggc++) #define stat_other_skip_bggc_count(sbi) ((sbi)->other_skip_bggc++) #define stat_inc_dirty_inode(sbi, type) ((sbi)->ndirty_inode[type]++) From 1a0bd289a5db1df8df8fab949633a0b8d3f235ee Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 25 Jun 2024 11:13:48 +0800 Subject: [PATCH 008/107] f2fs: atomic: fix to avoid racing w/ GC Case #1: SQLite App GC Thread Kworker Shrinker - f2fs_ioc_start_atomic_write - f2fs_ioc_commit_atomic_write - f2fs_commit_atomic_write - filemap_write_and_wait_range : write atomic_file's data to cow_inode echo 3 > drop_caches to drop atomic_file's cache. - f2fs_gc - gc_data_segment - move_data_page - set_page_dirty - writepages - f2fs_do_write_data_page : overwrite atomic_file's data to cow_inode - f2fs_down_write(&fi->i_gc_rwsem[WRITE]) - __f2fs_commit_atomic_write - f2fs_up_write(&fi->i_gc_rwsem[WRITE]) Case #2: SQLite App GC Thread Kworker - f2fs_ioc_start_atomic_write - __writeback_single_inode - do_writepages - f2fs_write_cache_pages - f2fs_write_single_data_page - f2fs_do_write_data_page : write atomic_file's data to cow_inode - f2fs_gc - gc_data_segment - move_data_page - set_page_dirty - writepages - f2fs_do_write_data_page : overwrite atomic_file's data to cow_inode - f2fs_ioc_commit_atomic_write In above cases racing in between atomic_write and GC, previous data in atomic_file may be overwrited to cow_file, result in data corruption. This patch introduces PAGE_PRIVATE_ATOMIC_WRITE bit flag in page.private, and use it to indicate that there is last dirty data in atomic file, and the data should be writebacked into cow_file, if the flag is not tagged in page, we should never write data across files. Fixes: 3db1de0e582c ("f2fs: change the current atomic write way") Cc: Daeho Jeong Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 10 +++++++++- fs/f2fs/f2fs.h | 8 +++++++- 2 files changed, 16 insertions(+), 2 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 6457e5bca9c9..be66b3a0e793 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2650,10 +2650,13 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio) struct dnode_of_data dn; struct node_info ni; bool ipu_force = false; + bool atomic_commit; int err = 0; /* Use COW inode to make dnode_of_data for atomic write */ - if (f2fs_is_atomic_file(inode)) + atomic_commit = f2fs_is_atomic_file(inode) && + page_private_atomic(fio->page); + if (atomic_commit) set_new_dnode(&dn, F2FS_I(inode)->cow_inode, NULL, NULL, 0); else set_new_dnode(&dn, inode, NULL, NULL, 0); @@ -2752,6 +2755,8 @@ got_it: f2fs_outplace_write_data(&dn, fio); trace_f2fs_do_write_data_page(page_folio(page), OPU); set_inode_flag(inode, FI_APPEND_WRITE); + if (atomic_commit) + clear_page_private_atomic(page); out_writepage: f2fs_put_dnode(&dn); out: @@ -3721,6 +3726,9 @@ static int f2fs_write_end(struct file *file, set_page_dirty(page); + if (f2fs_is_atomic_file(inode)) + set_page_private_atomic(page); + if (pos + copied > i_size_read(inode) && !f2fs_verity_in_progress(inode)) { f2fs_i_size_write(inode, pos + copied); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 0e181dde79c7..7b00846af915 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1424,7 +1424,8 @@ static inline void f2fs_clear_bit(unsigned int nr, char *addr); * bit 1 PAGE_PRIVATE_ONGOING_MIGRATION * bit 2 PAGE_PRIVATE_INLINE_INODE * bit 3 PAGE_PRIVATE_REF_RESOURCE - * bit 4- f2fs private data + * bit 4 PAGE_PRIVATE_ATOMIC_WRITE + * bit 5- f2fs private data * * Layout B: lowest bit should be 0 * page.private is a wrapped pointer. @@ -1434,6 +1435,7 @@ enum { PAGE_PRIVATE_ONGOING_MIGRATION, /* data page which is on-going migrating */ PAGE_PRIVATE_INLINE_INODE, /* inode page contains inline data */ PAGE_PRIVATE_REF_RESOURCE, /* dirty page has referenced resources */ + PAGE_PRIVATE_ATOMIC_WRITE, /* data page from atomic write path */ PAGE_PRIVATE_MAX }; @@ -2404,14 +2406,17 @@ static inline void clear_page_private_##name(struct page *page) \ PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER); PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION); +PAGE_PRIVATE_GET_FUNC(atomic, ATOMIC_WRITE); PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE); PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION); +PAGE_PRIVATE_SET_FUNC(atomic, ATOMIC_WRITE); PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE); PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION); +PAGE_PRIVATE_CLEAR_FUNC(atomic, ATOMIC_WRITE); static inline unsigned long get_page_private_data(struct page *page) { @@ -2443,6 +2448,7 @@ static inline void clear_page_private_all(struct page *page) clear_page_private_reference(page); clear_page_private_gcing(page); clear_page_private_inline(page); + clear_page_private_atomic(page); f2fs_bug_on(F2FS_P_SB(page), page_private(page)); } From d71973707efe8522b1f95189113273d3d5652808 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 30 Jul 2024 01:16:02 -0400 Subject: [PATCH 009/107] bpf: convert __bpf_prog_get() to CLASS(fd, ...) Irregularity here is fdput() not in the same scope as fdget(); just fold ____bpf_prog_get() into its (only) caller and that's it... Signed-off-by: Al Viro Acked-by: Andrii Nakryiko Reviewed-by: Christian Brauner Signed-off-by: Andrii Nakryiko --- kernel/bpf/syscall.c | 31 +++++++++---------------------- 1 file changed, 9 insertions(+), 22 deletions(-) diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 3093bf2cc266..4909e3f23065 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -2407,18 +2407,6 @@ int bpf_prog_new_fd(struct bpf_prog *prog) O_RDWR | O_CLOEXEC); } -static struct bpf_prog *____bpf_prog_get(struct fd f) -{ - if (!fd_file(f)) - return ERR_PTR(-EBADF); - if (fd_file(f)->f_op != &bpf_prog_fops) { - fdput(f); - return ERR_PTR(-EINVAL); - } - - return fd_file(f)->private_data; -} - void bpf_prog_add(struct bpf_prog *prog, int i) { atomic64_add(i, &prog->aux->refcnt); @@ -2474,20 +2462,19 @@ bool bpf_prog_get_ok(struct bpf_prog *prog, static struct bpf_prog *__bpf_prog_get(u32 ufd, enum bpf_prog_type *attach_type, bool attach_drv) { - struct fd f = fdget(ufd); + CLASS(fd, f)(ufd); struct bpf_prog *prog; - prog = ____bpf_prog_get(f); - if (IS_ERR(prog)) - return prog; - if (!bpf_prog_get_ok(prog, attach_type, attach_drv)) { - prog = ERR_PTR(-EINVAL); - goto out; - } + if (fd_empty(f)) + return ERR_PTR(-EBADF); + if (fd_file(f)->f_op != &bpf_prog_fops) + return ERR_PTR(-EINVAL); + + prog = fd_file(f)->private_data; + if (!bpf_prog_get_ok(prog, attach_type, attach_drv)) + return ERR_PTR(-EINVAL); bpf_prog_inc(prog); -out: - fdput(f); return prog; } From 51a1ca933f5dc9e764ac065ff0949130cc32a48b Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 13 Aug 2024 14:18:38 -0700 Subject: [PATCH 010/107] bpf: switch fdget_raw() uses to CLASS(fd_raw, ...) Swith fdget_raw() use cases in bpf_inode_storage.c to CLASS(fd_raw). Reviewed-by: Christian Brauner Signed-off-by: Al Viro Signed-off-by: Andrii Nakryiko --- kernel/bpf/bpf_inode_storage.c | 24 ++++++++---------------- 1 file changed, 8 insertions(+), 16 deletions(-) diff --git a/kernel/bpf/bpf_inode_storage.c b/kernel/bpf/bpf_inode_storage.c index 0a79aee6523d..29da6d3838f6 100644 --- a/kernel/bpf/bpf_inode_storage.c +++ b/kernel/bpf/bpf_inode_storage.c @@ -78,13 +78,12 @@ void bpf_inode_storage_free(struct inode *inode) static void *bpf_fd_inode_storage_lookup_elem(struct bpf_map *map, void *key) { struct bpf_local_storage_data *sdata; - struct fd f = fdget_raw(*(int *)key); + CLASS(fd_raw, f)(*(int *)key); - if (!fd_file(f)) + if (fd_empty(f)) return ERR_PTR(-EBADF); sdata = inode_storage_lookup(file_inode(fd_file(f)), map, true); - fdput(f); return sdata ? sdata->data : NULL; } @@ -92,19 +91,16 @@ static long bpf_fd_inode_storage_update_elem(struct bpf_map *map, void *key, void *value, u64 map_flags) { struct bpf_local_storage_data *sdata; - struct fd f = fdget_raw(*(int *)key); + CLASS(fd_raw, f)(*(int *)key); - if (!fd_file(f)) + if (fd_empty(f)) return -EBADF; - if (!inode_storage_ptr(file_inode(fd_file(f)))) { - fdput(f); + if (!inode_storage_ptr(file_inode(fd_file(f)))) return -EBADF; - } sdata = bpf_local_storage_update(file_inode(fd_file(f)), (struct bpf_local_storage_map *)map, value, map_flags, GFP_ATOMIC); - fdput(f); return PTR_ERR_OR_ZERO(sdata); } @@ -123,15 +119,11 @@ static int inode_storage_delete(struct inode *inode, struct bpf_map *map) static long bpf_fd_inode_storage_delete_elem(struct bpf_map *map, void *key) { - struct fd f = fdget_raw(*(int *)key); - int err; + CLASS(fd_raw, f)(*(int *)key); - if (!fd_file(f)) + if (fd_empty(f)) return -EBADF; - - err = inode_storage_delete(file_inode(fd_file(f)), map); - fdput(f); - return err; + return inode_storage_delete(file_inode(fd_file(f)), map); } /* *gfp_flags* is a hidden argument provided by the verifier */ From 535ead44ffd08479212e31729a7118bd4e9ac699 Mon Sep 17 00:00:00 2001 From: Andrii Nakryiko Date: Tue, 6 Aug 2024 14:31:34 -0700 Subject: [PATCH 011/107] bpf: factor out fetching bpf_map from FD and adding it to used_maps list Factor out the logic to extract bpf_map instances from FD embedded in bpf_insns, adding it to the list of used_maps (unless it's already there, in which case we just reuse map's index). This simplifies the logic in resolve_pseudo_ldimm64(), especially around `struct fd` handling, as all that is now neatly contained in the helper and doesn't leak into a dozen error handling paths. Signed-off-by: Andrii Nakryiko --- kernel/bpf/verifier.c | 119 ++++++++++++++++++++++++------------------ 1 file changed, 68 insertions(+), 51 deletions(-) diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index df3be12096cf..14e4ef687a59 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -18865,6 +18865,58 @@ static bool bpf_map_is_cgroup_storage(struct bpf_map *map) map->map_type == BPF_MAP_TYPE_PERCPU_CGROUP_STORAGE); } +/* Add map behind fd to used maps list, if it's not already there, and return + * its index. Also set *reused to true if this map was already in the list of + * used maps. + * Returns <0 on error, or >= 0 index, on success. + */ +static int add_used_map_from_fd(struct bpf_verifier_env *env, int fd, bool *reused) +{ + struct fd f = fdget(fd); + struct bpf_map *map; + int i; + + map = __bpf_map_get(f); + if (IS_ERR(map)) { + verbose(env, "fd %d is not pointing to valid bpf_map\n", fd); + return PTR_ERR(map); + } + + /* check whether we recorded this map already */ + for (i = 0; i < env->used_map_cnt; i++) { + if (env->used_maps[i] == map) { + *reused = true; + fdput(f); + return i; + } + } + + if (env->used_map_cnt >= MAX_USED_MAPS) { + verbose(env, "The total number of maps per program has reached the limit of %u\n", + MAX_USED_MAPS); + fdput(f); + return -E2BIG; + } + + if (env->prog->sleepable) + atomic64_inc(&map->sleepable_refcnt); + + /* hold the map. If the program is rejected by verifier, + * the map will be released by release_maps() or it + * will be used by the valid program until it's unloaded + * and all maps are released in bpf_free_used_maps() + */ + bpf_map_inc(map); + + *reused = false; + env->used_maps[env->used_map_cnt++] = map; + + fdput(f); + + return env->used_map_cnt - 1; + +} + /* find and rewrite pseudo imm in ld_imm64 instructions: * * 1. if it accesses map FD, replace it with actual map pointer. @@ -18876,7 +18928,7 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) { struct bpf_insn *insn = env->prog->insnsi; int insn_cnt = env->prog->len; - int i, j, err; + int i, err; err = bpf_prog_calc_tag(env->prog); if (err) @@ -18893,9 +18945,10 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) if (insn[0].code == (BPF_LD | BPF_IMM | BPF_DW)) { struct bpf_insn_aux_data *aux; struct bpf_map *map; - struct fd f; + int map_idx; u64 addr; u32 fd; + bool reused; if (i == insn_cnt - 1 || insn[1].code != 0 || insn[1].dst_reg != 0 || insn[1].src_reg != 0 || @@ -18956,20 +19009,18 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) break; } - f = fdget(fd); - map = __bpf_map_get(f); - if (IS_ERR(map)) { - verbose(env, "fd %d is not pointing to valid bpf_map\n", fd); - return PTR_ERR(map); - } - - err = check_map_prog_compatibility(env, map, env->prog); - if (err) { - fdput(f); - return err; - } + map_idx = add_used_map_from_fd(env, fd, &reused); + if (map_idx < 0) + return map_idx; + map = env->used_maps[map_idx]; aux = &env->insn_aux_data[i]; + aux->map_index = map_idx; + + err = check_map_prog_compatibility(env, map, env->prog); + if (err) + return err; + if (insn[0].src_reg == BPF_PSEUDO_MAP_FD || insn[0].src_reg == BPF_PSEUDO_MAP_IDX) { addr = (unsigned long)map; @@ -18978,13 +19029,11 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) if (off >= BPF_MAX_VAR_OFF) { verbose(env, "direct value offset of %u is not allowed\n", off); - fdput(f); return -EINVAL; } if (!map->ops->map_direct_value_addr) { verbose(env, "no direct value access support for this map type\n"); - fdput(f); return -EINVAL; } @@ -18992,7 +19041,6 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) if (err) { verbose(env, "invalid access to map value pointer, value_size=%u off=%u\n", map->value_size, off); - fdput(f); return err; } @@ -19003,70 +19051,39 @@ static int resolve_pseudo_ldimm64(struct bpf_verifier_env *env) insn[0].imm = (u32)addr; insn[1].imm = addr >> 32; - /* check whether we recorded this map already */ - for (j = 0; j < env->used_map_cnt; j++) { - if (env->used_maps[j] == map) { - aux->map_index = j; - fdput(f); - goto next_insn; - } - } - - if (env->used_map_cnt >= MAX_USED_MAPS) { - verbose(env, "The total number of maps per program has reached the limit of %u\n", - MAX_USED_MAPS); - fdput(f); - return -E2BIG; - } - - if (env->prog->sleepable) - atomic64_inc(&map->sleepable_refcnt); - /* hold the map. If the program is rejected by verifier, - * the map will be released by release_maps() or it - * will be used by the valid program until it's unloaded - * and all maps are released in bpf_free_used_maps() - */ - bpf_map_inc(map); - - aux->map_index = env->used_map_cnt; - env->used_maps[env->used_map_cnt++] = map; + /* proceed with extra checks only if its newly added used map */ + if (reused) + goto next_insn; if (bpf_map_is_cgroup_storage(map) && bpf_cgroup_storage_assign(env->prog->aux, map)) { verbose(env, "only one cgroup storage of each type is allowed\n"); - fdput(f); return -EBUSY; } if (map->map_type == BPF_MAP_TYPE_ARENA) { if (env->prog->aux->arena) { verbose(env, "Only one arena per program\n"); - fdput(f); return -EBUSY; } if (!env->allow_ptr_leaks || !env->bpf_capable) { verbose(env, "CAP_BPF and CAP_PERFMON are required to use arena\n"); - fdput(f); return -EPERM; } if (!env->prog->jit_requested) { verbose(env, "JIT is required to use arena\n"); - fdput(f); return -EOPNOTSUPP; } if (!bpf_jit_supports_arena()) { verbose(env, "JIT doesn't support arena\n"); - fdput(f); return -EOPNOTSUPP; } env->prog->aux->arena = (void *)map; if (!bpf_arena_get_user_vm_start(env->prog->aux->arena)) { verbose(env, "arena's user address must be set via map_extra or mmap()\n"); - fdput(f); return -EINVAL; } } - fdput(f); next_insn: insn++; i++; From 55f325958ccc41eaea43eb4546d4dc77c1b5ef8a Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 30 Jul 2024 01:16:04 -0400 Subject: [PATCH 012/107] bpf: switch maps to CLASS(fd, ...) Calling conventions for __bpf_map_get() would be more convenient if it left fpdut() on failure to callers. Makes for simpler logics in the callers. Among other things, the proof of memory safety no longer has to rely upon file->private_data never being ERR_PTR(...) for bpffs files. Original calling conventions made it impossible for the caller to tell whether __bpf_map_get() has returned ERR_PTR(-EINVAL) because it has found the file not be a bpf map one (in which case it would've done fdput()) or because it found that ERR_PTR(-EINVAL) in file->private_data of a bpf map file (in which case fdput() would _not_ have been done). Signed-off-by: Al Viro Reviewed-by: Christian Brauner Signed-off-by: Andrii Nakryiko --- include/linux/bpf.h | 11 +++- kernel/bpf/map_in_map.c | 38 ++++--------- kernel/bpf/syscall.c | 118 ++++++++++------------------------------ kernel/bpf/verifier.c | 7 +-- net/core/sock_map.c | 23 ++------ 5 files changed, 58 insertions(+), 139 deletions(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index b9425e410bcb..9f35df07e86d 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -2241,7 +2241,16 @@ void __bpf_obj_drop_impl(void *p, const struct btf_record *rec, bool percpu); struct bpf_map *bpf_map_get(u32 ufd); struct bpf_map *bpf_map_get_with_uref(u32 ufd); -struct bpf_map *__bpf_map_get(struct fd f); + +static inline struct bpf_map *__bpf_map_get(struct fd f) +{ + if (fd_empty(f)) + return ERR_PTR(-EBADF); + if (unlikely(fd_file(f)->f_op != &bpf_map_fops)) + return ERR_PTR(-EINVAL); + return fd_file(f)->private_data; +} + void bpf_map_inc(struct bpf_map *map); void bpf_map_inc_with_uref(struct bpf_map *map); struct bpf_map *__bpf_map_inc_not_zero(struct bpf_map *map, bool uref); diff --git a/kernel/bpf/map_in_map.c b/kernel/bpf/map_in_map.c index b4f18c85d7bc..645bd30bc9a9 100644 --- a/kernel/bpf/map_in_map.c +++ b/kernel/bpf/map_in_map.c @@ -11,24 +11,18 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd) { struct bpf_map *inner_map, *inner_map_meta; u32 inner_map_meta_size; - struct fd f; - int ret; + CLASS(fd, f)(inner_map_ufd); - f = fdget(inner_map_ufd); inner_map = __bpf_map_get(f); if (IS_ERR(inner_map)) return inner_map; /* Does not support >1 level map-in-map */ - if (inner_map->inner_map_meta) { - ret = -EINVAL; - goto put; - } + if (inner_map->inner_map_meta) + return ERR_PTR(-EINVAL); - if (!inner_map->ops->map_meta_equal) { - ret = -ENOTSUPP; - goto put; - } + if (!inner_map->ops->map_meta_equal) + return ERR_PTR(-ENOTSUPP); inner_map_meta_size = sizeof(*inner_map_meta); /* In some cases verifier needs to access beyond just base map. */ @@ -36,10 +30,8 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd) inner_map_meta_size = sizeof(struct bpf_array); inner_map_meta = kzalloc(inner_map_meta_size, GFP_USER); - if (!inner_map_meta) { - ret = -ENOMEM; - goto put; - } + if (!inner_map_meta) + return ERR_PTR(-ENOMEM); inner_map_meta->map_type = inner_map->map_type; inner_map_meta->key_size = inner_map->key_size; @@ -53,8 +45,9 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd) * invalid/empty/valid, but ERR_PTR in case of errors. During * equality NULL or IS_ERR is equivalent. */ - ret = PTR_ERR(inner_map_meta->record); - goto free; + struct bpf_map *ret = ERR_CAST(inner_map_meta->record); + kfree(inner_map_meta); + return ret; } /* Note: We must use the same BTF, as we also used btf_record_dup above * which relies on BTF being same for both maps, as some members like @@ -77,14 +70,7 @@ struct bpf_map *bpf_map_meta_alloc(int inner_map_ufd) inner_array_meta->elem_size = inner_array->elem_size; inner_map_meta->bypass_spec_v1 = inner_map->bypass_spec_v1; } - - fdput(f); return inner_map_meta; -free: - kfree(inner_map_meta); -put: - fdput(f); - return ERR_PTR(ret); } void bpf_map_meta_free(struct bpf_map *map_meta) @@ -110,9 +96,8 @@ void *bpf_map_fd_get_ptr(struct bpf_map *map, int ufd) { struct bpf_map *inner_map, *inner_map_meta; - struct fd f; + CLASS(fd, f)(ufd); - f = fdget(ufd); inner_map = __bpf_map_get(f); if (IS_ERR(inner_map)) return inner_map; @@ -123,7 +108,6 @@ void *bpf_map_fd_get_ptr(struct bpf_map *map, else inner_map = ERR_PTR(-EINVAL); - fdput(f); return inner_map; } diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index 4909e3f23065..ab0d94f41c48 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -1418,21 +1418,6 @@ put_token: return err; } -/* if error is returned, fd is released. - * On success caller should complete fd access with matching fdput() - */ -struct bpf_map *__bpf_map_get(struct fd f) -{ - if (!fd_file(f)) - return ERR_PTR(-EBADF); - if (fd_file(f)->f_op != &bpf_map_fops) { - fdput(f); - return ERR_PTR(-EINVAL); - } - - return fd_file(f)->private_data; -} - void bpf_map_inc(struct bpf_map *map) { atomic64_inc(&map->refcnt); @@ -1448,15 +1433,11 @@ EXPORT_SYMBOL_GPL(bpf_map_inc_with_uref); struct bpf_map *bpf_map_get(u32 ufd) { - struct fd f = fdget(ufd); - struct bpf_map *map; + CLASS(fd, f)(ufd); + struct bpf_map *map = __bpf_map_get(f); - map = __bpf_map_get(f); - if (IS_ERR(map)) - return map; - - bpf_map_inc(map); - fdput(f); + if (!IS_ERR(map)) + bpf_map_inc(map); return map; } @@ -1464,15 +1445,11 @@ EXPORT_SYMBOL(bpf_map_get); struct bpf_map *bpf_map_get_with_uref(u32 ufd) { - struct fd f = fdget(ufd); - struct bpf_map *map; + CLASS(fd, f)(ufd); + struct bpf_map *map = __bpf_map_get(f); - map = __bpf_map_get(f); - if (IS_ERR(map)) - return map; - - bpf_map_inc_with_uref(map); - fdput(f); + if (!IS_ERR(map)) + bpf_map_inc_with_uref(map); return map; } @@ -1537,11 +1514,9 @@ static int map_lookup_elem(union bpf_attr *attr) { void __user *ukey = u64_to_user_ptr(attr->key); void __user *uvalue = u64_to_user_ptr(attr->value); - int ufd = attr->map_fd; struct bpf_map *map; void *key, *value; u32 value_size; - struct fd f; int err; if (CHECK_ATTR(BPF_MAP_LOOKUP_ELEM)) @@ -1550,26 +1525,20 @@ static int map_lookup_elem(union bpf_attr *attr) if (attr->flags & ~BPF_F_LOCK) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); - if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) { - err = -EPERM; - goto err_put; - } + if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) + return -EPERM; if ((attr->flags & BPF_F_LOCK) && - !btf_record_has_field(map->record, BPF_SPIN_LOCK)) { - err = -EINVAL; - goto err_put; - } + !btf_record_has_field(map->record, BPF_SPIN_LOCK)) + return -EINVAL; key = __bpf_copy_key(ukey, map->key_size); - if (IS_ERR(key)) { - err = PTR_ERR(key); - goto err_put; - } + if (IS_ERR(key)) + return PTR_ERR(key); value_size = bpf_map_value_size(map); @@ -1600,8 +1569,6 @@ free_value: kvfree(value); free_key: kvfree(key); -err_put: - fdput(f); return err; } @@ -1612,17 +1579,15 @@ static int map_update_elem(union bpf_attr *attr, bpfptr_t uattr) { bpfptr_t ukey = make_bpfptr(attr->key, uattr.is_kernel); bpfptr_t uvalue = make_bpfptr(attr->value, uattr.is_kernel); - int ufd = attr->map_fd; struct bpf_map *map; void *key, *value; u32 value_size; - struct fd f; int err; if (CHECK_ATTR(BPF_MAP_UPDATE_ELEM)) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); @@ -1660,7 +1625,6 @@ free_key: kvfree(key); err_put: bpf_map_write_active_dec(map); - fdput(f); return err; } @@ -1669,16 +1633,14 @@ err_put: static int map_delete_elem(union bpf_attr *attr, bpfptr_t uattr) { bpfptr_t ukey = make_bpfptr(attr->key, uattr.is_kernel); - int ufd = attr->map_fd; struct bpf_map *map; - struct fd f; void *key; int err; if (CHECK_ATTR(BPF_MAP_DELETE_ELEM)) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); @@ -1715,7 +1677,6 @@ out: kvfree(key); err_put: bpf_map_write_active_dec(map); - fdput(f); return err; } @@ -1726,30 +1687,24 @@ static int map_get_next_key(union bpf_attr *attr) { void __user *ukey = u64_to_user_ptr(attr->key); void __user *unext_key = u64_to_user_ptr(attr->next_key); - int ufd = attr->map_fd; struct bpf_map *map; void *key, *next_key; - struct fd f; int err; if (CHECK_ATTR(BPF_MAP_GET_NEXT_KEY)) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); - if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) { - err = -EPERM; - goto err_put; - } + if (!(map_get_sys_perms(map, f) & FMODE_CAN_READ)) + return -EPERM; if (ukey) { key = __bpf_copy_key(ukey, map->key_size); - if (IS_ERR(key)) { - err = PTR_ERR(key); - goto err_put; - } + if (IS_ERR(key)) + return PTR_ERR(key); } else { key = NULL; } @@ -1781,8 +1736,6 @@ free_next_key: kvfree(next_key); free_key: kvfree(key); -err_put: - fdput(f); return err; } @@ -2011,11 +1964,9 @@ static int map_lookup_and_delete_elem(union bpf_attr *attr) { void __user *ukey = u64_to_user_ptr(attr->key); void __user *uvalue = u64_to_user_ptr(attr->value); - int ufd = attr->map_fd; struct bpf_map *map; void *key, *value; u32 value_size; - struct fd f; int err; if (CHECK_ATTR(BPF_MAP_LOOKUP_AND_DELETE_ELEM)) @@ -2024,7 +1975,7 @@ static int map_lookup_and_delete_elem(union bpf_attr *attr) if (attr->flags & ~BPF_F_LOCK) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); @@ -2094,7 +2045,6 @@ free_key: kvfree(key); err_put: bpf_map_write_active_dec(map); - fdput(f); return err; } @@ -2102,27 +2052,22 @@ err_put: static int map_freeze(const union bpf_attr *attr) { - int err = 0, ufd = attr->map_fd; + int err = 0; struct bpf_map *map; - struct fd f; if (CHECK_ATTR(BPF_MAP_FREEZE)) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->map_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); - if (map->map_type == BPF_MAP_TYPE_STRUCT_OPS || !IS_ERR_OR_NULL(map->record)) { - fdput(f); + if (map->map_type == BPF_MAP_TYPE_STRUCT_OPS || !IS_ERR_OR_NULL(map->record)) return -ENOTSUPP; - } - if (!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) { - fdput(f); + if (!(map_get_sys_perms(map, f) & FMODE_CAN_WRITE)) return -EPERM; - } mutex_lock(&map->freeze_mutex); if (bpf_map_write_active(map)) { @@ -2137,7 +2082,6 @@ static int map_freeze(const union bpf_attr *attr) WRITE_ONCE(map->frozen, true); err_put: mutex_unlock(&map->freeze_mutex); - fdput(f); return err; } @@ -5175,14 +5119,13 @@ static int bpf_map_do_batch(const union bpf_attr *attr, cmd == BPF_MAP_LOOKUP_AND_DELETE_BATCH; bool has_write = cmd != BPF_MAP_LOOKUP_BATCH; struct bpf_map *map; - int err, ufd; - struct fd f; + int err; if (CHECK_ATTR(BPF_MAP_BATCH)) return -EINVAL; - ufd = attr->batch.map_fd; - f = fdget(ufd); + CLASS(fd, f)(attr->batch.map_fd); + map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); @@ -5210,7 +5153,6 @@ err_put: maybe_wait_bpf_programs(map); bpf_map_write_active_dec(map); } - fdput(f); return err; } diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c index 14e4ef687a59..e3932f8ce10a 100644 --- a/kernel/bpf/verifier.c +++ b/kernel/bpf/verifier.c @@ -18872,7 +18872,7 @@ static bool bpf_map_is_cgroup_storage(struct bpf_map *map) */ static int add_used_map_from_fd(struct bpf_verifier_env *env, int fd, bool *reused) { - struct fd f = fdget(fd); + CLASS(fd, f)(fd); struct bpf_map *map; int i; @@ -18886,7 +18886,6 @@ static int add_used_map_from_fd(struct bpf_verifier_env *env, int fd, bool *reus for (i = 0; i < env->used_map_cnt; i++) { if (env->used_maps[i] == map) { *reused = true; - fdput(f); return i; } } @@ -18894,7 +18893,6 @@ static int add_used_map_from_fd(struct bpf_verifier_env *env, int fd, bool *reus if (env->used_map_cnt >= MAX_USED_MAPS) { verbose(env, "The total number of maps per program has reached the limit of %u\n", MAX_USED_MAPS); - fdput(f); return -E2BIG; } @@ -18911,10 +18909,7 @@ static int add_used_map_from_fd(struct bpf_verifier_env *env, int fd, bool *reus *reused = false; env->used_maps[env->used_map_cnt++] = map; - fdput(f); - return env->used_map_cnt - 1; - } /* find and rewrite pseudo imm in ld_imm64 instructions: diff --git a/net/core/sock_map.c b/net/core/sock_map.c index d3dbb92153f2..0f5f80f44d52 100644 --- a/net/core/sock_map.c +++ b/net/core/sock_map.c @@ -67,46 +67,39 @@ static struct bpf_map *sock_map_alloc(union bpf_attr *attr) int sock_map_get_from_fd(const union bpf_attr *attr, struct bpf_prog *prog) { - u32 ufd = attr->target_fd; struct bpf_map *map; - struct fd f; int ret; if (attr->attach_flags || attr->replace_bpf_fd) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->target_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); mutex_lock(&sockmap_mutex); ret = sock_map_prog_update(map, prog, NULL, NULL, attr->attach_type); mutex_unlock(&sockmap_mutex); - fdput(f); return ret; } int sock_map_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype) { - u32 ufd = attr->target_fd; struct bpf_prog *prog; struct bpf_map *map; - struct fd f; int ret; if (attr->attach_flags || attr->replace_bpf_fd) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->target_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); prog = bpf_prog_get(attr->attach_bpf_fd); - if (IS_ERR(prog)) { - ret = PTR_ERR(prog); - goto put_map; - } + if (IS_ERR(prog)) + return PTR_ERR(prog); if (prog->type != ptype) { ret = -EINVAL; @@ -118,8 +111,6 @@ int sock_map_prog_detach(const union bpf_attr *attr, enum bpf_prog_type ptype) mutex_unlock(&sockmap_mutex); put_prog: bpf_prog_put(prog); -put_map: - fdput(f); return ret; } @@ -1550,18 +1541,17 @@ int sock_map_bpf_prog_query(const union bpf_attr *attr, union bpf_attr __user *uattr) { __u32 __user *prog_ids = u64_to_user_ptr(attr->query.prog_ids); - u32 prog_cnt = 0, flags = 0, ufd = attr->target_fd; + u32 prog_cnt = 0, flags = 0; struct bpf_prog **pprog; struct bpf_prog *prog; struct bpf_map *map; - struct fd f; u32 id = 0; int ret; if (attr->query.query_flags) return -EINVAL; - f = fdget(ufd); + CLASS(fd, f)(attr->target_fd); map = __bpf_map_get(f); if (IS_ERR(map)) return PTR_ERR(map); @@ -1593,7 +1583,6 @@ end: copy_to_user(&uattr->query.prog_cnt, &prog_cnt, sizeof(prog_cnt))) ret = -EFAULT; - fdput(f); return ret; } From eb80ee85801cd8e3c6f39b08830867d2afecd8f5 Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 13 Aug 2024 14:34:10 -0700 Subject: [PATCH 013/107] bpf: trivial conversions for fdget() fdget() is the first thing done in scope, all matching fdput() are immediately followed by leaving the scope. Reviewed-by: Christian Brauner Signed-off-by: Al Viro Signed-off-by: Andrii Nakryiko --- kernel/bpf/btf.c | 11 +++-------- kernel/bpf/syscall.c | 10 +++------- kernel/bpf/token.c | 9 +++------ 3 files changed, 9 insertions(+), 21 deletions(-) diff --git a/kernel/bpf/btf.c b/kernel/bpf/btf.c index e34f58abc7e2..c4506d788c85 100644 --- a/kernel/bpf/btf.c +++ b/kernel/bpf/btf.c @@ -7676,21 +7676,16 @@ int btf_new_fd(const union bpf_attr *attr, bpfptr_t uattr, u32 uattr_size) struct btf *btf_get_by_fd(int fd) { struct btf *btf; - struct fd f; + CLASS(fd, f)(fd); - f = fdget(fd); - - if (!fd_file(f)) + if (fd_empty(f)) return ERR_PTR(-EBADF); - if (fd_file(f)->f_op != &btf_fops) { - fdput(f); + if (fd_file(f)->f_op != &btf_fops) return ERR_PTR(-EINVAL); - } btf = fd_file(f)->private_data; refcount_inc(&btf->refcnt); - fdput(f); return btf; } diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index ab0d94f41c48..d75e4e68801e 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -3187,20 +3187,16 @@ int bpf_link_new_fd(struct bpf_link *link) struct bpf_link *bpf_link_get_from_fd(u32 ufd) { - struct fd f = fdget(ufd); + CLASS(fd, f)(ufd); struct bpf_link *link; - if (!fd_file(f)) + if (fd_empty(f)) return ERR_PTR(-EBADF); - if (fd_file(f)->f_op != &bpf_link_fops && fd_file(f)->f_op != &bpf_link_fops_poll) { - fdput(f); + if (fd_file(f)->f_op != &bpf_link_fops && fd_file(f)->f_op != &bpf_link_fops_poll) return ERR_PTR(-EINVAL); - } link = fd_file(f)->private_data; bpf_link_inc(link); - fdput(f); - return link; } EXPORT_SYMBOL(bpf_link_get_from_fd); diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c index 9a1d356e79ed..9b92cb886d49 100644 --- a/kernel/bpf/token.c +++ b/kernel/bpf/token.c @@ -232,19 +232,16 @@ out_path: struct bpf_token *bpf_token_get_from_fd(u32 ufd) { - struct fd f = fdget(ufd); + CLASS(fd, f)(ufd); struct bpf_token *token; - if (!fd_file(f)) + if (fd_empty(f)) return ERR_PTR(-EBADF); - if (fd_file(f)->f_op != &bpf_token_fops) { - fdput(f); + if (fd_file(f)->f_op != &bpf_token_fops) return ERR_PTR(-EINVAL); - } token = fd_file(f)->private_data; bpf_token_inc(token); - fdput(f); return token; } From eceb7b33e5f3bc1e60047d8695cb937b93a08ced Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 30 Jul 2024 01:16:10 -0400 Subject: [PATCH 014/107] bpf: more trivial fdget() conversions All failure exits prior to fdget() leave the scope, all matching fdput() are immediately followed by leaving the scope. Reviewed-by: Christian Brauner Signed-off-by: Al Viro Signed-off-by: Andrii Nakryiko --- kernel/bpf/syscall.c | 22 +++++++--------------- 1 file changed, 7 insertions(+), 15 deletions(-) diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c index d75e4e68801e..65dcd92d0b2c 100644 --- a/kernel/bpf/syscall.c +++ b/kernel/bpf/syscall.c @@ -4901,33 +4901,25 @@ static int bpf_link_get_info_by_fd(struct file *file, static int bpf_obj_get_info_by_fd(const union bpf_attr *attr, union bpf_attr __user *uattr) { - int ufd = attr->info.bpf_fd; - struct fd f; - int err; - if (CHECK_ATTR(BPF_OBJ_GET_INFO_BY_FD)) return -EINVAL; - f = fdget(ufd); - if (!fd_file(f)) + CLASS(fd, f)(attr->info.bpf_fd); + if (fd_empty(f)) return -EBADFD; if (fd_file(f)->f_op == &bpf_prog_fops) - err = bpf_prog_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, + return bpf_prog_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, uattr); else if (fd_file(f)->f_op == &bpf_map_fops) - err = bpf_map_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, + return bpf_map_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, uattr); else if (fd_file(f)->f_op == &btf_fops) - err = bpf_btf_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, uattr); + return bpf_btf_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, uattr); else if (fd_file(f)->f_op == &bpf_link_fops || fd_file(f)->f_op == &bpf_link_fops_poll) - err = bpf_link_get_info_by_fd(fd_file(f), fd_file(f)->private_data, + return bpf_link_get_info_by_fd(fd_file(f), fd_file(f)->private_data, attr, uattr); - else - err = -EINVAL; - - fdput(f); - return err; + return -EINVAL; } #define BPF_BTF_LOAD_LAST_FIELD btf_token_fd From aaf8c0b9ae042494cb4585883b15c1332de77840 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 26 Jun 2024 09:47:27 +0800 Subject: [PATCH 015/107] f2fs: reduce expensive checkpoint trigger frequency We may trigger high frequent checkpoint for below case: 1. mkdir /mnt/dir1; set dir1 encrypted 2. touch /mnt/file1; fsync /mnt/file1 3. mkdir /mnt/dir2; set dir2 encrypted 4. touch /mnt/file2; fsync /mnt/file2 ... Although, newly created dir and file are not related, due to commit bbf156f7afa7 ("f2fs: fix lost xattrs of directories"), we will trigger checkpoint whenever fsync() comes after a new encrypted dir created. In order to avoid such performance regression issue, let's record an entry including directory's ino in global cache whenever we update directory's xattr data, and then triggerring checkpoint() only if xattr metadata of target file's parent was updated. This patch updates to cover below no encryption case as well: 1) parent is checkpointed 2) set_xattr(dir) w/ new xnid 3) create(file) 4) fsync(file) Fixes: bbf156f7afa7 ("f2fs: fix lost xattrs of directories") Reported-by: wangzijie Reported-by: Zhiguo Niu Tested-by: Zhiguo Niu Reported-by: Yunlei He Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 2 ++ fs/f2fs/file.c | 3 +++ fs/f2fs/xattr.c | 14 ++++++++++++-- include/trace/events/f2fs.h | 3 ++- 4 files changed, 19 insertions(+), 3 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 7b00846af915..f070a425a982 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -291,6 +291,7 @@ enum { APPEND_INO, /* for append ino list */ UPDATE_INO, /* for update ino list */ TRANS_DIR_INO, /* for transactions dir ino list */ + XATTR_DIR_INO, /* for xattr updated dir ino list */ FLUSH_INO, /* for multiple device flushing */ MAX_INO_ENTRY, /* max. list */ }; @@ -1161,6 +1162,7 @@ enum cp_reason_type { CP_FASTBOOT_MODE, CP_SPEC_LOG_NUM, CP_RECOVER_DIR, + CP_XATTR_DIR, }; enum iostat_type { diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index c598cfe5e0ed..e3bea575ddc3 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -218,6 +218,9 @@ static inline enum cp_reason_type need_do_checkpoint(struct inode *inode) f2fs_exist_written_data(sbi, F2FS_I(inode)->i_pino, TRANS_DIR_INO)) cp_reason = CP_RECOVER_DIR; + else if (f2fs_exist_written_data(sbi, F2FS_I(inode)->i_pino, + XATTR_DIR_INO)) + cp_reason = CP_XATTR_DIR; return cp_reason; } diff --git a/fs/f2fs/xattr.c b/fs/f2fs/xattr.c index f290fe9327c4..3f3874943679 100644 --- a/fs/f2fs/xattr.c +++ b/fs/f2fs/xattr.c @@ -629,6 +629,7 @@ static int __f2fs_setxattr(struct inode *inode, int index, const char *name, const void *value, size_t size, struct page *ipage, int flags) { + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct f2fs_xattr_entry *here, *last; void *base_addr, *last_base_addr; int found, newsize; @@ -772,9 +773,18 @@ retry: if (index == F2FS_XATTR_INDEX_ENCRYPTION && !strcmp(name, F2FS_XATTR_NAME_ENCRYPTION_CONTEXT)) f2fs_set_encrypted_inode(inode); - if (S_ISDIR(inode->i_mode)) - set_sbi_flag(F2FS_I_SB(inode), SBI_NEED_CP); + if (!S_ISDIR(inode->i_mode)) + goto same; + /* + * In restrict mode, fsync() always try to trigger checkpoint for all + * metadata consistency, in other mode, it triggers checkpoint when + * parent's xattr metadata was updated. + */ + if (F2FS_OPTION(sbi).fsync_mode == FSYNC_MODE_STRICT) + set_sbi_flag(sbi, SBI_NEED_CP); + else + f2fs_add_ino_entry(sbi, inode->i_ino, XATTR_DIR_INO); same: if (is_inode_flag_set(inode, FI_ACL_MODE)) { inode->i_mode = F2FS_I(inode)->i_acl_mode; diff --git a/include/trace/events/f2fs.h b/include/trace/events/f2fs.h index ed794b5fefbe..2851c823095b 100644 --- a/include/trace/events/f2fs.h +++ b/include/trace/events/f2fs.h @@ -139,7 +139,8 @@ TRACE_DEFINE_ENUM(EX_BLOCK_AGE); { CP_NODE_NEED_CP, "node needs cp" }, \ { CP_FASTBOOT_MODE, "fastboot mode" }, \ { CP_SPEC_LOG_NUM, "log type is 2" }, \ - { CP_RECOVER_DIR, "dir needs recovery" }) + { CP_RECOVER_DIR, "dir needs recovery" }, \ + { CP_XATTR_DIR, "dir's xattr updated" }) #define show_shutdown_mode(type) \ __print_symbolic(type, \ From 96cfeb0389530ae32ade8a48ae3ae1ac3b6c009d Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 27 Jun 2024 15:17:11 +0800 Subject: [PATCH 016/107] f2fs: fix to wait dio completion It should wait all existing dio write IOs before block removal, otherwise, previous direct write IO may overwrite data in the block which may be reused by other inode. Cc: stable@vger.kernel.org Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index e3bea575ddc3..67e5c6060403 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -1055,6 +1055,13 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry, return err; } + /* + * wait for inflight dio, blocks should be removed after + * IO completion. + */ + if (attr->ia_size < old_size) + inode_dio_wait(inode); + f2fs_down_write(&fi->i_gc_rwsem[WRITE]); filemap_invalidate_lock(inode->i_mapping); @@ -1891,6 +1898,12 @@ static long f2fs_fallocate(struct file *file, int mode, if (ret) goto out; + /* + * wait for inflight dio, blocks should be removed after IO + * completion. + */ + inode_dio_wait(inode); + if (mode & FALLOC_FL_PUNCH_HOLE) { if (offset >= inode->i_size) goto out; From 0cac51185e65dc2a20686184e02f3cafc99eb202 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Thu, 27 Jun 2024 15:15:21 +0800 Subject: [PATCH 017/107] f2fs: fix to avoid racing in between read and OPU dio write If lfs mode is on, buffered read may race w/ OPU dio write as below, it may cause buffered read hits unwritten data unexpectly, and for dio read, the race condition exists as well. Thread A Thread B - f2fs_file_write_iter - f2fs_dio_write_iter - __iomap_dio_rw - f2fs_iomap_begin - f2fs_map_blocks - __allocate_data_block - allocated blkaddr #x - iomap_dio_submit_bio - f2fs_file_read_iter - filemap_read - f2fs_read_data_folio - f2fs_mpage_readpages - f2fs_map_blocks : get blkaddr #x - f2fs_submit_read_bio IRQ - f2fs_read_end_io : read IO on blkaddr #x complete IRQ - iomap_dio_bio_end_io : direct write IO on blkaddr #x complete In LFS mode, if there is inflight dio, let's wait for its completion, this policy won't cover all race cases, however it is a tradeoff which avoids abusing lock around IO paths. Fixes: f847c699cff3 ("f2fs: allow out-place-update for direct IO in LFS mode") Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 67e5c6060403..0893a67811dd 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -4613,6 +4613,10 @@ static ssize_t f2fs_file_read_iter(struct kiocb *iocb, struct iov_iter *to) f2fs_trace_rw_file_path(iocb->ki_filp, iocb->ki_pos, iov_iter_count(to), READ); + /* In LFS mode, if there is inflight dio, wait for its completion */ + if (f2fs_lfs_mode(F2FS_I_SB(inode))) + inode_dio_wait(inode); + if (f2fs_should_use_dio(inode, iocb, to)) { ret = f2fs_dio_read_iter(iocb, to); } else { From 5bcde45578624c43eb35bb739b03c50459971865 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 18 Jun 2024 15:26:20 +0800 Subject: [PATCH 018/107] f2fs: get rid of buffer_head use Convert to use folio and related functionality. Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 1 - fs/f2fs/f2fs.h | 11 +++++- fs/f2fs/file.c | 1 - fs/f2fs/inode.c | 1 - fs/f2fs/super.c | 93 +++++++++++++++++++++++++++++-------------------- 5 files changed, 66 insertions(+), 41 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index be66b3a0e793..a2c3d39ddfed 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -7,7 +7,6 @@ */ #include #include -#include #include #include #include diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index f070a425a982..51fd5063a69c 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -11,7 +11,6 @@ #include #include #include -#include #include #include #include @@ -2006,6 +2005,16 @@ static inline struct f2fs_super_block *F2FS_RAW_SUPER(struct f2fs_sb_info *sbi) return (struct f2fs_super_block *)(sbi->raw_super); } +static inline struct f2fs_super_block *F2FS_SUPER_BLOCK(struct folio *folio, + pgoff_t index) +{ + pgoff_t idx_in_folio = index % (1 << folio_order(folio)); + + return (struct f2fs_super_block *) + (page_address(folio_page(folio, idx_in_folio)) + + F2FS_SUPER_OFFSET); +} + static inline struct f2fs_checkpoint *F2FS_CKPT(struct f2fs_sb_info *sbi) { return (struct f2fs_checkpoint *)(sbi->ckpt); diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 0893a67811dd..a458eb080198 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -8,7 +8,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index aef57172014f..5d7e4c7f5969 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -7,7 +7,6 @@ */ #include #include -#include #include #include #include diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index aa23e4cfd89e..4d2f3bb8e418 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -11,7 +11,6 @@ #include #include #include -#include #include #include #include @@ -3323,24 +3322,42 @@ loff_t max_file_blocks(struct inode *inode) return result; } -static int __f2fs_commit_super(struct buffer_head *bh, - struct f2fs_super_block *super) +static int __f2fs_commit_super(struct f2fs_sb_info *sbi, struct folio *folio, + pgoff_t index, bool update) { - lock_buffer(bh); - if (super) - memcpy(bh->b_data + F2FS_SUPER_OFFSET, super, sizeof(*super)); - set_buffer_dirty(bh); - unlock_buffer(bh); - + struct bio *bio; /* it's rare case, we can do fua all the time */ - return __sync_dirty_buffer(bh, REQ_SYNC | REQ_PREFLUSH | REQ_FUA); + blk_opf_t opf = REQ_OP_WRITE | REQ_SYNC | REQ_PREFLUSH | REQ_FUA; + int ret; + + folio_lock(folio); + folio_wait_writeback(folio); + if (update) + memcpy(F2FS_SUPER_BLOCK(folio, index), F2FS_RAW_SUPER(sbi), + sizeof(struct f2fs_super_block)); + folio_mark_dirty(folio); + folio_clear_dirty_for_io(folio); + folio_start_writeback(folio); + folio_unlock(folio); + + bio = bio_alloc(sbi->sb->s_bdev, 1, opf, GFP_NOFS); + + /* it doesn't need to set crypto context for superblock update */ + bio->bi_iter.bi_sector = SECTOR_FROM_BLOCK(folio_index(folio)); + + if (!bio_add_folio(bio, folio, folio_size(folio), 0)) + f2fs_bug_on(sbi, 1); + + ret = submit_bio_wait(bio); + folio_end_writeback(folio); + + return ret; } static inline bool sanity_check_area_boundary(struct f2fs_sb_info *sbi, - struct buffer_head *bh) + struct folio *folio, pgoff_t index) { - struct f2fs_super_block *raw_super = (struct f2fs_super_block *) - (bh->b_data + F2FS_SUPER_OFFSET); + struct f2fs_super_block *raw_super = F2FS_SUPER_BLOCK(folio, index); struct super_block *sb = sbi->sb; u32 segment0_blkaddr = le32_to_cpu(raw_super->segment0_blkaddr); u32 cp_blkaddr = le32_to_cpu(raw_super->cp_blkaddr); @@ -3415,7 +3432,7 @@ static inline bool sanity_check_area_boundary(struct f2fs_sb_info *sbi, set_sbi_flag(sbi, SBI_NEED_SB_WRITE); res = "internally"; } else { - err = __f2fs_commit_super(bh, NULL); + err = __f2fs_commit_super(sbi, folio, index, false); res = err ? "failed" : "done"; } f2fs_info(sbi, "Fix alignment : %s, start(%u) end(%llu) block(%u)", @@ -3428,12 +3445,11 @@ static inline bool sanity_check_area_boundary(struct f2fs_sb_info *sbi, } static int sanity_check_raw_super(struct f2fs_sb_info *sbi, - struct buffer_head *bh) + struct folio *folio, pgoff_t index) { block_t segment_count, segs_per_sec, secs_per_zone, segment_count_main; block_t total_sections, blocks_per_seg; - struct f2fs_super_block *raw_super = (struct f2fs_super_block *) - (bh->b_data + F2FS_SUPER_OFFSET); + struct f2fs_super_block *raw_super = F2FS_SUPER_BLOCK(folio, index); size_t crc_offset = 0; __u32 crc = 0; @@ -3591,7 +3607,7 @@ static int sanity_check_raw_super(struct f2fs_sb_info *sbi, } /* check CP/SIT/NAT/SSA/MAIN_AREA area boundary */ - if (sanity_check_area_boundary(sbi, bh)) + if (sanity_check_area_boundary(sbi, folio, index)) return -EFSCORRUPTED; return 0; @@ -3938,7 +3954,7 @@ static int read_raw_super_block(struct f2fs_sb_info *sbi, { struct super_block *sb = sbi->sb; int block; - struct buffer_head *bh; + struct folio *folio; struct f2fs_super_block *super; int err = 0; @@ -3947,32 +3963,32 @@ static int read_raw_super_block(struct f2fs_sb_info *sbi, return -ENOMEM; for (block = 0; block < 2; block++) { - bh = sb_bread(sb, block); - if (!bh) { + folio = read_mapping_folio(sb->s_bdev->bd_mapping, block, NULL); + if (IS_ERR(folio)) { f2fs_err(sbi, "Unable to read %dth superblock", block + 1); - err = -EIO; + err = PTR_ERR(folio); *recovery = 1; continue; } /* sanity checking of raw super */ - err = sanity_check_raw_super(sbi, bh); + err = sanity_check_raw_super(sbi, folio, block); if (err) { f2fs_err(sbi, "Can't find valid F2FS filesystem in %dth superblock", block + 1); - brelse(bh); + folio_put(folio); *recovery = 1; continue; } if (!*raw_super) { - memcpy(super, bh->b_data + F2FS_SUPER_OFFSET, + memcpy(super, F2FS_SUPER_BLOCK(folio, block), sizeof(*super)); *valid_super_block = block; *raw_super = super; } - brelse(bh); + folio_put(folio); } /* No valid superblock */ @@ -3986,7 +4002,8 @@ static int read_raw_super_block(struct f2fs_sb_info *sbi, int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover) { - struct buffer_head *bh; + struct folio *folio; + pgoff_t index; __u32 crc = 0; int err; @@ -4004,22 +4021,24 @@ int f2fs_commit_super(struct f2fs_sb_info *sbi, bool recover) } /* write back-up superblock first */ - bh = sb_bread(sbi->sb, sbi->valid_super_block ? 0 : 1); - if (!bh) - return -EIO; - err = __f2fs_commit_super(bh, F2FS_RAW_SUPER(sbi)); - brelse(bh); + index = sbi->valid_super_block ? 0 : 1; + folio = read_mapping_folio(sbi->sb->s_bdev->bd_mapping, index, NULL); + if (IS_ERR(folio)) + return PTR_ERR(folio); + err = __f2fs_commit_super(sbi, folio, index, true); + folio_put(folio); /* if we are in recovery path, skip writing valid superblock */ if (recover || err) return err; /* write current valid superblock */ - bh = sb_bread(sbi->sb, sbi->valid_super_block); - if (!bh) - return -EIO; - err = __f2fs_commit_super(bh, F2FS_RAW_SUPER(sbi)); - brelse(bh); + index = sbi->valid_super_block; + folio = read_mapping_folio(sbi->sb->s_bdev->bd_mapping, index, NULL); + if (IS_ERR(folio)) + return PTR_ERR(folio); + err = __f2fs_commit_super(sbi, folio, index, true); + folio_put(folio); return err; } From b722ff8ad66cf9beba971d9eb4bb7b5e6265ae5c Mon Sep 17 00:00:00 2001 From: Wenjie Cheng Date: Fri, 14 Jun 2024 00:48:41 +0000 Subject: [PATCH 019/107] Revert "f2fs: use flush command instead of FUA for zoned device" This reverts commit c550e25bca660ed2554cbb48d32b82d0bb98e4b1. Commit c550e25bca660ed2554cbb48d32b82d0bb98e4b1 ("f2fs: use flush command instead of FUA for zoned device") used additional flush command to keep write order. Since Commit dd291d77cc90eb6a86e9860ba8e6e38eebd57d12 ("block: Introduce zone write plugging") has enabled the block layer to handle this order issue, there is no need to use flush command. Signed-off-by: Wenjie Cheng Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 3 +-- fs/f2fs/node.c | 2 +- 2 files changed, 2 insertions(+), 3 deletions(-) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index a458eb080198..a7a3e6ac7f68 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -375,8 +375,7 @@ sync_nodes: f2fs_remove_ino_entry(sbi, ino, APPEND_INO); clear_inode_flag(inode, FI_APPEND_WRITE); flush_out: - if ((!atomic && F2FS_OPTION(sbi).fsync_mode != FSYNC_MODE_NOBARRIER) || - (atomic && !test_opt(sbi, NOBARRIER) && f2fs_sb_has_blkzoned(sbi))) + if (!atomic && F2FS_OPTION(sbi).fsync_mode != FSYNC_MODE_NOBARRIER) ret = f2fs_issue_flush(sbi, inode->i_ino); if (!ret) { f2fs_remove_ino_entry(sbi, ino, UPDATE_INO); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 58721a55f173..2fe182f73599 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1674,7 +1674,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, goto redirty_out; } - if (atomic && !test_opt(sbi, NOBARRIER) && !f2fs_sb_has_blkzoned(sbi)) + if (atomic && !test_opt(sbi, NOBARRIER)) fio.op_flags |= REQ_PREFLUSH | REQ_FUA; /* should add to global list before clearing PAGECACHE status */ From 6f092b55e1ad2d5c6a82d91dcf93534aeefcf1b9 Mon Sep 17 00:00:00 2001 From: liujinbao1 Date: Wed, 21 Feb 2024 15:32:49 +0800 Subject: [PATCH 020/107] f2fs: sysfs: support atgc_enabled When we add "atgc" to the fstab table, ATGC is not immediately enabled. There is a 7-day time threshold, and we can use "atgc_enabled" to show whether ATGC is enabled. Signed-off-by: liujinbao1 Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++++ fs/f2fs/sysfs.c | 8 ++++++++ 2 files changed, 14 insertions(+) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index 3500920ab7ce..ca61ffcae126 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -579,6 +579,12 @@ Description: When ATGC is on, it controls age threshold to bypass GCing young candidates whose age is not beyond the threshold, by default it was initialized as 604800 seconds (equals to 7 days). +What: /sys/fs/f2fs//atgc_enabled +Date: Feb 2024 +Contact: "Jinbao Liu" +Description: It represents whether ATGC is on or off. The value is 1 which + indicates that ATGC is on, and 0 indicates that it is off. + What: /sys/fs/f2fs//gc_reclaimed_segments Date: July 2021 Contact: "Daeho Jeong" diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index 63ff2d1647eb..ed4bf434207a 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -170,6 +170,12 @@ static ssize_t undiscard_blks_show(struct f2fs_attr *a, SM_I(sbi)->dcc_info->undiscard_blks); } +static ssize_t atgc_enabled_show(struct f2fs_attr *a, + struct f2fs_sb_info *sbi, char *buf) +{ + return sysfs_emit(buf, "%d\n", sbi->am.atgc_enabled ? 1 : 0); +} + static ssize_t gc_mode_show(struct f2fs_attr *a, struct f2fs_sb_info *sbi, char *buf) { @@ -1082,6 +1088,7 @@ F2FS_GENERAL_RO_ATTR(encoding); F2FS_GENERAL_RO_ATTR(mounted_time_sec); F2FS_GENERAL_RO_ATTR(main_blkaddr); F2FS_GENERAL_RO_ATTR(pending_discard); +F2FS_GENERAL_RO_ATTR(atgc_enabled); F2FS_GENERAL_RO_ATTR(gc_mode); #ifdef CONFIG_F2FS_STAT_FS F2FS_GENERAL_RO_ATTR(moved_blocks_background); @@ -1211,6 +1218,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(atgc_candidate_count), ATTR_LIST(atgc_age_weight), ATTR_LIST(atgc_age_threshold), + ATTR_LIST(atgc_enabled), ATTR_LIST(seq_file_ra_mul), ATTR_LIST(gc_segment_mode), ATTR_LIST(gc_reclaimed_segments), From f97a11c86cc38da14fd1d44fde767ae2ee1b6ebe Mon Sep 17 00:00:00 2001 From: Zijie Wang Date: Mon, 1 Apr 2024 21:24:08 +0800 Subject: [PATCH 021/107] f2fs: use f2fs_get_node_page when write inline data We just need inode page when write inline data, use f2fs_get_node_page() to get it instead of using dnode_of_data, which can eliminate unnecessary struct use. Signed-off-by: Zijie Wang Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/inline.c | 23 +++++++++++------------ 1 file changed, 11 insertions(+), 12 deletions(-) diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index cca7d448e55c..d76a3bed7093 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -262,33 +262,32 @@ out: int f2fs_write_inline_data(struct inode *inode, struct page *page) { - struct dnode_of_data dn; - int err; + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); + struct page *ipage; - set_new_dnode(&dn, inode, NULL, NULL, 0); - err = f2fs_get_dnode_of_data(&dn, 0, LOOKUP_NODE); - if (err) - return err; + ipage = f2fs_get_node_page(sbi, inode->i_ino); + if (IS_ERR(ipage)) + return PTR_ERR(ipage); if (!f2fs_has_inline_data(inode)) { - f2fs_put_dnode(&dn); + f2fs_put_page(ipage, 1); return -EAGAIN; } f2fs_bug_on(F2FS_I_SB(inode), page->index); - f2fs_wait_on_page_writeback(dn.inode_page, NODE, true, true); - memcpy_from_page(inline_data_addr(inode, dn.inode_page), + f2fs_wait_on_page_writeback(ipage, NODE, true, true); + memcpy_from_page(inline_data_addr(inode, ipage), page, 0, MAX_INLINE_DATA(inode)); - set_page_dirty(dn.inode_page); + set_page_dirty(ipage); f2fs_clear_page_cache_dirty_tag(page); set_inode_flag(inode, FI_APPEND_WRITE); set_inode_flag(inode, FI_DATA_EXIST); - clear_page_private_inline(dn.inode_page); - f2fs_put_dnode(&dn); + clear_page_private_inline(ipage); + f2fs_put_page(ipage, 1); return 0; } From d33ebd57b987bb8dc499f4468170c8f85f64fe72 Mon Sep 17 00:00:00 2001 From: Zhiguo Niu Date: Fri, 26 Jul 2024 14:27:26 +0800 Subject: [PATCH 022/107] f2fs: fix to use per-inode maxbytes and cleanup This is a supplement to commit 6d1451bf7f84 ("f2fs: fix to use per-inode maxbytes") for some missed cases, also cleanup redundant code in f2fs_llseek. Cc: Chengguang Xu Signed-off-by: Zhiguo Niu Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 4 ++-- fs/f2fs/file.c | 7 ++----- fs/f2fs/verity.c | 5 +++-- 3 files changed, 7 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index a2c3d39ddfed..e9570f4e0f21 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1938,7 +1938,7 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, inode_lock_shared(inode); - maxbytes = max_file_blocks(inode) << F2FS_BLKSIZE_BITS; + maxbytes = F2FS_BLK_TO_BYTES(max_file_blocks(inode)); if (start > maxbytes) { ret = -EFBIG; goto out; @@ -2063,7 +2063,7 @@ out: static inline loff_t f2fs_readpage_limit(struct inode *inode) { if (IS_ENABLED(CONFIG_FS_VERITY) && IS_VERITY(inode)) - return inode->i_sb->s_maxbytes; + return F2FS_BLK_TO_BYTES(max_file_blocks(inode)); return i_size_read(inode); } diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index a7a3e6ac7f68..a2d466ca2579 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -432,7 +432,7 @@ static bool __found_offset(struct address_space *mapping, static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence) { struct inode *inode = file->f_mapping->host; - loff_t maxbytes = inode->i_sb->s_maxbytes; + loff_t maxbytes = F2FS_BLK_TO_BYTES(max_file_blocks(inode)); struct dnode_of_data dn; pgoff_t pgofs, end_offset; loff_t data_ofs = offset; @@ -514,10 +514,7 @@ fail: static loff_t f2fs_llseek(struct file *file, loff_t offset, int whence) { struct inode *inode = file->f_mapping->host; - loff_t maxbytes = inode->i_sb->s_maxbytes; - - if (f2fs_compressed_file(inode)) - maxbytes = max_file_blocks(inode) << F2FS_BLKSIZE_BITS; + loff_t maxbytes = F2FS_BLK_TO_BYTES(max_file_blocks(inode)); switch (whence) { case SEEK_SET: diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c index f7bb0c54502c..83310b5e0422 100644 --- a/fs/f2fs/verity.c +++ b/fs/f2fs/verity.c @@ -74,7 +74,7 @@ static int pagecache_write(struct inode *inode, const void *buf, size_t count, struct address_space *mapping = inode->i_mapping; const struct address_space_operations *aops = mapping->a_ops; - if (pos + count > inode->i_sb->s_maxbytes) + if (pos + count > F2FS_BLK_TO_BYTES(max_file_blocks(inode))) return -EFBIG; while (count) { @@ -237,7 +237,8 @@ static int f2fs_get_verity_descriptor(struct inode *inode, void *buf, pos = le64_to_cpu(dloc.pos); /* Get the descriptor */ - if (pos + size < pos || pos + size > inode->i_sb->s_maxbytes || + if (pos + size < pos || + pos + size > F2FS_BLK_TO_BYTES(max_file_blocks(inode)) || pos < f2fs_verity_metadata_pos(inode) || size > INT_MAX) { f2fs_warn(F2FS_I_SB(inode), "invalid verity xattr"); f2fs_handle_error(F2FS_I_SB(inode), From 8fb9f31984bdc0aaa1b0f7c7e7a27bd3137f8744 Mon Sep 17 00:00:00 2001 From: Zhiguo Niu Date: Thu, 1 Aug 2024 09:33:51 +0800 Subject: [PATCH 023/107] f2fs: clean up val{>>,<<}F2FS_BLKSIZE_BITS Use F2FS_BYTES_TO_BLK(bytes) and F2FS_BLK_TO_BYTES(blk) for cleanup Signed-off-by: Zhiguo Niu Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/debug.c | 2 +- fs/f2fs/file.c | 6 +++--- fs/f2fs/node.c | 4 ++-- fs/f2fs/super.c | 2 +- include/linux/f2fs_fs.h | 2 +- 6 files changed, 9 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index bdd96329dddd..f3d22b8ef2ff 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -1551,7 +1551,7 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) blk = start_blk + BLKS_PER_SEG(sbi) - nm_i->nat_bits_blocks; for (i = 0; i < nm_i->nat_bits_blocks; i++) f2fs_update_meta_page(sbi, nm_i->nat_bits + - (i << F2FS_BLKSIZE_BITS), blk + i); + F2FS_BLK_TO_BYTES(i), blk + i); } /* write out checkpoint buffer at block 0 */ diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index 8b0e1e71b667..546b8ba91261 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -275,7 +275,7 @@ static void update_mem_info(struct f2fs_sb_info *sbi) /* build nm */ si->base_mem += sizeof(struct f2fs_nm_info); si->base_mem += __bitmap_size(sbi, NAT_BITMAP); - si->base_mem += (NM_I(sbi)->nat_bits_blocks << F2FS_BLKSIZE_BITS); + si->base_mem += F2FS_BLK_TO_BYTES(NM_I(sbi)->nat_bits_blocks); si->base_mem += NM_I(sbi)->nat_blocks * f2fs_bitmap_size(NAT_ENTRY_PER_BLOCK); si->base_mem += NM_I(sbi)->nat_blocks / 8; diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index a2d466ca2579..bddcb2cd945a 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2979,9 +2979,9 @@ static int f2fs_move_file_range(struct file *file_in, loff_t pos_in, } f2fs_lock_op(sbi); - ret = __exchange_data_block(src, dst, pos_in >> F2FS_BLKSIZE_BITS, - pos_out >> F2FS_BLKSIZE_BITS, - len >> F2FS_BLKSIZE_BITS, false); + ret = __exchange_data_block(src, dst, F2FS_BYTES_TO_BLK(pos_in), + F2FS_BYTES_TO_BLK(pos_out), + F2FS_BYTES_TO_BLK(len), false); if (!ret) { if (dst_max_i_size) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 2fe182f73599..9e7a6e21f30c 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -3166,7 +3166,7 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) nm_i->nat_bits_blocks = F2FS_BLK_ALIGN((nat_bits_bytes << 1) + 8); nm_i->nat_bits = f2fs_kvzalloc(sbi, - nm_i->nat_bits_blocks << F2FS_BLKSIZE_BITS, GFP_KERNEL); + F2FS_BLK_TO_BYTES(nm_i->nat_bits_blocks), GFP_KERNEL); if (!nm_i->nat_bits) return -ENOMEM; @@ -3185,7 +3185,7 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) if (IS_ERR(page)) return PTR_ERR(page); - memcpy(nm_i->nat_bits + (i << F2FS_BLKSIZE_BITS), + memcpy(nm_i->nat_bits + F2FS_BLK_TO_BYTES(i), page_address(page), F2FS_BLKSIZE); f2fs_put_page(page, 1); } diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 4d2f3bb8e418..977f038f3557 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -3317,7 +3317,7 @@ loff_t max_file_blocks(struct inode *inode) * fit within U32_MAX + 1 data units. */ - result = min(result, (((loff_t)U32_MAX + 1) * 4096) >> F2FS_BLKSIZE_BITS); + result = min(result, F2FS_BYTES_TO_BLK(((loff_t)U32_MAX + 1) * 4096)); return result; } diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index 01bee2b289c2..f17f89a259e2 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -19,7 +19,6 @@ #define F2FS_BLKSIZE_BITS PAGE_SHIFT /* bits for F2FS_BLKSIZE */ #define F2FS_MAX_EXTENSION 64 /* # of extension entries */ #define F2FS_EXTENSION_LEN 8 /* max size of extension */ -#define F2FS_BLK_ALIGN(x) (((x) + F2FS_BLKSIZE - 1) >> F2FS_BLKSIZE_BITS) #define NULL_ADDR ((block_t)0) /* used as block_t addresses */ #define NEW_ADDR ((block_t)-1) /* used as block_t addresses */ @@ -28,6 +27,7 @@ #define F2FS_BYTES_TO_BLK(bytes) ((bytes) >> F2FS_BLKSIZE_BITS) #define F2FS_BLK_TO_BYTES(blk) ((blk) << F2FS_BLKSIZE_BITS) #define F2FS_BLK_END_BYTES(blk) (F2FS_BLK_TO_BYTES(blk + 1) - 1) +#define F2FS_BLK_ALIGN(x) (F2FS_BYTES_TO_BLK((x) + F2FS_BLKSIZE - 1)) /* 0, 1(node nid), 2(meta nid) are reserved node id */ #define F2FS_RESERVED_NODE_NUM 3 From 4f5a100f87f32cb65d4bb1ad282a08c92f6f591e Mon Sep 17 00:00:00 2001 From: Jann Horn Date: Tue, 6 Aug 2024 16:07:16 +0200 Subject: [PATCH 024/107] f2fs: Require FMODE_WRITE for atomic write ioctls The F2FS ioctls for starting and committing atomic writes check for inode_owner_or_capable(), but this does not give LSMs like SELinux or Landlock an opportunity to deny the write access - if the caller's FSUID matches the inode's UID, inode_owner_or_capable() immediately returns true. There are scenarios where LSMs want to deny a process the ability to write particular files, even files that the FSUID of the process owns; but this can currently partially be bypassed using atomic write ioctls in two ways: - F2FS_IOC_START_ATOMIC_REPLACE + F2FS_IOC_COMMIT_ATOMIC_WRITE can truncate an inode to size 0 - F2FS_IOC_START_ATOMIC_WRITE + F2FS_IOC_ABORT_ATOMIC_WRITE can revert changes another process concurrently made to a file Fix it by requiring FMODE_WRITE for these operations, just like for F2FS_IOC_MOVE_RANGE. Since any legitimate caller should only be using these ioctls when intending to write into the file, that seems unlikely to break anything. Fixes: 88b88a667971 ("f2fs: support atomic writes") Cc: stable@vger.kernel.org Signed-off-by: Jann Horn Reviewed-by: Chao Yu Reviewed-by: Eric Biggers Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index bddcb2cd945a..2c591fbc75a9 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2131,6 +2131,9 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) loff_t isize; int ret; + if (!(filp->f_mode & FMODE_WRITE)) + return -EBADF; + if (!inode_owner_or_capable(idmap, inode)) return -EACCES; @@ -2239,6 +2242,9 @@ static int f2fs_ioc_commit_atomic_write(struct file *filp) struct mnt_idmap *idmap = file_mnt_idmap(filp); int ret; + if (!(filp->f_mode & FMODE_WRITE)) + return -EBADF; + if (!inode_owner_or_capable(idmap, inode)) return -EACCES; @@ -2271,6 +2277,9 @@ static int f2fs_ioc_abort_atomic_write(struct file *filp) struct mnt_idmap *idmap = file_mnt_idmap(filp); int ret; + if (!(filp->f_mode & FMODE_WRITE)) + return -EBADF; + if (!inode_owner_or_capable(idmap, inode)) return -EACCES; From 8c1b787938fd86bab27a1492fa887408c75fec2b Mon Sep 17 00:00:00 2001 From: Yeongjin Gil Date: Tue, 13 Aug 2024 16:32:44 +0900 Subject: [PATCH 025/107] f2fs: Create COW inode from parent dentry for atomic write The i_pino in f2fs_inode_info has the previous parent's i_ino when inode was renamed, which may cause f2fs_ioc_start_atomic_write to fail. If file_wrong_pino is true and i_nlink is 1, then to find a valid pino, we should refer to the dentry from inode. To resolve this issue, let's get parent inode using parent dentry directly. Fixes: 3db1de0e582c ("f2fs: change the current atomic write way") Reviewed-by: Sungjong Seo Reviewed-by: Sunmin Jeong Signed-off-by: Yeongjin Gil Reviewed-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 12 +++--------- 1 file changed, 3 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 2c591fbc75a9..082e81ba09b9 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2127,7 +2127,6 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) struct mnt_idmap *idmap = file_mnt_idmap(filp); struct f2fs_inode_info *fi = F2FS_I(inode); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - struct inode *pinode; loff_t isize; int ret; @@ -2180,15 +2179,10 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) /* Check if the inode already has a COW inode */ if (fi->cow_inode == NULL) { /* Create a COW inode for atomic write */ - pinode = f2fs_iget(inode->i_sb, fi->i_pino); - if (IS_ERR(pinode)) { - f2fs_up_write(&fi->i_gc_rwsem[WRITE]); - ret = PTR_ERR(pinode); - goto out; - } + struct dentry *dentry = file_dentry(filp); + struct inode *dir = d_inode(dentry->d_parent); - ret = f2fs_get_tmpfile(idmap, pinode, &fi->cow_inode); - iput(pinode); + ret = f2fs_get_tmpfile(idmap, dir, &fi->cow_inode); if (ret) { f2fs_up_write(&fi->i_gc_rwsem[WRITE]); goto out; From a4d7f2b3238fd5f76b9e6434a0bd5d2e29049cff Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Mon, 12 Aug 2024 22:12:42 +0800 Subject: [PATCH 026/107] f2fs: fix to wait page writeback before setting gcing flag Soft IRQ Thread - f2fs_write_end_io - f2fs_defragment_range - set_page_private_gcing - type = WB_DATA_TYPE(page, false); : assign type w/ F2FS_WB_CP_DATA due to page_private_gcing() is true - dec_page_count() w/ wrong type - end_page_writeback() Value of F2FS_WB_CP_DATA reference count may become negative under above race condition, the root cause is we missed to wait page writeback before setting gcing page private flag, let's fix it. Fixes: 2d1fe8a86bf5 ("f2fs: fix to tag gcing flag on page during file defragment") Fixes: 4961acdd65c9 ("f2fs: fix to tag gcing flag on page during block migration") Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 082e81ba09b9..73f70a7a2567 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2807,6 +2807,8 @@ do_map: goto clear_out; } + f2fs_wait_on_page_writeback(page, DATA, true, true); + set_page_dirty(page); set_page_private_gcing(page); f2fs_put_page(page, 1); @@ -4207,6 +4209,8 @@ static int redirty_blocks(struct inode *inode, pgoff_t page_idx, int len) /* It will never fail, when page has pinned above */ f2fs_bug_on(F2FS_I_SB(inode), !page); + f2fs_wait_on_page_writeback(page, DATA, true, true); + set_page_dirty(page); set_page_private_gcing(page); f2fs_put_page(page, 1); From ebd3309aec6271c4616573b0cb83ea25e623070a Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 7 Aug 2024 18:24:35 +0800 Subject: [PATCH 027/107] f2fs: atomic: fix to truncate pagecache before on-disk metadata truncation We should always truncate pagecache while truncating on-disk data. Fixes: a46bebd502fe ("f2fs: synchronize atomic write aborts") Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 73f70a7a2567..5fb8e4242c7a 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2195,6 +2195,10 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) F2FS_I(fi->cow_inode)->atomic_inode = inode; } else { /* Reuse the already created COW inode */ + f2fs_bug_on(sbi, get_dirty_pages(fi->cow_inode)); + + invalidate_mapping_pages(fi->cow_inode->i_mapping, 0, -1); + ret = f2fs_do_truncate_blocks(fi->cow_inode, 0, true); if (ret) { f2fs_up_write(&fi->i_gc_rwsem[WRITE]); From c7f114d864ac91515bb07ac271e9824a20f5ed95 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 30 Jul 2024 09:08:55 +0800 Subject: [PATCH 028/107] f2fs: fix to avoid use-after-free in f2fs_stop_gc_thread() syzbot reports a f2fs bug as below: __dump_stack lib/dump_stack.c:88 [inline] dump_stack_lvl+0x241/0x360 lib/dump_stack.c:114 print_report+0xe8/0x550 mm/kasan/report.c:491 kasan_report+0x143/0x180 mm/kasan/report.c:601 kasan_check_range+0x282/0x290 mm/kasan/generic.c:189 instrument_atomic_read_write include/linux/instrumented.h:96 [inline] atomic_fetch_add_relaxed include/linux/atomic/atomic-instrumented.h:252 [inline] __refcount_add include/linux/refcount.h:184 [inline] __refcount_inc include/linux/refcount.h:241 [inline] refcount_inc include/linux/refcount.h:258 [inline] get_task_struct include/linux/sched/task.h:118 [inline] kthread_stop+0xca/0x630 kernel/kthread.c:704 f2fs_stop_gc_thread+0x65/0xb0 fs/f2fs/gc.c:210 f2fs_do_shutdown+0x192/0x540 fs/f2fs/file.c:2283 f2fs_ioc_shutdown fs/f2fs/file.c:2325 [inline] __f2fs_ioctl+0x443a/0xbe60 fs/f2fs/file.c:4325 vfs_ioctl fs/ioctl.c:51 [inline] __do_sys_ioctl fs/ioctl.c:907 [inline] __se_sys_ioctl+0xfc/0x170 fs/ioctl.c:893 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f The root cause is below race condition, it may cause use-after-free issue in sbi->gc_th pointer. - remount - f2fs_remount - f2fs_stop_gc_thread - kfree(gc_th) - f2fs_ioc_shutdown - f2fs_do_shutdown - f2fs_stop_gc_thread - kthread_stop(gc_th->f2fs_gc_task) : sbi->gc_thread = NULL; We will call f2fs_do_shutdown() in two paths: - for f2fs_ioc_shutdown() path, we should grab sb->s_umount semaphore for fixing. - for f2fs_shutdown() path, it's safe since caller has already grabbed sb->s_umount semaphore. Reported-by: syzbot+1a8e2b31f2ac9bd3d148@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-f2fs-devel/0000000000005c7ccb061e032b9b@google.com Fixes: 7950e9ac638e ("f2fs: stop gc/discard thread after fs shutdown") Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 2 +- fs/f2fs/file.c | 11 +++++++++-- fs/f2fs/super.c | 2 +- 3 files changed, 11 insertions(+), 4 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 51fd5063a69c..0e0189084462 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3508,7 +3508,7 @@ int f2fs_setattr(struct mnt_idmap *idmap, struct dentry *dentry, int f2fs_truncate_hole(struct inode *inode, pgoff_t pg_start, pgoff_t pg_end); void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count); int f2fs_do_shutdown(struct f2fs_sb_info *sbi, unsigned int flag, - bool readonly); + bool readonly, bool need_lock); int f2fs_precache_extents(struct inode *inode); int f2fs_fileattr_get(struct dentry *dentry, struct fileattr *fa); int f2fs_fileattr_set(struct mnt_idmap *idmap, diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 5fb8e4242c7a..50075f8d5e9b 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2297,7 +2297,7 @@ static int f2fs_ioc_abort_atomic_write(struct file *filp) } int f2fs_do_shutdown(struct f2fs_sb_info *sbi, unsigned int flag, - bool readonly) + bool readonly, bool need_lock) { struct super_block *sb = sbi->sb; int ret = 0; @@ -2344,12 +2344,19 @@ int f2fs_do_shutdown(struct f2fs_sb_info *sbi, unsigned int flag, if (readonly) goto out; + /* grab sb->s_umount to avoid racing w/ remount() */ + if (need_lock) + down_read(&sbi->sb->s_umount); + f2fs_stop_gc_thread(sbi); f2fs_stop_discard_thread(sbi); f2fs_drop_discard_cmd(sbi); clear_opt(sbi, DISCARD); + if (need_lock) + up_read(&sbi->sb->s_umount); + f2fs_update_time(sbi, REQ_TIME); out: @@ -2386,7 +2393,7 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) } } - ret = f2fs_do_shutdown(sbi, in, readonly); + ret = f2fs_do_shutdown(sbi, in, readonly, true); if (need_drop) mnt_drop_write_file(filp); diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 977f038f3557..1c9316788a16 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -2560,7 +2560,7 @@ restore_opts: static void f2fs_shutdown(struct super_block *sb) { - f2fs_do_shutdown(F2FS_SB(sb), F2FS_GOING_DOWN_NOSYNC, false); + f2fs_do_shutdown(F2FS_SB(sb), F2FS_GOING_DOWN_NOSYNC, false, false); } #ifdef CONFIG_QUOTA From 43aec4d01bd2ce961817a777b3846f8318f398e4 Mon Sep 17 00:00:00 2001 From: Shin'ichiro Kawasaki Date: Fri, 16 Aug 2024 13:07:03 +0900 Subject: [PATCH 029/107] f2fs: check discard support for conventional zones As the helper function f2fs_bdev_support_discard() shows, f2fs checks if the target block devices support discard by calling bdev_max_discard_sectors() and bdev_is_zoned(). This check works well for most cases, but it does not work for conventional zones on zoned block devices. F2fs assumes that zoned block devices support discard, and calls __submit_discard_cmd(). When __submit_discard_cmd() is called for sequential write required zones, it works fine since __submit_discard_cmd() issues zone reset commands instead of discard commands. However, when __submit_discard_cmd() is called for conventional zones, __blkdev_issue_discard() is called even when the devices do not support discard. The inappropriate __blkdev_issue_discard() call was not a problem before the commit 30f1e7241422 ("block: move discard checks into the ioctl handler") because __blkdev_issue_discard() checked if the target devices support discard or not. If not, it returned EOPNOTSUPP. After the commit, __blkdev_issue_discard() no longer checks it. It always returns zero and sets NULL to the given bio pointer. This NULL pointer triggers f2fs_bug_on() in __submit_discard_cmd(). The BUG is recreated with the commands below at the umount step, where /dev/nullb0 is a zoned null_blk with 5GB total size, 128MB zone size and 10 conventional zones. $ mkfs.f2fs -f -m /dev/nullb0 $ mount /dev/nullb0 /mnt $ for ((i=0;i<5;i++)); do dd if=/dev/zero of=/mnt/test bs=65536 count=1600 conv=fsync; done $ umount /mnt To fix the BUG, avoid the inappropriate __blkdev_issue_discard() call. When discard is requested for conventional zones, check if the device supports discard or not. If not, return EOPNOTSUPP. Fixes: 30f1e7241422 ("block: move discard checks into the ioctl handler") Cc: stable@vger.kernel.org Signed-off-by: Shin'ichiro Kawasaki Reviewed-by: Damien Le Moal Reviewed-by: Chao Yu Reviewed-by: Christoph Hellwig Signed-off-by: Jaegeuk Kim --- fs/f2fs/segment.c | 7 +++++++ 1 file changed, 7 insertions(+) diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 3aee71c9f3c6..ed070b9ef91b 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -1282,6 +1282,13 @@ static int __submit_discard_cmd(struct f2fs_sb_info *sbi, wait_list, issued); return 0; } + + /* + * Issue discard for conventional zones only if the device + * supports discard. + */ + if (!bdev_max_discard_sectors(bdev)) + return -EOPNOTSUPP; } #endif From f785cec298c95d00058560c0715233294a04b8f3 Mon Sep 17 00:00:00 2001 From: Yeongjin Gil Date: Mon, 19 Aug 2024 17:34:30 +0900 Subject: [PATCH 030/107] f2fs: compress: don't redirty sparse cluster during {,de}compress In f2fs_do_write_data_page, when the data block is NULL_ADDR, it skips writepage considering that it has been already truncated. This results in an infinite loop as the PAGECACHE_TAG_TOWRITE tag is not cleared during the writeback process for a compressed file including NULL_ADDR in compress_mode=user. This is the reproduction process: 1. dd if=/dev/zero bs=4096 count=1024 seek=1024 of=testfile 2. f2fs_io compress testfile 3. dd if=/dev/zero bs=4096 count=1 conv=notrunc of=testfile 4. f2fs_io decompress testfile To prevent the problem, let's check whether the cluster is fully allocated before redirty its pages. Fixes: 5fdb322ff2c2 ("f2fs: add F2FS_IOC_DECOMPRESS_FILE and F2FS_IOC_COMPRESS_FILE") Reviewed-by: Sungjong Seo Reviewed-by: Sunmin Jeong Tested-by: Jaewook Kim Signed-off-by: Yeongjin Gil Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/compress.c | 36 ++++++++++++++++++++++++++++-------- fs/f2fs/f2fs.h | 12 ++++++++++++ fs/f2fs/file.c | 39 +++++++++++++++++++++------------------ 3 files changed, 61 insertions(+), 26 deletions(-) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index 990b93689b46..f55d54bb12f4 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -945,7 +945,7 @@ static int __f2fs_get_cluster_blocks(struct inode *inode, unsigned int cluster_size = F2FS_I(inode)->i_cluster_size; int count, i; - for (i = 1, count = 1; i < cluster_size; i++) { + for (i = 0, count = 0; i < cluster_size; i++) { block_t blkaddr = data_blkaddr(dn->inode, dn->node_page, dn->ofs_in_node + i); @@ -956,8 +956,8 @@ static int __f2fs_get_cluster_blocks(struct inode *inode, return count; } -static int __f2fs_cluster_blocks(struct inode *inode, - unsigned int cluster_idx, bool compr_blks) +static int __f2fs_cluster_blocks(struct inode *inode, unsigned int cluster_idx, + enum cluster_check_type type) { struct dnode_of_data dn; unsigned int start_idx = cluster_idx << @@ -978,10 +978,12 @@ static int __f2fs_cluster_blocks(struct inode *inode, } if (dn.data_blkaddr == COMPRESS_ADDR) { - if (compr_blks) - ret = __f2fs_get_cluster_blocks(inode, &dn); - else + if (type == CLUSTER_COMPR_BLKS) + ret = 1 + __f2fs_get_cluster_blocks(inode, &dn); + else if (type == CLUSTER_IS_COMPR) ret = 1; + } else if (type == CLUSTER_RAW_BLKS) { + ret = __f2fs_get_cluster_blocks(inode, &dn); } fail: f2fs_put_dnode(&dn); @@ -991,7 +993,16 @@ fail: /* return # of compressed blocks in compressed cluster */ static int f2fs_compressed_blocks(struct compress_ctx *cc) { - return __f2fs_cluster_blocks(cc->inode, cc->cluster_idx, true); + return __f2fs_cluster_blocks(cc->inode, cc->cluster_idx, + CLUSTER_COMPR_BLKS); +} + +/* return # of raw blocks in non-compressed cluster */ +static int f2fs_decompressed_blocks(struct inode *inode, + unsigned int cluster_idx) +{ + return __f2fs_cluster_blocks(inode, cluster_idx, + CLUSTER_RAW_BLKS); } /* return whether cluster is compressed one or not */ @@ -999,7 +1010,16 @@ int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index) { return __f2fs_cluster_blocks(inode, index >> F2FS_I(inode)->i_log_cluster_size, - false); + CLUSTER_IS_COMPR); +} + +/* return whether cluster contains non raw blocks or not */ +bool f2fs_is_sparse_cluster(struct inode *inode, pgoff_t index) +{ + unsigned int cluster_idx = index >> F2FS_I(inode)->i_log_cluster_size; + + return f2fs_decompressed_blocks(inode, cluster_idx) != + F2FS_I(inode)->i_cluster_size; } static bool cluster_may_compress(struct compress_ctx *cc) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 0e0189084462..ec82bde0c51d 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -4302,6 +4302,11 @@ static inline bool f2fs_meta_inode_gc_required(struct inode *inode) * compress.c */ #ifdef CONFIG_F2FS_FS_COMPRESSION +enum cluster_check_type { + CLUSTER_IS_COMPR, /* check only if compressed cluster */ + CLUSTER_COMPR_BLKS, /* return # of compressed blocks in a cluster */ + CLUSTER_RAW_BLKS /* return # of raw blocks in a cluster */ +}; bool f2fs_is_compressed_page(struct page *page); struct page *f2fs_compress_control_page(struct page *page); int f2fs_prepare_compress_overwrite(struct inode *inode, @@ -4328,6 +4333,7 @@ int f2fs_write_multi_pages(struct compress_ctx *cc, struct writeback_control *wbc, enum iostat_type io_type); int f2fs_is_compressed_cluster(struct inode *inode, pgoff_t index); +bool f2fs_is_sparse_cluster(struct inode *inode, pgoff_t index); void f2fs_update_read_extent_tree_range_compressed(struct inode *inode, pgoff_t fofs, block_t blkaddr, unsigned int llen, unsigned int c_len); @@ -4414,6 +4420,12 @@ static inline bool f2fs_load_compressed_page(struct f2fs_sb_info *sbi, static inline void f2fs_invalidate_compress_pages(struct f2fs_sb_info *sbi, nid_t ino) { } #define inc_compr_inode_stat(inode) do { } while (0) +static inline int f2fs_is_compressed_cluster( + struct inode *inode, + pgoff_t index) { return 0; } +static inline bool f2fs_is_sparse_cluster( + struct inode *inode, + pgoff_t index) { return true; } static inline void f2fs_update_read_extent_tree_range_compressed( struct inode *inode, pgoff_t fofs, block_t blkaddr, diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 50075f8d5e9b..0bfcc1bf5832 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -4236,9 +4236,8 @@ static int f2fs_ioc_decompress_file(struct file *filp) struct inode *inode = file_inode(filp); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct f2fs_inode_info *fi = F2FS_I(inode); - pgoff_t page_idx = 0, last_idx; - int cluster_size = fi->i_cluster_size; - int count, ret; + pgoff_t page_idx = 0, last_idx, cluster_idx; + int ret; if (!f2fs_sb_has_compression(sbi) || F2FS_OPTION(sbi).compress_mode != COMPR_MODE_USER) @@ -4273,10 +4272,15 @@ static int f2fs_ioc_decompress_file(struct file *filp) goto out; last_idx = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE); + last_idx >>= fi->i_log_cluster_size; - count = last_idx - page_idx; - while (count && count >= cluster_size) { - ret = redirty_blocks(inode, page_idx, cluster_size); + for (cluster_idx = 0; cluster_idx < last_idx; cluster_idx++) { + page_idx = cluster_idx << fi->i_log_cluster_size; + + if (!f2fs_is_compressed_cluster(inode, page_idx)) + continue; + + ret = redirty_blocks(inode, page_idx, fi->i_cluster_size); if (ret < 0) break; @@ -4286,9 +4290,6 @@ static int f2fs_ioc_decompress_file(struct file *filp) break; } - count -= cluster_size; - page_idx += cluster_size; - cond_resched(); if (fatal_signal_pending(current)) { ret = -EINTR; @@ -4315,9 +4316,9 @@ static int f2fs_ioc_compress_file(struct file *filp) { struct inode *inode = file_inode(filp); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); - pgoff_t page_idx = 0, last_idx; - int cluster_size = F2FS_I(inode)->i_cluster_size; - int count, ret; + struct f2fs_inode_info *fi = F2FS_I(inode); + pgoff_t page_idx = 0, last_idx, cluster_idx; + int ret; if (!f2fs_sb_has_compression(sbi) || F2FS_OPTION(sbi).compress_mode != COMPR_MODE_USER) @@ -4351,10 +4352,15 @@ static int f2fs_ioc_compress_file(struct file *filp) set_inode_flag(inode, FI_ENABLE_COMPRESS); last_idx = DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE); + last_idx >>= fi->i_log_cluster_size; - count = last_idx - page_idx; - while (count && count >= cluster_size) { - ret = redirty_blocks(inode, page_idx, cluster_size); + for (cluster_idx = 0; cluster_idx < last_idx; cluster_idx++) { + page_idx = cluster_idx << fi->i_log_cluster_size; + + if (f2fs_is_sparse_cluster(inode, page_idx)) + continue; + + ret = redirty_blocks(inode, page_idx, fi->i_cluster_size); if (ret < 0) break; @@ -4364,9 +4370,6 @@ static int f2fs_ioc_compress_file(struct file *filp) break; } - count -= cluster_size; - page_idx += cluster_size; - cond_resched(); if (fatal_signal_pending(current)) { ret = -EINTR; From b2c160f4f3cfe9334b3e6bb52cb89a3a20ef0385 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Fri, 16 Aug 2024 09:13:42 +0800 Subject: [PATCH 031/107] f2fs: atomic: fix to forbid dio in atomic_file atomic write can only be used via buffered IO, let's fail direct IO on atomic_file and return -EOPNOTSUPP. Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 36 ++++++++++++++++++++++++------------ 1 file changed, 24 insertions(+), 12 deletions(-) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 0bfcc1bf5832..81d0b86c5718 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2162,6 +2162,7 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) goto out; f2fs_down_write(&fi->i_gc_rwsem[WRITE]); + f2fs_down_write(&fi->i_gc_rwsem[READ]); /* * Should wait end_io to count F2FS_WB_CP_DATA correctly by @@ -2171,10 +2172,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) f2fs_warn(sbi, "Unexpected flush for atomic writes: ino=%lu, npages=%u", inode->i_ino, get_dirty_pages(inode)); ret = filemap_write_and_wait_range(inode->i_mapping, 0, LLONG_MAX); - if (ret) { - f2fs_up_write(&fi->i_gc_rwsem[WRITE]); - goto out; - } + if (ret) + goto out_unlock; /* Check if the inode already has a COW inode */ if (fi->cow_inode == NULL) { @@ -2183,10 +2182,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) struct inode *dir = d_inode(dentry->d_parent); ret = f2fs_get_tmpfile(idmap, dir, &fi->cow_inode); - if (ret) { - f2fs_up_write(&fi->i_gc_rwsem[WRITE]); - goto out; - } + if (ret) + goto out_unlock; set_inode_flag(fi->cow_inode, FI_COW_FILE); clear_inode_flag(fi->cow_inode, FI_INLINE_DATA); @@ -2200,10 +2197,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) invalidate_mapping_pages(fi->cow_inode->i_mapping, 0, -1); ret = f2fs_do_truncate_blocks(fi->cow_inode, 0, true); - if (ret) { - f2fs_up_write(&fi->i_gc_rwsem[WRITE]); - goto out; - } + if (ret) + goto out_unlock; } f2fs_write_inode(inode, NULL); @@ -2222,7 +2217,11 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) } f2fs_i_size_write(fi->cow_inode, isize); +out_unlock: + f2fs_up_write(&fi->i_gc_rwsem[READ]); f2fs_up_write(&fi->i_gc_rwsem[WRITE]); + if (ret) + goto out; f2fs_update_time(sbi, REQ_TIME); fi->atomic_write_task = current; @@ -4570,6 +4569,13 @@ static ssize_t f2fs_dio_read_iter(struct kiocb *iocb, struct iov_iter *to) f2fs_down_read(&fi->i_gc_rwsem[READ]); } + /* dio is not compatible w/ atomic file */ + if (f2fs_is_atomic_file(inode)) { + f2fs_up_read(&fi->i_gc_rwsem[READ]); + ret = -EOPNOTSUPP; + goto out; + } + /* * We have to use __iomap_dio_rw() and iomap_dio_complete() instead of * the higher-level function iomap_dio_rw() in order to ensure that the @@ -4985,6 +4991,12 @@ static ssize_t f2fs_file_write_iter(struct kiocb *iocb, struct iov_iter *from) /* Determine whether we will do a direct write or a buffered write. */ dio = f2fs_should_use_dio(inode, iocb, from); + /* dio is not compatible w/ atomic write */ + if (dio && f2fs_is_atomic_file(inode)) { + ret = -EOPNOTSUPP; + goto out_unlock; + } + /* Possibly preallocate the blocks for the write. */ target_size = iocb->ki_pos + iov_iter_count(from); preallocated = f2fs_preallocate_blocks(iocb, from, dio); From f7a678bbe5a8f22cfcef5369757cc9b95f73e027 Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Thu, 15 Aug 2024 08:47:28 +0800 Subject: [PATCH 032/107] f2fs: Use sysfs_emit_at() to simplify code This file already uses sysfs_emit(). So be consistent and also use sysfs_emit_at(). This slightly simplifies the code and makes it more readable. Reviewed-by: Chao Yu Signed-off-by: Christophe JAILLET Signed-off-by: Jaegeuk Kim --- fs/f2fs/sysfs.c | 45 +++++++++++++++++++++------------------------ 1 file changed, 21 insertions(+), 24 deletions(-) diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index ed4bf434207a..654a541dcc45 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -188,50 +188,50 @@ static ssize_t features_show(struct f2fs_attr *a, int len = 0; if (f2fs_sb_has_encrypt(sbi)) - len += scnprintf(buf, PAGE_SIZE - len, "%s", + len += sysfs_emit_at(buf, len, "%s", "encryption"); if (f2fs_sb_has_blkzoned(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "blkzoned"); if (f2fs_sb_has_extra_attr(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "extra_attr"); if (f2fs_sb_has_project_quota(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "projquota"); if (f2fs_sb_has_inode_chksum(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "inode_checksum"); if (f2fs_sb_has_flexible_inline_xattr(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "flexible_inline_xattr"); if (f2fs_sb_has_quota_ino(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "quota_ino"); if (f2fs_sb_has_inode_crtime(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "inode_crtime"); if (f2fs_sb_has_lost_found(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "lost_found"); if (f2fs_sb_has_verity(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "verity"); if (f2fs_sb_has_sb_chksum(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "sb_checksum"); if (f2fs_sb_has_casefold(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "casefold"); if (f2fs_sb_has_readonly(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "readonly"); if (f2fs_sb_has_compression(sbi)) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "compression"); - len += scnprintf(buf + len, PAGE_SIZE - len, "%s%s", + len += sysfs_emit_at(buf, len, "%s%s", len ? ", " : "", "pin_file"); - len += scnprintf(buf + len, PAGE_SIZE - len, "\n"); + len += sysfs_emit_at(buf, len, "\n"); return len; } @@ -329,17 +329,14 @@ static ssize_t f2fs_sbi_show(struct f2fs_attr *a, int hot_count = sbi->raw_super->hot_ext_count; int len = 0, i; - len += scnprintf(buf + len, PAGE_SIZE - len, - "cold file extension:\n"); + len += sysfs_emit_at(buf, len, "cold file extension:\n"); for (i = 0; i < cold_count; i++) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s\n", - extlist[i]); + len += sysfs_emit_at(buf, len, "%s\n", extlist[i]); - len += scnprintf(buf + len, PAGE_SIZE - len, - "hot file extension:\n"); + len += sysfs_emit_at(buf, len, "hot file extension:\n"); for (i = cold_count; i < cold_count + hot_count; i++) - len += scnprintf(buf + len, PAGE_SIZE - len, "%s\n", - extlist[i]); + len += sysfs_emit_at(buf, len, "%s\n", extlist[i]); + return len; } From 5de8acb41c86f1d335d165e0a350441ea3a1f480 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Wed, 29 May 2024 17:09:07 +0200 Subject: [PATCH 033/107] fuse: cleanup request queuing towards virtiofs Virtiofs has its own queuing mechanism, but still requests are first queued on fiq->pending to be immediately dequeued and queued onto the virtio queue. The queuing on fiq->pending is unnecessary and might even have some performance impact due to being a contention point. Forget requests are handled similarly. Move the queuing of requests and forgets into the fiq->ops->*. fuse_iqueue_ops are renamed to reflect the new semantics. Reviewed-by: Stefan Hajnoczi Fixed-by: Jingbo Xu Reviewed-by: Jingbo Xu Tested-by: Peter-Jan Gootzen Reviewed-by: Peter-Jan Gootzen Signed-off-by: Miklos Szeredi --- fs/fuse/dev.c | 159 ++++++++++++++++++++++++-------------------- fs/fuse/fuse_i.h | 19 ++---- fs/fuse/virtio_fs.c | 41 ++++-------- 3 files changed, 106 insertions(+), 113 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 67443ef07285..2364df0324f7 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -194,11 +194,22 @@ unsigned int fuse_len_args(unsigned int numargs, struct fuse_arg *args) } EXPORT_SYMBOL_GPL(fuse_len_args); -u64 fuse_get_unique(struct fuse_iqueue *fiq) +static u64 fuse_get_unique_locked(struct fuse_iqueue *fiq) { fiq->reqctr += FUSE_REQ_ID_STEP; return fiq->reqctr; } + +u64 fuse_get_unique(struct fuse_iqueue *fiq) +{ + u64 ret; + + spin_lock(&fiq->lock); + ret = fuse_get_unique_locked(fiq); + spin_unlock(&fiq->lock); + + return ret; +} EXPORT_SYMBOL_GPL(fuse_get_unique); static unsigned int fuse_req_hash(u64 unique) @@ -217,22 +228,68 @@ __releases(fiq->lock) spin_unlock(&fiq->lock); } +static void fuse_dev_queue_forget(struct fuse_iqueue *fiq, struct fuse_forget_link *forget) +{ + spin_lock(&fiq->lock); + if (fiq->connected) { + fiq->forget_list_tail->next = forget; + fiq->forget_list_tail = forget; + fuse_dev_wake_and_unlock(fiq); + } else { + kfree(forget); + spin_unlock(&fiq->lock); + } +} + +static void fuse_dev_queue_interrupt(struct fuse_iqueue *fiq, struct fuse_req *req) +{ + spin_lock(&fiq->lock); + if (list_empty(&req->intr_entry)) { + list_add_tail(&req->intr_entry, &fiq->interrupts); + /* + * Pairs with smp_mb() implied by test_and_set_bit() + * from fuse_request_end(). + */ + smp_mb(); + if (test_bit(FR_FINISHED, &req->flags)) { + list_del_init(&req->intr_entry); + spin_unlock(&fiq->lock); + } else { + fuse_dev_wake_and_unlock(fiq); + } + } else { + spin_unlock(&fiq->lock); + } +} + +static void fuse_dev_queue_req(struct fuse_iqueue *fiq, struct fuse_req *req) +{ + spin_lock(&fiq->lock); + if (fiq->connected) { + if (req->in.h.opcode != FUSE_NOTIFY_REPLY) + req->in.h.unique = fuse_get_unique_locked(fiq); + list_add_tail(&req->list, &fiq->pending); + fuse_dev_wake_and_unlock(fiq); + } else { + spin_unlock(&fiq->lock); + req->out.h.error = -ENOTCONN; + fuse_request_end(req); + } +} + const struct fuse_iqueue_ops fuse_dev_fiq_ops = { - .wake_forget_and_unlock = fuse_dev_wake_and_unlock, - .wake_interrupt_and_unlock = fuse_dev_wake_and_unlock, - .wake_pending_and_unlock = fuse_dev_wake_and_unlock, + .send_forget = fuse_dev_queue_forget, + .send_interrupt = fuse_dev_queue_interrupt, + .send_req = fuse_dev_queue_req, }; EXPORT_SYMBOL_GPL(fuse_dev_fiq_ops); -static void queue_request_and_unlock(struct fuse_iqueue *fiq, - struct fuse_req *req) -__releases(fiq->lock) +static void fuse_send_one(struct fuse_iqueue *fiq, struct fuse_req *req) { req->in.h.len = sizeof(struct fuse_in_header) + fuse_len_args(req->args->in_numargs, (struct fuse_arg *) req->args->in_args); - list_add_tail(&req->list, &fiq->pending); - fiq->ops->wake_pending_and_unlock(fiq); + fiq->ops->send_req(fiq, req); } void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget, @@ -243,15 +300,7 @@ void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget, forget->forget_one.nodeid = nodeid; forget->forget_one.nlookup = nlookup; - spin_lock(&fiq->lock); - if (fiq->connected) { - fiq->forget_list_tail->next = forget; - fiq->forget_list_tail = forget; - fiq->ops->wake_forget_and_unlock(fiq); - } else { - kfree(forget); - spin_unlock(&fiq->lock); - } + fiq->ops->send_forget(fiq, forget); } static void flush_bg_queue(struct fuse_conn *fc) @@ -265,9 +314,7 @@ static void flush_bg_queue(struct fuse_conn *fc) req = list_first_entry(&fc->bg_queue, struct fuse_req, list); list_del(&req->list); fc->active_background++; - spin_lock(&fiq->lock); - req->in.h.unique = fuse_get_unique(fiq); - queue_request_and_unlock(fiq, req); + fuse_send_one(fiq, req); } } @@ -337,29 +384,12 @@ static int queue_interrupt(struct fuse_req *req) { struct fuse_iqueue *fiq = &req->fm->fc->iq; - spin_lock(&fiq->lock); /* Check for we've sent request to interrupt this req */ - if (unlikely(!test_bit(FR_INTERRUPTED, &req->flags))) { - spin_unlock(&fiq->lock); + if (unlikely(!test_bit(FR_INTERRUPTED, &req->flags))) return -EINVAL; - } - if (list_empty(&req->intr_entry)) { - list_add_tail(&req->intr_entry, &fiq->interrupts); - /* - * Pairs with smp_mb() implied by test_and_set_bit() - * from fuse_request_end(). - */ - smp_mb(); - if (test_bit(FR_FINISHED, &req->flags)) { - list_del_init(&req->intr_entry); - spin_unlock(&fiq->lock); - return 0; - } - fiq->ops->wake_interrupt_and_unlock(fiq); - } else { - spin_unlock(&fiq->lock); - } + fiq->ops->send_interrupt(fiq, req); + return 0; } @@ -414,21 +444,15 @@ static void __fuse_request_send(struct fuse_req *req) struct fuse_iqueue *fiq = &req->fm->fc->iq; BUG_ON(test_bit(FR_BACKGROUND, &req->flags)); - spin_lock(&fiq->lock); - if (!fiq->connected) { - spin_unlock(&fiq->lock); - req->out.h.error = -ENOTCONN; - } else { - req->in.h.unique = fuse_get_unique(fiq); - /* acquire extra reference, since request is still needed - after fuse_request_end() */ - __fuse_get_request(req); - queue_request_and_unlock(fiq, req); - request_wait_answer(req); - /* Pairs with smp_wmb() in fuse_request_end() */ - smp_rmb(); - } + /* acquire extra reference, since request is still needed after + fuse_request_end() */ + __fuse_get_request(req); + fuse_send_one(fiq, req); + + request_wait_answer(req); + /* Pairs with smp_wmb() in fuse_request_end() */ + smp_rmb(); } static void fuse_adjust_compat(struct fuse_conn *fc, struct fuse_args *args) @@ -583,7 +607,6 @@ static int fuse_simple_notify_reply(struct fuse_mount *fm, { struct fuse_req *req; struct fuse_iqueue *fiq = &fm->fc->iq; - int err = 0; req = fuse_get_req(fm, false); if (IS_ERR(req)) @@ -594,16 +617,9 @@ static int fuse_simple_notify_reply(struct fuse_mount *fm, fuse_args_to_req(req, args); - spin_lock(&fiq->lock); - if (fiq->connected) { - queue_request_and_unlock(fiq, req); - } else { - err = -ENODEV; - spin_unlock(&fiq->lock); - fuse_put_request(req); - } + fuse_send_one(fiq, req); - return err; + return 0; } /* @@ -1075,9 +1091,9 @@ __releases(fiq->lock) return err ? err : reqsize; } -struct fuse_forget_link *fuse_dequeue_forget(struct fuse_iqueue *fiq, - unsigned int max, - unsigned int *countp) +static struct fuse_forget_link *fuse_dequeue_forget(struct fuse_iqueue *fiq, + unsigned int max, + unsigned int *countp) { struct fuse_forget_link *head = fiq->forget_list_head.next; struct fuse_forget_link **newhead = &head; @@ -1096,7 +1112,6 @@ struct fuse_forget_link *fuse_dequeue_forget(struct fuse_iqueue *fiq, return head; } -EXPORT_SYMBOL(fuse_dequeue_forget); static int fuse_read_single_forget(struct fuse_iqueue *fiq, struct fuse_copy_state *cs, @@ -1111,7 +1126,7 @@ __releases(fiq->lock) struct fuse_in_header ih = { .opcode = FUSE_FORGET, .nodeid = forget->forget_one.nodeid, - .unique = fuse_get_unique(fiq), + .unique = fuse_get_unique_locked(fiq), .len = sizeof(ih) + sizeof(arg), }; @@ -1142,7 +1157,7 @@ __releases(fiq->lock) struct fuse_batch_forget_in arg = { .count = 0 }; struct fuse_in_header ih = { .opcode = FUSE_BATCH_FORGET, - .unique = fuse_get_unique(fiq), + .unique = fuse_get_unique_locked(fiq), .len = sizeof(ih) + sizeof(arg), }; @@ -1828,7 +1843,7 @@ static void fuse_resend(struct fuse_conn *fc) } /* iq and pq requests are both oldest to newest */ list_splice(&to_queue, &fiq->pending); - fiq->ops->wake_pending_and_unlock(fiq); + fuse_dev_wake_and_unlock(fiq); } static int fuse_notify_resend(struct fuse_conn *fc) diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index f23919610313..33b21255817e 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -449,22 +449,19 @@ struct fuse_iqueue; */ struct fuse_iqueue_ops { /** - * Signal that a forget has been queued + * Send one forget */ - void (*wake_forget_and_unlock)(struct fuse_iqueue *fiq) - __releases(fiq->lock); + void (*send_forget)(struct fuse_iqueue *fiq, struct fuse_forget_link *link); /** - * Signal that an INTERRUPT request has been queued + * Send interrupt for request */ - void (*wake_interrupt_and_unlock)(struct fuse_iqueue *fiq) - __releases(fiq->lock); + void (*send_interrupt)(struct fuse_iqueue *fiq, struct fuse_req *req); /** - * Signal that a request has been queued + * Send one request */ - void (*wake_pending_and_unlock)(struct fuse_iqueue *fiq) - __releases(fiq->lock); + void (*send_req)(struct fuse_iqueue *fiq, struct fuse_req *req); /** * Clean up when fuse_iqueue is destroyed @@ -1053,10 +1050,6 @@ void fuse_queue_forget(struct fuse_conn *fc, struct fuse_forget_link *forget, struct fuse_forget_link *fuse_alloc_forget(void); -struct fuse_forget_link *fuse_dequeue_forget(struct fuse_iqueue *fiq, - unsigned int max, - unsigned int *countp); - /* * Initialize READ or READDIR request */ diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c index dd5260141615..2fee9eb5ad0b 100644 --- a/fs/fuse/virtio_fs.c +++ b/fs/fuse/virtio_fs.c @@ -1091,22 +1091,13 @@ static struct virtio_driver virtio_fs_driver = { #endif }; -static void virtio_fs_wake_forget_and_unlock(struct fuse_iqueue *fiq) -__releases(fiq->lock) +static void virtio_fs_send_forget(struct fuse_iqueue *fiq, struct fuse_forget_link *link) { - struct fuse_forget_link *link; struct virtio_fs_forget *forget; struct virtio_fs_forget_req *req; - struct virtio_fs *fs; - struct virtio_fs_vq *fsvq; - u64 unique; - - link = fuse_dequeue_forget(fiq, 1, NULL); - unique = fuse_get_unique(fiq); - - fs = fiq->priv; - fsvq = &fs->vqs[VQ_HIPRIO]; - spin_unlock(&fiq->lock); + struct virtio_fs *fs = fiq->priv; + struct virtio_fs_vq *fsvq = &fs->vqs[VQ_HIPRIO]; + u64 unique = fuse_get_unique(fiq); /* Allocate a buffer for the request */ forget = kmalloc(sizeof(*forget), GFP_NOFS | __GFP_NOFAIL); @@ -1126,8 +1117,7 @@ __releases(fiq->lock) kfree(link); } -static void virtio_fs_wake_interrupt_and_unlock(struct fuse_iqueue *fiq) -__releases(fiq->lock) +static void virtio_fs_send_interrupt(struct fuse_iqueue *fiq, struct fuse_req *req) { /* * TODO interrupts. @@ -1136,7 +1126,6 @@ __releases(fiq->lock) * Exceptions are blocking lock operations; for example fcntl(F_SETLKW) * with shared lock between host and guest. */ - spin_unlock(&fiq->lock); } /* Count number of scatter-gather elements required */ @@ -1341,21 +1330,17 @@ out: return ret; } -static void virtio_fs_wake_pending_and_unlock(struct fuse_iqueue *fiq) -__releases(fiq->lock) +static void virtio_fs_send_req(struct fuse_iqueue *fiq, struct fuse_req *req) { unsigned int queue_id; struct virtio_fs *fs; - struct fuse_req *req; struct virtio_fs_vq *fsvq; int ret; - WARN_ON(list_empty(&fiq->pending)); - req = list_last_entry(&fiq->pending, struct fuse_req, list); + if (req->in.h.opcode != FUSE_NOTIFY_REPLY) + req->in.h.unique = fuse_get_unique(fiq); + clear_bit(FR_PENDING, &req->flags); - list_del_init(&req->list); - WARN_ON(!list_empty(&fiq->pending)); - spin_unlock(&fiq->lock); fs = fiq->priv; queue_id = VQ_REQUEST + fs->mq_map[raw_smp_processor_id()]; @@ -1393,10 +1378,10 @@ __releases(fiq->lock) } static const struct fuse_iqueue_ops virtio_fs_fiq_ops = { - .wake_forget_and_unlock = virtio_fs_wake_forget_and_unlock, - .wake_interrupt_and_unlock = virtio_fs_wake_interrupt_and_unlock, - .wake_pending_and_unlock = virtio_fs_wake_pending_and_unlock, - .release = virtio_fs_fiq_release, + .send_forget = virtio_fs_send_forget, + .send_interrupt = virtio_fs_send_interrupt, + .send_req = virtio_fs_send_req, + .release = virtio_fs_fiq_release, }; static inline void virtio_fs_ctx_set_defaults(struct fuse_fs_context *ctx) From ac5cffec53be0b0231b89470a357bd3a5814f599 Mon Sep 17 00:00:00 2001 From: yangyun Date: Wed, 14 Aug 2024 17:36:00 +0800 Subject: [PATCH 034/107] fuse: add fast path for fuse_range_is_writeback In some cases, the fi->writepages may be empty. And there is no need to check fi->writepages with spin_lock, which may have an impact on performance due to lock contention. For example, in scenarios where multiple readers read the same file without any writers, or where the page cache is not enabled. Also remove the outdated comment since commit 6b2fb79963fb ("fuse: optimize writepages search") has optimize the situation by replacing list with rb-tree. Signed-off-by: yangyun Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index ed76121f73f2..b41c032d1c6d 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -448,9 +448,6 @@ static struct fuse_writepage_args *fuse_find_writeback(struct fuse_inode *fi, /* * Check if any page in a range is under writeback - * - * This is currently done by walking the list of writepage requests - * for the inode, which can be pretty inefficient. */ static bool fuse_range_is_writeback(struct inode *inode, pgoff_t idx_from, pgoff_t idx_to) @@ -458,6 +455,9 @@ static bool fuse_range_is_writeback(struct inode *inode, pgoff_t idx_from, struct fuse_inode *fi = get_fuse_inode(inode); bool found; + if (RB_EMPTY_ROOT(&fi->writepages)) + return false; + spin_lock(&fi->lock); found = fuse_find_writeback(fi, idx_from, idx_to); spin_unlock(&fi->lock); From 509a6458b44f72bb6854854c89cf76e56f11c9f1 Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:02 -0700 Subject: [PATCH 035/107] fuse: drop unused fuse_mount arg in fuse_writepage_finish() Drop the unused "struct fuse_mount *fm" arg in fuse_writepage_finish(). No functional changes added. Signed-off-by: Joanne Koong Reviewed-by: Jingbo Xu Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index b41c032d1c6d..39d20c88283b 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1769,8 +1769,7 @@ static void fuse_writepage_free(struct fuse_writepage_args *wpa) kfree(wpa); } -static void fuse_writepage_finish(struct fuse_mount *fm, - struct fuse_writepage_args *wpa) +static void fuse_writepage_finish(struct fuse_writepage_args *wpa) { struct fuse_args_pages *ap = &wpa->ia.ap; struct inode *inode = wpa->inode; @@ -1829,7 +1828,7 @@ __acquires(fi->lock) out_free: fi->writectr--; rb_erase(&wpa->writepages_entry, &fi->writepages); - fuse_writepage_finish(fm, wpa); + fuse_writepage_finish(wpa); spin_unlock(&fi->lock); /* After rb_erase() aux request list is private */ @@ -1965,7 +1964,7 @@ static void fuse_writepage_end(struct fuse_mount *fm, struct fuse_args *args, fuse_send_writepage(fm, next, inarg->offset + inarg->size); } fi->writectr--; - fuse_writepage_finish(fm, wpa); + fuse_writepage_finish(wpa); spin_unlock(&fi->lock); fuse_writepage_free(wpa); } From c04e3b2118192384153b4eac595768e2ffb7ac4a Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:03 -0700 Subject: [PATCH 036/107] fuse: refactor finished writeback stats updates into helper function Move the logic for updating the bdi and page stats for a finished writeback into a separate helper function, where it can be called from both fuse_writepage_finish() and fuse_writepage_add() (in the case where there is already an auxiliary write request for the page). No functional changes added. Suggested by: Jingbo Xu Signed-off-by: Joanne Koong Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 31 ++++++++++++++----------------- 1 file changed, 14 insertions(+), 17 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 39d20c88283b..e0bc5528d6b5 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1769,19 +1769,25 @@ static void fuse_writepage_free(struct fuse_writepage_args *wpa) kfree(wpa); } +static void fuse_writepage_finish_stat(struct inode *inode, struct page *page) +{ + struct backing_dev_info *bdi = inode_to_bdi(inode); + + dec_wb_stat(&bdi->wb, WB_WRITEBACK); + dec_node_page_state(page, NR_WRITEBACK_TEMP); + wb_writeout_inc(&bdi->wb); +} + static void fuse_writepage_finish(struct fuse_writepage_args *wpa) { struct fuse_args_pages *ap = &wpa->ia.ap; struct inode *inode = wpa->inode; struct fuse_inode *fi = get_fuse_inode(inode); - struct backing_dev_info *bdi = inode_to_bdi(inode); int i; - for (i = 0; i < ap->num_pages; i++) { - dec_wb_stat(&bdi->wb, WB_WRITEBACK); - dec_node_page_state(ap->pages[i], NR_WRITEBACK_TEMP); - wb_writeout_inc(&bdi->wb); - } + for (i = 0; i < ap->num_pages; i++) + fuse_writepage_finish_stat(inode, ap->pages[i]); + wake_up(&fi->page_waitq); } @@ -1833,14 +1839,9 @@ __acquires(fi->lock) /* After rb_erase() aux request list is private */ for (aux = wpa->next; aux; aux = next) { - struct backing_dev_info *bdi = inode_to_bdi(aux->inode); - next = aux->next; aux->next = NULL; - - dec_wb_stat(&bdi->wb, WB_WRITEBACK); - dec_node_page_state(aux->ia.ap.pages[0], NR_WRITEBACK_TEMP); - wb_writeout_inc(&bdi->wb); + fuse_writepage_finish_stat(aux->inode, aux->ia.ap.pages[0]); fuse_writepage_free(aux); } @@ -2209,11 +2210,7 @@ static bool fuse_writepage_add(struct fuse_writepage_args *new_wpa, spin_unlock(&fi->lock); if (tmp) { - struct backing_dev_info *bdi = inode_to_bdi(new_wpa->inode); - - dec_wb_stat(&bdi->wb, WB_WRITEBACK); - dec_node_page_state(new_ap->pages[0], NR_WRITEBACK_TEMP); - wb_writeout_inc(&bdi->wb); + fuse_writepage_finish_stat(new_wpa->inode, new_ap->pages[0]); fuse_writepage_free(new_wpa); } From 672c3b7457fcee9656c36a29a4b21ec4a652433e Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:05 -0700 Subject: [PATCH 037/107] fuse: move initialization of fuse_file to fuse_writepages() instead of in callback Prior to this change, data->ff is checked and if not initialized then initialized in the fuse_writepages_fill() callback, which gets called for every dirty page in the address space mapping. This logic is better placed in the main fuse_writepages() caller where data.ff is initialized before walking the dirty pages. No functional changes added. Signed-off-by: Joanne Koong Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index e0bc5528d6b5..9d4f8bdc3865 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -2263,13 +2263,6 @@ static int fuse_writepages_fill(struct folio *folio, struct page *tmp_page; int err; - if (!data->ff) { - err = -EIO; - data->ff = fuse_write_file_get(fi); - if (!data->ff) - goto out_unlock; - } - if (wpa && fuse_writepage_need_send(fc, &folio->page, ap, data)) { fuse_writepages_send(data); data->wpa = NULL; @@ -2348,13 +2341,13 @@ static int fuse_writepages(struct address_space *mapping, struct writeback_control *wbc) { struct inode *inode = mapping->host; + struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_fill_wb_data data; int err; - err = -EIO; if (fuse_is_bad(inode)) - goto out; + return -EIO; if (wbc->sync_mode == WB_SYNC_NONE && fc->num_background >= fc->congestion_threshold) @@ -2362,7 +2355,9 @@ static int fuse_writepages(struct address_space *mapping, data.inode = inode; data.wpa = NULL; - data.ff = NULL; + data.ff = fuse_write_file_get(fi); + if (!data.ff) + return -EIO; err = -ENOMEM; data.orig_pages = kcalloc(fc->max_pages, @@ -2376,11 +2371,10 @@ static int fuse_writepages(struct address_space *mapping, WARN_ON(!data.wpa->ia.ap.num_pages); fuse_writepages_send(&data); } - if (data.ff) - fuse_file_put(data.ff, false); kfree(data.orig_pages); out: + fuse_file_put(data.ff, false); return err; } From 9a8ebcf5e04e6cc9472bfcdd90b2aeef35a2f8f6 Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:06 -0700 Subject: [PATCH 038/107] fuse: convert fuse_writepages_fill() to use a folio for its tmp page To pave the way for refactoring out the shared logic in fuse_writepages_fill() and fuse_writepage_locked(), this change converts the temporary page in fuse_writepages_fill() to use the folio API. This is similar to the change in commit e0887e095a80 ("fuse: Convert fuse_writepage_locked to take a folio"), which converted the tmp page in fuse_writepage_locked() to use the folio API. inc_node_page_state() is intentionally preserved here instead of converting to node_stat_add_folio() since it is updating the stat of the underlying page and to better maintain API symmetry with dec_node_page_stat() in fuse_writepage_finish_stat(). No functional changes added. Signed-off-by: Joanne Koong Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 9d4f8bdc3865..0653fac40782 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -2260,7 +2260,7 @@ static int fuse_writepages_fill(struct folio *folio, struct inode *inode = data->inode; struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_conn *fc = get_fuse_conn(inode); - struct page *tmp_page; + struct folio *tmp_folio; int err; if (wpa && fuse_writepage_need_send(fc, &folio->page, ap, data)) { @@ -2269,8 +2269,8 @@ static int fuse_writepages_fill(struct folio *folio, } err = -ENOMEM; - tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM); - if (!tmp_page) + tmp_folio = folio_alloc(GFP_NOFS | __GFP_HIGHMEM, 0); + if (!tmp_folio) goto out_unlock; /* @@ -2290,7 +2290,7 @@ static int fuse_writepages_fill(struct folio *folio, err = -ENOMEM; wpa = fuse_writepage_args_alloc(); if (!wpa) { - __free_page(tmp_page); + folio_put(tmp_folio); goto out_unlock; } fuse_writepage_add_to_bucket(fc, wpa); @@ -2308,14 +2308,14 @@ static int fuse_writepages_fill(struct folio *folio, } folio_start_writeback(folio); - copy_highpage(tmp_page, &folio->page); - ap->pages[ap->num_pages] = tmp_page; + folio_copy(tmp_folio, folio); + ap->pages[ap->num_pages] = &tmp_folio->page; ap->descs[ap->num_pages].offset = 0; ap->descs[ap->num_pages].length = PAGE_SIZE; data->orig_pages[ap->num_pages] = &folio->page; inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK); - inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP); + inc_node_page_state(&tmp_folio->page, NR_WRITEBACK_TEMP); err = 0; if (data->wpa) { From 4046d3adcca42b7678f11c71e46bd32bafb4dad1 Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:07 -0700 Subject: [PATCH 039/107] fuse: move fuse file initialization to wpa allocation time Before this change, wpa->ia.ff is initialized with an acquired reference on the fuse file right before it submits the writeback request. If there are auxiliary writebacks, then the initialization and reference acquisition needs to also be set before we submit the auxiliary writeback request. To make the logic simpler and to pave the way for a subsequent refactoring of fuse_writepages_fill() and fuse_writepage_locked(), this change initializes and acquires wpa->ia.ff when the wpa is allocated. No functional changes added. Signed-off-by: Joanne Koong Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 0653fac40782..be0d7fdf04d0 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1762,8 +1762,7 @@ static void fuse_writepage_free(struct fuse_writepage_args *wpa) for (i = 0; i < ap->num_pages; i++) __free_page(ap->pages[i]); - if (wpa->ia.ff) - fuse_file_put(wpa->ia.ff, false); + fuse_file_put(wpa->ia.ff, false); kfree(ap->pages); kfree(wpa); @@ -1936,7 +1935,6 @@ static void fuse_writepage_end(struct fuse_mount *fm, struct fuse_args *args, wpa->next = next->next; next->next = NULL; - next->ia.ff = fuse_file_get(wpa->ia.ff); tree_insert(&fi->writepages, next); /* @@ -2155,7 +2153,6 @@ static void fuse_writepages_send(struct fuse_fill_wb_data *data) int num_pages = wpa->ia.ap.num_pages; int i; - wpa->ia.ff = fuse_file_get(data->ff); spin_lock(&fi->lock); list_add_tail(&wpa->queue_entry, &fi->queued_writes); fuse_flush_writepages(inode); @@ -2300,6 +2297,7 @@ static int fuse_writepages_fill(struct folio *folio, ap = &wpa->ia.ap; fuse_write_args_fill(&wpa->ia, data->ff, folio_pos(folio), 0); wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE; + wpa->ia.ff = fuse_file_get(data->ff); wpa->next = NULL; ap->args.in_pages = true; ap->args.end = fuse_writepage_end; From 0acad9289be33d324537d6c51988be0541b1139d Mon Sep 17 00:00:00 2001 From: Joanne Koong Date: Mon, 26 Aug 2024 14:19:08 -0700 Subject: [PATCH 040/107] fuse: refactor out shared logic in fuse_writepages_fill() and fuse_writepage_locked() This change refactors the shared logic in fuse_writepages_fill() and fuse_writepages_locked() into two separate helper functions, fuse_writepage_args_page_fill() and fuse_writepage_args_setup(). No functional changes added. Signed-off-by: Joanne Koong Reviewed-by: Josef Bacik Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 103 +++++++++++++++++++++++++++---------------------- 1 file changed, 57 insertions(+), 46 deletions(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index be0d7fdf04d0..fc587f3bc56f 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -2047,49 +2047,77 @@ static void fuse_writepage_add_to_bucket(struct fuse_conn *fc, rcu_read_unlock(); } +static void fuse_writepage_args_page_fill(struct fuse_writepage_args *wpa, struct folio *folio, + struct folio *tmp_folio, uint32_t page_index) +{ + struct inode *inode = folio->mapping->host; + struct fuse_args_pages *ap = &wpa->ia.ap; + + folio_copy(tmp_folio, folio); + + ap->pages[page_index] = &tmp_folio->page; + ap->descs[page_index].offset = 0; + ap->descs[page_index].length = PAGE_SIZE; + + inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK); + inc_node_page_state(&tmp_folio->page, NR_WRITEBACK_TEMP); +} + +static struct fuse_writepage_args *fuse_writepage_args_setup(struct folio *folio, + struct fuse_file *ff) +{ + struct inode *inode = folio->mapping->host; + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_writepage_args *wpa; + struct fuse_args_pages *ap; + + wpa = fuse_writepage_args_alloc(); + if (!wpa) + return NULL; + + fuse_writepage_add_to_bucket(fc, wpa); + fuse_write_args_fill(&wpa->ia, ff, folio_pos(folio), 0); + wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE; + wpa->inode = inode; + wpa->ia.ff = ff; + + ap = &wpa->ia.ap; + ap->args.in_pages = true; + ap->args.end = fuse_writepage_end; + + return wpa; +} + static int fuse_writepage_locked(struct folio *folio) { struct address_space *mapping = folio->mapping; struct inode *inode = mapping->host; - struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_writepage_args *wpa; struct fuse_args_pages *ap; struct folio *tmp_folio; + struct fuse_file *ff; int error = -ENOMEM; - folio_start_writeback(folio); - - wpa = fuse_writepage_args_alloc(); - if (!wpa) - goto err; - ap = &wpa->ia.ap; - tmp_folio = folio_alloc(GFP_NOFS | __GFP_HIGHMEM, 0); if (!tmp_folio) - goto err_free; + goto err; error = -EIO; - wpa->ia.ff = fuse_write_file_get(fi); - if (!wpa->ia.ff) + ff = fuse_write_file_get(fi); + if (!ff) goto err_nofile; - fuse_writepage_add_to_bucket(fc, wpa); - fuse_write_args_fill(&wpa->ia, wpa->ia.ff, folio_pos(folio), 0); + wpa = fuse_writepage_args_setup(folio, ff); + error = -ENOMEM; + if (!wpa) + goto err_writepage_args; - folio_copy(tmp_folio, folio); - wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE; - wpa->next = NULL; - ap->args.in_pages = true; + ap = &wpa->ia.ap; ap->num_pages = 1; - ap->pages[0] = &tmp_folio->page; - ap->descs[0].offset = 0; - ap->descs[0].length = PAGE_SIZE; - ap->args.end = fuse_writepage_end; - wpa->inode = inode; - inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK); - node_stat_add_folio(tmp_folio, NR_WRITEBACK_TEMP); + folio_start_writeback(folio); + fuse_writepage_args_page_fill(wpa, folio, tmp_folio, 0); spin_lock(&fi->lock); tree_insert(&fi->writepages, wpa); @@ -2101,13 +2129,12 @@ static int fuse_writepage_locked(struct folio *folio) return 0; +err_writepage_args: + fuse_file_put(ff, false); err_nofile: folio_put(tmp_folio); -err_free: - kfree(wpa); err: mapping_set_error(folio->mapping, error); - folio_end_writeback(folio); return error; } @@ -2285,36 +2312,20 @@ static int fuse_writepages_fill(struct folio *folio, */ if (data->wpa == NULL) { err = -ENOMEM; - wpa = fuse_writepage_args_alloc(); + wpa = fuse_writepage_args_setup(folio, data->ff); if (!wpa) { folio_put(tmp_folio); goto out_unlock; } - fuse_writepage_add_to_bucket(fc, wpa); - + fuse_file_get(wpa->ia.ff); data->max_pages = 1; - ap = &wpa->ia.ap; - fuse_write_args_fill(&wpa->ia, data->ff, folio_pos(folio), 0); - wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE; - wpa->ia.ff = fuse_file_get(data->ff); - wpa->next = NULL; - ap->args.in_pages = true; - ap->args.end = fuse_writepage_end; - ap->num_pages = 0; - wpa->inode = inode; } folio_start_writeback(folio); - folio_copy(tmp_folio, folio); - ap->pages[ap->num_pages] = &tmp_folio->page; - ap->descs[ap->num_pages].offset = 0; - ap->descs[ap->num_pages].length = PAGE_SIZE; + fuse_writepage_args_page_fill(wpa, folio, tmp_folio, ap->num_pages); data->orig_pages[ap->num_pages] = &folio->page; - inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK); - inc_node_page_state(&tmp_folio->page, NR_WRITEBACK_TEMP); - err = 0; if (data->wpa) { /* From 396b209e405a571ce8e06d3760ffc3e389a944f1 Mon Sep 17 00:00:00 2001 From: Josef Bacik Date: Wed, 3 Jul 2024 10:38:42 -0400 Subject: [PATCH 041/107] fuse: add simple request tracepoints I've been timing various fuse operations and it's quite annoying to do with kprobes. Add two tracepoints for sending and ending fuse requests to make it easier to debug and time various operations. Signed-off-by: Josef Bacik Reviewed-by: Bernd Schubert Signed-off-by: Miklos Szeredi --- fs/fuse/Makefile | 3 + fs/fuse/dev.c | 5 ++ fs/fuse/fuse_trace.h | 132 +++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 140 insertions(+) create mode 100644 fs/fuse/fuse_trace.h diff --git a/fs/fuse/Makefile b/fs/fuse/Makefile index 6e0228c6d0cb..ce0ff7a9007b 100644 --- a/fs/fuse/Makefile +++ b/fs/fuse/Makefile @@ -3,6 +3,9 @@ # Makefile for the FUSE filesystem. # +# Needed for trace events +ccflags-y = -I$(src) + obj-$(CONFIG_FUSE_FS) += fuse.o obj-$(CONFIG_CUSE) += cuse.o obj-$(CONFIG_VIRTIO_FS) += virtiofs.o diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 2364df0324f7..1b6088288a16 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -22,6 +22,9 @@ #include #include +#define CREATE_TRACE_POINTS +#include "fuse_trace.h" + MODULE_ALIAS_MISCDEV(FUSE_MINOR); MODULE_ALIAS("devname:fuse"); @@ -289,6 +292,7 @@ static void fuse_send_one(struct fuse_iqueue *fiq, struct fuse_req *req) req->in.h.len = sizeof(struct fuse_in_header) + fuse_len_args(req->args->in_numargs, (struct fuse_arg *) req->args->in_args); + trace_fuse_request_send(req); fiq->ops->send_req(fiq, req); } @@ -335,6 +339,7 @@ void fuse_request_end(struct fuse_req *req) if (test_and_set_bit(FR_FINISHED, &req->flags)) goto put_request; + trace_fuse_request_end(req); /* * test_and_set_bit() implies smp_mb() between bit * changing and below FR_INTERRUPTED check. Pairs with diff --git a/fs/fuse/fuse_trace.h b/fs/fuse/fuse_trace.h new file mode 100644 index 000000000000..bbe9ddd8c716 --- /dev/null +++ b/fs/fuse/fuse_trace.h @@ -0,0 +1,132 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM fuse + +#if !defined(_TRACE_FUSE_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_FUSE_H + +#include + +#define OPCODES \ + EM( FUSE_LOOKUP, "FUSE_LOOKUP") \ + EM( FUSE_FORGET, "FUSE_FORGET") \ + EM( FUSE_GETATTR, "FUSE_GETATTR") \ + EM( FUSE_SETATTR, "FUSE_SETATTR") \ + EM( FUSE_READLINK, "FUSE_READLINK") \ + EM( FUSE_SYMLINK, "FUSE_SYMLINK") \ + EM( FUSE_MKNOD, "FUSE_MKNOD") \ + EM( FUSE_MKDIR, "FUSE_MKDIR") \ + EM( FUSE_UNLINK, "FUSE_UNLINK") \ + EM( FUSE_RMDIR, "FUSE_RMDIR") \ + EM( FUSE_RENAME, "FUSE_RENAME") \ + EM( FUSE_LINK, "FUSE_LINK") \ + EM( FUSE_OPEN, "FUSE_OPEN") \ + EM( FUSE_READ, "FUSE_READ") \ + EM( FUSE_WRITE, "FUSE_WRITE") \ + EM( FUSE_STATFS, "FUSE_STATFS") \ + EM( FUSE_RELEASE, "FUSE_RELEASE") \ + EM( FUSE_FSYNC, "FUSE_FSYNC") \ + EM( FUSE_SETXATTR, "FUSE_SETXATTR") \ + EM( FUSE_GETXATTR, "FUSE_GETXATTR") \ + EM( FUSE_LISTXATTR, "FUSE_LISTXATTR") \ + EM( FUSE_REMOVEXATTR, "FUSE_REMOVEXATTR") \ + EM( FUSE_FLUSH, "FUSE_FLUSH") \ + EM( FUSE_INIT, "FUSE_INIT") \ + EM( FUSE_OPENDIR, "FUSE_OPENDIR") \ + EM( FUSE_READDIR, "FUSE_READDIR") \ + EM( FUSE_RELEASEDIR, "FUSE_RELEASEDIR") \ + EM( FUSE_FSYNCDIR, "FUSE_FSYNCDIR") \ + EM( FUSE_GETLK, "FUSE_GETLK") \ + EM( FUSE_SETLK, "FUSE_SETLK") \ + EM( FUSE_SETLKW, "FUSE_SETLKW") \ + EM( FUSE_ACCESS, "FUSE_ACCESS") \ + EM( FUSE_CREATE, "FUSE_CREATE") \ + EM( FUSE_INTERRUPT, "FUSE_INTERRUPT") \ + EM( FUSE_BMAP, "FUSE_BMAP") \ + EM( FUSE_DESTROY, "FUSE_DESTROY") \ + EM( FUSE_IOCTL, "FUSE_IOCTL") \ + EM( FUSE_POLL, "FUSE_POLL") \ + EM( FUSE_NOTIFY_REPLY, "FUSE_NOTIFY_REPLY") \ + EM( FUSE_BATCH_FORGET, "FUSE_BATCH_FORGET") \ + EM( FUSE_FALLOCATE, "FUSE_FALLOCATE") \ + EM( FUSE_READDIRPLUS, "FUSE_READDIRPLUS") \ + EM( FUSE_RENAME2, "FUSE_RENAME2") \ + EM( FUSE_LSEEK, "FUSE_LSEEK") \ + EM( FUSE_COPY_FILE_RANGE, "FUSE_COPY_FILE_RANGE") \ + EM( FUSE_SETUPMAPPING, "FUSE_SETUPMAPPING") \ + EM( FUSE_REMOVEMAPPING, "FUSE_REMOVEMAPPING") \ + EM( FUSE_SYNCFS, "FUSE_SYNCFS") \ + EM( FUSE_TMPFILE, "FUSE_TMPFILE") \ + EM( FUSE_STATX, "FUSE_STATX") \ + EMe(CUSE_INIT, "CUSE_INIT") + +/* + * This will turn the above table into TRACE_DEFINE_ENUM() for each of the + * entries. + */ +#undef EM +#undef EMe +#define EM(a, b) TRACE_DEFINE_ENUM(a); +#define EMe(a, b) TRACE_DEFINE_ENUM(a); + +OPCODES + +/* Now we redfine it with the table that __print_symbolic needs. */ +#undef EM +#undef EMe +#define EM(a, b) {a, b}, +#define EMe(a, b) {a, b} + +TRACE_EVENT(fuse_request_send, + TP_PROTO(const struct fuse_req *req), + + TP_ARGS(req), + + TP_STRUCT__entry( + __field(dev_t, connection) + __field(uint64_t, unique) + __field(enum fuse_opcode, opcode) + __field(uint32_t, len) + ), + + TP_fast_assign( + __entry->connection = req->fm->fc->dev; + __entry->unique = req->in.h.unique; + __entry->opcode = req->in.h.opcode; + __entry->len = req->in.h.len; + ), + + TP_printk("connection %u req %llu opcode %u (%s) len %u ", + __entry->connection, __entry->unique, __entry->opcode, + __print_symbolic(__entry->opcode, OPCODES), __entry->len) +); + +TRACE_EVENT(fuse_request_end, + TP_PROTO(const struct fuse_req *req), + + TP_ARGS(req), + + TP_STRUCT__entry( + __field(dev_t, connection) + __field(uint64_t, unique) + __field(uint32_t, len) + __field(int32_t, error) + ), + + TP_fast_assign( + __entry->connection = req->fm->fc->dev; + __entry->unique = req->in.h.unique; + __entry->len = req->out.h.len; + __entry->error = req->out.h.error; + ), + + TP_printk("connection %u req %llu len %u error %d", __entry->connection, + __entry->unique, __entry->len, __entry->error) +); + +#endif /* _TRACE_FUSE_H */ + +#undef TRACE_INCLUDE_PATH +#define TRACE_INCLUDE_PATH . +#define TRACE_INCLUDE_FILE fuse_trace +#include From 506b21c945b9716a1e092189c260d9400c52fa14 Mon Sep 17 00:00:00 2001 From: Aurelien Aptel Date: Fri, 17 May 2024 16:10:28 +0000 Subject: [PATCH 042/107] fuse: use correct name fuse_conn_list in docstring fuse_mount_list doesn't exist, use fuse_conn_list. Signed-off-by: Aurelien Aptel Reviewed-by: Bernd Schubert Signed-off-by: Miklos Szeredi --- fs/fuse/fuse_i.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 33b21255817e..bdbf9a8f3fc5 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -866,7 +866,7 @@ struct fuse_conn { /** Negotiated minor version */ unsigned minor; - /** Entry on the fuse_mount_list */ + /** Entry on the fuse_conn_list */ struct list_head entry; /** Device ID from the root super block */ From 2097154a10c6ee78be8796411e5d0ad81ee06ed6 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:12 +0200 Subject: [PATCH 043/107] namespace: introduce SB_I_NOIDMAP flag Right now we determine if filesystem support vfs idmappings or not basing on the FS_ALLOW_IDMAP flag presence. This "static" way works perfecly well for local filesystems like ext4, xfs, btrfs, etc. But for network-like filesystems like fuse, cephfs this approach is not ideal, because sometimes proper support of vfs idmaps requires some extensions for the on-wire protocol, which implies that changes have to be made not only in the Linux kernel code but also in the 3rd party components like libfuse, cephfs MDS server and so on. We have seen that issue during our work on cephfs idmapped mounts [1] with Christian, but right now I'm working on the idmapped mounts support for fuse/virtiofs and I think that it is a right time for this extension. [1] 5ccd8530dd7 ("ceph: handle idmapped mounts in create_request_message()") Suggested-by: Christian Brauner Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/namespace.c | 4 ++++ include/linux/fs.h | 1 + 2 files changed, 5 insertions(+) diff --git a/fs/namespace.c b/fs/namespace.c index 328087a4df8a..d1702285c915 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -4436,6 +4436,10 @@ static int can_idmap_mount(const struct mount_kattr *kattr, struct mount *mnt) if (!(m->mnt_sb->s_type->fs_flags & FS_ALLOW_IDMAP)) return -EINVAL; + /* The filesystem has turned off idmapped mounts. */ + if (m->mnt_sb->s_iflags & SB_I_NOIDMAP) + return -EINVAL; + /* We're not controlling the superblock. */ if (!ns_capable(fs_userns, CAP_SYS_ADMIN)) return -EPERM; diff --git a/include/linux/fs.h b/include/linux/fs.h index fd34b5755c0b..6ff547ef21f2 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1189,6 +1189,7 @@ extern int send_sigurg(struct fown_struct *fown); #define SB_I_TS_EXPIRY_WARNED 0x00000400 /* warned about timestamp range expiry */ #define SB_I_RETIRED 0x00000800 /* superblock shouldn't be reused */ #define SB_I_NOUMASK 0x00001000 /* VFS does not apply umask */ +#define SB_I_NOIDMAP 0x00002000 /* No idmapped mounts on this superblock */ /* Possible states of 'frozen' field */ enum { From aa16880d9f13c6490e80ad614402c8a6fe6f3efa Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:13 +0200 Subject: [PATCH 044/107] fuse: add basic infrastructure to support idmappings Add some preparational changes in fuse_get_req/fuse_force_creds to handle idmappings. Miklos suggested [1], [2] to change the meaning of in.h.uid/in.h.gid fields when daemon declares support for idmapped mounts. In a new semantic, we fill uid/gid values in fuse header with a id-mapped caller uid/gid (for requests which create new inodes), for all the rest cases we just send -1 to userspace. No functional changes intended. Link: https://lore.kernel.org/all/CAJfpegsVY97_5mHSc06mSw79FehFWtoXT=hhTUK_E-Yhr7OAuQ@mail.gmail.com/ [1] Link: https://lore.kernel.org/all/CAJfpegtHQsEUuFq1k4ZbTD3E1h-GsrN3PWyv7X8cg6sfU_W2Yw@mail.gmail.com/ [2] Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/dev.c | 50 +++++++++++++++++++++++++++++---------- fs/fuse/inode.c | 1 + include/uapi/linux/fuse.h | 2 ++ 3 files changed, 41 insertions(+), 12 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 1b6088288a16..b4092b3b4b5a 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -108,7 +108,9 @@ static void fuse_drop_waiting(struct fuse_conn *fc) static void fuse_put_request(struct fuse_req *req); -static struct fuse_req *fuse_get_req(struct fuse_mount *fm, bool for_background) +static struct fuse_req *fuse_get_req(struct mnt_idmap *idmap, + struct fuse_mount *fm, + bool for_background) { struct fuse_conn *fc = fm->fc; struct fuse_req *req; @@ -140,19 +142,37 @@ static struct fuse_req *fuse_get_req(struct fuse_mount *fm, bool for_background) goto out; } - req->in.h.uid = from_kuid(fc->user_ns, current_fsuid()); - req->in.h.gid = from_kgid(fc->user_ns, current_fsgid()); req->in.h.pid = pid_nr_ns(task_pid(current), fc->pid_ns); __set_bit(FR_WAITING, &req->flags); if (for_background) __set_bit(FR_BACKGROUND, &req->flags); - if (unlikely(req->in.h.uid == ((uid_t)-1) || - req->in.h.gid == ((gid_t)-1))) { - fuse_put_request(req); - return ERR_PTR(-EOVERFLOW); + if ((fm->sb->s_iflags & SB_I_NOIDMAP) || idmap) { + kuid_t idmapped_fsuid; + kgid_t idmapped_fsgid; + + /* + * Note, that when + * (fm->sb->s_iflags & SB_I_NOIDMAP) is true, then + * (idmap == &nop_mnt_idmap) is always true and therefore, + * mapped_fsuid(idmap, fc->user_ns) == current_fsuid(). + */ + idmapped_fsuid = idmap ? mapped_fsuid(idmap, fc->user_ns) : current_fsuid(); + idmapped_fsgid = idmap ? mapped_fsgid(idmap, fc->user_ns) : current_fsgid(); + req->in.h.uid = from_kuid(fc->user_ns, idmapped_fsuid); + req->in.h.gid = from_kgid(fc->user_ns, idmapped_fsgid); + + if (unlikely(req->in.h.uid == ((uid_t)-1) || + req->in.h.gid == ((gid_t)-1))) { + fuse_put_request(req); + return ERR_PTR(-EOVERFLOW); + } + } else { + req->in.h.uid = FUSE_INVALID_UIDGID; + req->in.h.gid = FUSE_INVALID_UIDGID; } + return req; out: @@ -497,8 +517,14 @@ static void fuse_force_creds(struct fuse_req *req) { struct fuse_conn *fc = req->fm->fc; - req->in.h.uid = from_kuid_munged(fc->user_ns, current_fsuid()); - req->in.h.gid = from_kgid_munged(fc->user_ns, current_fsgid()); + if (req->fm->sb->s_iflags & SB_I_NOIDMAP) { + req->in.h.uid = from_kuid_munged(fc->user_ns, current_fsuid()); + req->in.h.gid = from_kgid_munged(fc->user_ns, current_fsgid()); + } else { + req->in.h.uid = FUSE_INVALID_UIDGID; + req->in.h.gid = FUSE_INVALID_UIDGID; + } + req->in.h.pid = pid_nr_ns(task_pid(current), fc->pid_ns); } @@ -530,7 +556,7 @@ ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args) __set_bit(FR_FORCE, &req->flags); } else { WARN_ON(args->nocreds); - req = fuse_get_req(fm, false); + req = fuse_get_req(NULL, fm, false); if (IS_ERR(req)) return PTR_ERR(req); } @@ -591,7 +617,7 @@ int fuse_simple_background(struct fuse_mount *fm, struct fuse_args *args, __set_bit(FR_BACKGROUND, &req->flags); } else { WARN_ON(args->nocreds); - req = fuse_get_req(fm, true); + req = fuse_get_req(NULL, fm, true); if (IS_ERR(req)) return PTR_ERR(req); } @@ -613,7 +639,7 @@ static int fuse_simple_notify_reply(struct fuse_mount *fm, struct fuse_req *req; struct fuse_iqueue *fiq = &fm->fc->iq; - req = fuse_get_req(fm, false); + req = fuse_get_req(NULL, fm, false); if (IS_ERR(req)) return PTR_ERR(req); diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index bebd89002328..ed53e173337b 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -1572,6 +1572,7 @@ static void fuse_sb_defaults(struct super_block *sb) sb->s_time_gran = 1; sb->s_export_op = &fuse_export_operations; sb->s_iflags |= SB_I_IMA_UNVERIFIABLE_SIGNATURE; + sb->s_iflags |= SB_I_NOIDMAP; if (sb->s_user_ns != &init_user_ns) sb->s_iflags |= SB_I_UNTRUSTED_MOUNTER; sb->s_flags &= ~(SB_NOSEC | SB_I_VERSION); diff --git a/include/uapi/linux/fuse.h b/include/uapi/linux/fuse.h index d08b99d60f6f..2ccf38181df2 100644 --- a/include/uapi/linux/fuse.h +++ b/include/uapi/linux/fuse.h @@ -984,6 +984,8 @@ struct fuse_fallocate_in { */ #define FUSE_UNIQUE_RESEND (1ULL << 63) +#define FUSE_INVALID_UIDGID ((uint32_t)(-1)) + struct fuse_in_header { uint32_t len; uint32_t opcode; From 10dc721836c0c968990a519e147d4cb7decdae5c Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:14 +0200 Subject: [PATCH 045/107] fuse: add an idmap argument to fuse_simple_request If idmap == NULL *and* filesystem daemon declared idmapped mounts support, then uid/gid values in a fuse header will be -1. No functional changes intended. Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/dax.c | 4 ++-- fs/fuse/dev.c | 6 ++++-- fs/fuse/dir.c | 26 +++++++++++++------------- fs/fuse/file.c | 32 ++++++++++++++++---------------- fs/fuse/fuse_i.h | 3 ++- fs/fuse/inode.c | 6 +++--- fs/fuse/ioctl.c | 2 +- fs/fuse/readdir.c | 4 ++-- fs/fuse/xattr.c | 8 ++++---- 9 files changed, 47 insertions(+), 44 deletions(-) diff --git a/fs/fuse/dax.c b/fs/fuse/dax.c index 12ef91d170bb..6d8368d66dd4 100644 --- a/fs/fuse/dax.c +++ b/fs/fuse/dax.c @@ -207,7 +207,7 @@ static int fuse_setup_one_mapping(struct inode *inode, unsigned long start_idx, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err < 0) return err; dmap->writable = writable; @@ -245,7 +245,7 @@ static int fuse_send_removemapping(struct inode *inode, args.in_args[0].value = inargp; args.in_args[1].size = inargp->count * sizeof(*remove_one); args.in_args[1].value = remove_one; - return fuse_simple_request(fm, &args); + return fuse_simple_request(NULL, fm, &args); } static int dmap_removemapping_list(struct inode *inode, unsigned int num, diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index b4092b3b4b5a..5ca2ab48527e 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -539,7 +539,9 @@ static void fuse_args_to_req(struct fuse_req *req, struct fuse_args *args) __set_bit(FR_ASYNC, &req->flags); } -ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args) +ssize_t fuse_simple_request(struct mnt_idmap *idmap, + struct fuse_mount *fm, + struct fuse_args *args) { struct fuse_conn *fc = fm->fc; struct fuse_req *req; @@ -556,7 +558,7 @@ ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args) __set_bit(FR_FORCE, &req->flags); } else { WARN_ON(args->nocreds); - req = fuse_get_req(NULL, fm, false); + req = fuse_get_req(idmap, fm, false); if (IS_ERR(req)) return PTR_ERR(req); } diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 8e96df9fd76c..6ce7968365e7 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -230,7 +230,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags) parent = dget_parent(entry); fuse_lookup_init(fm->fc, &args, get_node_id(d_inode(parent)), &entry->d_name, &outarg); - ret = fuse_simple_request(fm, &args); + ret = fuse_simple_request(NULL, fm, &args); dput(parent); /* Zero nodeid is same as -ENOENT */ if (!ret && !outarg.nodeid) @@ -383,7 +383,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, const struct qstr *name attr_version = fuse_get_attr_version(fm->fc); fuse_lookup_init(fm->fc, &args, nodeid, name, outarg); - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); /* Zero nodeid is same as -ENOENT, but with valid timeout */ if (err || !outarg->nodeid) goto out_put_forget; @@ -672,7 +672,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, if (err) goto out_free_ff; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); free_ext_value(&args); if (err) goto out_free_ff; @@ -803,7 +803,7 @@ static int create_new_entry(struct fuse_mount *fm, struct fuse_args *args, goto out_put_forget_req; } - err = fuse_simple_request(fm, args); + err = fuse_simple_request(NULL, fm, args); free_ext_value(args); if (err) goto out_put_forget_req; @@ -987,7 +987,7 @@ static int fuse_unlink(struct inode *dir, struct dentry *entry) args.in_numargs = 1; args.in_args[0].size = entry->d_name.len + 1; args.in_args[0].value = entry->d_name.name; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) { fuse_dir_changed(dir); fuse_entry_unlinked(entry); @@ -1010,7 +1010,7 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry) args.in_numargs = 1; args.in_args[0].size = entry->d_name.len + 1; args.in_args[0].value = entry->d_name.name; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) { fuse_dir_changed(dir); fuse_entry_unlinked(entry); @@ -1040,7 +1040,7 @@ static int fuse_rename_common(struct inode *olddir, struct dentry *oldent, args.in_args[1].value = oldent->d_name.name; args.in_args[2].size = newent->d_name.len + 1; args.in_args[2].value = newent->d_name.name; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) { /* ctime changes */ fuse_update_ctime(d_inode(oldent)); @@ -1210,7 +1210,7 @@ static int fuse_do_statx(struct inode *inode, struct file *file, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err) return err; @@ -1268,7 +1268,7 @@ static int fuse_do_getattr(struct inode *inode, struct kstat *stat, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) { if (fuse_invalid_attr(&outarg.attr) || inode_wrong_type(inode, outarg.attr.mode)) { @@ -1472,7 +1472,7 @@ static int fuse_access(struct inode *inode, int mask) args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fm->fc->no_access = 1; err = 0; @@ -1584,7 +1584,7 @@ static int fuse_readlink_page(struct inode *inode, struct page *page) ap.args.page_zeroing = true; ap.args.out_numargs = 1; ap.args.out_args[0].size = desc.length; - res = fuse_simple_request(fm, &ap.args); + res = fuse_simple_request(NULL, fm, &ap.args); fuse_invalidate_atime(inode); @@ -1857,7 +1857,7 @@ int fuse_flush_times(struct inode *inode, struct fuse_file *ff) } fuse_setattr_fill(fm->fc, &args, inode, &inarg, &outarg); - return fuse_simple_request(fm, &args); + return fuse_simple_request(NULL, fm, &args); } /* @@ -1970,7 +1970,7 @@ int fuse_do_setattr(struct dentry *dentry, struct iattr *attr, inarg.valid |= FATTR_KILL_SUIDGID; } fuse_setattr_fill(fc, &args, inode, &inarg, &outarg); - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err) { if (err == -EINTR) fuse_invalidate_attr(inode); diff --git a/fs/fuse/file.c b/fs/fuse/file.c index fc587f3bc56f..3c18e90fff07 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -48,7 +48,7 @@ static int fuse_send_open(struct fuse_mount *fm, u64 nodeid, args.out_args[0].size = sizeof(*outargp); args.out_args[0].value = outargp; - return fuse_simple_request(fm, &args); + return fuse_simple_request(NULL, fm, &args); } struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release) @@ -111,7 +111,7 @@ static void fuse_file_put(struct fuse_file *ff, bool sync) if (!args) { /* Do nothing when server does not implement 'open' */ } else if (sync) { - fuse_simple_request(ff->fm, args); + fuse_simple_request(NULL, ff->fm, args); fuse_release_end(ff->fm, args, 0); } else { args->end = fuse_release_end; @@ -539,7 +539,7 @@ static int fuse_flush(struct file *file, fl_owner_t id) args.in_args[0].value = &inarg; args.force = true; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fm->fc->no_flush = 1; err = 0; @@ -572,7 +572,7 @@ int fuse_fsync_common(struct file *file, loff_t start, loff_t end, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - return fuse_simple_request(fm, &args); + return fuse_simple_request(NULL, fm, &args); } static int fuse_fsync(struct file *file, loff_t start, loff_t end, @@ -814,7 +814,7 @@ static ssize_t fuse_send_read(struct fuse_io_args *ia, loff_t pos, size_t count, if (ia->io->async) return fuse_async_req_send(fm, ia, count); - return fuse_simple_request(fm, &ia->ap.args); + return fuse_simple_request(NULL, fm, &ia->ap.args); } static void fuse_read_update_size(struct inode *inode, loff_t size, @@ -878,7 +878,7 @@ static int fuse_do_readpage(struct file *file, struct page *page) desc.length--; fuse_read_args_fill(&ia, file, pos, desc.length, FUSE_READ); - res = fuse_simple_request(fm, &ia.ap.args); + res = fuse_simple_request(NULL, fm, &ia.ap.args); if (res < 0) return res; /* @@ -976,7 +976,7 @@ static void fuse_send_readpages(struct fuse_io_args *ia, struct file *file) if (!err) return; } else { - res = fuse_simple_request(fm, &ap->args); + res = fuse_simple_request(NULL, fm, &ap->args); err = res < 0 ? res : 0; } fuse_readpages_end(fm, &ap->args, err); @@ -1101,7 +1101,7 @@ static ssize_t fuse_send_write(struct fuse_io_args *ia, loff_t pos, if (ia->io->async) return fuse_async_req_send(fm, ia, count); - err = fuse_simple_request(fm, &ia->ap.args); + err = fuse_simple_request(NULL, fm, &ia->ap.args); if (!err && ia->write.out.size > count) err = -EIO; @@ -1147,7 +1147,7 @@ static ssize_t fuse_send_write_pages(struct fuse_io_args *ia, if (fm->fc->handle_killpriv_v2 && !capable(CAP_FSETID)) ia->write.in.write_flags |= FUSE_WRITE_KILL_SUIDGID; - err = fuse_simple_request(fm, &ap->args); + err = fuse_simple_request(NULL, fm, &ap->args); if (!err && ia->write.out.size > count) err = -EIO; @@ -2661,7 +2661,7 @@ static int fuse_getlk(struct file *file, struct file_lock *fl) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) err = convert_fuse_file_lock(fm->fc, &outarg.lk, fl); @@ -2685,7 +2685,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) } fuse_lk_fill(&args, file, fl, opcode, pid_nr, flock, &inarg); - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); /* locking is restartable */ if (err == -EINTR) @@ -2759,7 +2759,7 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) fm->fc->no_bmap = 1; @@ -2791,7 +2791,7 @@ static loff_t fuse_lseek(struct file *file, loff_t offset, int whence) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err) { if (err == -ENOSYS) { fm->fc->no_lseek = 1; @@ -2924,7 +2924,7 @@ __poll_t fuse_file_poll(struct file *file, poll_table *wait) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) return demangle_poll(outarg.revents); @@ -3146,7 +3146,7 @@ static long fuse_file_fallocate(struct file *file, int mode, loff_t offset, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fm->fc->no_fallocate = 1; err = -EOPNOTSUPP; @@ -3258,7 +3258,7 @@ static ssize_t __fuse_copy_file_range(struct file *file_in, loff_t pos_in, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fc->no_copy_file_range = 1; err = -EOPNOTSUPP; diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index bdbf9a8f3fc5..407ee2542ff5 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -1147,7 +1147,8 @@ void __exit fuse_ctl_cleanup(void); /** * Simple request sending that does request allocation and freeing */ -ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args); +ssize_t fuse_simple_request(struct mnt_idmap *idmap, struct fuse_mount *fm, + struct fuse_args *args); int fuse_simple_background(struct fuse_mount *fm, struct fuse_args *args, gfp_t gfp_flags); diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index ed53e173337b..b08bc6081612 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -586,7 +586,7 @@ static void fuse_send_destroy(struct fuse_mount *fm) args.opcode = FUSE_DESTROY; args.force = true; args.nocreds = true; - fuse_simple_request(fm, &args); + fuse_simple_request(NULL, fm, &args); } } @@ -624,7 +624,7 @@ static int fuse_statfs(struct dentry *dentry, struct kstatfs *buf) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (!err) convert_fuse_statfs(buf, &outarg.st); return err; @@ -713,7 +713,7 @@ static int fuse_sync_fs(struct super_block *sb, int wait) args.nodeid = get_node_id(sb->s_root->d_inode); args.out_numargs = 0; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fc->sync_fs = 0; err = 0; diff --git a/fs/fuse/ioctl.c b/fs/fuse/ioctl.c index 572ce8a82ceb..b40dd931167d 100644 --- a/fs/fuse/ioctl.c +++ b/fs/fuse/ioctl.c @@ -18,7 +18,7 @@ static ssize_t fuse_send_ioctl(struct fuse_mount *fm, struct fuse_args *args, args->out_args[0].size = sizeof(*outarg); args->out_args[0].value = outarg; - ret = fuse_simple_request(fm, args); + ret = fuse_simple_request(NULL, fm, args); /* Translate ENOSYS, which shouldn't be returned from fs */ if (ret == -ENOSYS) diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c index 0377b6dc24c8..e8a093289421 100644 --- a/fs/fuse/readdir.c +++ b/fs/fuse/readdir.c @@ -279,7 +279,7 @@ static void fuse_force_forget(struct file *file, u64 nodeid) args.force = true; args.noreply = true; - fuse_simple_request(fm, &args); + fuse_simple_request(NULL, fm, &args); /* ignore errors */ } @@ -358,7 +358,7 @@ static int fuse_readdir_uncached(struct file *file, struct dir_context *ctx) FUSE_READDIR); } locked = fuse_lock_inode(inode); - res = fuse_simple_request(fm, &ap->args); + res = fuse_simple_request(NULL, fm, &ap->args); fuse_unlock_inode(inode, locked); if (res >= 0) { if (!res) { diff --git a/fs/fuse/xattr.c b/fs/fuse/xattr.c index 9f568d345c51..0a9b60de3668 100644 --- a/fs/fuse/xattr.c +++ b/fs/fuse/xattr.c @@ -37,7 +37,7 @@ int fuse_setxattr(struct inode *inode, const char *name, const void *value, args.in_args[1].value = name; args.in_args[2].size = size; args.in_args[2].value = value; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fm->fc->no_setxattr = 1; err = -EOPNOTSUPP; @@ -79,7 +79,7 @@ ssize_t fuse_getxattr(struct inode *inode, const char *name, void *value, args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; } - ret = fuse_simple_request(fm, &args); + ret = fuse_simple_request(NULL, fm, &args); if (!ret && !size) ret = min_t(size_t, outarg.size, XATTR_SIZE_MAX); if (ret == -ENOSYS) { @@ -141,7 +141,7 @@ ssize_t fuse_listxattr(struct dentry *entry, char *list, size_t size) args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; } - ret = fuse_simple_request(fm, &args); + ret = fuse_simple_request(NULL, fm, &args); if (!ret && !size) ret = min_t(size_t, outarg.size, XATTR_LIST_MAX); if (ret > 0 && size) @@ -167,7 +167,7 @@ int fuse_removexattr(struct inode *inode, const char *name) args.in_numargs = 1; args.in_args[0].size = strlen(name) + 1; args.in_args[0].value = name; - err = fuse_simple_request(fm, &args); + err = fuse_simple_request(NULL, fm, &args); if (err == -ENOSYS) { fm->fc->no_removexattr = 1; err = -EOPNOTSUPP; From d561254fb7ba451f580becb21ab3472bc0265080 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:15 +0200 Subject: [PATCH 046/107] fuse: support idmapped FUSE_EXT_GROUPS We don't need to remap parent_gid, but have to adjust group membership checks and take idmapping into account. Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 19 ++++++++++++------- 1 file changed, 12 insertions(+), 7 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 6ce7968365e7..8a936dc0072b 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -545,17 +545,21 @@ static u32 fuse_ext_size(size_t size) /* * This adds just a single supplementary group that matches the parent's group. */ -static int get_create_supp_group(struct inode *dir, struct fuse_in_arg *ext) +static int get_create_supp_group(struct mnt_idmap *idmap, + struct inode *dir, + struct fuse_in_arg *ext) { struct fuse_conn *fc = get_fuse_conn(dir); struct fuse_ext_header *xh; struct fuse_supp_groups *sg; kgid_t kgid = dir->i_gid; + vfsgid_t vfsgid = make_vfsgid(idmap, fc->user_ns, kgid); gid_t parent_gid = from_kgid(fc->user_ns, kgid); + u32 sg_len = fuse_ext_size(sizeof(*sg) + sizeof(sg->groups[0])); - if (parent_gid == (gid_t) -1 || gid_eq(kgid, current_fsgid()) || - !in_group_p(kgid)) + if (parent_gid == (gid_t) -1 || vfsgid_eq_kgid(vfsgid, current_fsgid()) || + !vfsgid_in_group_p(vfsgid)) return 0; xh = extend_arg(ext, sg_len); @@ -572,7 +576,8 @@ static int get_create_supp_group(struct inode *dir, struct fuse_in_arg *ext) return 0; } -static int get_create_ext(struct fuse_args *args, +static int get_create_ext(struct mnt_idmap *idmap, + struct fuse_args *args, struct inode *dir, struct dentry *dentry, umode_t mode) { @@ -583,7 +588,7 @@ static int get_create_ext(struct fuse_args *args, if (fc->init_security) err = get_security_context(dentry, mode, &ext); if (!err && fc->create_supp_group) - err = get_create_supp_group(dir, &ext); + err = get_create_supp_group(idmap, dir, &ext); if (!err && ext.size) { WARN_ON(args->in_numargs >= ARRAY_SIZE(args->in_args)); @@ -668,7 +673,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, args.out_args[1].size = sizeof(*outopenp); args.out_args[1].value = outopenp; - err = get_create_ext(&args, dir, entry, mode); + err = get_create_ext(&nop_mnt_idmap, &args, dir, entry, mode); if (err) goto out_free_ff; @@ -798,7 +803,7 @@ static int create_new_entry(struct fuse_mount *fm, struct fuse_args *args, args->out_args[0].value = &outarg; if (args->opcode != FUSE_LINK) { - err = get_create_ext(args, dir, entry, mode); + err = get_create_ext(&nop_mnt_idmap, args, dir, entry, mode); if (err) goto out_put_forget_req; } From 556208e139e19e0b308e104c6b0e42a9e0a54c3a Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:16 +0200 Subject: [PATCH 047/107] fuse: support idmap for mkdir/mknod/symlink/create/tmpfile We have all the infrastructure in place, we just need to pass an idmapping here. Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 38 ++++++++++++++++++++------------------ 1 file changed, 20 insertions(+), 18 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 8a936dc0072b..9da8feede181 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -614,9 +614,9 @@ static void free_ext_value(struct fuse_args *args) * If the filesystem doesn't support this, then fall back to separate * 'mknod' + 'open' requests. */ -static int fuse_create_open(struct inode *dir, struct dentry *entry, - struct file *file, unsigned int flags, - umode_t mode, u32 opcode) +static int fuse_create_open(struct mnt_idmap *idmap, struct inode *dir, + struct dentry *entry, struct file *file, + unsigned int flags, umode_t mode, u32 opcode) { int err; struct inode *inode; @@ -673,11 +673,11 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, args.out_args[1].size = sizeof(*outopenp); args.out_args[1].value = outopenp; - err = get_create_ext(&nop_mnt_idmap, &args, dir, entry, mode); + err = get_create_ext(idmap, &args, dir, entry, mode); if (err) goto out_free_ff; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(idmap, fm, &args); free_ext_value(&args); if (err) goto out_free_ff; @@ -734,6 +734,7 @@ static int fuse_atomic_open(struct inode *dir, struct dentry *entry, umode_t mode) { int err; + struct mnt_idmap *idmap = file_mnt_idmap(file); struct fuse_conn *fc = get_fuse_conn(dir); struct dentry *res = NULL; @@ -758,7 +759,7 @@ static int fuse_atomic_open(struct inode *dir, struct dentry *entry, if (fc->no_create) goto mknod; - err = fuse_create_open(dir, entry, file, flags, mode, FUSE_CREATE); + err = fuse_create_open(idmap, dir, entry, file, flags, mode, FUSE_CREATE); if (err == -ENOSYS) { fc->no_create = 1; goto mknod; @@ -769,7 +770,7 @@ out_dput: return err; mknod: - err = fuse_mknod(&nop_mnt_idmap, dir, entry, mode, 0); + err = fuse_mknod(idmap, dir, entry, mode, 0); if (err) goto out_dput; no_open: @@ -779,9 +780,9 @@ no_open: /* * Code shared between mknod, mkdir, symlink and link */ -static int create_new_entry(struct fuse_mount *fm, struct fuse_args *args, - struct inode *dir, struct dentry *entry, - umode_t mode) +static int create_new_entry(struct mnt_idmap *idmap, struct fuse_mount *fm, + struct fuse_args *args, struct inode *dir, + struct dentry *entry, umode_t mode) { struct fuse_entry_out outarg; struct inode *inode; @@ -803,12 +804,12 @@ static int create_new_entry(struct fuse_mount *fm, struct fuse_args *args, args->out_args[0].value = &outarg; if (args->opcode != FUSE_LINK) { - err = get_create_ext(&nop_mnt_idmap, args, dir, entry, mode); + err = get_create_ext(idmap, args, dir, entry, mode); if (err) goto out_put_forget_req; } - err = fuse_simple_request(NULL, fm, args); + err = fuse_simple_request(idmap, fm, args); free_ext_value(args); if (err) goto out_put_forget_req; @@ -869,13 +870,13 @@ static int fuse_mknod(struct mnt_idmap *idmap, struct inode *dir, args.in_args[0].value = &inarg; args.in_args[1].size = entry->d_name.len + 1; args.in_args[1].value = entry->d_name.name; - return create_new_entry(fm, &args, dir, entry, mode); + return create_new_entry(idmap, fm, &args, dir, entry, mode); } static int fuse_create(struct mnt_idmap *idmap, struct inode *dir, struct dentry *entry, umode_t mode, bool excl) { - return fuse_mknod(&nop_mnt_idmap, dir, entry, mode, 0); + return fuse_mknod(idmap, dir, entry, mode, 0); } static int fuse_tmpfile(struct mnt_idmap *idmap, struct inode *dir, @@ -887,7 +888,8 @@ static int fuse_tmpfile(struct mnt_idmap *idmap, struct inode *dir, if (fc->no_tmpfile) return -EOPNOTSUPP; - err = fuse_create_open(dir, file->f_path.dentry, file, file->f_flags, mode, FUSE_TMPFILE); + err = fuse_create_open(idmap, dir, file->f_path.dentry, file, + file->f_flags, mode, FUSE_TMPFILE); if (err == -ENOSYS) { fc->no_tmpfile = 1; err = -EOPNOTSUPP; @@ -914,7 +916,7 @@ static int fuse_mkdir(struct mnt_idmap *idmap, struct inode *dir, args.in_args[0].value = &inarg; args.in_args[1].size = entry->d_name.len + 1; args.in_args[1].value = entry->d_name.name; - return create_new_entry(fm, &args, dir, entry, S_IFDIR); + return create_new_entry(idmap, fm, &args, dir, entry, S_IFDIR); } static int fuse_symlink(struct mnt_idmap *idmap, struct inode *dir, @@ -930,7 +932,7 @@ static int fuse_symlink(struct mnt_idmap *idmap, struct inode *dir, args.in_args[0].value = entry->d_name.name; args.in_args[1].size = len; args.in_args[1].value = link; - return create_new_entry(fm, &args, dir, entry, S_IFLNK); + return create_new_entry(idmap, fm, &args, dir, entry, S_IFLNK); } void fuse_flush_time_update(struct inode *inode) @@ -1124,7 +1126,7 @@ static int fuse_link(struct dentry *entry, struct inode *newdir, args.in_args[0].value = &inarg; args.in_args[1].size = newent->d_name.len + 1; args.in_args[1].value = newent->d_name.name; - err = create_new_entry(fm, &args, newdir, newent, inode->i_mode); + err = create_new_entry(NULL, fm, &args, newdir, newent, inode->i_mode); if (!err) fuse_update_ctime_in_cache(inode); else if (err == -EINTR) From 2a8c810d5e492906fe0af94a6e8b2398714e80e0 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:17 +0200 Subject: [PATCH 048/107] fuse: support idmapped getattr inode op We have to: - pass an idmapping to the generic_fillattr() to properly handle UIG/GID mapping for the userspace. - pass -/- to fuse_fillattr() (analog of generic_fillattr() in fuse). Difference between these two is that generic_fillattr() takes all the stat() data from the inode directly, while fuse_fillattr() codepath takes a fresh data just from the userspace reply on the FUSE_GETATTR request. In some cases we can just pass &nop_mnt_idmap, because idmapping won't be used in these codepaths. For example, when 3rd argument of fuse_do_getattr() is NULL then idmap argument is not used. Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 44 ++++++++++++++++++++++++-------------------- 1 file changed, 24 insertions(+), 20 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 9da8feede181..2bc3ae40619a 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1135,18 +1135,22 @@ static int fuse_link(struct dentry *entry, struct inode *newdir, return err; } -static void fuse_fillattr(struct inode *inode, struct fuse_attr *attr, - struct kstat *stat) +static void fuse_fillattr(struct mnt_idmap *idmap, struct inode *inode, + struct fuse_attr *attr, struct kstat *stat) { unsigned int blkbits; struct fuse_conn *fc = get_fuse_conn(inode); + vfsuid_t vfsuid = make_vfsuid(idmap, fc->user_ns, + make_kuid(fc->user_ns, attr->uid)); + vfsgid_t vfsgid = make_vfsgid(idmap, fc->user_ns, + make_kgid(fc->user_ns, attr->gid)); stat->dev = inode->i_sb->s_dev; stat->ino = attr->ino; stat->mode = (inode->i_mode & S_IFMT) | (attr->mode & 07777); stat->nlink = attr->nlink; - stat->uid = make_kuid(fc->user_ns, attr->uid); - stat->gid = make_kgid(fc->user_ns, attr->gid); + stat->uid = vfsuid_into_kuid(vfsuid); + stat->gid = vfsgid_into_kgid(vfsgid); stat->rdev = inode->i_rdev; stat->atime.tv_sec = attr->atime; stat->atime.tv_nsec = attr->atimensec; @@ -1185,8 +1189,8 @@ static void fuse_statx_to_attr(struct fuse_statx *sx, struct fuse_attr *attr) attr->blksize = sx->blksize; } -static int fuse_do_statx(struct inode *inode, struct file *file, - struct kstat *stat) +static int fuse_do_statx(struct mnt_idmap *idmap, struct inode *inode, + struct file *file, struct kstat *stat) { int err; struct fuse_attr attr; @@ -1239,15 +1243,15 @@ static int fuse_do_statx(struct inode *inode, struct file *file, stat->result_mask = sx->mask & (STATX_BASIC_STATS | STATX_BTIME); stat->btime.tv_sec = sx->btime.tv_sec; stat->btime.tv_nsec = min_t(u32, sx->btime.tv_nsec, NSEC_PER_SEC - 1); - fuse_fillattr(inode, &attr, stat); + fuse_fillattr(idmap, inode, &attr, stat); stat->result_mask |= STATX_TYPE; } return 0; } -static int fuse_do_getattr(struct inode *inode, struct kstat *stat, - struct file *file) +static int fuse_do_getattr(struct mnt_idmap *idmap, struct inode *inode, + struct kstat *stat, struct file *file) { int err; struct fuse_getattr_in inarg; @@ -1286,15 +1290,15 @@ static int fuse_do_getattr(struct inode *inode, struct kstat *stat, ATTR_TIMEOUT(&outarg), attr_version); if (stat) - fuse_fillattr(inode, &outarg.attr, stat); + fuse_fillattr(idmap, inode, &outarg.attr, stat); } } return err; } -static int fuse_update_get_attr(struct inode *inode, struct file *file, - struct kstat *stat, u32 request_mask, - unsigned int flags) +static int fuse_update_get_attr(struct mnt_idmap *idmap, struct inode *inode, + struct file *file, struct kstat *stat, + u32 request_mask, unsigned int flags) { struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_conn *fc = get_fuse_conn(inode); @@ -1325,17 +1329,17 @@ retry: forget_all_cached_acls(inode); /* Try statx if BTIME is requested */ if (!fc->no_statx && (request_mask & ~STATX_BASIC_STATS)) { - err = fuse_do_statx(inode, file, stat); + err = fuse_do_statx(idmap, inode, file, stat); if (err == -ENOSYS) { fc->no_statx = 1; err = 0; goto retry; } } else { - err = fuse_do_getattr(inode, stat, file); + err = fuse_do_getattr(idmap, inode, stat, file); } } else if (stat) { - generic_fillattr(&nop_mnt_idmap, request_mask, inode, stat); + generic_fillattr(idmap, request_mask, inode, stat); stat->mode = fi->orig_i_mode; stat->ino = fi->orig_ino; if (test_bit(FUSE_I_BTIME, &fi->state)) { @@ -1349,7 +1353,7 @@ retry: int fuse_update_attributes(struct inode *inode, struct file *file, u32 mask) { - return fuse_update_get_attr(inode, file, NULL, mask, 0); + return fuse_update_get_attr(&nop_mnt_idmap, inode, file, NULL, mask, 0); } int fuse_reverse_inval_entry(struct fuse_conn *fc, u64 parent_nodeid, @@ -1493,7 +1497,7 @@ static int fuse_perm_getattr(struct inode *inode, int mask) return -ECHILD; forget_all_cached_acls(inode); - return fuse_do_getattr(inode, NULL, NULL); + return fuse_do_getattr(&nop_mnt_idmap, inode, NULL, NULL); } /* @@ -2072,7 +2076,7 @@ static int fuse_setattr(struct mnt_idmap *idmap, struct dentry *entry, * ia_mode calculation may have used stale i_mode. * Refresh and recalculate. */ - ret = fuse_do_getattr(inode, NULL, file); + ret = fuse_do_getattr(idmap, inode, NULL, file); if (ret) return ret; @@ -2129,7 +2133,7 @@ static int fuse_getattr(struct mnt_idmap *idmap, return -EACCES; } - return fuse_update_get_attr(inode, NULL, stat, request_mask, flags); + return fuse_update_get_attr(idmap, inode, NULL, stat, request_mask, flags); } static const struct inode_operations fuse_dir_inode_operations = { From c1d82215d391ded456e7bdd1c0893c9cb6322272 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:18 +0200 Subject: [PATCH 049/107] fuse: support idmapped ->permission inode op We only cover the case when "default_permissions" flag is used. A reason for that is that otherwise all the permission checks are done in the userspace and we have to deal with VFS idmapping in the userspace (which is bad), alternatively we have to provide the userspace with idmapped req->in.h.uid/req->in.h.gid which is also not align with VFS idmaps philosophy. Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 2bc3ae40619a..70ed1c47d7ce 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1545,7 +1545,7 @@ static int fuse_permission(struct mnt_idmap *idmap, } if (fc->default_permissions) { - err = generic_permission(&nop_mnt_idmap, inode, mask); + err = generic_permission(idmap, inode, mask); /* If permission is denied, try to refresh file attributes. This is also needed, because the root @@ -1553,7 +1553,7 @@ static int fuse_permission(struct mnt_idmap *idmap, if (err == -EACCES && !refreshed) { err = fuse_perm_getattr(inode, mask); if (!err) - err = generic_permission(&nop_mnt_idmap, + err = generic_permission(idmap, inode, mask); } From 276a02569920f793047479fe7ef5da2c3e285a80 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:19 +0200 Subject: [PATCH 050/107] fuse: support idmapped ->setattr op Need to translate uid and gid in case of chown(2). Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 34 +++++++++++++++++++++++----------- fs/fuse/file.c | 2 +- fs/fuse/fuse_i.h | 4 ++-- 3 files changed, 26 insertions(+), 14 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 70ed1c47d7ce..5fa9c99f4d70 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1749,17 +1749,29 @@ static bool update_mtime(unsigned ivalid, bool trust_local_mtime) return true; } -static void iattr_to_fattr(struct fuse_conn *fc, struct iattr *iattr, - struct fuse_setattr_in *arg, bool trust_local_cmtime) +static void iattr_to_fattr(struct mnt_idmap *idmap, struct fuse_conn *fc, + struct iattr *iattr, struct fuse_setattr_in *arg, + bool trust_local_cmtime) { unsigned ivalid = iattr->ia_valid; if (ivalid & ATTR_MODE) arg->valid |= FATTR_MODE, arg->mode = iattr->ia_mode; - if (ivalid & ATTR_UID) - arg->valid |= FATTR_UID, arg->uid = from_kuid(fc->user_ns, iattr->ia_uid); - if (ivalid & ATTR_GID) - arg->valid |= FATTR_GID, arg->gid = from_kgid(fc->user_ns, iattr->ia_gid); + + if (ivalid & ATTR_UID) { + kuid_t fsuid = from_vfsuid(idmap, fc->user_ns, iattr->ia_vfsuid); + + arg->valid |= FATTR_UID; + arg->uid = from_kuid(fc->user_ns, fsuid); + } + + if (ivalid & ATTR_GID) { + kgid_t fsgid = from_vfsgid(idmap, fc->user_ns, iattr->ia_vfsgid); + + arg->valid |= FATTR_GID; + arg->gid = from_kgid(fc->user_ns, fsgid); + } + if (ivalid & ATTR_SIZE) arg->valid |= FATTR_SIZE, arg->size = iattr->ia_size; if (ivalid & ATTR_ATIME) { @@ -1879,8 +1891,8 @@ int fuse_flush_times(struct inode *inode, struct fuse_file *ff) * vmtruncate() doesn't allow for this case, so do the rlimit checking * and the actual truncation by hand. */ -int fuse_do_setattr(struct dentry *dentry, struct iattr *attr, - struct file *file) +int fuse_do_setattr(struct mnt_idmap *idmap, struct dentry *dentry, + struct iattr *attr, struct file *file) { struct inode *inode = d_inode(dentry); struct fuse_mount *fm = get_fuse_mount(inode); @@ -1900,7 +1912,7 @@ int fuse_do_setattr(struct dentry *dentry, struct iattr *attr, if (!fc->default_permissions) attr->ia_valid |= ATTR_FORCE; - err = setattr_prepare(&nop_mnt_idmap, dentry, attr); + err = setattr_prepare(idmap, dentry, attr); if (err) return err; @@ -1959,7 +1971,7 @@ int fuse_do_setattr(struct dentry *dentry, struct iattr *attr, memset(&inarg, 0, sizeof(inarg)); memset(&outarg, 0, sizeof(outarg)); - iattr_to_fattr(fc, attr, &inarg, trust_local_cmtime); + iattr_to_fattr(idmap, fc, attr, &inarg, trust_local_cmtime); if (file) { struct fuse_file *ff = file->private_data; inarg.valid |= FATTR_FH; @@ -2094,7 +2106,7 @@ static int fuse_setattr(struct mnt_idmap *idmap, struct dentry *entry, if (!attr->ia_valid) return 0; - ret = fuse_do_setattr(entry, attr, file); + ret = fuse_do_setattr(idmap, entry, attr, file); if (!ret) { /* * If filesystem supports acls it may have updated acl xattrs in diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 3c18e90fff07..0c33ddf97864 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -2971,7 +2971,7 @@ static void fuse_do_truncate(struct file *file) attr.ia_file = file; attr.ia_valid |= ATTR_FILE; - fuse_do_setattr(file_dentry(file), &attr, file); + fuse_do_setattr(file_mnt_idmap(file), file_dentry(file), &attr, file); } static inline loff_t fuse_round_up(struct fuse_conn *fc, loff_t off) diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 407ee2542ff5..7dd0661aeab6 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -1324,8 +1324,8 @@ bool fuse_write_update_attr(struct inode *inode, loff_t pos, ssize_t written); int fuse_flush_times(struct inode *inode, struct fuse_file *ff); int fuse_write_inode(struct inode *inode, struct writeback_control *wbc); -int fuse_do_setattr(struct dentry *dentry, struct iattr *attr, - struct file *file); +int fuse_do_setattr(struct mnt_idmap *idmap, struct dentry *dentry, + struct iattr *attr, struct file *file); void fuse_set_initialized(struct fuse_conn *fc); From 4d833befa20253049e4f48ece9cb82f12800405d Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:20 +0200 Subject: [PATCH 051/107] fuse: drop idmap argument from __fuse_get_acl We don't need to have idmap in the __fuse_get_acl as we don't have any use for it. In the current POSIX ACL implementation, idmapped mounts are taken into account on the userspace/kernel border (see vfs_set_acl_idmapped_mnt() and vfs_posix_acl_to_xattr()). Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/acl.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/fs/fuse/acl.c b/fs/fuse/acl.c index 04cfd8fee992..897d813c5e92 100644 --- a/fs/fuse/acl.c +++ b/fs/fuse/acl.c @@ -12,7 +12,6 @@ #include static struct posix_acl *__fuse_get_acl(struct fuse_conn *fc, - struct mnt_idmap *idmap, struct inode *inode, int type, bool rcu) { int size; @@ -74,7 +73,7 @@ struct posix_acl *fuse_get_acl(struct mnt_idmap *idmap, if (fuse_no_acl(fc, inode)) return ERR_PTR(-EOPNOTSUPP); - return __fuse_get_acl(fc, idmap, inode, type, false); + return __fuse_get_acl(fc, inode, type, false); } struct posix_acl *fuse_get_inode_acl(struct inode *inode, int type, bool rcu) @@ -90,8 +89,7 @@ struct posix_acl *fuse_get_inode_acl(struct inode *inode, int type, bool rcu) */ if (!fc->posix_acl) return NULL; - - return __fuse_get_acl(fc, &nop_mnt_idmap, inode, type, rcu); + return __fuse_get_acl(fc, inode, type, rcu); } int fuse_set_acl(struct mnt_idmap *idmap, struct dentry *dentry, From d395d0a5d2544b639cd7b0055e4de85b0efc2762 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:21 +0200 Subject: [PATCH 052/107] fuse: support idmapped ->set_acl It's just a matter of adjusting a permission check condition for S_ISGID flag. All the rest is already handled in the generic VFS code. Notice that this permission check is the analog of what we have in posix_acl_update_mode() generic helper, but fuse doesn't use this helper as on the kernel side we don't care about ensuring that POSIX ACL and CHMOD permissions are in sync as it is a responsibility of a userspace daemon to handle that. For the same reason we don't have a calls to posix_acl_chmod(), while most of other filesystem do. Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/acl.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/fuse/acl.c b/fs/fuse/acl.c index 897d813c5e92..8f484b105f13 100644 --- a/fs/fuse/acl.c +++ b/fs/fuse/acl.c @@ -144,8 +144,8 @@ int fuse_set_acl(struct mnt_idmap *idmap, struct dentry *dentry, * be stripped. */ if (fc->posix_acl && - !in_group_or_capable(&nop_mnt_idmap, inode, - i_gid_into_vfsgid(&nop_mnt_idmap, inode))) + !in_group_or_capable(idmap, inode, + i_gid_into_vfsgid(idmap, inode))) extra_flags |= FUSE_SETXATTR_ACL_KILL_SGID; ret = fuse_setxattr(inode, name, value, size, 0, extra_flags); From 4be75ffe721cbf925478ebfbff872b02833899b7 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:22 +0200 Subject: [PATCH 053/107] fuse: support idmapped ->rename op RENAME_WHITEOUT is a special case of ->rename and we need to take idmappings into account there. Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 5fa9c99f4d70..6f289fa9cc62 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1026,7 +1026,7 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry) return err; } -static int fuse_rename_common(struct inode *olddir, struct dentry *oldent, +static int fuse_rename_common(struct mnt_idmap *idmap, struct inode *olddir, struct dentry *oldent, struct inode *newdir, struct dentry *newent, unsigned int flags, int opcode, size_t argsize) { @@ -1047,7 +1047,7 @@ static int fuse_rename_common(struct inode *olddir, struct dentry *oldent, args.in_args[1].value = oldent->d_name.name; args.in_args[2].size = newent->d_name.len + 1; args.in_args[2].value = newent->d_name.name; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(idmap, fm, &args); if (!err) { /* ctime changes */ fuse_update_ctime(d_inode(oldent)); @@ -1093,7 +1093,8 @@ static int fuse_rename2(struct mnt_idmap *idmap, struct inode *olddir, if (fc->no_rename2 || fc->minor < 23) return -EINVAL; - err = fuse_rename_common(olddir, oldent, newdir, newent, flags, + err = fuse_rename_common((flags & RENAME_WHITEOUT) ? idmap : NULL, + olddir, oldent, newdir, newent, flags, FUSE_RENAME2, sizeof(struct fuse_rename2_in)); if (err == -ENOSYS) { @@ -1101,7 +1102,7 @@ static int fuse_rename2(struct mnt_idmap *idmap, struct inode *olddir, err = -EINVAL; } } else { - err = fuse_rename_common(olddir, oldent, newdir, newent, 0, + err = fuse_rename_common(NULL, olddir, oldent, newdir, newent, 0, FUSE_RENAME, sizeof(struct fuse_rename_in)); } From 5b8ca5a54cb89ab07b0389f50e038e533cdfdd86 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:23 +0200 Subject: [PATCH 054/107] fuse: handle idmappings properly in ->write_iter() This is needed to properly clear suid/sgid. Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index 0c33ddf97864..ca553d7a7c9e 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1398,6 +1398,7 @@ static void fuse_dio_unlock(struct kiocb *iocb, bool exclusive) static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from) { struct file *file = iocb->ki_filp; + struct mnt_idmap *idmap = file_mnt_idmap(file); struct address_space *mapping = file->f_mapping; ssize_t written = 0; struct inode *inode = mapping->host; @@ -1412,7 +1413,7 @@ static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from) return err; if (fc->handle_killpriv_v2 && - setattr_should_drop_suidgid(&nop_mnt_idmap, + setattr_should_drop_suidgid(idmap, file_inode(file))) { goto writethrough; } From 6d14b18596ca69719cfe1af87dbf3c5e763d29b5 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:24 +0200 Subject: [PATCH 055/107] fuse: warn if fuse_access is called when idmapped mounts are allowed It is not possible with the current fuse code, but let's protect ourselves from regressions in the future. Signed-off-by: Alexander Mikhalitsyn Signed-off-by: Miklos Szeredi --- fs/fuse/dir.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 6f289fa9cc62..99f9948bf68b 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1474,6 +1474,14 @@ static int fuse_access(struct inode *inode, int mask) BUG_ON(mask & MAY_NOT_BLOCK); + /* + * We should not send FUSE_ACCESS to the userspace + * when idmapped mounts are enabled as for this case + * we have fc->default_permissions = 1 and access + * permission checks are done on the kernel side. + */ + WARN_ON_ONCE(!(fm->sb->s_iflags & SB_I_NOIDMAP)); + if (fm->fc->no_access) return 0; From 16e1503eaf329129170e4e7a078aee17686967a5 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:25 +0200 Subject: [PATCH 056/107] fuse: allow idmapped mounts Now we have everything in place and we can allow idmapped mounts by setting the FS_ALLOW_IDMAP flag. Notice that real availability of idmapped mounts will depend on the fuse daemon. Fuse daemon have to set FUSE_ALLOW_IDMAP flag in the FUSE_INIT reply. To discuss: - we enable idmapped mounts support only if "default_permissions" mode is enabled, because otherwise we would need to deal with UID/GID mappings in the userspace side OR provide the userspace with idmapped req->in.h.uid/req->in.h.gid values which is not something that we probably want to. Idmapped mounts philosophy is not about faking caller uid/gid. Some extra links and examples: - libfuse support https://github.com/mihalicyn/libfuse/commits/idmap_support - fuse-overlayfs support: https://github.com/mihalicyn/fuse-overlayfs/commits/idmap_support - cephfs-fuse conversion example https://github.com/mihalicyn/ceph/commits/fuse_idmap - glusterfs conversion example https://github.com/mihalicyn/glusterfs/commits/fuse_idmap Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/inode.c | 12 +++++++++--- include/uapi/linux/fuse.h | 20 +++++++++++++++++++- 2 files changed, 28 insertions(+), 4 deletions(-) diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index b08bc6081612..d7edb3fb829f 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -1348,6 +1348,12 @@ static void process_init_reply(struct fuse_mount *fm, struct fuse_args *args, } if (flags & FUSE_NO_EXPORT_SUPPORT) fm->sb->s_export_op = &fuse_export_fid_operations; + if (flags & FUSE_ALLOW_IDMAP) { + if (fc->default_permissions) + fm->sb->s_iflags &= ~SB_I_NOIDMAP; + else + ok = false; + } } else { ra_pages = fc->max_read / PAGE_SIZE; fc->no_lock = 1; @@ -1395,7 +1401,7 @@ void fuse_send_init(struct fuse_mount *fm) FUSE_HANDLE_KILLPRIV_V2 | FUSE_SETXATTR_EXT | FUSE_INIT_EXT | FUSE_SECURITY_CTX | FUSE_CREATE_SUPP_GROUP | FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP | - FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND; + FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND | FUSE_ALLOW_IDMAP; #ifdef CONFIG_FUSE_DAX if (fm->fc->dax) flags |= FUSE_MAP_ALIGNMENT; @@ -1985,7 +1991,7 @@ static void fuse_kill_sb_anon(struct super_block *sb) static struct file_system_type fuse_fs_type = { .owner = THIS_MODULE, .name = "fuse", - .fs_flags = FS_HAS_SUBTYPE | FS_USERNS_MOUNT, + .fs_flags = FS_HAS_SUBTYPE | FS_USERNS_MOUNT | FS_ALLOW_IDMAP, .init_fs_context = fuse_init_fs_context, .parameters = fuse_fs_parameters, .kill_sb = fuse_kill_sb_anon, @@ -2006,7 +2012,7 @@ static struct file_system_type fuseblk_fs_type = { .init_fs_context = fuse_init_fs_context, .parameters = fuse_fs_parameters, .kill_sb = fuse_kill_sb_blk, - .fs_flags = FS_REQUIRES_DEV | FS_HAS_SUBTYPE, + .fs_flags = FS_REQUIRES_DEV | FS_HAS_SUBTYPE | FS_ALLOW_IDMAP, }; MODULE_ALIAS_FS("fuseblk"); diff --git a/include/uapi/linux/fuse.h b/include/uapi/linux/fuse.h index 2ccf38181df2..f1e99458e29e 100644 --- a/include/uapi/linux/fuse.h +++ b/include/uapi/linux/fuse.h @@ -217,6 +217,9 @@ * - add backing_id to fuse_open_out, add FOPEN_PASSTHROUGH open flag * - add FUSE_NO_EXPORT_SUPPORT init flag * - add FUSE_NOTIFY_RESEND, add FUSE_HAS_RESEND init flag + * + * 7.41 + * - add FUSE_ALLOW_IDMAP */ #ifndef _LINUX_FUSE_H @@ -252,7 +255,7 @@ #define FUSE_KERNEL_VERSION 7 /** Minor version number of this interface */ -#define FUSE_KERNEL_MINOR_VERSION 40 +#define FUSE_KERNEL_MINOR_VERSION 41 /** The node ID of the root inode */ #define FUSE_ROOT_ID 1 @@ -421,6 +424,7 @@ struct fuse_file_lock { * FUSE_NO_EXPORT_SUPPORT: explicitly disable export support * FUSE_HAS_RESEND: kernel supports resending pending requests, and the high bit * of the request ID indicates resend requests + * FUSE_ALLOW_IDMAP: allow creation of idmapped mounts */ #define FUSE_ASYNC_READ (1 << 0) #define FUSE_POSIX_LOCKS (1 << 1) @@ -466,6 +470,7 @@ struct fuse_file_lock { /* Obsolete alias for FUSE_DIRECT_IO_ALLOW_MMAP */ #define FUSE_DIRECT_IO_RELAX FUSE_DIRECT_IO_ALLOW_MMAP +#define FUSE_ALLOW_IDMAP (1ULL << 40) /** * CUSE INIT request/reply flags @@ -984,6 +989,19 @@ struct fuse_fallocate_in { */ #define FUSE_UNIQUE_RESEND (1ULL << 63) +/** + * This value will be set by the kernel to + * (struct fuse_in_header).{uid,gid} fields in + * case when: + * - fuse daemon enabled FUSE_ALLOW_IDMAP + * - idmapping information is not available and uid/gid + * can not be mapped in accordance with an idmapping. + * + * Note: an idmapping information always available + * for inode creation operations like: + * FUSE_MKNOD, FUSE_SYMLINK, FUSE_MKDIR, FUSE_TMPFILE, + * FUSE_CREATE and FUSE_RENAME2 (with RENAME_WHITEOUT). + */ #define FUSE_INVALID_UIDGID ((uint32_t)(-1)) struct fuse_in_header { From 10de741a5d7ab110eb026ad214eea2c5e5fe7e9c Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:23 +0800 Subject: [PATCH 057/107] f2fs: convert f2fs_compress_ctx_add_page() to use folio onvert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/compress.c | 14 +++++++------- fs/f2fs/data.c | 4 ++-- fs/f2fs/f2fs.h | 2 +- 3 files changed, 10 insertions(+), 10 deletions(-) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index f55d54bb12f4..a96c7b16fe19 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -160,17 +160,17 @@ void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse) cc->cluster_idx = NULL_CLUSTER; } -void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page) +void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct folio *folio) { unsigned int cluster_ofs; - if (!f2fs_cluster_can_merge_page(cc, page->index)) + if (!f2fs_cluster_can_merge_page(cc, folio->index)) f2fs_bug_on(F2FS_I_SB(cc->inode), 1); - cluster_ofs = offset_in_cluster(cc, page->index); - cc->rpages[cluster_ofs] = page; + cluster_ofs = offset_in_cluster(cc, folio->index); + cc->rpages[cluster_ofs] = folio_page(folio, 0); cc->nr_rpages++; - cc->cluster_idx = cluster_idx(cc, page->index); + cc->cluster_idx = cluster_idx(cc, folio->index); } #ifdef CONFIG_F2FS_FS_LZO @@ -1113,7 +1113,7 @@ retry: if (PageUptodate(page)) f2fs_put_page(page, 1); else - f2fs_compress_ctx_add_page(cc, page); + f2fs_compress_ctx_add_page(cc, page_folio(page)); } if (!f2fs_cluster_is_empty(cc)) { @@ -1143,7 +1143,7 @@ retry: } f2fs_wait_on_page_writeback(page, DATA, true, true); - f2fs_compress_ctx_add_page(cc, page); + f2fs_compress_ctx_add_page(cc, page_folio(page)); if (!PageUptodate(page)) { release_and_retry: diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e9570f4e0f21..100b6526717f 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2429,7 +2429,7 @@ static int f2fs_mpage_readpages(struct inode *inode, if (ret) goto set_error_page; - f2fs_compress_ctx_add_page(&cc, &folio->page); + f2fs_compress_ctx_add_page(&cc, folio); goto next_page; read_single_page: @@ -3161,7 +3161,7 @@ continue_unlock: #ifdef CONFIG_F2FS_FS_COMPRESSION if (f2fs_compressed_file(inode)) { folio_get(folio); - f2fs_compress_ctx_add_page(&cc, &folio->page); + f2fs_compress_ctx_add_page(&cc, folio); continue; } #endif diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index ec82bde0c51d..569d9d2faae1 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -4327,7 +4327,7 @@ bool f2fs_cluster_can_merge_page(struct compress_ctx *cc, pgoff_t index); bool f2fs_all_cluster_page_ready(struct compress_ctx *cc, struct page **pages, int index, int nr_pages, bool uptodate); bool f2fs_sanity_check_cluster(struct dnode_of_data *dn); -void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct page *page); +void f2fs_compress_ctx_add_page(struct compress_ctx *cc, struct folio *folio); int f2fs_write_multi_pages(struct compress_ctx *cc, int *submitted, struct writeback_control *wbc, From aec5755951b74e3bbb5ddee39ac142a788547854 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:24 +0800 Subject: [PATCH 058/107] f2fs: convert f2fs_vm_page_mkwrite() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 81d0b86c5718..a8d153eb0a95 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -53,7 +53,7 @@ static vm_fault_t f2fs_filemap_fault(struct vm_fault *vmf) static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) { - struct page *page = vmf->page; + struct folio *folio = page_folio(vmf->page); struct inode *inode = file_inode(vmf->vma->vm_file); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct dnode_of_data dn; @@ -85,7 +85,7 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) #ifdef CONFIG_F2FS_FS_COMPRESSION if (f2fs_compressed_file(inode)) { - int ret = f2fs_is_compressed_cluster(inode, page->index); + int ret = f2fs_is_compressed_cluster(inode, folio->index); if (ret < 0) { err = ret; @@ -105,11 +105,11 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) file_update_time(vmf->vma->vm_file); filemap_invalidate_lock_shared(inode->i_mapping); - lock_page(page); - if (unlikely(page->mapping != inode->i_mapping || - page_offset(page) > i_size_read(inode) || - !PageUptodate(page))) { - unlock_page(page); + folio_lock(folio); + if (unlikely(folio->mapping != inode->i_mapping || + folio_pos(folio) > i_size_read(inode) || + !folio_test_uptodate(folio))) { + folio_unlock(folio); err = -EFAULT; goto out_sem; } @@ -117,9 +117,9 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) set_new_dnode(&dn, inode, NULL, NULL, 0); if (need_alloc) { /* block allocation */ - err = f2fs_get_block_locked(&dn, page->index); + err = f2fs_get_block_locked(&dn, folio->index); } else { - err = f2fs_get_dnode_of_data(&dn, page->index, LOOKUP_NODE); + err = f2fs_get_dnode_of_data(&dn, folio->index, LOOKUP_NODE); f2fs_put_dnode(&dn); if (f2fs_is_pinned_file(inode) && !__is_valid_data_blkaddr(dn.data_blkaddr)) @@ -127,11 +127,11 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) } if (err) { - unlock_page(page); + folio_unlock(folio); goto out_sem; } - f2fs_wait_on_page_writeback(page, DATA, false, true); + f2fs_wait_on_page_writeback(folio_page(folio, 0), DATA, false, true); /* wait for GCed page writeback via META_MAPPING */ f2fs_wait_on_block_writeback(inode, dn.data_blkaddr); @@ -139,18 +139,18 @@ static vm_fault_t f2fs_vm_page_mkwrite(struct vm_fault *vmf) /* * check to see if the page is mapped already (no holes) */ - if (PageMappedToDisk(page)) + if (folio_test_mappedtodisk(folio)) goto out_sem; /* page is wholly or partially inside EOF */ - if (((loff_t)(page->index + 1) << PAGE_SHIFT) > + if (((loff_t)(folio->index + 1) << PAGE_SHIFT) > i_size_read(inode)) { loff_t offset; offset = i_size_read(inode) & ~PAGE_MASK; - zero_user_segment(page, offset, PAGE_SIZE); + folio_zero_segment(folio, offset, folio_size(folio)); } - set_page_dirty(page); + folio_mark_dirty(folio); f2fs_update_iostat(sbi, inode, APP_MAPPED_IO, F2FS_BLKSIZE); f2fs_update_time(sbi, REQ_TIME); @@ -162,7 +162,7 @@ out_sem: out: ret = vmf_fs_error(err); - trace_f2fs_vm_page_mkwrite(inode, page->index, vmf->vma->vm_flags, ret); + trace_f2fs_vm_page_mkwrite(inode, folio->index, vmf->vma->vm_flags, ret); return ret; } From fd3a11aff4bef4df7dff25c71e4c51fe8d93e2e7 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:25 +0800 Subject: [PATCH 059/107] f2fs: convert f2fs_clear_page_cache_dirty_tag() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 3 +-- fs/f2fs/dir.c | 2 +- fs/f2fs/f2fs.h | 2 +- fs/f2fs/inline.c | 2 +- fs/f2fs/node.c | 2 +- 5 files changed, 5 insertions(+), 6 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 100b6526717f..0df5a0c66ede 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -4117,9 +4117,8 @@ const struct address_space_operations f2fs_dblock_aops = { .swap_deactivate = f2fs_swap_deactivate, }; -void f2fs_clear_page_cache_dirty_tag(struct page *page) +void f2fs_clear_page_cache_dirty_tag(struct folio *folio) { - struct folio *folio = page_folio(page); struct address_space *mapping = folio->mapping; unsigned long flags; diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 14900ca8a9ff..5fcc952107e9 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -868,7 +868,7 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page, if (bit_pos == NR_DENTRY_IN_BLOCK && !f2fs_truncate_hole(dir, page->index, page->index + 1)) { - f2fs_clear_page_cache_dirty_tag(page); + f2fs_clear_page_cache_dirty_tag(page_folio(page)); clear_page_dirty_for_io(page); ClearPageUptodate(page); clear_page_private_all(page); diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 569d9d2faae1..fafc648a6642 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3890,7 +3890,7 @@ void f2fs_write_failed(struct inode *inode, loff_t to); void f2fs_invalidate_folio(struct folio *folio, size_t offset, size_t length); bool f2fs_release_folio(struct folio *folio, gfp_t wait); bool f2fs_overwrite_io(struct inode *inode, loff_t pos, size_t len); -void f2fs_clear_page_cache_dirty_tag(struct page *page); +void f2fs_clear_page_cache_dirty_tag(struct folio *folio); int f2fs_init_post_read_processing(void); void f2fs_destroy_post_read_processing(void); int f2fs_init_post_read_wq(struct f2fs_sb_info *sbi); diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index d76a3bed7093..98d122164cd1 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -281,7 +281,7 @@ int f2fs_write_inline_data(struct inode *inode, struct page *page) page, 0, MAX_INLINE_DATA(inode)); set_page_dirty(ipage); - f2fs_clear_page_cache_dirty_tag(page); + f2fs_clear_page_cache_dirty_tag(page_folio(page)); set_inode_flag(inode, FI_APPEND_WRITE); set_inode_flag(inode, FI_DATA_EXIST); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 9e7a6e21f30c..20518c12f5b7 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -123,7 +123,7 @@ bool f2fs_available_free_memory(struct f2fs_sb_info *sbi, int type) static void clear_node_page_dirty(struct page *page) { if (PageDirty(page)) { - f2fs_clear_page_cache_dirty_tag(page); + f2fs_clear_page_cache_dirty_tag(page_folio(page)); clear_page_dirty_for_io(page); dec_page_count(F2FS_P_SB(page), F2FS_DIRTY_NODES); } From b084662139ae964073bfdbcf633f3ca9fa1a6d27 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:26 +0800 Subject: [PATCH 060/107] f2fs: convert f2fs_write_inline_data() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 2 +- fs/f2fs/f2fs.h | 2 +- fs/f2fs/inline.c | 10 +++++----- fs/f2fs/node.c | 2 +- 4 files changed, 8 insertions(+), 8 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 0df5a0c66ede..6d2a37db0c9e 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2866,7 +2866,7 @@ write: err = -EAGAIN; if (f2fs_has_inline_data(inode)) { - err = f2fs_write_inline_data(inode, page); + err = f2fs_write_inline_data(inode, folio); if (!err) goto out; } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index fafc648a6642..12f6589c83b9 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -4185,7 +4185,7 @@ int f2fs_read_inline_data(struct inode *inode, struct folio *folio); int f2fs_convert_inline_page(struct dnode_of_data *dn, struct page *page); int f2fs_convert_inline_inode(struct inode *inode); int f2fs_try_convert_inline_dir(struct inode *dir, struct dentry *dentry); -int f2fs_write_inline_data(struct inode *inode, struct page *page); +int f2fs_write_inline_data(struct inode *inode, struct folio *folio); int f2fs_recover_inline_data(struct inode *inode, struct page *npage); struct f2fs_dir_entry *f2fs_find_in_inline_dir(struct inode *dir, const struct f2fs_filename *fname, diff --git a/fs/f2fs/inline.c b/fs/f2fs/inline.c index 98d122164cd1..005babf1bed1 100644 --- a/fs/f2fs/inline.c +++ b/fs/f2fs/inline.c @@ -260,7 +260,7 @@ out: return err; } -int f2fs_write_inline_data(struct inode *inode, struct page *page) +int f2fs_write_inline_data(struct inode *inode, struct folio *folio) { struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct page *ipage; @@ -274,14 +274,14 @@ int f2fs_write_inline_data(struct inode *inode, struct page *page) return -EAGAIN; } - f2fs_bug_on(F2FS_I_SB(inode), page->index); + f2fs_bug_on(F2FS_I_SB(inode), folio->index); f2fs_wait_on_page_writeback(ipage, NODE, true, true); - memcpy_from_page(inline_data_addr(inode, ipage), - page, 0, MAX_INLINE_DATA(inode)); + memcpy_from_folio(inline_data_addr(inode, ipage), + folio, 0, MAX_INLINE_DATA(inode)); set_page_dirty(ipage); - f2fs_clear_page_cache_dirty_tag(page_folio(page)); + f2fs_clear_page_cache_dirty_tag(folio); set_inode_flag(inode, FI_APPEND_WRITE); set_inode_flag(inode, FI_DATA_EXIST); diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 20518c12f5b7..2b764c86aacc 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1535,7 +1535,7 @@ static void flush_inline_data(struct f2fs_sb_info *sbi, nid_t ino) if (!clear_page_dirty_for_io(page)) goto page_out; - ret = f2fs_write_inline_data(inode, page); + ret = f2fs_write_inline_data(inode, page_folio(page)); inode_dec_dirty_pages(inode); f2fs_remove_dirty_inode(inode); if (ret) From 46a75ef866030b3500bf15fd3400c6779eb33752 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:27 +0800 Subject: [PATCH 061/107] f2fs: convert f2fs_write_single_data_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/compress.c | 3 ++- fs/f2fs/data.c | 29 +++++++++++++++-------------- fs/f2fs/f2fs.h | 2 +- 3 files changed, 18 insertions(+), 16 deletions(-) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index a96c7b16fe19..630bb78655d5 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -1543,7 +1543,8 @@ continue_unlock: if (!clear_page_dirty_for_io(cc->rpages[i])) goto continue_unlock; - ret = f2fs_write_single_data_page(cc->rpages[i], &submitted, + ret = f2fs_write_single_data_page(page_folio(cc->rpages[i]), + &submitted, NULL, NULL, wbc, io_type, compr_blocks, false); if (ret) { diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 6d2a37db0c9e..0779e222f709 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2764,7 +2764,7 @@ out: return err; } -int f2fs_write_single_data_page(struct page *page, int *submitted, +int f2fs_write_single_data_page(struct folio *folio, int *submitted, struct bio **bio, sector_t *last_block, struct writeback_control *wbc, @@ -2772,12 +2772,13 @@ int f2fs_write_single_data_page(struct page *page, int *submitted, int compr_blocks, bool allow_balance) { - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; + struct page *page = folio_page(folio, 0); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); loff_t i_size = i_size_read(inode); const pgoff_t end_index = ((unsigned long long)i_size) >> PAGE_SHIFT; - loff_t psize = (loff_t)(page->index + 1) << PAGE_SHIFT; + loff_t psize = (loff_t)(folio->index + 1) << PAGE_SHIFT; unsigned offset = 0; bool need_balance_fs = false; bool quota_inode = IS_NOQUOTA(inode); @@ -2801,11 +2802,11 @@ int f2fs_write_single_data_page(struct page *page, int *submitted, .last_block = last_block, }; - trace_f2fs_writepage(page_folio(page), DATA); + trace_f2fs_writepage(folio, DATA); /* we should bypass data pages to proceed the kworker jobs */ if (unlikely(f2fs_cp_error(sbi))) { - mapping_set_error(page->mapping, -EIO); + mapping_set_error(folio->mapping, -EIO); /* * don't drop any dirty dentry pages for keeping lastest * directory structure. @@ -2823,7 +2824,7 @@ int f2fs_write_single_data_page(struct page *page, int *submitted, if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) goto redirty_out; - if (page->index < end_index || + if (folio->index < end_index || f2fs_verity_in_progress(inode) || compr_blocks) goto write; @@ -2833,10 +2834,10 @@ int f2fs_write_single_data_page(struct page *page, int *submitted, * this page does not have to be written to disk. */ offset = i_size & (PAGE_SIZE - 1); - if ((page->index >= end_index + 1) || !offset) + if ((folio->index >= end_index + 1) || !offset) goto out; - zero_user_segment(page, offset, PAGE_SIZE); + folio_zero_segment(folio, offset, folio_size(folio)); write: /* Dentry/quota blocks are controlled by checkpoint */ if (S_ISDIR(inode->i_mode) || quota_inode) { @@ -2896,7 +2897,7 @@ done: out: inode_dec_dirty_pages(inode); if (err) { - ClearPageUptodate(page); + folio_clear_uptodate(folio); clear_page_private_gcing(page); } @@ -2906,7 +2907,7 @@ out: f2fs_remove_dirty_inode(inode); submitted = NULL; } - unlock_page(page); + folio_unlock(folio); if (!S_ISDIR(inode->i_mode) && !IS_NOQUOTA(inode) && !F2FS_I(inode)->wb_task && allow_balance) f2fs_balance_fs(sbi, need_balance_fs); @@ -2924,7 +2925,7 @@ out: return 0; redirty_out: - redirty_page_for_writepage(wbc, page); + folio_redirty_for_writepage(wbc, folio); /* * pageout() in MM translates EAGAIN, so calls handle_write_error() * -> mapping_set_error() -> set_bit(AS_EIO, ...). @@ -2933,7 +2934,7 @@ redirty_out: */ if (!err || wbc->for_reclaim) return AOP_WRITEPAGE_ACTIVATE; - unlock_page(page); + folio_unlock(folio); return err; } @@ -2955,7 +2956,7 @@ static int f2fs_write_data_page(struct page *page, out: #endif - return f2fs_write_single_data_page(page, NULL, NULL, NULL, + return f2fs_write_single_data_page(page_folio(page), NULL, NULL, NULL, wbc, FS_DATA_IO, 0, true); } @@ -3165,7 +3166,7 @@ continue_unlock: continue; } #endif - ret = f2fs_write_single_data_page(&folio->page, + ret = f2fs_write_single_data_page(folio, &submitted, &bio, &last_block, wbc, io_type, 0, true); if (ret == AOP_WRITEPAGE_ACTIVATE) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 12f6589c83b9..994661d97635 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3881,7 +3881,7 @@ int f2fs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, int f2fs_encrypt_one_page(struct f2fs_io_info *fio); bool f2fs_should_update_inplace(struct inode *inode, struct f2fs_io_info *fio); bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio); -int f2fs_write_single_data_page(struct page *page, int *submitted, +int f2fs_write_single_data_page(struct folio *folio, int *submitted, struct bio **bio, sector_t *last_block, struct writeback_control *wbc, enum iostat_type io_type, From e55561d278f2b03c0131e5408df4ab9fed2ae51d Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:28 +0800 Subject: [PATCH 062/107] f2fs: convert f2fs_do_write_meta_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/f2fs.h | 2 +- fs/f2fs/segment.c | 14 +++++++------- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index f3d22b8ef2ff..d73d4d05313e 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -362,7 +362,7 @@ static int __f2fs_write_meta_page(struct page *page, if (wbc->for_reclaim && page->index < GET_SUM_BLOCK(sbi, 0)) goto redirty_out; - f2fs_do_write_meta_page(sbi, page, io_type); + f2fs_do_write_meta_page(sbi, page_folio(page), io_type); dec_page_count(sbi, F2FS_DIRTY_META); if (wbc->for_reclaim) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 994661d97635..2c1664f951e9 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3732,7 +3732,7 @@ bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi, struct page *f2fs_get_sum_page(struct f2fs_sb_info *sbi, unsigned int segno); void f2fs_update_meta_page(struct f2fs_sb_info *sbi, void *src, block_t blk_addr); -void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page, +void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct folio *folio, enum iostat_type io_type); void f2fs_do_write_node_page(unsigned int nid, struct f2fs_io_info *fio); void f2fs_outplace_write_data(struct dnode_of_data *dn, diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index ed070b9ef91b..e3cc2706f544 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -3811,7 +3811,7 @@ out: f2fs_up_read(&fio->sbi->io_order_lock); } -void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page, +void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct folio *folio, enum iostat_type io_type) { struct f2fs_io_info fio = { @@ -3820,20 +3820,20 @@ void f2fs_do_write_meta_page(struct f2fs_sb_info *sbi, struct page *page, .temp = HOT, .op = REQ_OP_WRITE, .op_flags = REQ_SYNC | REQ_META | REQ_PRIO, - .old_blkaddr = page->index, - .new_blkaddr = page->index, - .page = page, + .old_blkaddr = folio->index, + .new_blkaddr = folio->index, + .page = folio_page(folio, 0), .encrypted_page = NULL, .in_list = 0, }; - if (unlikely(page->index >= MAIN_BLKADDR(sbi))) + if (unlikely(folio->index >= MAIN_BLKADDR(sbi))) fio.op_flags &= ~REQ_META; - set_page_writeback(page); + folio_start_writeback(folio); f2fs_submit_page_write(&fio); - stat_inc_meta_count(sbi, page->index); + stat_inc_meta_count(sbi, folio->index); f2fs_update_iostat(sbi, NULL, io_type, F2FS_BLKSIZE); } From 138a762e7f45474c33dfb269f3e6641741be1508 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:29 +0800 Subject: [PATCH 063/107] f2fs: convert __f2fs_write_meta_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index d73d4d05313e..ffaf295a198a 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -345,30 +345,31 @@ static int __f2fs_write_meta_page(struct page *page, enum iostat_type io_type) { struct f2fs_sb_info *sbi = F2FS_P_SB(page); + struct folio *folio = page_folio(page); - trace_f2fs_writepage(page_folio(page), META); + trace_f2fs_writepage(folio, META); if (unlikely(f2fs_cp_error(sbi))) { if (is_sbi_flag_set(sbi, SBI_IS_CLOSE)) { - ClearPageUptodate(page); + folio_clear_uptodate(folio); dec_page_count(sbi, F2FS_DIRTY_META); - unlock_page(page); + folio_unlock(folio); return 0; } goto redirty_out; } if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) goto redirty_out; - if (wbc->for_reclaim && page->index < GET_SUM_BLOCK(sbi, 0)) + if (wbc->for_reclaim && folio->index < GET_SUM_BLOCK(sbi, 0)) goto redirty_out; - f2fs_do_write_meta_page(sbi, page_folio(page), io_type); + f2fs_do_write_meta_page(sbi, folio, io_type); dec_page_count(sbi, F2FS_DIRTY_META); if (wbc->for_reclaim) f2fs_submit_merged_write_cond(sbi, NULL, page, 0, META); - unlock_page(page); + folio_unlock(folio); if (unlikely(f2fs_cp_error(sbi))) f2fs_submit_merged_write(sbi, META); From 3981e94285886cd712b563a82a2e99f902510c78 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 14 Aug 2024 21:50:59 +0800 Subject: [PATCH 064/107] f2fs: convert f2fs_read_multi_pages() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 28 ++++++++++++++++------------ 1 file changed, 16 insertions(+), 12 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 0779e222f709..4f4e76c33611 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2207,19 +2207,22 @@ int f2fs_read_multi_pages(struct compress_ctx *cc, struct bio **bio_ret, /* get rid of pages beyond EOF */ for (i = 0; i < cc->cluster_size; i++) { struct page *page = cc->rpages[i]; + struct folio *folio; if (!page) continue; - if ((sector_t)page->index >= last_block_in_file) { - zero_user_segment(page, 0, PAGE_SIZE); - if (!PageUptodate(page)) - SetPageUptodate(page); - } else if (!PageUptodate(page)) { + + folio = page_folio(page); + if ((sector_t)folio->index >= last_block_in_file) { + folio_zero_segment(folio, 0, folio_size(folio)); + if (!folio_test_uptodate(folio)) + folio_mark_uptodate(folio); + } else if (!folio_test_uptodate(folio)) { continue; } - unlock_page(page); + folio_unlock(folio); if (for_write) - put_page(page); + folio_put(folio); cc->rpages[i] = NULL; cc->nr_rpages--; } @@ -2279,7 +2282,7 @@ skip_reading_dnode: } for (i = 0; i < cc->nr_cpages; i++) { - struct page *page = dic->cpages[i]; + struct folio *folio = page_folio(dic->cpages[i]); block_t blkaddr; struct bio_post_read_ctx *ctx; @@ -2289,7 +2292,8 @@ skip_reading_dnode: f2fs_wait_on_block_writeback(inode, blkaddr); - if (f2fs_load_compressed_page(sbi, page, blkaddr)) { + if (f2fs_load_compressed_page(sbi, folio_page(folio, 0), + blkaddr)) { if (atomic_dec_and_test(&dic->remaining_pages)) { f2fs_decompress_cluster(dic, true); break; @@ -2299,7 +2303,7 @@ skip_reading_dnode: if (bio && (!page_is_mergeable(sbi, bio, *last_block_in_bio, blkaddr) || - !f2fs_crypt_mergeable_bio(bio, inode, page->index, NULL))) { + !f2fs_crypt_mergeable_bio(bio, inode, folio->index, NULL))) { submit_and_realloc: f2fs_submit_read_bio(sbi, bio, DATA); bio = NULL; @@ -2308,7 +2312,7 @@ submit_and_realloc: if (!bio) { bio = f2fs_grab_read_bio(inode, blkaddr, nr_pages, f2fs_ra_op_flags(rac), - page->index, for_write); + folio->index, for_write); if (IS_ERR(bio)) { ret = PTR_ERR(bio); f2fs_decompress_end_io(dic, ret, true); @@ -2318,7 +2322,7 @@ submit_and_realloc: } } - if (bio_add_page(bio, page, blocksize, 0) < blocksize) + if (!bio_add_folio(bio, folio, blocksize, 0)) goto submit_and_realloc; ctx = get_post_read_ctx(bio); From 763a0dc792f9d7c79c1d85415c6cebc1f79bba16 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 13 Aug 2024 22:13:31 +0800 Subject: [PATCH 065/107] f2fs: convert f2fs_handle_page_eio() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/checkpoint.c | 2 +- fs/f2fs/f2fs.h | 6 ++++-- fs/f2fs/node.c | 2 +- 3 files changed, 6 insertions(+), 4 deletions(-) diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index ffaf295a198a..7f76460b721f 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -99,7 +99,7 @@ repeat: } if (unlikely(!PageUptodate(page))) { - f2fs_handle_page_eio(sbi, page->index, META); + f2fs_handle_page_eio(sbi, page_folio(page), META); f2fs_put_page(page, 1); return ERR_PTR(-EIO); } diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 2c1664f951e9..460c05191c28 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -4678,9 +4678,11 @@ static inline void f2fs_io_schedule_timeout(long timeout) io_schedule_timeout(timeout); } -static inline void f2fs_handle_page_eio(struct f2fs_sb_info *sbi, pgoff_t ofs, - enum page_type type) +static inline void f2fs_handle_page_eio(struct f2fs_sb_info *sbi, + struct folio *folio, enum page_type type) { + pgoff_t ofs = folio->index; + if (unlikely(f2fs_cp_error(sbi))) return; diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 2b764c86aacc..7beaa3690e03 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1492,7 +1492,7 @@ out_err: out_put_err: /* ENOENT comes from read_node_page which is not an error. */ if (err != -ENOENT) - f2fs_handle_page_eio(sbi, page->index, NODE); + f2fs_handle_page_eio(sbi, page_folio(page), NODE); f2fs_put_page(page, 1); return ERR_PTR(err); } From 1e5df24c1721d077d9a397982c8b68e486243daf Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:54:59 +0800 Subject: [PATCH 066/107] f2fs: convert f2fs_submit_page_read() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 4f4e76c33611..0655fddfc4ba 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1086,7 +1086,7 @@ static struct bio *f2fs_grab_read_bio(struct inode *inode, block_t blkaddr, } /* This can handle encryption stuffs */ -static int f2fs_submit_page_read(struct inode *inode, struct page *page, +static int f2fs_submit_page_read(struct inode *inode, struct folio *folio, block_t blkaddr, blk_opf_t op_flags, bool for_write) { @@ -1094,14 +1094,14 @@ static int f2fs_submit_page_read(struct inode *inode, struct page *page, struct bio *bio; bio = f2fs_grab_read_bio(inode, blkaddr, 1, op_flags, - page->index, for_write); + folio->index, for_write); if (IS_ERR(bio)) return PTR_ERR(bio); /* wait for GCed page writeback via META_MAPPING */ f2fs_wait_on_block_writeback(inode, blkaddr); - if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) { + if (!bio_add_folio(bio, folio, PAGE_SIZE, 0)) { iostat_update_and_unbind_ctx(bio); if (bio->bi_private) mempool_free(bio->bi_private, bio_post_read_ctx_pool); @@ -1269,7 +1269,7 @@ got_it: return page; } - err = f2fs_submit_page_read(inode, page, dn.data_blkaddr, + err = f2fs_submit_page_read(inode, page_folio(page), dn.data_blkaddr, op_flags, for_write); if (err) goto put_err; @@ -3668,8 +3668,8 @@ repeat: goto fail; } err = f2fs_submit_page_read(use_cow ? - F2FS_I(inode)->cow_inode : inode, page, - blkaddr, 0, true); + F2FS_I(inode)->cow_inode : inode, + page_folio(page), blkaddr, 0, true); if (err) goto fail; From 12e61ce27137dc42c1f2fbd704d0285f8582df62 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:00 +0800 Subject: [PATCH 067/107] f2fs: convert f2fs_write_begin() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 44 +++++++++++++++++++++++--------------------- 1 file changed, 23 insertions(+), 21 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 0655fddfc4ba..85ac05c3655a 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -3378,11 +3378,11 @@ void f2fs_write_failed(struct inode *inode, loff_t to) } static int prepare_write_begin(struct f2fs_sb_info *sbi, - struct page *page, loff_t pos, unsigned len, + struct folio *folio, loff_t pos, unsigned int len, block_t *blk_addr, bool *node_changed) { - struct inode *inode = page->mapping->host; - pgoff_t index = page->index; + struct inode *inode = folio->mapping->host; + pgoff_t index = folio->index; struct dnode_of_data dn; struct page *ipage; bool locked = false; @@ -3419,13 +3419,13 @@ restart: if (f2fs_has_inline_data(inode)) { if (pos + len <= MAX_INLINE_DATA(inode)) { - f2fs_do_read_inline_data(page_folio(page), ipage); + f2fs_do_read_inline_data(folio, ipage); set_inode_flag(inode, FI_DATA_EXIST); if (inode->i_nlink) set_page_private_inline(ipage); goto out; } - err = f2fs_convert_inline_page(&dn, page); + err = f2fs_convert_inline_page(&dn, folio_page(folio, 0)); if (err || dn.data_blkaddr != NULL_ADDR) goto out; } @@ -3518,12 +3518,12 @@ unlock_out: } static int prepare_atomic_write_begin(struct f2fs_sb_info *sbi, - struct page *page, loff_t pos, unsigned int len, + struct folio *folio, loff_t pos, unsigned int len, block_t *blk_addr, bool *node_changed, bool *use_cow) { - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; struct inode *cow_inode = F2FS_I(inode)->cow_inode; - pgoff_t index = page->index; + pgoff_t index = folio->index; int err = 0; block_t ori_blk_addr = NULL_ADDR; @@ -3566,6 +3566,7 @@ static int f2fs_write_begin(struct file *file, struct address_space *mapping, struct inode *inode = mapping->host; struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct page *page = NULL; + struct folio *folio; pgoff_t index = ((unsigned long long) pos) >> PAGE_SHIFT; bool need_balance = false; bool use_cow = false; @@ -3625,22 +3626,23 @@ repeat: /* TODO: cluster can be compressed due to race with .writepage */ *pagep = page; + folio = page_folio(page); if (f2fs_is_atomic_file(inode)) - err = prepare_atomic_write_begin(sbi, page, pos, len, + err = prepare_atomic_write_begin(sbi, folio, pos, len, &blkaddr, &need_balance, &use_cow); else - err = prepare_write_begin(sbi, page, pos, len, + err = prepare_write_begin(sbi, folio, pos, len, &blkaddr, &need_balance); if (err) goto fail; if (need_balance && !IS_NOQUOTA(inode) && has_not_enough_free_secs(sbi, 0, 0)) { - unlock_page(page); + folio_unlock(folio); f2fs_balance_fs(sbi, true); - lock_page(page); - if (page->mapping != mapping) { + folio_lock(folio); + if (folio->mapping != mapping) { /* The page got truncated from under us */ f2fs_put_page(page, 1); goto repeat; @@ -3649,18 +3651,18 @@ repeat: f2fs_wait_on_page_writeback(page, DATA, false, true); - if (len == PAGE_SIZE || PageUptodate(page)) + if (len == PAGE_SIZE || folio_test_uptodate(folio)) return 0; if (!(pos & (PAGE_SIZE - 1)) && (pos + len) >= i_size_read(inode) && !f2fs_verity_in_progress(inode)) { - zero_user_segment(page, len, PAGE_SIZE); + folio_zero_segment(folio, len, folio_size(folio)); return 0; } if (blkaddr == NEW_ADDR) { - zero_user_segment(page, 0, PAGE_SIZE); - SetPageUptodate(page); + folio_zero_segment(folio, 0, folio_size(folio)); + folio_mark_uptodate(folio); } else { if (!f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE_READ)) { @@ -3669,16 +3671,16 @@ repeat: } err = f2fs_submit_page_read(use_cow ? F2FS_I(inode)->cow_inode : inode, - page_folio(page), blkaddr, 0, true); + folio, blkaddr, 0, true); if (err) goto fail; - lock_page(page); - if (unlikely(page->mapping != mapping)) { + folio_lock(folio); + if (unlikely(folio->mapping != mapping)) { f2fs_put_page(page, 1); goto repeat; } - if (unlikely(!PageUptodate(page))) { + if (unlikely(!folio_test_uptodate(folio))) { err = -EIO; goto fail; } From 3c86d1eec7891c53ea9c71d7a2970c248413a855 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:01 +0800 Subject: [PATCH 068/107] f2fs: convert f2fs_write_end() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 13 +++++++------ 1 file changed, 7 insertions(+), 6 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 85ac05c3655a..e114d738b6b4 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -3698,7 +3698,8 @@ static int f2fs_write_end(struct file *file, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) { - struct inode *inode = page->mapping->host; + struct folio *folio = page_folio(page); + struct inode *inode = folio->mapping->host; trace_f2fs_write_end(inode, pos, len, copied); @@ -3707,17 +3708,17 @@ static int f2fs_write_end(struct file *file, * should be PAGE_SIZE. Otherwise, we treat it with zero copied and * let generic_perform_write() try to copy data again through copied=0. */ - if (!PageUptodate(page)) { + if (!folio_test_uptodate(folio)) { if (unlikely(copied != len)) copied = 0; else - SetPageUptodate(page); + folio_mark_uptodate(folio); } #ifdef CONFIG_F2FS_FS_COMPRESSION /* overwrite compressed file */ if (f2fs_compressed_file(inode) && fsdata) { - f2fs_compress_write_end(inode, fsdata, page->index, copied); + f2fs_compress_write_end(inode, fsdata, folio->index, copied); f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); if (pos + copied > i_size_read(inode) && @@ -3730,10 +3731,10 @@ static int f2fs_write_end(struct file *file, if (!copied) goto unlock_out; - set_page_dirty(page); + folio_mark_dirty(folio); if (f2fs_is_atomic_file(inode)) - set_page_private_atomic(page); + set_page_private_atomic(folio_page(folio, 0)); if (pos + copied > i_size_read(inode) && !f2fs_verity_in_progress(inode)) { From 1efc78310c9d25bdc2020202a50c012025602dd3 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:02 +0800 Subject: [PATCH 069/107] f2fs: convert f2fs_set_compressed_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/compress.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index 630bb78655d5..e642c869db4a 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -90,11 +90,13 @@ bool f2fs_is_compressed_page(struct page *page) static void f2fs_set_compressed_page(struct page *page, struct inode *inode, pgoff_t index, void *data) { - attach_page_private(page, (void *)data); + struct folio *folio = page_folio(page); + + folio_attach_private(folio, (void *)data); /* i_crypto_info and iv index */ - page->index = index; - page->mapping = inode->i_mapping; + folio->index = index; + folio->mapping = inode->i_mapping; } static void f2fs_drop_rpages(struct compress_ctx *cc, int len, bool unlock) From bcf4d2dd840e5f4401a2d9425bd0ce94cab12336 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:03 +0800 Subject: [PATCH 070/107] f2fs: convert f2fs_do_write_data_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 26 +++++++++++++------------- 1 file changed, 13 insertions(+), 13 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e114d738b6b4..c57ddee2c7c5 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2648,8 +2648,8 @@ static inline bool need_inplace_update(struct f2fs_io_info *fio) int f2fs_do_write_data_page(struct f2fs_io_info *fio) { - struct page *page = fio->page; - struct inode *inode = page->mapping->host; + struct folio *folio = page_folio(fio->page); + struct inode *inode = folio->mapping->host; struct dnode_of_data dn; struct node_info ni; bool ipu_force = false; @@ -2658,14 +2658,14 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio) /* Use COW inode to make dnode_of_data for atomic write */ atomic_commit = f2fs_is_atomic_file(inode) && - page_private_atomic(fio->page); + page_private_atomic(folio_page(folio, 0)); if (atomic_commit) set_new_dnode(&dn, F2FS_I(inode)->cow_inode, NULL, NULL, 0); else set_new_dnode(&dn, inode, NULL, NULL, 0); if (need_inplace_update(fio) && - f2fs_lookup_read_extent_cache_block(inode, page->index, + f2fs_lookup_read_extent_cache_block(inode, folio->index, &fio->old_blkaddr)) { if (!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr, DATA_GENERIC_ENHANCE)) @@ -2680,7 +2680,7 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio) if (fio->need_lock == LOCK_REQ && !f2fs_trylock_op(fio->sbi)) return -EAGAIN; - err = f2fs_get_dnode_of_data(&dn, page->index, LOOKUP_NODE); + err = f2fs_get_dnode_of_data(&dn, folio->index, LOOKUP_NODE); if (err) goto out; @@ -2688,8 +2688,8 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio) /* This page is already truncated */ if (fio->old_blkaddr == NULL_ADDR) { - ClearPageUptodate(page); - clear_page_private_gcing(page); + folio_clear_uptodate(folio); + clear_page_private_gcing(folio_page(folio, 0)); goto out_writepage; } got_it: @@ -2715,7 +2715,7 @@ got_it: if (err) goto out_writepage; - set_page_writeback(page); + folio_start_writeback(folio); f2fs_put_dnode(&dn); if (fio->need_lock == LOCK_REQ) f2fs_unlock_op(fio->sbi); @@ -2723,11 +2723,11 @@ got_it: if (err) { if (fscrypt_inode_uses_fs_layer_crypto(inode)) fscrypt_finalize_bounce_page(&fio->encrypted_page); - end_page_writeback(page); + folio_end_writeback(folio); } else { set_inode_flag(inode, FI_UPDATE_WRITE); } - trace_f2fs_do_write_data_page(page_folio(page), IPU); + trace_f2fs_do_write_data_page(folio, IPU); return err; } @@ -2749,17 +2749,17 @@ got_it: if (err) goto out_writepage; - set_page_writeback(page); + folio_start_writeback(folio); if (fio->compr_blocks && fio->old_blkaddr == COMPRESS_ADDR) f2fs_i_compr_blocks_update(inode, fio->compr_blocks - 1, false); /* LFS mode write path */ f2fs_outplace_write_data(&dn, fio); - trace_f2fs_do_write_data_page(page_folio(page), OPU); + trace_f2fs_do_write_data_page(folio, OPU); set_inode_flag(inode, FI_APPEND_WRITE); if (atomic_commit) - clear_page_private_atomic(page); + clear_page_private_atomic(folio_page(folio, 0)); out_writepage: f2fs_put_dnode(&dn); out: From 2ac0aa31a15132b88de25f4b687ea24efd81c823 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:04 +0800 Subject: [PATCH 071/107] f2fs: convert f2fs_write_data_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index c57ddee2c7c5..e69097267b99 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -2945,22 +2945,23 @@ redirty_out: static int f2fs_write_data_page(struct page *page, struct writeback_control *wbc) { + struct folio *folio = page_folio(page); #ifdef CONFIG_F2FS_FS_COMPRESSION - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; if (unlikely(f2fs_cp_error(F2FS_I_SB(inode)))) goto out; if (f2fs_compressed_file(inode)) { - if (f2fs_is_compressed_cluster(inode, page->index)) { - redirty_page_for_writepage(wbc, page); + if (f2fs_is_compressed_cluster(inode, folio->index)) { + folio_redirty_for_writepage(wbc, folio); return AOP_WRITEPAGE_ACTIVATE; } } out: #endif - return f2fs_write_single_data_page(page_folio(page), NULL, NULL, NULL, + return f2fs_write_single_data_page(folio, NULL, NULL, NULL, wbc, FS_DATA_IO, 0, true); } From 4deccfbdc4464b66fe50d71e9d64240483efb07d Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:05 +0800 Subject: [PATCH 072/107] f2fs: convert __write_node_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 19 ++++++++++--------- 1 file changed, 10 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 7beaa3690e03..aeda3eecc9e2 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1608,6 +1608,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, enum iostat_type io_type, unsigned int *seq_id) { struct f2fs_sb_info *sbi = F2FS_P_SB(page); + struct folio *folio = page_folio(page); nid_t nid; struct node_info ni; struct f2fs_io_info fio = { @@ -1624,15 +1625,15 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, }; unsigned int seq; - trace_f2fs_writepage(page_folio(page), NODE); + trace_f2fs_writepage(folio, NODE); if (unlikely(f2fs_cp_error(sbi))) { /* keep node pages in remount-ro mode */ if (F2FS_OPTION(sbi).errors == MOUNT_ERRORS_READONLY) goto redirty_out; - ClearPageUptodate(page); + folio_clear_uptodate(folio); dec_page_count(sbi, F2FS_DIRTY_NODES); - unlock_page(page); + folio_unlock(folio); return 0; } @@ -1646,7 +1647,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, /* get old block addr of this node page */ nid = nid_of_node(page); - f2fs_bug_on(sbi, page->index != nid); + f2fs_bug_on(sbi, folio->index != nid); if (f2fs_get_node_info(sbi, nid, &ni, !do_balance)) goto redirty_out; @@ -1660,10 +1661,10 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, /* This page is already truncated */ if (unlikely(ni.blk_addr == NULL_ADDR)) { - ClearPageUptodate(page); + folio_clear_uptodate(folio); dec_page_count(sbi, F2FS_DIRTY_NODES); f2fs_up_read(&sbi->node_write); - unlock_page(page); + folio_unlock(folio); return 0; } @@ -1684,7 +1685,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, *seq_id = seq; } - set_page_writeback(page); + folio_start_writeback(folio); fio.old_blkaddr = ni.blk_addr; f2fs_do_write_node_page(nid, &fio); @@ -1697,7 +1698,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, submitted = NULL; } - unlock_page(page); + folio_unlock(folio); if (unlikely(f2fs_cp_error(sbi))) { f2fs_submit_merged_write(sbi, NODE); @@ -1711,7 +1712,7 @@ static int __write_node_page(struct page *page, bool atomic, bool *submitted, return 0; redirty_out: - redirty_page_for_writepage(wbc, page); + folio_redirty_for_writepage(wbc, folio); return AOP_WRITEPAGE_ACTIVATE; } From 2eaa98e5203979fd5fa6386f307f4f8277a72593 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:06 +0800 Subject: [PATCH 073/107] f2fs: convert read_node_page() to use folio Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/node.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index aeda3eecc9e2..f5e5abce695b 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -1369,6 +1369,7 @@ fail: */ static int read_node_page(struct page *page, blk_opf_t op_flags) { + struct folio *folio = page_folio(page); struct f2fs_sb_info *sbi = F2FS_P_SB(page); struct node_info ni; struct f2fs_io_info fio = { @@ -1381,21 +1382,21 @@ static int read_node_page(struct page *page, blk_opf_t op_flags) }; int err; - if (PageUptodate(page)) { + if (folio_test_uptodate(folio)) { if (!f2fs_inode_chksum_verify(sbi, page)) { - ClearPageUptodate(page); + folio_clear_uptodate(folio); return -EFSBADCRC; } return LOCKED_PAGE; } - err = f2fs_get_node_info(sbi, page->index, &ni, false); + err = f2fs_get_node_info(sbi, folio->index, &ni, false); if (err) return err; /* NEW_ADDR can be seen, after cp_error drops some dirty node pages */ if (unlikely(ni.blk_addr == NULL_ADDR || ni.blk_addr == NEW_ADDR)) { - ClearPageUptodate(page); + folio_clear_uptodate(folio); return -ENOENT; } From 5697e94daab9e40cbf1fb368781d110f22fb6760 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 20 Aug 2024 22:55:07 +0800 Subject: [PATCH 074/107] f2fs: get rid of page->index Convert to use folio, so that we can get rid of 'page->index' to prepare for removal of 'index' field in structure page [1]. [1] https://lore.kernel.org/all/Zp8fgUSIBGQ1TN0D@casper.infradead.org/ Cc: Matthew Wilcox Signed-off-by: Chao Yu Reviewed-by: Li Zetao Signed-off-by: Jaegeuk Kim --- fs/f2fs/compress.c | 2 +- fs/f2fs/data.c | 12 ++++++------ fs/f2fs/dir.c | 3 ++- fs/f2fs/inode.c | 3 ++- fs/f2fs/node.c | 4 ++-- fs/f2fs/segment.c | 3 ++- 6 files changed, 15 insertions(+), 12 deletions(-) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index e642c869db4a..7f26440e8595 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -881,7 +881,7 @@ static bool cluster_has_invalid_data(struct compress_ctx *cc) f2fs_bug_on(F2FS_I_SB(cc->inode), !page); /* beyond EOF */ - if (page->index >= nr_pages) + if (page_folio(page)->index >= nr_pages) return true; } return false; diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index e69097267b99..c6d688208f8b 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -354,7 +354,7 @@ static void f2fs_write_end_io(struct bio *bio) } f2fs_bug_on(sbi, page->mapping == NODE_MAPPING(sbi) && - page->index != nid_of_node(page)); + page_folio(page)->index != nid_of_node(page)); dec_page_count(sbi, type); if (f2fs_in_warm_node_list(sbi, page)) @@ -703,7 +703,7 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) bio = __bio_alloc(fio, 1); f2fs_set_bio_crypt_ctx(bio, fio->page->mapping->host, - fio->page->index, fio, GFP_NOIO); + page_folio(fio->page)->index, fio, GFP_NOIO); if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) { bio_put(bio); @@ -802,7 +802,7 @@ static int add_ipu_page(struct f2fs_io_info *fio, struct bio **bio, fio->new_blkaddr)); if (f2fs_crypt_mergeable_bio(*bio, fio->page->mapping->host, - fio->page->index, fio) && + page_folio(fio->page)->index, fio) && bio_add_page(*bio, page, PAGE_SIZE, 0) == PAGE_SIZE) { ret = 0; @@ -902,7 +902,7 @@ alloc_new: if (!bio) { bio = __bio_alloc(fio, BIO_MAX_VECS); f2fs_set_bio_crypt_ctx(bio, fio->page->mapping->host, - fio->page->index, fio, GFP_NOIO); + page_folio(fio->page)->index, fio, GFP_NOIO); add_bio_entry(fio->sbi, bio, page, fio->temp); } else { @@ -995,13 +995,13 @@ next: (!io_is_mergeable(sbi, io->bio, io, fio, io->last_block_in_bio, fio->new_blkaddr) || !f2fs_crypt_mergeable_bio(io->bio, fio->page->mapping->host, - bio_page->index, fio))) + page_folio(bio_page)->index, fio))) __submit_merged_bio(io); alloc_new: if (io->bio == NULL) { io->bio = __bio_alloc(fio, BIO_MAX_VECS); f2fs_set_bio_crypt_ctx(io->bio, fio->page->mapping->host, - bio_page->index, fio, GFP_NOIO); + page_folio(bio_page)->index, fio, GFP_NOIO); io->fio = *fio; } diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 5fcc952107e9..1136539a57a8 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -842,6 +842,7 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page, struct f2fs_dentry_block *dentry_blk; unsigned int bit_pos; int slots = GET_DENTRY_SLOTS(le16_to_cpu(dentry->name_len)); + pgoff_t index = page_folio(page)->index; int i; f2fs_update_time(F2FS_I_SB(dir), REQ_TIME); @@ -867,7 +868,7 @@ void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page, set_page_dirty(page); if (bit_pos == NR_DENTRY_IN_BLOCK && - !f2fs_truncate_hole(dir, page->index, page->index + 1)) { + !f2fs_truncate_hole(dir, index, index + 1)) { f2fs_clear_page_cache_dirty_tag(page_folio(page)); clear_page_dirty_for_io(page); ClearPageUptodate(page); diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index 5d7e4c7f5969..b2d5c3ef8e24 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -174,7 +174,8 @@ bool f2fs_inode_chksum_verify(struct f2fs_sb_info *sbi, struct page *page) if (provided != calculated) f2fs_warn(sbi, "checksum invalid, nid = %lu, ino_of_node = %x, %x vs. %x", - page->index, ino_of_node(page), provided, calculated); + page_folio(page)->index, ino_of_node(page), + provided, calculated); return provided == calculated; } diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index f5e5abce695b..59b13ff243fa 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -919,7 +919,7 @@ static int truncate_node(struct dnode_of_data *dn) clear_node_page_dirty(dn->node_page); set_sbi_flag(sbi, SBI_IS_DIRTY); - index = dn->node_page->index; + index = page_folio(dn->node_page)->index; f2fs_put_page(dn->node_page, 1); invalidate_mapping_pages(NODE_MAPPING(sbi), @@ -1869,7 +1869,7 @@ continue_unlock: } if (!ret && atomic && !marked) { f2fs_debug(sbi, "Retry to write fsync mark: ino=%u, idx=%lx", - ino, last_page->index); + ino, page_folio(last_page)->index); lock_page(last_page); f2fs_wait_on_page_writeback(last_page, NODE, true, true); set_page_dirty(last_page); diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index e3cc2706f544..314be0d7816d 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -3552,7 +3552,8 @@ static int __get_segment_type_6(struct f2fs_io_info *fio) if (file_is_cold(inode) || f2fs_need_compress_data(inode)) return CURSEG_COLD_DATA; - type = __get_age_segment_type(inode, fio->page->index); + type = __get_age_segment_type(inode, + page_folio(fio->page)->index); if (type != NO_CHECK_TYPE) return type; From fccaa81de87e80b1809906f7e438e5766fbdc172 Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Wed, 4 Sep 2024 08:33:06 -0700 Subject: [PATCH 075/107] f2fs: prevent atomic file from being dirtied before commit Keep atomic file clean while updating and make it dirtied during commit in order to avoid unnecessary and excessive inode updates in the previous fix. Fixes: 4bf78322346f ("f2fs: mark inode dirty for FI_ATOMIC_COMMITTED flag") Signed-off-by: Daeho Jeong Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 2 +- fs/f2fs/inode.c | 5 +++++ fs/f2fs/segment.c | 8 ++++++++ 3 files changed, 14 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 460c05191c28..8d850192dfdf 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -808,6 +808,7 @@ enum { FI_ALIGNED_WRITE, /* enable aligned write */ FI_COW_FILE, /* indicate COW file */ FI_ATOMIC_COMMITTED, /* indicate atomic commit completed except disk sync */ + FI_ATOMIC_DIRTIED, /* indicate atomic file is dirtied */ FI_ATOMIC_REPLACE, /* indicate atomic replace */ FI_OPENED_FILE, /* indicate file has been opened */ FI_MAX, /* max flag, never be used */ @@ -3067,7 +3068,6 @@ static inline void __mark_inode_dirty_flag(struct inode *inode, case FI_INLINE_DOTS: case FI_PIN_FILE: case FI_COMPRESS_RELEASED: - case FI_ATOMIC_COMMITTED: f2fs_mark_inode_dirty_sync(inode, true); } } diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c index b2d5c3ef8e24..1ed86df343a5 100644 --- a/fs/f2fs/inode.c +++ b/fs/f2fs/inode.c @@ -34,6 +34,11 @@ void f2fs_mark_inode_dirty_sync(struct inode *inode, bool sync) if (f2fs_inode_dirtied(inode, sync)) return; + if (f2fs_is_atomic_file(inode)) { + set_inode_flag(inode, FI_ATOMIC_DIRTIED); + return; + } + mark_inode_dirty_sync(inode); } diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 314be0d7816d..5ec0d9f655f8 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -199,6 +199,10 @@ void f2fs_abort_atomic_write(struct inode *inode, bool clean) clear_inode_flag(inode, FI_ATOMIC_COMMITTED); clear_inode_flag(inode, FI_ATOMIC_REPLACE); clear_inode_flag(inode, FI_ATOMIC_FILE); + if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) { + clear_inode_flag(inode, FI_ATOMIC_DIRTIED); + f2fs_mark_inode_dirty_sync(inode, true); + } stat_dec_atomic_inode(inode); F2FS_I(inode)->atomic_write_task = NULL; @@ -366,6 +370,10 @@ out: } else { sbi->committed_atomic_block += fi->atomic_write_cnt; set_inode_flag(inode, FI_ATOMIC_COMMITTED); + if (is_inode_flag_set(inode, FI_ATOMIC_DIRTIED)) { + clear_inode_flag(inode, FI_ATOMIC_DIRTIED); + f2fs_mark_inode_dirty_sync(inode, true); + } } __complete_revoke_list(inode, &revoke_list, ret ? true : false); From 884ee6dc85b959bc152f15bca80c30f06069e6c4 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Fri, 6 Sep 2024 14:27:24 +0800 Subject: [PATCH 076/107] f2fs: get rid of online repaire on corrupted directory syzbot reports a f2fs bug as below: kernel BUG at fs/f2fs/inode.c:896! RIP: 0010:f2fs_evict_inode+0x1598/0x15c0 fs/f2fs/inode.c:896 Call Trace: evict+0x532/0x950 fs/inode.c:704 dispose_list fs/inode.c:747 [inline] evict_inodes+0x5f9/0x690 fs/inode.c:797 generic_shutdown_super+0x9d/0x2d0 fs/super.c:627 kill_block_super+0x44/0x90 fs/super.c:1696 kill_f2fs_super+0x344/0x690 fs/f2fs/super.c:4898 deactivate_locked_super+0xc4/0x130 fs/super.c:473 cleanup_mnt+0x41f/0x4b0 fs/namespace.c:1373 task_work_run+0x24f/0x310 kernel/task_work.c:228 ptrace_notify+0x2d2/0x380 kernel/signal.c:2402 ptrace_report_syscall include/linux/ptrace.h:415 [inline] ptrace_report_syscall_exit include/linux/ptrace.h:477 [inline] syscall_exit_work+0xc6/0x190 kernel/entry/common.c:173 syscall_exit_to_user_mode_prepare kernel/entry/common.c:200 [inline] __syscall_exit_to_user_mode_work kernel/entry/common.c:205 [inline] syscall_exit_to_user_mode+0x279/0x370 kernel/entry/common.c:218 do_syscall_64+0x100/0x230 arch/x86/entry/common.c:89 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0010:f2fs_evict_inode+0x1598/0x15c0 fs/f2fs/inode.c:896 Online repaire on corrupted directory in f2fs_lookup() can generate dirty data/meta while racing w/ readonly remount, it may leave dirty inode after filesystem becomes readonly, however, checkpoint() will skips flushing dirty inode in a state of readonly mode, result in above panic. Let's get rid of online repaire in f2fs_lookup(), and leave the work to fsck.f2fs. Fixes: 510022a85839 ("f2fs: add F2FS_INLINE_DOTS to recover missing dot dentries") Reported-by: syzbot+ebea2790904673d7c618@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/000000000000a7b20f061ff2d56a@google.com Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 11 ------- fs/f2fs/namei.c | 68 ----------------------------------------- include/linux/f2fs_fs.h | 2 +- 3 files changed, 1 insertion(+), 80 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 8d850192dfdf..0abb7bb39296 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -790,7 +790,6 @@ enum { FI_NEED_IPU, /* used for ipu per file */ FI_ATOMIC_FILE, /* indicate atomic file */ FI_DATA_EXIST, /* indicate data exists */ - FI_INLINE_DOTS, /* indicate inline dot dentries */ FI_SKIP_WRITES, /* should skip data page writeback */ FI_OPU_WRITE, /* used for opu per file */ FI_DIRTY_FILE, /* indicate regular/symlink has dirty pages */ @@ -3065,7 +3064,6 @@ static inline void __mark_inode_dirty_flag(struct inode *inode, return; fallthrough; case FI_DATA_EXIST: - case FI_INLINE_DOTS: case FI_PIN_FILE: case FI_COMPRESS_RELEASED: f2fs_mark_inode_dirty_sync(inode, true); @@ -3189,8 +3187,6 @@ static inline void get_inline_info(struct inode *inode, struct f2fs_inode *ri) set_bit(FI_INLINE_DENTRY, fi->flags); if (ri->i_inline & F2FS_DATA_EXIST) set_bit(FI_DATA_EXIST, fi->flags); - if (ri->i_inline & F2FS_INLINE_DOTS) - set_bit(FI_INLINE_DOTS, fi->flags); if (ri->i_inline & F2FS_EXTRA_ATTR) set_bit(FI_EXTRA_ATTR, fi->flags); if (ri->i_inline & F2FS_PIN_FILE) @@ -3211,8 +3207,6 @@ static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri) ri->i_inline |= F2FS_INLINE_DENTRY; if (is_inode_flag_set(inode, FI_DATA_EXIST)) ri->i_inline |= F2FS_DATA_EXIST; - if (is_inode_flag_set(inode, FI_INLINE_DOTS)) - ri->i_inline |= F2FS_INLINE_DOTS; if (is_inode_flag_set(inode, FI_EXTRA_ATTR)) ri->i_inline |= F2FS_EXTRA_ATTR; if (is_inode_flag_set(inode, FI_PIN_FILE)) @@ -3293,11 +3287,6 @@ static inline int f2fs_exist_data(struct inode *inode) return is_inode_flag_set(inode, FI_DATA_EXIST); } -static inline int f2fs_has_inline_dots(struct inode *inode) -{ - return is_inode_flag_set(inode, FI_INLINE_DOTS); -} - static inline int f2fs_is_mmap_file(struct inode *inode) { return is_inode_flag_set(inode, FI_MMAP_FILE); diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c index 38b4750475db..57d46e1439de 100644 --- a/fs/f2fs/namei.c +++ b/fs/f2fs/namei.c @@ -457,62 +457,6 @@ struct dentry *f2fs_get_parent(struct dentry *child) return d_obtain_alias(f2fs_iget(child->d_sb, ino)); } -static int __recover_dot_dentries(struct inode *dir, nid_t pino) -{ - struct f2fs_sb_info *sbi = F2FS_I_SB(dir); - struct qstr dot = QSTR_INIT(".", 1); - struct f2fs_dir_entry *de; - struct page *page; - int err = 0; - - if (f2fs_readonly(sbi->sb)) { - f2fs_info(sbi, "skip recovering inline_dots inode (ino:%lu, pino:%u) in readonly mountpoint", - dir->i_ino, pino); - return 0; - } - - if (!S_ISDIR(dir->i_mode)) { - f2fs_err(sbi, "inconsistent inode status, skip recovering inline_dots inode (ino:%lu, i_mode:%u, pino:%u)", - dir->i_ino, dir->i_mode, pino); - set_sbi_flag(sbi, SBI_NEED_FSCK); - return -ENOTDIR; - } - - err = f2fs_dquot_initialize(dir); - if (err) - return err; - - f2fs_balance_fs(sbi, true); - - f2fs_lock_op(sbi); - - de = f2fs_find_entry(dir, &dot, &page); - if (de) { - f2fs_put_page(page, 0); - } else if (IS_ERR(page)) { - err = PTR_ERR(page); - goto out; - } else { - err = f2fs_do_add_link(dir, &dot, NULL, dir->i_ino, S_IFDIR); - if (err) - goto out; - } - - de = f2fs_find_entry(dir, &dotdot_name, &page); - if (de) - f2fs_put_page(page, 0); - else if (IS_ERR(page)) - err = PTR_ERR(page); - else - err = f2fs_do_add_link(dir, &dotdot_name, NULL, pino, S_IFDIR); -out: - if (!err) - clear_inode_flag(dir, FI_INLINE_DOTS); - - f2fs_unlock_op(sbi); - return err; -} - static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry, unsigned int flags) { @@ -522,7 +466,6 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry, struct dentry *new; nid_t ino = -1; int err = 0; - unsigned int root_ino = F2FS_ROOT_INO(F2FS_I_SB(dir)); struct f2fs_filename fname; trace_f2fs_lookup_start(dir, dentry, flags); @@ -558,17 +501,6 @@ static struct dentry *f2fs_lookup(struct inode *dir, struct dentry *dentry, goto out; } - if ((dir->i_ino == root_ino) && f2fs_has_inline_dots(dir)) { - err = __recover_dot_dentries(dir, root_ino); - if (err) - goto out_iput; - } - - if (f2fs_has_inline_dots(inode)) { - err = __recover_dot_dentries(inode, dir->i_ino); - if (err) - goto out_iput; - } if (IS_ENCRYPTED(dir) && (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode)) && !fscrypt_has_permitted_context(dir, inode)) { diff --git a/include/linux/f2fs_fs.h b/include/linux/f2fs_fs.h index f17f89a259e2..b0b821edfd97 100644 --- a/include/linux/f2fs_fs.h +++ b/include/linux/f2fs_fs.h @@ -278,7 +278,7 @@ struct node_footer { #define F2FS_INLINE_DATA 0x02 /* file inline data flag */ #define F2FS_INLINE_DENTRY 0x04 /* file inline dentry flag */ #define F2FS_DATA_EXIST 0x08 /* file inline data exist flag */ -#define F2FS_INLINE_DOTS 0x10 /* file having implicit dot dentries */ +#define F2FS_INLINE_DOTS 0x10 /* file having implicit dot dentries (obsolete) */ #define F2FS_EXTRA_ATTR 0x20 /* file having extra attribute */ #define F2FS_PIN_FILE 0x40 /* file should not be gced */ #define F2FS_COMPRESS_RELEASED 0x80 /* file released compressed blocks */ From bfe5c02654261bfb8bd9cb174a67f3279ea99e58 Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Wed, 4 Sep 2024 11:20:47 +0800 Subject: [PATCH 077/107] f2fs: fix to check atomic_file in f2fs ioctl interfaces Some f2fs ioctl interfaces like f2fs_ioc_set_pin_file(), f2fs_move_file_range(), and f2fs_defragment_range() missed to check atomic_write status, which may cause potential race issue, fix it. Cc: stable@vger.kernel.org Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/file.c | 13 ++++++++++++- 1 file changed, 12 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index a8d153eb0a95..99903eafa7fe 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -2710,7 +2710,8 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi, (range->start + range->len) >> PAGE_SHIFT, DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE)); - if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED)) { + if (is_inode_flag_set(inode, FI_COMPRESS_RELEASED) || + f2fs_is_atomic_file(inode)) { err = -EINVAL; goto unlock_out; } @@ -2943,6 +2944,11 @@ static int f2fs_move_file_range(struct file *file_in, loff_t pos_in, goto out_unlock; } + if (f2fs_is_atomic_file(src) || f2fs_is_atomic_file(dst)) { + ret = -EINVAL; + goto out_unlock; + } + ret = -EINVAL; if (pos_in + len > src->i_size || pos_in + len < pos_in) goto out_unlock; @@ -3326,6 +3332,11 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg) inode_lock(inode); + if (f2fs_is_atomic_file(inode)) { + ret = -EINVAL; + goto out; + } + if (!pin) { clear_inode_flag(inode, FI_PIN_FILE); f2fs_i_gc_failures_write(inode, 0); From 0638a3197c194bed837c157c3574685e36febc7b Mon Sep 17 00:00:00 2001 From: Daejun Park Date: Thu, 5 Sep 2024 14:24:33 +0900 Subject: [PATCH 078/107] f2fs: avoid unused block when dio write in LFS mode This patch addresses the problem that when using LFS mode, unused blocks may occur in f2fs_map_blocks() during block allocation for dio writes. If a new section is allocated during block allocation, it will not be included in the map struct by map_is_mergeable() if the LBA of the allocated block is not contiguous. However, the block already allocated in this process will remain unused due to the LFS mode. This patch avoids the possibility of unused blocks by escaping f2fs_map_blocks() when allocating the last block in a section. Signed-off-by: Daejun Park Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/data.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index c6d688208f8b..b94cf6eea2f9 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -1712,6 +1712,14 @@ skip: dn.ofs_in_node = end_offset; } + if (flag == F2FS_GET_BLOCK_DIO && f2fs_lfs_mode(sbi) && + map->m_may_create) { + /* the next block to be allocated may not be contiguous. */ + if (GET_SEGOFF_FROM_SEG0(sbi, blkaddr) % BLKS_PER_SEC(sbi) == + CAP_BLKS_PER_SEC(sbi) - 1) + goto sync_out; + } + if (pgofs >= end) goto sync_out; else if (dn.ofs_in_node < end_offset) From 5062b5bed4323275f2f89bc185c6a28d62cfcfd5 Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:40 -0700 Subject: [PATCH 079/107] f2fs: make BG GC more aggressive for zoned devices Since we don't have any GC on device side for zoned devices, need more aggressive BG GC. So, tune the parameters for that. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 20 ++++++++++++++++++-- fs/f2fs/gc.c | 25 +++++++++++++++++++++---- fs/f2fs/gc.h | 21 +++++++++++++++++++++ fs/f2fs/super.c | 5 +++++ 4 files changed, 65 insertions(+), 6 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 0abb7bb39296..5031e7fa287a 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -2879,13 +2879,26 @@ static inline bool is_inflight_io(struct f2fs_sb_info *sbi, int type) return false; } +static inline bool is_inflight_read_io(struct f2fs_sb_info *sbi) +{ + return get_pages(sbi, F2FS_RD_DATA) || get_pages(sbi, F2FS_DIO_READ); +} + static inline bool is_idle(struct f2fs_sb_info *sbi, int type) { + bool zoned_gc = (type == GC_TIME && + F2FS_HAS_FEATURE(sbi, F2FS_FEATURE_BLKZONED)); + if (sbi->gc_mode == GC_URGENT_HIGH) return true; - if (is_inflight_io(sbi, type)) - return false; + if (zoned_gc) { + if (is_inflight_read_io(sbi)) + return false; + } else { + if (is_inflight_io(sbi, type)) + return false; + } if (sbi->gc_mode == GC_URGENT_MID) return true; @@ -2894,6 +2907,9 @@ static inline bool is_idle(struct f2fs_sb_info *sbi, int type) (type == DISCARD_TIME || type == GC_TIME)) return true; + if (zoned_gc) + return true; + return f2fs_time_over(sbi, type); } diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 724bbcb447d3..46e3bc26b78a 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -116,7 +116,17 @@ static int gc_thread_func(void *data) goto next; } - if (has_enough_invalid_blocks(sbi)) + if (f2fs_sb_has_blkzoned(sbi)) { + if (has_enough_free_blocks(sbi, LIMIT_NO_ZONED_GC)) { + wait_ms = gc_th->no_gc_sleep_time; + f2fs_up_write(&sbi->gc_lock); + goto next; + } + if (wait_ms == gc_th->no_gc_sleep_time) + wait_ms = gc_th->max_sleep_time; + } + + if (need_to_boost_gc(sbi)) decrease_sleep_time(gc_th, &wait_ms); else increase_sleep_time(gc_th, &wait_ms); @@ -179,9 +189,16 @@ int f2fs_start_gc_thread(struct f2fs_sb_info *sbi) return -ENOMEM; gc_th->urgent_sleep_time = DEF_GC_THREAD_URGENT_SLEEP_TIME; - gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME; - gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME; - gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME; + + if (f2fs_sb_has_blkzoned(sbi)) { + gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME_ZONED; + gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME_ZONED; + gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME_ZONED; + } else { + gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME; + gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME; + gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME; + } gc_th->gc_wake = false; diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index a8ea3301b815..55c4ba73362e 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -15,6 +15,11 @@ #define DEF_GC_THREAD_MAX_SLEEP_TIME 60000 #define DEF_GC_THREAD_NOGC_SLEEP_TIME 300000 /* wait 5 min */ +/* GC sleep parameters for zoned deivces */ +#define DEF_GC_THREAD_MIN_SLEEP_TIME_ZONED 10 +#define DEF_GC_THREAD_MAX_SLEEP_TIME_ZONED 20 +#define DEF_GC_THREAD_NOGC_SLEEP_TIME_ZONED 60000 + /* choose candidates from sections which has age of more than 7 days */ #define DEF_GC_THREAD_AGE_THRESHOLD (60 * 60 * 24 * 7) #define DEF_GC_THREAD_CANDIDATE_RATIO 20 /* select 20% oldest sections as candidates */ @@ -25,6 +30,9 @@ #define LIMIT_INVALID_BLOCK 40 /* percentage over total user space */ #define LIMIT_FREE_BLOCK 40 /* percentage over invalid + free space */ +#define LIMIT_NO_ZONED_GC 60 /* percentage over total user space of no gc for zoned devices */ +#define LIMIT_BOOST_ZONED_GC 25 /* percentage over total user space of boosted gc for zoned devices */ + #define DEF_GC_FAILED_PINNED_FILES 2048 #define MAX_GC_FAILED_PINNED_FILES USHRT_MAX @@ -152,6 +160,12 @@ static inline void decrease_sleep_time(struct f2fs_gc_kthread *gc_th, *wait -= min_time; } +static inline bool has_enough_free_blocks(struct f2fs_sb_info *sbi, + unsigned int limit_perc) +{ + return free_sections(sbi) > ((sbi->total_sections * limit_perc) / 100); +} + static inline bool has_enough_invalid_blocks(struct f2fs_sb_info *sbi) { block_t user_block_count = sbi->user_block_count; @@ -167,3 +181,10 @@ static inline bool has_enough_invalid_blocks(struct f2fs_sb_info *sbi) free_user_blocks(sbi) < limit_free_user_blocks(invalid_user_blocks)); } + +static inline bool need_to_boost_gc(struct f2fs_sb_info *sbi) +{ + if (f2fs_sb_has_blkzoned(sbi)) + return !has_enough_free_blocks(sbi, LIMIT_BOOST_ZONED_GC); + return has_enough_invalid_blocks(sbi); +} diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 1c9316788a16..69aee4c3d97f 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -706,6 +706,11 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount) if (!strcmp(name, "on")) { F2FS_OPTION(sbi).bggc_mode = BGGC_MODE_ON; } else if (!strcmp(name, "off")) { + if (f2fs_sb_has_blkzoned(sbi)) { + f2fs_warn(sbi, "zoned devices need bggc"); + kfree(name); + return -EINVAL; + } F2FS_OPTION(sbi).bggc_mode = BGGC_MODE_OFF; } else if (!strcmp(name, "sync")) { F2FS_OPTION(sbi).bggc_mode = BGGC_MODE_SYNC; From 8c890c4c60342719526520133fb1b6f69f196ab8 Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:41 -0700 Subject: [PATCH 080/107] f2fs: introduce migration_window_granularity We can control the scanning window granularity for GC migration. For more frequent scanning and GC on zoned devices, we need a fine grained control knob for it. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 8 +++++++ fs/f2fs/f2fs.h | 2 ++ fs/f2fs/gc.c | 31 +++++++++++++++++-------- fs/f2fs/gc.h | 1 + fs/f2fs/super.c | 2 ++ fs/f2fs/sysfs.c | 7 ++++++ 6 files changed, 41 insertions(+), 10 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index ca61ffcae126..e6fa55037daf 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -783,3 +783,11 @@ Description: The zone UFS we are currently using consists of two parts: blkzone_alloc_policy = 1 Only allow writing to sequential zones blkzone_alloc_policy = 2 Prioritize writing to conventional zones ======================== ========================================= + +What: /sys/fs/f2fs//migration_window_granularity +Date: September 2024 +Contact: "Daeho Jeong" +Description: Controls migration window granularity of garbage collection on large + section. it can control the scanning window granularity for GC migration + in a unit of segment, while migration_granularity controls the number + of segments which can be migrated at the same turn. diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 5031e7fa287a..635318af9c43 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1696,6 +1696,8 @@ struct f2fs_sb_info { unsigned int max_victim_search; /* migration granularity of garbage collection, unit: segment */ unsigned int migration_granularity; + /* migration window granularity of garbage collection, unit: segment */ + unsigned int migration_window_granularity; /* * for stat information. diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 46e3bc26b78a..5cd316d2102d 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -1708,24 +1708,34 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, struct blk_plug plug; unsigned int segno = start_segno; unsigned int end_segno = start_segno + SEGS_PER_SEC(sbi); + unsigned int sec_end_segno; int seg_freed = 0, migrated = 0; unsigned char type = IS_DATASEG(get_seg_entry(sbi, segno)->type) ? SUM_TYPE_DATA : SUM_TYPE_NODE; unsigned char data_type = (type == SUM_TYPE_DATA) ? DATA : NODE; int submitted = 0; - if (__is_large_section(sbi)) - end_segno = rounddown(end_segno, SEGS_PER_SEC(sbi)); + if (__is_large_section(sbi)) { + sec_end_segno = rounddown(end_segno, SEGS_PER_SEC(sbi)); - /* - * zone-capacity can be less than zone-size in zoned devices, - * resulting in less than expected usable segments in the zone, - * calculate the end segno in the zone which can be garbage collected - */ - if (f2fs_sb_has_blkzoned(sbi)) - end_segno -= SEGS_PER_SEC(sbi) - + /* + * zone-capacity can be less than zone-size in zoned devices, + * resulting in less than expected usable segments in the zone, + * calculate the end segno in the zone which can be garbage + * collected + */ + if (f2fs_sb_has_blkzoned(sbi)) + sec_end_segno -= SEGS_PER_SEC(sbi) - f2fs_usable_segs_in_sec(sbi, segno); + if (gc_type == BG_GC) + end_segno = start_segno + + sbi->migration_window_granularity; + + if (end_segno > sec_end_segno) + end_segno = sec_end_segno; + } + sanity_check_seg_type(sbi, get_seg_entry(sbi, segno)->type); /* readahead multi ssa blocks those have contiguous address */ @@ -1803,7 +1813,8 @@ freed: if (__is_large_section(sbi)) sbi->next_victim_seg[gc_type] = - (segno + 1 < end_segno) ? segno + 1 : NULL_SEGNO; + (segno + 1 < sec_end_segno) ? + segno + 1 : NULL_SEGNO; skip: f2fs_put_page(sum_page, 0); } diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index 55c4ba73362e..245f93663745 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -32,6 +32,7 @@ #define LIMIT_NO_ZONED_GC 60 /* percentage over total user space of no gc for zoned devices */ #define LIMIT_BOOST_ZONED_GC 25 /* percentage over total user space of boosted gc for zoned devices */ +#define DEF_MIGRATION_WINDOW_GRANULARITY_ZONED 3 #define DEF_GC_FAILED_PINNED_FILES 2048 #define MAX_GC_FAILED_PINNED_FILES USHRT_MAX diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index 69aee4c3d97f..e68f9c38af7d 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -3807,6 +3807,8 @@ static void init_sb_info(struct f2fs_sb_info *sbi) sbi->next_victim_seg[FG_GC] = NULL_SEGNO; sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH; sbi->migration_granularity = SEGS_PER_SEC(sbi); + sbi->migration_window_granularity = f2fs_sb_has_blkzoned(sbi) ? + DEF_MIGRATION_WINDOW_GRANULARITY_ZONED : SEGS_PER_SEC(sbi); sbi->seq_file_ra_mul = MIN_RA_MUL; sbi->max_fragment_chunk = DEF_FRAGMENT_SIZE; sbi->max_fragment_hole = DEF_FRAGMENT_SIZE; diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index 654a541dcc45..02b35edc9e1d 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -564,6 +564,11 @@ out: return -EINVAL; } + if (!strcmp(a->attr.name, "migration_window_granularity")) { + if (t == 0 || t > SEGS_PER_SEC(sbi)) + return -EINVAL; + } + if (!strcmp(a->attr.name, "gc_urgent")) { if (t == 0) { sbi->gc_mode = GC_NORMAL; @@ -1013,6 +1018,7 @@ F2FS_SBI_RW_ATTR(gc_pin_file_thresh, gc_pin_file_threshold); F2FS_SBI_RW_ATTR(gc_reclaimed_segments, gc_reclaimed_segs); F2FS_SBI_GENERAL_RW_ATTR(max_victim_search); F2FS_SBI_GENERAL_RW_ATTR(migration_granularity); +F2FS_SBI_GENERAL_RW_ATTR(migration_window_granularity); F2FS_SBI_GENERAL_RW_ATTR(dir_level); #ifdef CONFIG_F2FS_IOSTAT F2FS_SBI_GENERAL_RW_ATTR(iostat_enable); @@ -1154,6 +1160,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(min_ssr_sections), ATTR_LIST(max_victim_search), ATTR_LIST(migration_granularity), + ATTR_LIST(migration_window_granularity), ATTR_LIST(dir_level), ATTR_LIST(ram_thresh), ATTR_LIST(ra_nid_pages), From 4cdca5a904b1c72a0ba6ac51a121351128c02fd8 Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:42 -0700 Subject: [PATCH 081/107] f2fs: add reserved_segments sysfs node For the fine tuning of GC behavior, add reserved_segments sysfs node. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 6 ++++++ fs/f2fs/sysfs.c | 2 ++ 2 files changed, 8 insertions(+) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index e6fa55037daf..6527a14082d2 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -791,3 +791,9 @@ Description: Controls migration window granularity of garbage collection on larg section. it can control the scanning window granularity for GC migration in a unit of segment, while migration_granularity controls the number of segments which can be migrated at the same turn. + +What: /sys/fs/f2fs//reserved_segments +Date: September 2024 +Contact: "Daeho Jeong" +Description: In order to fine tune GC behavior, we can control the number of + reserved segments. diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index 02b35edc9e1d..8eedba8ae710 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -986,6 +986,7 @@ SM_INFO_GENERAL_RW_ATTR(min_fsync_blocks); SM_INFO_GENERAL_RW_ATTR(min_seq_blocks); SM_INFO_GENERAL_RW_ATTR(min_hot_blocks); SM_INFO_GENERAL_RW_ATTR(min_ssr_sections); +SM_INFO_GENERAL_RW_ATTR(reserved_segments); /* DCC_INFO ATTR */ DCC_INFO_RW_ATTR(max_small_discards, max_discards); @@ -1158,6 +1159,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(min_seq_blocks), ATTR_LIST(min_hot_blocks), ATTR_LIST(min_ssr_sections), + ATTR_LIST(reserved_segments), ATTR_LIST(max_victim_search), ATTR_LIST(migration_granularity), ATTR_LIST(migration_window_granularity), From 2223fe652f759649ae1d520e47e5f06727c0acbd Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:43 -0700 Subject: [PATCH 082/107] f2fs: increase BG GC migration window granularity when boosted for zoned devices Need bigger BG GC migration window granularity when free section is running low. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/gc.c | 12 ++++++++++-- fs/f2fs/gc.h | 1 + 2 files changed, 11 insertions(+), 2 deletions(-) diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 5cd316d2102d..9a3d3994cf2b 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -1728,10 +1728,18 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, sec_end_segno -= SEGS_PER_SEC(sbi) - f2fs_usable_segs_in_sec(sbi, segno); - if (gc_type == BG_GC) - end_segno = start_segno + + if (gc_type == BG_GC) { + unsigned int window_granularity = sbi->migration_window_granularity; + if (f2fs_sb_has_blkzoned(sbi) && + !has_enough_free_blocks(sbi, + LIMIT_BOOST_ZONED_GC)) + window_granularity *= BOOST_GC_MULTIPLE; + + end_segno = start_segno + window_granularity; + } + if (end_segno > sec_end_segno) end_segno = sec_end_segno; } diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index 245f93663745..78abeebd68b5 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -33,6 +33,7 @@ #define LIMIT_NO_ZONED_GC 60 /* percentage over total user space of no gc for zoned devices */ #define LIMIT_BOOST_ZONED_GC 25 /* percentage over total user space of boosted gc for zoned devices */ #define DEF_MIGRATION_WINDOW_GRANULARITY_ZONED 3 +#define BOOST_GC_MULTIPLE 5 #define DEF_GC_FAILED_PINNED_FILES 2048 #define MAX_GC_FAILED_PINNED_FILES USHRT_MAX From 9748c2ddea4a3f46a498bff4cf2bf9a5629e3f8b Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:44 -0700 Subject: [PATCH 083/107] f2fs: do FG_GC when GC boosting is required for zoned devices Under low free section count, we need to use FG_GC instead of BG_GC to recover free sections. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 1 + fs/f2fs/gc.c | 24 +++++++++++++++++------- 2 files changed, 18 insertions(+), 7 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 635318af9c43..3e4371d51042 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -1300,6 +1300,7 @@ struct f2fs_gc_control { bool no_bg_gc; /* check the space and stop bg_gc */ bool should_migrate_blocks; /* should migrate blocks */ bool err_gc_skipped; /* return EAGAIN if GC skipped */ + bool one_time; /* require one time GC in one migration unit */ unsigned int nr_free_secs; /* # of free sections to do GC */ }; diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 9a3d3994cf2b..a59fec64eccf 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -81,6 +81,8 @@ static int gc_thread_func(void *data) continue; } + gc_control.one_time = false; + /* * [GC triggering condition] * 0. GC is not conducted currently. @@ -126,15 +128,19 @@ static int gc_thread_func(void *data) wait_ms = gc_th->max_sleep_time; } - if (need_to_boost_gc(sbi)) + if (need_to_boost_gc(sbi)) { decrease_sleep_time(gc_th, &wait_ms); - else + if (f2fs_sb_has_blkzoned(sbi)) + gc_control.one_time = true; + } else { increase_sleep_time(gc_th, &wait_ms); + } do_gc: stat_inc_gc_call_count(sbi, foreground ? FOREGROUND : BACKGROUND); - sync_mode = F2FS_OPTION(sbi).bggc_mode == BGGC_MODE_SYNC; + sync_mode = (F2FS_OPTION(sbi).bggc_mode == BGGC_MODE_SYNC) || + gc_control.one_time; /* foreground GC was been triggered via f2fs_balance_fs() */ if (foreground) @@ -1701,7 +1707,7 @@ static int __get_victim(struct f2fs_sb_info *sbi, unsigned int *victim, static int do_garbage_collect(struct f2fs_sb_info *sbi, unsigned int start_segno, struct gc_inode_list *gc_list, int gc_type, - bool force_migrate) + bool force_migrate, bool one_time) { struct page *sum_page; struct f2fs_summary_block *sum; @@ -1728,7 +1734,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, sec_end_segno -= SEGS_PER_SEC(sbi) - f2fs_usable_segs_in_sec(sbi, segno); - if (gc_type == BG_GC) { + if (gc_type == BG_GC || one_time) { unsigned int window_granularity = sbi->migration_window_granularity; @@ -1911,7 +1917,8 @@ retry: } seg_freed = do_garbage_collect(sbi, segno, &gc_list, gc_type, - gc_control->should_migrate_blocks); + gc_control->should_migrate_blocks, + gc_control->one_time); if (seg_freed < 0) goto stop; @@ -1922,6 +1929,9 @@ retry: total_sec_freed++; } + if (gc_control->one_time) + goto stop; + if (gc_type == FG_GC) { sbi->cur_victim_sec = NULL_SEGNO; @@ -2047,7 +2057,7 @@ int f2fs_gc_range(struct f2fs_sb_info *sbi, }; do_garbage_collect(sbi, segno, &gc_list, FG_GC, - dry_run_sections == 0); + dry_run_sections == 0, false); put_gc_inode(&gc_list); if (!dry_run && get_valid_blocks(sbi, segno, true)) From 9a481a1c16f4653c3414d996c3603eb42926e28b Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:45 -0700 Subject: [PATCH 084/107] f2fs: create gc_no_zoned_gc_percent and gc_boost_zoned_gc_percent Added control knobs for gc_no_zoned_gc_percent and gc_boost_zoned_gc_percent. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 14 ++++++++++++++ fs/f2fs/gc.c | 12 +++++++++--- fs/f2fs/gc.h | 4 ++++ fs/f2fs/sysfs.c | 4 ++++ 4 files changed, 31 insertions(+), 3 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index 6527a14082d2..0be1a53925c4 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -797,3 +797,17 @@ Date: September 2024 Contact: "Daeho Jeong" Description: In order to fine tune GC behavior, we can control the number of reserved segments. + +What: /sys/fs/f2fs//gc_no_zoned_gc_percent +Date: September 2024 +Contact: "Daeho Jeong" +Description: If the percentage of free sections over total sections is above this + number, F2FS do not garbage collection for zoned devices through the + background GC thread. the default number is "60". + +What: /sys/fs/f2fs//gc_boost_zoned_gc_percent +Date: September 2024 +Contact: "Daeho Jeong" +Description: If the percentage of free sections over total sections is under this + number, F2FS boosts garbage collection for zoned devices through the + background GC thread. the default number is "25". diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index a59fec64eccf..319b4b80ba72 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -119,7 +119,8 @@ static int gc_thread_func(void *data) } if (f2fs_sb_has_blkzoned(sbi)) { - if (has_enough_free_blocks(sbi, LIMIT_NO_ZONED_GC)) { + if (has_enough_free_blocks(sbi, + gc_th->no_zoned_gc_percent)) { wait_ms = gc_th->no_gc_sleep_time; f2fs_up_write(&sbi->gc_lock); goto next; @@ -200,10 +201,14 @@ int f2fs_start_gc_thread(struct f2fs_sb_info *sbi) gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME_ZONED; gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME_ZONED; gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME_ZONED; + gc_th->no_zoned_gc_percent = LIMIT_NO_ZONED_GC; + gc_th->boost_zoned_gc_percent = LIMIT_BOOST_ZONED_GC; } else { gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME; gc_th->max_sleep_time = DEF_GC_THREAD_MAX_SLEEP_TIME; gc_th->no_gc_sleep_time = DEF_GC_THREAD_NOGC_SLEEP_TIME; + gc_th->no_zoned_gc_percent = 0; + gc_th->boost_zoned_gc_percent = 0; } gc_th->gc_wake = false; @@ -1740,8 +1745,9 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, if (f2fs_sb_has_blkzoned(sbi) && !has_enough_free_blocks(sbi, - LIMIT_BOOST_ZONED_GC)) - window_granularity *= BOOST_GC_MULTIPLE; + sbi->gc_thread->boost_zoned_gc_percent)) + window_granularity *= + BOOST_GC_MULTIPLE; end_segno = start_segno + window_granularity; } diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index 78abeebd68b5..c063310c8dde 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -61,6 +61,10 @@ struct f2fs_gc_kthread { * caller of f2fs_balance_fs() * will wait on this wait queue. */ + + /* for gc control for zoned devices */ + unsigned int no_zoned_gc_percent; + unsigned int boost_zoned_gc_percent; }; struct gc_inode_list { diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index 8eedba8ae710..a65c356a2769 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -977,6 +977,8 @@ GC_THREAD_RW_ATTR(gc_urgent_sleep_time, urgent_sleep_time); GC_THREAD_RW_ATTR(gc_min_sleep_time, min_sleep_time); GC_THREAD_RW_ATTR(gc_max_sleep_time, max_sleep_time); GC_THREAD_RW_ATTR(gc_no_gc_sleep_time, no_gc_sleep_time); +GC_THREAD_RW_ATTR(gc_no_zoned_gc_percent, no_zoned_gc_percent); +GC_THREAD_RW_ATTR(gc_boost_zoned_gc_percent, boost_zoned_gc_percent); /* SM_INFO ATTR */ SM_INFO_RW_ATTR(reclaim_segments, rec_prefree_segments); @@ -1137,6 +1139,8 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(gc_min_sleep_time), ATTR_LIST(gc_max_sleep_time), ATTR_LIST(gc_no_gc_sleep_time), + ATTR_LIST(gc_no_zoned_gc_percent), + ATTR_LIST(gc_boost_zoned_gc_percent), ATTR_LIST(gc_idle), ATTR_LIST(gc_urgent), ATTR_LIST(reclaim_segments), From e791d00bd06cb2d3a10afa43e57f6364931d0ada Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Mon, 9 Sep 2024 15:19:46 -0700 Subject: [PATCH 085/107] f2fs: add valid block ratio not to do excessive GC for one time GC We need to introduce a valid block ratio threshold not to trigger excessive GC for zoned deivces. The initial value of it is 95%. So, F2FS will stop the thread from intiating GC for sections having valid blocks exceeding the ratio. Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- Documentation/ABI/testing/sysfs-fs-f2fs | 8 ++++++++ fs/f2fs/f2fs.h | 2 +- fs/f2fs/gc.c | 16 ++++++++++++---- fs/f2fs/gc.h | 2 ++ fs/f2fs/segment.c | 6 ++++-- fs/f2fs/segment.h | 1 + fs/f2fs/sysfs.c | 2 ++ 7 files changed, 30 insertions(+), 7 deletions(-) diff --git a/Documentation/ABI/testing/sysfs-fs-f2fs b/Documentation/ABI/testing/sysfs-fs-f2fs index 0be1a53925c4..fdedf1ea944b 100644 --- a/Documentation/ABI/testing/sysfs-fs-f2fs +++ b/Documentation/ABI/testing/sysfs-fs-f2fs @@ -811,3 +811,11 @@ Contact: "Daeho Jeong" Description: If the percentage of free sections over total sections is under this number, F2FS boosts garbage collection for zoned devices through the background GC thread. the default number is "25". + +What: /sys/fs/f2fs//gc_valid_thresh_ratio +Date: September 2024 +Contact: "Daeho Jeong" +Description: It controls the valid block ratio threshold not to trigger excessive GC + for zoned deivces. The initial value of it is 95(%). F2FS will stop the + background GC thread from intiating GC for sections having valid blocks + exceeding the ratio. diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 3e4371d51042..9a226d36a36c 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3922,7 +3922,7 @@ void f2fs_destroy_garbage_collection_cache(void); /* victim selection function for cleaning and SSR */ int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result, int gc_type, int type, char alloc_mode, - unsigned long long age); + unsigned long long age, bool one_time); /* * recovery.c diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 319b4b80ba72..d3e3104c4492 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -196,6 +196,7 @@ int f2fs_start_gc_thread(struct f2fs_sb_info *sbi) return -ENOMEM; gc_th->urgent_sleep_time = DEF_GC_THREAD_URGENT_SLEEP_TIME; + gc_th->valid_thresh_ratio = DEF_GC_THREAD_VALID_THRESH_RATIO; if (f2fs_sb_has_blkzoned(sbi)) { gc_th->min_sleep_time = DEF_GC_THREAD_MIN_SLEEP_TIME_ZONED; @@ -396,6 +397,11 @@ static inline unsigned int get_gc_cost(struct f2fs_sb_info *sbi, if (p->alloc_mode == SSR) return get_seg_entry(sbi, segno)->ckpt_valid_blocks; + if (p->one_time_gc && (get_valid_blocks(sbi, segno, true) >= + CAP_BLKS_PER_SEC(sbi) * sbi->gc_thread->valid_thresh_ratio / + 100)) + return UINT_MAX; + /* alloc_mode == LFS */ if (p->gc_mode == GC_GREEDY) return get_valid_blocks(sbi, segno, true); @@ -770,7 +776,7 @@ static int f2fs_gc_pinned_control(struct inode *inode, int gc_type, */ int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result, int gc_type, int type, char alloc_mode, - unsigned long long age) + unsigned long long age, bool one_time) { struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); struct sit_info *sm = SIT_I(sbi); @@ -787,6 +793,7 @@ int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result, p.alloc_mode = alloc_mode; p.age = age; p.age_threshold = sbi->am.age_threshold; + p.one_time_gc = one_time; retry: select_policy(sbi, gc_type, type, &p); @@ -1698,13 +1705,14 @@ next_step: } static int __get_victim(struct f2fs_sb_info *sbi, unsigned int *victim, - int gc_type) + int gc_type, bool one_time) { struct sit_info *sit_i = SIT_I(sbi); int ret; down_write(&sit_i->sentry_lock); - ret = f2fs_get_victim(sbi, victim, gc_type, NO_CHECK_TYPE, LFS, 0); + ret = f2fs_get_victim(sbi, victim, gc_type, NO_CHECK_TYPE, + LFS, 0, one_time); up_write(&sit_i->sentry_lock); return ret; } @@ -1911,7 +1919,7 @@ gc_more: goto stop; } retry: - ret = __get_victim(sbi, &segno, gc_type); + ret = __get_victim(sbi, &segno, gc_type, gc_control->one_time); if (ret) { /* allow to search victim from sections has pinned data */ if (ret == -ENODATA && gc_type == FG_GC && diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index c063310c8dde..2914b678bf8f 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -25,6 +25,7 @@ #define DEF_GC_THREAD_CANDIDATE_RATIO 20 /* select 20% oldest sections as candidates */ #define DEF_GC_THREAD_MAX_CANDIDATE_COUNT 10 /* select at most 10 sections as candidates */ #define DEF_GC_THREAD_AGE_WEIGHT 60 /* age weight */ +#define DEF_GC_THREAD_VALID_THRESH_RATIO 95 /* do not GC over 95% valid block ratio for one time GC */ #define DEFAULT_ACCURACY_CLASS 10000 /* accuracy class */ #define LIMIT_INVALID_BLOCK 40 /* percentage over total user space */ @@ -65,6 +66,7 @@ struct f2fs_gc_kthread { /* for gc control for zoned devices */ unsigned int no_zoned_gc_percent; unsigned int boost_zoned_gc_percent; + unsigned int valid_thresh_ratio; }; struct gc_inode_list { diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 5ec0d9f655f8..112d58d566d4 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -3090,7 +3090,8 @@ static int get_ssr_segment(struct f2fs_sb_info *sbi, int type, sanity_check_seg_type(sbi, seg_type); /* f2fs_need_SSR() already forces to do this */ - if (!f2fs_get_victim(sbi, &segno, BG_GC, seg_type, alloc_mode, age)) { + if (!f2fs_get_victim(sbi, &segno, BG_GC, seg_type, + alloc_mode, age, false)) { curseg->next_segno = segno; return 1; } @@ -3117,7 +3118,8 @@ static int get_ssr_segment(struct f2fs_sb_info *sbi, int type, for (; cnt-- > 0; reversed ? i-- : i++) { if (i == seg_type) continue; - if (!f2fs_get_victim(sbi, &segno, BG_GC, i, alloc_mode, age)) { + if (!f2fs_get_victim(sbi, &segno, BG_GC, i, + alloc_mode, age, false)) { curseg->next_segno = segno; return 1; } diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h index bfc01a521cb9..43db2d3e8c85 100644 --- a/fs/f2fs/segment.h +++ b/fs/f2fs/segment.h @@ -188,6 +188,7 @@ struct victim_sel_policy { unsigned int min_segno; /* segment # having min. cost */ unsigned long long age; /* mtime of GCed section*/ unsigned long long age_threshold;/* age threshold */ + bool one_time_gc; /* one time GC */ }; struct seg_entry { diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index a65c356a2769..d18931b7b094 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -979,6 +979,7 @@ GC_THREAD_RW_ATTR(gc_max_sleep_time, max_sleep_time); GC_THREAD_RW_ATTR(gc_no_gc_sleep_time, no_gc_sleep_time); GC_THREAD_RW_ATTR(gc_no_zoned_gc_percent, no_zoned_gc_percent); GC_THREAD_RW_ATTR(gc_boost_zoned_gc_percent, boost_zoned_gc_percent); +GC_THREAD_RW_ATTR(gc_valid_thresh_ratio, valid_thresh_ratio); /* SM_INFO ATTR */ SM_INFO_RW_ATTR(reclaim_segments, rec_prefree_segments); @@ -1141,6 +1142,7 @@ static struct attribute *f2fs_attrs[] = { ATTR_LIST(gc_no_gc_sleep_time), ATTR_LIST(gc_no_zoned_gc_percent), ATTR_LIST(gc_boost_zoned_gc_percent), + ATTR_LIST(gc_valid_thresh_ratio), ATTR_LIST(gc_idle), ATTR_LIST(gc_urgent), ATTR_LIST(reclaim_segments), From 930c6ab93492c4b15436524e704950b364b2930c Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 10 Sep 2024 11:07:13 +0800 Subject: [PATCH 086/107] f2fs: fix to don't set SB_RDONLY in f2fs_handle_critical_error() syzbot reports a f2fs bug as below: ------------[ cut here ]------------ WARNING: CPU: 1 PID: 58 at kernel/rcu/sync.c:177 rcu_sync_dtor+0xcd/0x180 kernel/rcu/sync.c:177 CPU: 1 UID: 0 PID: 58 Comm: kworker/1:2 Not tainted 6.10.0-syzkaller-12562-g1722389b0d86 #0 Workqueue: events destroy_super_work RIP: 0010:rcu_sync_dtor+0xcd/0x180 kernel/rcu/sync.c:177 Call Trace: percpu_free_rwsem+0x41/0x80 kernel/locking/percpu-rwsem.c:42 destroy_super_work+0xec/0x130 fs/super.c:282 process_one_work kernel/workqueue.c:3231 [inline] process_scheduled_works+0xa2c/0x1830 kernel/workqueue.c:3312 worker_thread+0x86d/0xd40 kernel/workqueue.c:3390 kthread+0x2f0/0x390 kernel/kthread.c:389 ret_from_fork+0x4b/0x80 arch/x86/kernel/process.c:147 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244 As Christian Brauner pointed out [1]: the root cause is f2fs sets SB_RDONLY flag in internal function, rather than setting the flag covered w/ sb->s_umount semaphore via remount procedure, then below race condition causes this bug: - freeze_super() - sb_wait_write(sb, SB_FREEZE_WRITE) - sb_wait_write(sb, SB_FREEZE_PAGEFAULT) - sb_wait_write(sb, SB_FREEZE_FS) - f2fs_handle_critical_error - sb->s_flags |= SB_RDONLY - thaw_super - thaw_super_locked - sb_rdonly() is true, so it skips sb_freeze_unlock(sb, SB_FREEZE_FS) - deactivate_locked_super Since f2fs has almost the same logic as ext4 [2] when handling critical error in filesystem if it mounts w/ errors=remount-ro option: - set CP_ERROR_FLAG flag which indicates filesystem is stopped - record errors to superblock - set SB_RDONLY falg Once we set CP_ERROR_FLAG flag, all writable interfaces can detect the flag and stop any further updates on filesystem. So, it is safe to not set SB_RDONLY flag, let's remove the logic and keep in line w/ ext4 [3]. [1] https://lore.kernel.org/all/20240729-himbeeren-funknetz-96e62f9c7aee@brauner [2] https://lore.kernel.org/all/20240729132721.hxih6ehigadqf7wx@quack3 [3] https://lore.kernel.org/linux-ext4/20240805201241.27286-1-jack@suse.cz Fixes: b62e71be2110 ("f2fs: support errors=remount-ro|continue|panic mountoption") Reported-by: syzbot+20d7e439f76bbbd863a7@syzkaller.appspotmail.com Closes: https://lore.kernel.org/all/000000000000b90a8e061e21d12f@google.com/ Cc: Jan Kara Cc: Christian Brauner Signed-off-by: Chao Yu Reviewed-by: Christian Brauner Signed-off-by: Jaegeuk Kim --- fs/f2fs/super.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index e68f9c38af7d..fc2c586c7619 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -4199,12 +4199,14 @@ void f2fs_handle_critical_error(struct f2fs_sb_info *sbi, unsigned char reason, } f2fs_warn(sbi, "Remounting filesystem read-only"); + /* - * Make sure updated value of ->s_mount_flags will be visible before - * ->s_flags update + * We have already set CP_ERROR_FLAG flag to stop all updates + * to filesystem, so it doesn't need to set SB_RDONLY flag here + * because the flag should be set covered w/ sb->s_umount semaphore + * via remount procedure, otherwise, it will confuse code like + * freeze_super() which will lead to deadlocks and other problems. */ - smp_wmb(); - sb->s_flags |= SB_RDONLY; } static void f2fs_record_error_work(struct work_struct *work) From 65a6ce4726c27b45600303f06496fef46d00b57f Mon Sep 17 00:00:00 2001 From: Chao Yu Date: Tue, 10 Sep 2024 09:16:19 +0800 Subject: [PATCH 087/107] f2fs: fix to don't panic system for no free segment fault injection f2fs: fix to don't panic system for no free segment fault injection syzbot reports a f2fs bug as below: F2FS-fs (loop0): inject no free segment in get_new_segment of __allocate_new_segment+0x1ce/0x940 fs/f2fs/segment.c:3167 F2FS-fs (loop0): Stopped filesystem due to reason: 7 ------------[ cut here ]------------ kernel BUG at fs/f2fs/segment.c:2748! CPU: 0 UID: 0 PID: 5109 Comm: syz-executor304 Not tainted 6.11.0-rc6-syzkaller-00363-g89f5e14d05b4 #0 RIP: 0010:get_new_segment fs/f2fs/segment.c:2748 [inline] RIP: 0010:new_curseg+0x1f61/0x1f70 fs/f2fs/segment.c:2836 Call Trace: __allocate_new_segment+0x1ce/0x940 fs/f2fs/segment.c:3167 f2fs_allocate_new_section fs/f2fs/segment.c:3181 [inline] f2fs_allocate_pinning_section+0xfa/0x4e0 fs/f2fs/segment.c:3195 f2fs_expand_inode_data+0x5d6/0xbb0 fs/f2fs/file.c:1799 f2fs_fallocate+0x448/0x960 fs/f2fs/file.c:1903 vfs_fallocate+0x553/0x6c0 fs/open.c:334 do_vfs_ioctl+0x2592/0x2e50 fs/ioctl.c:886 __do_sys_ioctl fs/ioctl.c:905 [inline] __se_sys_ioctl+0x81/0x170 fs/ioctl.c:893 do_syscall_x64 arch/x86/entry/common.c:52 [inline] do_syscall_64+0xf3/0x230 arch/x86/entry/common.c:83 entry_SYSCALL_64_after_hwframe+0x77/0x7f RIP: 0010:get_new_segment fs/f2fs/segment.c:2748 [inline] RIP: 0010:new_curseg+0x1f61/0x1f70 fs/f2fs/segment.c:2836 The root cause is when we inject no free segment fault into f2fs, we should not panic system, fix it. Fixes: 8b10d3653735 ("f2fs: introduce FAULT_NO_SEGMENT") Reported-by: syzbot+341e5f32ebafbb46b81c@syzkaller.appspotmail.com Closes: https://lore.kernel.org/linux-f2fs-devel/000000000000f0ee5b0621ab694b@google.com Signed-off-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/segment.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index 112d58d566d4..b7c1ca51ea32 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -2730,6 +2730,7 @@ find_other_zone: MAIN_SECS(sbi)); if (secno >= MAIN_SECS(sbi)) { ret = -ENOSPC; + f2fs_bug_on(sbi, 1); goto out_unlock; } } @@ -2740,6 +2741,7 @@ find_other_zone: MAIN_SECS(sbi)); if (secno >= MAIN_SECS(sbi)) { ret = -ENOSPC; + f2fs_bug_on(sbi, 1); goto out_unlock; } } @@ -2781,10 +2783,8 @@ got_it: out_unlock: spin_unlock(&free_i->segmap_lock); - if (ret == -ENOSPC) { + if (ret == -ENOSPC) f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_NO_SEGMENT); - f2fs_bug_on(sbi, 1); - } return ret; } From 2af583afcf9d54b988b68d1033101e73edd72a9b Mon Sep 17 00:00:00 2001 From: liuderong Date: Wed, 11 Sep 2024 10:40:20 +0800 Subject: [PATCH 088/107] f2fs: remove unused parameters Remove unused parameter segno from f2fs_usable_segs_in_sec. Signed-off-by: liuderong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/f2fs.h | 3 +-- fs/f2fs/gc.c | 6 +++--- fs/f2fs/segment.c | 3 +-- fs/f2fs/segment.h | 4 ++-- 4 files changed, 7 insertions(+), 9 deletions(-) diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 9a226d36a36c..33f5449dc22d 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -3780,8 +3780,7 @@ void f2fs_destroy_segment_manager_caches(void); int f2fs_rw_hint_to_seg_type(struct f2fs_sb_info *sbi, enum rw_hint hint); enum rw_hint f2fs_io_type_to_rw_hint(struct f2fs_sb_info *sbi, enum page_type type, enum temp_type temp); -unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi, - unsigned int segno); +unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi); unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi, unsigned int segno); diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index d3e3104c4492..2903c4c4a76c 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -368,7 +368,7 @@ static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno) unsigned char age = 0; unsigned char u; unsigned int i; - unsigned int usable_segs_per_sec = f2fs_usable_segs_in_sec(sbi, segno); + unsigned int usable_segs_per_sec = f2fs_usable_segs_in_sec(sbi); for (i = 0; i < usable_segs_per_sec; i++) mtime += get_seg_entry(sbi, start + i)->mtime; @@ -1745,7 +1745,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, */ if (f2fs_sb_has_blkzoned(sbi)) sec_end_segno -= SEGS_PER_SEC(sbi) - - f2fs_usable_segs_in_sec(sbi, segno); + f2fs_usable_segs_in_sec(sbi); if (gc_type == BG_GC || one_time) { unsigned int window_granularity = @@ -1938,7 +1938,7 @@ retry: total_freed += seg_freed; - if (seg_freed == f2fs_usable_segs_in_sec(sbi, segno)) { + if (seg_freed == f2fs_usable_segs_in_sec(sbi)) { sec_freed++; total_sec_freed++; } diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index b7c1ca51ea32..1766254279d2 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -5422,8 +5422,7 @@ unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi, return BLKS_PER_SEG(sbi); } -unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi, - unsigned int segno) +unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi) { if (f2fs_sb_has_blkzoned(sbi)) return CAP_SEGS_PER_SEC(sbi); diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h index 43db2d3e8c85..71adb4a43bec 100644 --- a/fs/f2fs/segment.h +++ b/fs/f2fs/segment.h @@ -431,7 +431,7 @@ static inline void __set_free(struct f2fs_sb_info *sbi, unsigned int segno) unsigned int secno = GET_SEC_FROM_SEG(sbi, segno); unsigned int start_segno = GET_SEG_FROM_SEC(sbi, secno); unsigned int next; - unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi, segno); + unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi); spin_lock(&free_i->segmap_lock); clear_bit(segno, free_i->free_segmap); @@ -465,7 +465,7 @@ static inline void __set_test_and_free(struct f2fs_sb_info *sbi, unsigned int secno = GET_SEC_FROM_SEG(sbi, segno); unsigned int start_segno = GET_SEG_FROM_SEC(sbi, secno); unsigned int next; - unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi, segno); + unsigned int usable_segs = f2fs_usable_segs_in_sec(sbi); spin_lock(&free_i->segmap_lock); if (test_and_clear_bit(segno, free_i->free_segmap)) { From 433d7ce2d86d21274838c9e8c796f4232cd13cdb Mon Sep 17 00:00:00 2001 From: Andrii Nakryiko Date: Tue, 6 Aug 2024 15:38:12 -0700 Subject: [PATCH 089/107] security,bpf: constify struct path in bpf_token_create() LSM hook There is no reason why struct path pointer shouldn't be const-qualified when being passed into bpf_token_create() LSM hook. Add that const. Acked-by: Paul Moore (LSM/SELinux) Suggested-by: Al Viro Signed-off-by: Andrii Nakryiko --- include/linux/lsm_hook_defs.h | 2 +- include/linux/security.h | 4 ++-- security/security.c | 2 +- security/selinux/hooks.c | 2 +- 4 files changed, 5 insertions(+), 5 deletions(-) diff --git a/include/linux/lsm_hook_defs.h b/include/linux/lsm_hook_defs.h index 855db460e08b..462b55378241 100644 --- a/include/linux/lsm_hook_defs.h +++ b/include/linux/lsm_hook_defs.h @@ -431,7 +431,7 @@ LSM_HOOK(int, 0, bpf_prog_load, struct bpf_prog *prog, union bpf_attr *attr, struct bpf_token *token) LSM_HOOK(void, LSM_RET_VOID, bpf_prog_free, struct bpf_prog *prog) LSM_HOOK(int, 0, bpf_token_create, struct bpf_token *token, union bpf_attr *attr, - struct path *path) + const struct path *path) LSM_HOOK(void, LSM_RET_VOID, bpf_token_free, struct bpf_token *token) LSM_HOOK(int, 0, bpf_token_cmd, const struct bpf_token *token, enum bpf_cmd cmd) LSM_HOOK(int, 0, bpf_token_capable, const struct bpf_token *token, int cap) diff --git a/include/linux/security.h b/include/linux/security.h index 1390f1efb4f0..31523a2c71c4 100644 --- a/include/linux/security.h +++ b/include/linux/security.h @@ -2137,7 +2137,7 @@ extern int security_bpf_prog_load(struct bpf_prog *prog, union bpf_attr *attr, struct bpf_token *token); extern void security_bpf_prog_free(struct bpf_prog *prog); extern int security_bpf_token_create(struct bpf_token *token, union bpf_attr *attr, - struct path *path); + const struct path *path); extern void security_bpf_token_free(struct bpf_token *token); extern int security_bpf_token_cmd(const struct bpf_token *token, enum bpf_cmd cmd); extern int security_bpf_token_capable(const struct bpf_token *token, int cap); @@ -2177,7 +2177,7 @@ static inline void security_bpf_prog_free(struct bpf_prog *prog) { } static inline int security_bpf_token_create(struct bpf_token *token, union bpf_attr *attr, - struct path *path) + const struct path *path) { return 0; } diff --git a/security/security.c b/security/security.c index 8cee5b6c6e6d..d8d0b67ced25 100644 --- a/security/security.c +++ b/security/security.c @@ -5510,7 +5510,7 @@ int security_bpf_prog_load(struct bpf_prog *prog, union bpf_attr *attr, * Return: Returns 0 on success, error on failure. */ int security_bpf_token_create(struct bpf_token *token, union bpf_attr *attr, - struct path *path) + const struct path *path) { return call_int_hook(bpf_token_create, token, attr, path); } diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index 55c78c318ccd..0eec141a8f37 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -6965,7 +6965,7 @@ static void selinux_bpf_prog_free(struct bpf_prog *prog) } static int selinux_bpf_token_create(struct bpf_token *token, union bpf_attr *attr, - struct path *path) + const struct path *path) { struct bpf_security_struct *bpfsec; From 37d3dd663f7485bf3e444f40abee3c68f53158cb Mon Sep 17 00:00:00 2001 From: Al Viro Date: Tue, 30 Jul 2024 01:16:21 -0400 Subject: [PATCH 090/107] bpf: convert bpf_token_create() to CLASS(fd, ...) Keep file reference through the entire thing, don't bother with grabbing struct path reference and while we are at it, don't confuse the hell out of readers by random mix of path.dentry->d_sb and path.mnt->mnt_sb uses - these two are equal, so just put one of those into a local variable and use that. Reviewed-by: Christian Brauner Acked-by: Andrii Nakryiko Signed-off-by: Al Viro Signed-off-by: Andrii Nakryiko --- kernel/bpf/token.c | 65 ++++++++++++++++------------------------------ 1 file changed, 23 insertions(+), 42 deletions(-) diff --git a/kernel/bpf/token.c b/kernel/bpf/token.c index 9b92cb886d49..dcbec1a0dfb3 100644 --- a/kernel/bpf/token.c +++ b/kernel/bpf/token.c @@ -116,67 +116,52 @@ int bpf_token_create(union bpf_attr *attr) struct user_namespace *userns; struct inode *inode; struct file *file; + CLASS(fd, f)(attr->token_create.bpffs_fd); struct path path; - struct fd f; + struct super_block *sb; umode_t mode; int err, fd; - f = fdget(attr->token_create.bpffs_fd); - if (!fd_file(f)) + if (fd_empty(f)) return -EBADF; path = fd_file(f)->f_path; - path_get(&path); - fdput(f); + sb = path.dentry->d_sb; - if (path.dentry != path.mnt->mnt_sb->s_root) { - err = -EINVAL; - goto out_path; - } - if (path.mnt->mnt_sb->s_op != &bpf_super_ops) { - err = -EINVAL; - goto out_path; - } + if (path.dentry != sb->s_root) + return -EINVAL; + if (sb->s_op != &bpf_super_ops) + return -EINVAL; err = path_permission(&path, MAY_ACCESS); if (err) - goto out_path; + return err; - userns = path.dentry->d_sb->s_user_ns; + userns = sb->s_user_ns; /* * Enforce that creators of BPF tokens are in the same user * namespace as the BPF FS instance. This makes reasoning about * permissions a lot easier and we can always relax this later. */ - if (current_user_ns() != userns) { - err = -EPERM; - goto out_path; - } - if (!ns_capable(userns, CAP_BPF)) { - err = -EPERM; - goto out_path; - } + if (current_user_ns() != userns) + return -EPERM; + if (!ns_capable(userns, CAP_BPF)) + return -EPERM; /* Creating BPF token in init_user_ns doesn't make much sense. */ - if (current_user_ns() == &init_user_ns) { - err = -EOPNOTSUPP; - goto out_path; - } + if (current_user_ns() == &init_user_ns) + return -EOPNOTSUPP; - mnt_opts = path.dentry->d_sb->s_fs_info; + mnt_opts = sb->s_fs_info; if (mnt_opts->delegate_cmds == 0 && mnt_opts->delegate_maps == 0 && mnt_opts->delegate_progs == 0 && - mnt_opts->delegate_attachs == 0) { - err = -ENOENT; /* no BPF token delegation is set up */ - goto out_path; - } + mnt_opts->delegate_attachs == 0) + return -ENOENT; /* no BPF token delegation is set up */ mode = S_IFREG | ((S_IRUSR | S_IWUSR) & ~current_umask()); - inode = bpf_get_inode(path.mnt->mnt_sb, NULL, mode); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - goto out_path; - } + inode = bpf_get_inode(sb, NULL, mode); + if (IS_ERR(inode)) + return PTR_ERR(inode); inode->i_op = &bpf_token_iops; inode->i_fop = &bpf_token_fops; @@ -185,8 +170,7 @@ int bpf_token_create(union bpf_attr *attr) file = alloc_file_pseudo(inode, path.mnt, BPF_TOKEN_INODE_NAME, O_RDWR, &bpf_token_fops); if (IS_ERR(file)) { iput(inode); - err = PTR_ERR(file); - goto out_path; + return PTR_ERR(file); } token = kzalloc(sizeof(*token), GFP_USER); @@ -218,15 +202,12 @@ int bpf_token_create(union bpf_attr *attr) file->private_data = token; fd_install(fd, file); - path_put(&path); return fd; out_token: bpf_token_free(token); out_file: fput(file); -out_path: - path_put(&path); return err; } From 5cc69a27abfa91abbb39fc584f82d6c867b60f47 Mon Sep 17 00:00:00 2001 From: Daeho Jeong Date: Thu, 12 Sep 2024 09:59:58 -0700 Subject: [PATCH 091/107] f2fs: forcibly migrate to secure space for zoned device file pinning We need to migrate data blocks even though it is full to secure space for zoned device file pinning. Fixes: 9703d69d9d15 ("f2fs: support file pinning for zoned devices") Signed-off-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/gc.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 2903c4c4a76c..9322a7200e31 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -2070,8 +2070,7 @@ int f2fs_gc_range(struct f2fs_sb_info *sbi, .iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS), }; - do_garbage_collect(sbi, segno, &gc_list, FG_GC, - dry_run_sections == 0, false); + do_garbage_collect(sbi, segno, &gc_list, FG_GC, true, false); put_gc_inode(&gc_list); if (!dry_run && get_valid_blocks(sbi, segno, true)) From fba27cf005038a5fedf001b49636594819dbb1d5 Mon Sep 17 00:00:00 2001 From: Yuezhang Mo Date: Fri, 12 Jan 2024 10:32:58 +0800 Subject: [PATCH 092/107] exfat: drop ->i_size_ondisk ->i_size_ondisk is no longer used by exfat_write_begin() after commit(11a347fb6cef exfat: change to get file size from DataLength), drop it. Signed-off-by: Yuezhang Mo Signed-off-by: Namjae Jeon --- fs/exfat/exfat_fs.h | 10 +++++----- fs/exfat/file.c | 8 ++------ fs/exfat/inode.c | 17 +++++------------ fs/exfat/namei.c | 1 - fs/exfat/super.c | 1 - 5 files changed, 12 insertions(+), 25 deletions(-) diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h index ecc5db952deb..77cc62b026d3 100644 --- a/fs/exfat/exfat_fs.h +++ b/fs/exfat/exfat_fs.h @@ -309,11 +309,6 @@ struct exfat_inode_info { /* for avoiding the race between alloc and free */ unsigned int cache_valid_id; - /* - * NOTE: i_size_ondisk is 64bits, so must hold ->inode_lock to access. - * physically allocated size. - */ - loff_t i_size_ondisk; /* block-aligned i_size (used in cont_write_begin) */ loff_t i_size_aligned; /* on-disk position of directory entry or 0 */ @@ -417,6 +412,11 @@ static inline bool is_valid_cluster(struct exfat_sb_info *sbi, return clus >= EXFAT_FIRST_CLUSTER && clus < sbi->num_clusters; } +static inline loff_t exfat_ondisk_size(const struct inode *inode) +{ + return ((loff_t)inode->i_blocks) << 9; +} + /* super.c */ int exfat_set_volume_dirty(struct super_block *sb); int exfat_clear_volume_dirty(struct super_block *sb); diff --git a/fs/exfat/file.c b/fs/exfat/file.c index e19469e88000..fd4a8a170c1f 100644 --- a/fs/exfat/file.c +++ b/fs/exfat/file.c @@ -29,7 +29,7 @@ static int exfat_cont_expand(struct inode *inode, loff_t size) if (ret) return ret; - num_clusters = EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi); + num_clusters = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi); new_num_clusters = EXFAT_B_TO_CLU_ROUND_UP(size, sbi); if (new_num_clusters == num_clusters) @@ -75,7 +75,6 @@ out: i_size_write(inode, size); ei->i_size_aligned = round_up(size, sb->s_blocksize); - ei->i_size_ondisk = ei->i_size_aligned; inode->i_blocks = round_up(size, sbi->cluster_size) >> 9; mark_inode_dirty(inode); @@ -159,7 +158,7 @@ int __exfat_truncate(struct inode *inode) exfat_set_volume_dirty(sb); num_clusters_new = EXFAT_B_TO_CLU_ROUND_UP(i_size_read(inode), sbi); - num_clusters_phys = EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi); + num_clusters_phys = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi); exfat_chain_set(&clu, ei->start_clu, num_clusters_phys, ei->flags); @@ -270,9 +269,6 @@ write_size: aligned_size++; } - if (ei->i_size_ondisk > i_size_read(inode)) - ei->i_size_ondisk = aligned_size; - if (ei->i_size_aligned > i_size_read(inode)) ei->i_size_aligned = aligned_size; mutex_unlock(&sbi->s_lock); diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c index 05f0e07b01d0..ed737ef288d2 100644 --- a/fs/exfat/inode.c +++ b/fs/exfat/inode.c @@ -130,11 +130,9 @@ static int exfat_map_cluster(struct inode *inode, unsigned int clu_offset, struct exfat_sb_info *sbi = EXFAT_SB(sb); struct exfat_inode_info *ei = EXFAT_I(inode); unsigned int local_clu_offset = clu_offset; - unsigned int num_to_be_allocated = 0, num_clusters = 0; + unsigned int num_to_be_allocated = 0, num_clusters; - if (ei->i_size_ondisk > 0) - num_clusters = - EXFAT_B_TO_CLU_ROUND_UP(ei->i_size_ondisk, sbi); + num_clusters = EXFAT_B_TO_CLU(exfat_ondisk_size(inode), sbi); if (clu_offset >= num_clusters) num_to_be_allocated = clu_offset - num_clusters + 1; @@ -268,10 +266,10 @@ static int exfat_map_new_buffer(struct exfat_inode_info *ei, set_buffer_new(bh); /* - * Adjust i_size_aligned if i_size_ondisk is bigger than it. + * Adjust i_size_aligned if ondisk_size is bigger than it. */ - if (ei->i_size_ondisk > ei->i_size_aligned) - ei->i_size_aligned = ei->i_size_ondisk; + if (exfat_ondisk_size(&ei->vfs_inode) > ei->i_size_aligned) + ei->i_size_aligned = exfat_ondisk_size(&ei->vfs_inode); return 0; } @@ -317,10 +315,6 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, max_blocks = min(mapped_blocks, max_blocks); pos = EXFAT_BLK_TO_B((iblock + 1), sb); - if ((create && iblock >= last_block) || buffer_delay(bh_result)) { - if (ei->i_size_ondisk < pos) - ei->i_size_ondisk = pos; - } map_bh(bh_result, sb, phys); if (buffer_delay(bh_result)) @@ -674,7 +668,6 @@ static int exfat_fill_inode(struct inode *inode, struct exfat_dir_entry *info) } ei->i_size_aligned = size; - ei->i_size_ondisk = size; exfat_save_attr(inode, info->attr); diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 631ad9e8e32a..6313dee5c9bb 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -372,7 +372,6 @@ static int exfat_find_empty_entry(struct inode *inode, /* directory inode should be updated in here */ i_size_write(inode, size); - ei->i_size_ondisk += sbi->cluster_size; ei->i_size_aligned += sbi->cluster_size; ei->valid_size += sbi->cluster_size; ei->flags = p_dir->flags; diff --git a/fs/exfat/super.c b/fs/exfat/super.c index 323ecebe6f0e..6516edff38fc 100644 --- a/fs/exfat/super.c +++ b/fs/exfat/super.c @@ -371,7 +371,6 @@ static int exfat_read_root(struct inode *inode) inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9; ei->i_pos = ((loff_t)sbi->root_dir << 32) | 0xffffffff; ei->i_size_aligned = i_size_read(inode); - ei->i_size_ondisk = i_size_read(inode); exfat_save_attr(inode, EXFAT_ATTR_SUBDIR); ei->i_crtime = simple_inode_init_ts(inode); From 231eb762bbe8fa232e672e5a399eed16a0c0c45a Mon Sep 17 00:00:00 2001 From: Yuezhang Mo Date: Tue, 17 Sep 2024 14:58:56 +0900 Subject: [PATCH 093/107] exfat: do not fallback to buffered write After commit(11a347fb6cef exfat: change to get file size from DataLength), the remaining area or hole had been filled with zeros before calling exfat_direct_IO(), so there is no need to fallback to buffered write, and ->i_size_aligned is no longer needed, drop it. Signed-off-by: Yuezhang Mo Signed-off-by: Namjae Jeon --- fs/exfat/exfat_fs.h | 2 -- fs/exfat/file.c | 11 ------- fs/exfat/inode.c | 74 +++++++++------------------------------------ fs/exfat/namei.c | 1 - fs/exfat/super.c | 1 - 5 files changed, 15 insertions(+), 74 deletions(-) diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h index 77cc62b026d3..5be214288a05 100644 --- a/fs/exfat/exfat_fs.h +++ b/fs/exfat/exfat_fs.h @@ -309,8 +309,6 @@ struct exfat_inode_info { /* for avoiding the race between alloc and free */ unsigned int cache_valid_id; - /* block-aligned i_size (used in cont_write_begin) */ - loff_t i_size_aligned; /* on-disk position of directory entry or 0 */ loff_t i_pos; loff_t valid_size; diff --git a/fs/exfat/file.c b/fs/exfat/file.c index fd4a8a170c1f..860c1f3b2cd7 100644 --- a/fs/exfat/file.c +++ b/fs/exfat/file.c @@ -74,7 +74,6 @@ out: /* Expanded range not zeroed, do not update valid_size */ i_size_write(inode, size); - ei->i_size_aligned = round_up(size, sb->s_blocksize); inode->i_blocks = round_up(size, sbi->cluster_size) >> 9; mark_inode_dirty(inode); @@ -244,8 +243,6 @@ void exfat_truncate(struct inode *inode) struct super_block *sb = inode->i_sb; struct exfat_sb_info *sbi = EXFAT_SB(sb); struct exfat_inode_info *ei = EXFAT_I(inode); - unsigned int blocksize = i_blocksize(inode); - loff_t aligned_size; int err; mutex_lock(&sbi->s_lock); @@ -263,14 +260,6 @@ void exfat_truncate(struct inode *inode) inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9; write_size: - aligned_size = i_size_read(inode); - if (aligned_size & (blocksize - 1)) { - aligned_size |= (blocksize - 1); - aligned_size++; - } - - if (ei->i_size_aligned > i_size_read(inode)) - ei->i_size_aligned = aligned_size; mutex_unlock(&sbi->s_lock); } diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c index ed737ef288d2..e7a74a948be1 100644 --- a/fs/exfat/inode.c +++ b/fs/exfat/inode.c @@ -258,21 +258,6 @@ static int exfat_map_cluster(struct inode *inode, unsigned int clu_offset, return 0; } -static int exfat_map_new_buffer(struct exfat_inode_info *ei, - struct buffer_head *bh, loff_t pos) -{ - if (buffer_delay(bh) && pos > ei->i_size_aligned) - return -EIO; - set_buffer_new(bh); - - /* - * Adjust i_size_aligned if ondisk_size is bigger than it. - */ - if (exfat_ondisk_size(&ei->vfs_inode) > ei->i_size_aligned) - ei->i_size_aligned = exfat_ondisk_size(&ei->vfs_inode); - return 0; -} - static int exfat_get_block(struct inode *inode, sector_t iblock, struct buffer_head *bh_result, int create) { @@ -286,7 +271,6 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, sector_t last_block; sector_t phys = 0; sector_t valid_blks; - loff_t pos; mutex_lock(&sbi->s_lock); last_block = EXFAT_B_TO_BLK_ROUND_UP(i_size_read(inode), sb); @@ -314,8 +298,6 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, mapped_blocks = sbi->sect_per_clus - sec_offset; max_blocks = min(mapped_blocks, max_blocks); - pos = EXFAT_BLK_TO_B((iblock + 1), sb); - map_bh(bh_result, sb, phys); if (buffer_delay(bh_result)) clear_buffer_delay(bh_result); @@ -336,13 +318,7 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, } /* The area has not been written, map and mark as new. */ - err = exfat_map_new_buffer(ei, bh_result, pos); - if (err) { - exfat_fs_error(sb, - "requested for bmap out of range(pos : (%llu) > i_size_aligned(%llu)\n", - pos, ei->i_size_aligned); - goto unlock_ret; - } + set_buffer_new(bh_result); ei->valid_size = EXFAT_BLK_TO_B(iblock + max_blocks, sb); mark_inode_dirty(inode); @@ -365,7 +341,7 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, * The block has been partially written, * zero the unwritten part and map the block. */ - loff_t size, off; + loff_t size, off, pos; max_blocks = 1; @@ -376,7 +352,7 @@ static int exfat_get_block(struct inode *inode, sector_t iblock, if (!bh_result->b_folio) goto done; - pos -= sb->s_blocksize; + pos = EXFAT_BLK_TO_B(iblock, sb); size = ei->valid_size - pos; off = pos & (PAGE_SIZE - 1); @@ -463,14 +439,6 @@ static int exfat_write_end(struct file *file, struct address_space *mapping, int err; err = generic_write_end(file, mapping, pos, len, copied, folio, fsdata); - - if (ei->i_size_aligned < i_size_read(inode)) { - exfat_fs_error(inode->i_sb, - "invalid size(size(%llu) > aligned(%llu)\n", - i_size_read(inode), ei->i_size_aligned); - return -EIO; - } - if (err < len) exfat_write_failed(mapping, pos+len); @@ -498,20 +466,6 @@ static ssize_t exfat_direct_IO(struct kiocb *iocb, struct iov_iter *iter) int rw = iov_iter_rw(iter); ssize_t ret; - if (rw == WRITE) { - /* - * FIXME: blockdev_direct_IO() doesn't use ->write_begin(), - * so we need to update the ->i_size_aligned to block boundary. - * - * But we must fill the remaining area or hole by nul for - * updating ->i_size_aligned - * - * Return 0, and fallback to normal buffered write. - */ - if (EXFAT_I(inode)->i_size_aligned < size) - return 0; - } - /* * Need to use the DIO_LOCKING for avoiding the race * condition of exfat_get_block() and ->truncate(). @@ -525,8 +479,18 @@ static ssize_t exfat_direct_IO(struct kiocb *iocb, struct iov_iter *iter) } else size = pos + ret; - /* zero the unwritten part in the partially written block */ - if (rw == READ && pos < ei->valid_size && ei->valid_size < size) { + if (rw == WRITE) { + /* + * If the block had been partially written before this write, + * ->valid_size will not be updated in exfat_get_block(), + * update it here. + */ + if (ei->valid_size < size) { + ei->valid_size = size; + mark_inode_dirty(inode); + } + } else if (pos < ei->valid_size && ei->valid_size < size) { + /* zero the unwritten part in the partially written block */ iov_iter_revert(iter, size - ei->valid_size); iov_iter_zero(size - ei->valid_size, iter); } @@ -661,14 +625,6 @@ static int exfat_fill_inode(struct inode *inode, struct exfat_dir_entry *info) i_size_write(inode, size); - /* ondisk and aligned size should be aligned with block size */ - if (size & (inode->i_sb->s_blocksize - 1)) { - size |= (inode->i_sb->s_blocksize - 1); - size++; - } - - ei->i_size_aligned = size; - exfat_save_attr(inode, info->attr); inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9; diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 6313dee5c9bb..3e6c789a72c4 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -372,7 +372,6 @@ static int exfat_find_empty_entry(struct inode *inode, /* directory inode should be updated in here */ i_size_write(inode, size); - ei->i_size_aligned += sbi->cluster_size; ei->valid_size += sbi->cluster_size; ei->flags = p_dir->flags; inode->i_blocks += sbi->cluster_size >> 9; diff --git a/fs/exfat/super.c b/fs/exfat/super.c index 6516edff38fc..904583f87433 100644 --- a/fs/exfat/super.c +++ b/fs/exfat/super.c @@ -370,7 +370,6 @@ static int exfat_read_root(struct inode *inode) inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9; ei->i_pos = ((loff_t)sbi->root_dir << 32) | 0xffffffff; - ei->i_size_aligned = i_size_read(inode); exfat_save_attr(inode, EXFAT_ATTR_SUBDIR); ei->i_crtime = simple_inode_init_ts(inode); From f761fcdd289d07e8547fef7ac76c3760fc7803f2 Mon Sep 17 00:00:00 2001 From: Dongliang Cui Date: Wed, 18 Sep 2024 07:40:05 +0900 Subject: [PATCH 094/107] exfat: Implement sops->shutdown and ioctl We found that when writing a large file through buffer write, if the disk is inaccessible, exFAT does not return an error normally, which leads to the writing process not stopping properly. To easily reproduce this issue, you can follow the steps below: 1. format a device to exFAT and then mount (with a full disk erase) 2. dd if=/dev/zero of=/exfat_mount/test.img bs=1M count=8192 3. eject the device You may find that the dd process does not stop immediately and may continue for a long time. The root cause of this issue is that during buffer write process, exFAT does not need to access the disk to look up directory entries or the FAT table (whereas FAT would do) every time data is written. Instead, exFAT simply marks the buffer as dirty and returns, delegating the writeback operation to the writeback process. If the disk cannot be accessed at this time, the error will only be returned to the writeback process, and the original process will not receive the error, so it cannot be returned to the user side. When the disk cannot be accessed normally, an error should be returned to stop the writing process. Implement sops->shutdown and ioctl to shut down the file system when underlying block device is marked dead. Signed-off-by: Dongliang Cui Signed-off-by: Zhiguo Niu Signed-off-by: Namjae Jeon --- fs/exfat/exfat_fs.h | 12 ++++++++++++ fs/exfat/file.c | 21 ++++++++++++++++++++ fs/exfat/inode.c | 9 +++++++++ fs/exfat/namei.c | 15 +++++++++++++++ fs/exfat/super.c | 39 ++++++++++++++++++++++++++++++++++++++ include/uapi/linux/exfat.h | 25 ++++++++++++++++++++++++ 6 files changed, 121 insertions(+) create mode 100644 include/uapi/linux/exfat.h diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h index 5be214288a05..3cdc1de362a9 100644 --- a/fs/exfat/exfat_fs.h +++ b/fs/exfat/exfat_fs.h @@ -10,6 +10,7 @@ #include #include #include +#include #define EXFAT_ROOT_INO 1 @@ -148,6 +149,9 @@ enum { #define DIR_CACHE_SIZE \ (DIV_ROUND_UP(EXFAT_DEN_TO_B(ES_MAX_ENTRY_NUM), SECTOR_SIZE) + 1) +/* Superblock flags */ +#define EXFAT_FLAGS_SHUTDOWN 1 + struct exfat_dentry_namebuf { char *lfn; int lfnbuf_len; /* usually MAX_UNINAME_BUF_SIZE */ @@ -267,6 +271,8 @@ struct exfat_sb_info { unsigned int clu_srch_ptr; /* cluster search pointer */ unsigned int used_clusters; /* number of used clusters */ + unsigned long s_exfat_flags; /* Exfat superblock flags */ + struct mutex s_lock; /* superblock lock */ struct mutex bitmap_lock; /* bitmap lock */ struct exfat_mount_options options; @@ -331,6 +337,11 @@ static inline struct exfat_inode_info *EXFAT_I(struct inode *inode) return container_of(inode, struct exfat_inode_info, vfs_inode); } +static inline int exfat_forced_shutdown(struct super_block *sb) +{ + return test_bit(EXFAT_FLAGS_SHUTDOWN, &EXFAT_SB(sb)->s_exfat_flags); +} + /* * If ->i_mode can't hold 0222 (i.e. ATTR_RO), we use ->i_attrs to * save ATTR_RO instead of ->i_mode. @@ -459,6 +470,7 @@ int exfat_file_fsync(struct file *file, loff_t start, loff_t end, int datasync); long exfat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); long exfat_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); +int exfat_force_shutdown(struct super_block *sb, u32 flags); /* namei.c */ extern const struct dentry_operations exfat_dentry_ops; diff --git a/fs/exfat/file.c b/fs/exfat/file.c index 860c1f3b2cd7..a945ede3bf97 100644 --- a/fs/exfat/file.c +++ b/fs/exfat/file.c @@ -287,6 +287,9 @@ int exfat_setattr(struct mnt_idmap *idmap, struct dentry *dentry, unsigned int ia_valid; int error; + if (unlikely(exfat_forced_shutdown(inode->i_sb))) + return -EIO; + if ((attr->ia_valid & ATTR_SIZE) && attr->ia_size > i_size_read(inode)) { error = exfat_cont_expand(inode, attr->ia_size); @@ -470,6 +473,19 @@ static int exfat_ioctl_fitrim(struct inode *inode, unsigned long arg) return 0; } +static int exfat_ioctl_shutdown(struct super_block *sb, unsigned long arg) +{ + u32 flags; + + if (!capable(CAP_SYS_ADMIN)) + return -EPERM; + + if (get_user(flags, (__u32 __user *)arg)) + return -EFAULT; + + return exfat_force_shutdown(sb, flags); +} + long exfat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) { struct inode *inode = file_inode(filp); @@ -480,6 +496,8 @@ long exfat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) return exfat_ioctl_get_attributes(inode, user_attr); case FAT_IOCTL_SET_ATTRIBUTES: return exfat_ioctl_set_attributes(filp, user_attr); + case EXFAT_IOC_SHUTDOWN: + return exfat_ioctl_shutdown(inode->i_sb, arg); case FITRIM: return exfat_ioctl_fitrim(inode, arg); default: @@ -500,6 +518,9 @@ int exfat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync) struct inode *inode = filp->f_mapping->host; int err; + if (unlikely(exfat_forced_shutdown(inode->i_sb))) + return -EIO; + err = __generic_file_fsync(filp, start, end, datasync); if (err) return err; diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c index e7a74a948be1..d724de8f57bf 100644 --- a/fs/exfat/inode.c +++ b/fs/exfat/inode.c @@ -102,6 +102,9 @@ int exfat_write_inode(struct inode *inode, struct writeback_control *wbc) { int ret; + if (unlikely(exfat_forced_shutdown(inode->i_sb))) + return -EIO; + mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock); ret = __exfat_write_inode(inode, wbc->sync_mode == WB_SYNC_ALL); mutex_unlock(&EXFAT_SB(inode->i_sb)->s_lock); @@ -402,6 +405,9 @@ static void exfat_readahead(struct readahead_control *rac) static int exfat_writepages(struct address_space *mapping, struct writeback_control *wbc) { + if (unlikely(exfat_forced_shutdown(mapping->host->i_sb))) + return -EIO; + return mpage_writepages(mapping, wbc, exfat_get_block); } @@ -422,6 +428,9 @@ static int exfat_write_begin(struct file *file, struct address_space *mapping, { int ret; + if (unlikely(exfat_forced_shutdown(mapping->host->i_sb))) + return -EIO; + ret = block_write_begin(mapping, pos, len, foliop, exfat_get_block); if (ret < 0) diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 3e6c789a72c4..2c4c44229352 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -547,6 +547,9 @@ static int exfat_create(struct mnt_idmap *idmap, struct inode *dir, int err; loff_t size = i_size_read(dir); + if (unlikely(exfat_forced_shutdown(sb))) + return -EIO; + mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_set_volume_dirty(sb); err = exfat_add_entry(dir, dentry->d_name.name, &cdir, TYPE_FILE, @@ -770,6 +773,9 @@ static int exfat_unlink(struct inode *dir, struct dentry *dentry) struct exfat_entry_set_cache es; int entry, err = 0; + if (unlikely(exfat_forced_shutdown(sb))) + return -EIO; + mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_chain_dup(&cdir, &ei->dir); entry = ei->entry; @@ -823,6 +829,9 @@ static int exfat_mkdir(struct mnt_idmap *idmap, struct inode *dir, int err; loff_t size = i_size_read(dir); + if (unlikely(exfat_forced_shutdown(sb))) + return -EIO; + mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_set_volume_dirty(sb); err = exfat_add_entry(dir, dentry->d_name.name, &cdir, TYPE_DIR, @@ -913,6 +922,9 @@ static int exfat_rmdir(struct inode *dir, struct dentry *dentry) struct exfat_entry_set_cache es; int entry, err; + if (unlikely(exfat_forced_shutdown(sb))) + return -EIO; + mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock); exfat_chain_dup(&cdir, &ei->dir); @@ -980,6 +992,9 @@ static int exfat_rename_file(struct inode *inode, struct exfat_chain *p_dir, struct exfat_entry_set_cache old_es, new_es; int sync = IS_DIRSYNC(inode); + if (unlikely(exfat_forced_shutdown(sb))) + return -EIO; + num_new_entries = exfat_calc_num_entries(p_uniname); if (num_new_entries < 0) return num_new_entries; diff --git a/fs/exfat/super.c b/fs/exfat/super.c index 904583f87433..bd57844414aa 100644 --- a/fs/exfat/super.c +++ b/fs/exfat/super.c @@ -46,6 +46,9 @@ static int exfat_sync_fs(struct super_block *sb, int wait) struct exfat_sb_info *sbi = EXFAT_SB(sb); int err = 0; + if (unlikely(exfat_forced_shutdown(sb))) + return 0; + if (!wait) return 0; @@ -167,6 +170,41 @@ static int exfat_show_options(struct seq_file *m, struct dentry *root) return 0; } +int exfat_force_shutdown(struct super_block *sb, u32 flags) +{ + int ret; + struct exfat_sb_info *sbi = sb->s_fs_info; + struct exfat_mount_options *opts = &sbi->options; + + if (exfat_forced_shutdown(sb)) + return 0; + + switch (flags) { + case EXFAT_GOING_DOWN_DEFAULT: + case EXFAT_GOING_DOWN_FULLSYNC: + ret = bdev_freeze(sb->s_bdev); + if (ret) + return ret; + bdev_thaw(sb->s_bdev); + set_bit(EXFAT_FLAGS_SHUTDOWN, &sbi->s_exfat_flags); + break; + case EXFAT_GOING_DOWN_NOSYNC: + set_bit(EXFAT_FLAGS_SHUTDOWN, &sbi->s_exfat_flags); + break; + default: + return -EINVAL; + } + + if (opts->discard) + opts->discard = 0; + return 0; +} + +static void exfat_shutdown(struct super_block *sb) +{ + exfat_force_shutdown(sb, EXFAT_GOING_DOWN_NOSYNC); +} + static struct inode *exfat_alloc_inode(struct super_block *sb) { struct exfat_inode_info *ei; @@ -193,6 +231,7 @@ static const struct super_operations exfat_sops = { .sync_fs = exfat_sync_fs, .statfs = exfat_statfs, .show_options = exfat_show_options, + .shutdown = exfat_shutdown, }; enum { diff --git a/include/uapi/linux/exfat.h b/include/uapi/linux/exfat.h new file mode 100644 index 000000000000..46d95b16fc4b --- /dev/null +++ b/include/uapi/linux/exfat.h @@ -0,0 +1,25 @@ +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ +/* + * Copyright (C) 2024 Unisoc Technologies Co., Ltd. + */ + +#ifndef _UAPI_LINUX_EXFAT_H +#define _UAPI_LINUX_EXFAT_H +#include +#include + +/* + * exfat-specific ioctl commands + */ + +#define EXFAT_IOC_SHUTDOWN _IOR('X', 125, __u32) + +/* + * Flags used by EXFAT_IOC_SHUTDOWN + */ + +#define EXFAT_GOING_DOWN_DEFAULT 0x0 /* default with full sync */ +#define EXFAT_GOING_DOWN_FULLSYNC 0x1 /* going down with full sync*/ +#define EXFAT_GOING_DOWN_NOSYNC 0x2 /* going down */ + +#endif /* _UAPI_LINUX_EXFAT_H */ From d2b537b3e533f28e0d97293fe9293161fe8cd137 Mon Sep 17 00:00:00 2001 From: Yuezhang Mo Date: Tue, 3 Sep 2024 15:01:09 +0800 Subject: [PATCH 095/107] exfat: fix memory leak in exfat_load_bitmap() If the first directory entry in the root directory is not a bitmap directory entry, 'bh' will not be released and reassigned, which will cause a memory leak. Fixes: 1e49a94cf707 ("exfat: add bitmap operations") Cc: stable@vger.kernel.org Signed-off-by: Yuezhang Mo Reviewed-by: Aoyama Wataru Signed-off-by: Namjae Jeon --- fs/exfat/balloc.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/fs/exfat/balloc.c b/fs/exfat/balloc.c index 0356c88252bd..ce9be95c9172 100644 --- a/fs/exfat/balloc.c +++ b/fs/exfat/balloc.c @@ -91,11 +91,8 @@ int exfat_load_bitmap(struct super_block *sb) return -EIO; type = exfat_get_entry_type(ep); - if (type == TYPE_UNUSED) - break; - if (type != TYPE_BITMAP) - continue; - if (ep->dentry.bitmap.flags == 0x0) { + if (type == TYPE_BITMAP && + ep->dentry.bitmap.flags == 0x0) { int err; err = exfat_allocate_bitmap(sb, ep); @@ -103,6 +100,9 @@ int exfat_load_bitmap(struct super_block *sb) return err; } brelse(bh); + + if (type == TYPE_UNUSED) + return -EINVAL; } if (exfat_get_next_cluster(sb, &clu.dir)) From 862b9a8eb900d347af5be0eb1aeef9b161a83e77 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Tue, 3 Sep 2024 17:16:26 +0200 Subject: [PATCH 096/107] virtio_fs: allow idmapped mounts Allow idmapped mounts for virtiofs. It's absolutely safe as for virtiofs we have the same feature negotiation mechanism as for classical fuse filesystems. This does not affect any existing setups anyhow. virtiofsd support: https://gitlab.com/virtio-fs/virtiofsd/-/merge_requests/245 Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Reviewed-by: Stefan Hajnoczi Signed-off-by: Miklos Szeredi --- fs/fuse/virtio_fs.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c index 2fee9eb5ad0b..b6bd6ab71f5d 100644 --- a/fs/fuse/virtio_fs.c +++ b/fs/fuse/virtio_fs.c @@ -1613,6 +1613,7 @@ static struct file_system_type virtio_fs_type = { .name = "virtiofs", .init_fs_context = virtio_fs_init_fs_context, .kill_sb = virtio_kill_sb, + .fs_flags = FS_ALLOW_IDMAP, }; static int virtio_fs_uevent(const struct kobject *kobj, struct kobj_uevent_env *env) From efad7153bf93db8565128f7567aab1d23e221098 Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 13 Sep 2024 12:47:01 +0200 Subject: [PATCH 097/107] fuse: allow O_PATH fd for FUSE_DEV_IOC_BACKING_OPEN Only f_path is used from backing files registered with FUSE_DEV_IOC_BACKING_OPEN, so it makes sense to allow O_PATH descriptors. O_PATH files have an empty f_op, so don't check read_iter/write_iter. Reviewed-by: Amir Goldstein Signed-off-by: Miklos Szeredi --- fs/fuse/passthrough.c | 7 ++----- 1 file changed, 2 insertions(+), 5 deletions(-) diff --git a/fs/fuse/passthrough.c b/fs/fuse/passthrough.c index 9666d13884ce..62aee8289d11 100644 --- a/fs/fuse/passthrough.c +++ b/fs/fuse/passthrough.c @@ -228,16 +228,13 @@ int fuse_backing_open(struct fuse_conn *fc, struct fuse_backing_map *map) if (map->flags || map->padding) goto out; - file = fget(map->fd); + file = fget_raw(map->fd); res = -EBADF; if (!file) goto out; - res = -EOPNOTSUPP; - if (!file->f_op->read_iter || !file->f_op->write_iter) - goto out_fput; - backing_sb = file_inode(file)->i_sb; + pr_info("%s: %x:%pD %i\n", __func__, backing_sb->s_dev, file, backing_sb->s_stack_depth); res = -ELOOP; if (backing_sb->s_stack_depth >= fc->max_stack_depth) goto out_fput; From 3988a60d3aaabd6cca64fbd8f7be65c0c878d87b Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Thu, 12 Sep 2024 16:58:24 +0200 Subject: [PATCH 098/107] fs/fuse: fix null-ptr-deref when checking SB_I_NOIDMAP flag It was reported [1] that on linux-next/fs-next the following crash is reproducible: [ 42.659136] Oops: general protection fault, probably for non-canonical address 0xdffffc000000000b: 0000 [#1] PREEMPT SMP KASAN NOPTI [ 42.660501] fbcon: Taking over console [ 42.660930] KASAN: null-ptr-deref in range [0x0000000000000058-0x000000000000005f] [ 42.661752] CPU: 1 UID: 0 PID: 1589 Comm: dtprobed Not tainted 6.11.0-rc6+ #1 [ 42.662565] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.6.6 08/22/2023 [ 42.663472] RIP: 0010:fuse_get_req+0x36b/0x990 [fuse] [ 42.664046] Code: 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 8c 05 00 00 48 b8 00 00 00 00 00 fc ff df 48 8b 6d 08 48 8d 7d 58 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 4d 05 00 00 f6 45 59 20 0f 85 06 03 00 00 48 83 [ 42.666945] RSP: 0018:ffffc900009a7730 EFLAGS: 00010212 [ 42.668837] RAX: dffffc0000000000 RBX: 1ffff92000134eed RCX: ffffffffc20dec9a [ 42.670122] RDX: 000000000000000b RSI: 0000000000000008 RDI: 0000000000000058 [ 42.672154] RBP: 0000000000000000 R08: 0000000000000001 R09: ffffed1022110172 [ 42.672160] R10: ffff888110880b97 R11: ffffc900009a737a R12: 0000000000000001 [ 42.672179] R13: ffff888110880b60 R14: ffff888110880b90 R15: ffff888169973840 [ 42.672186] FS: 00007f28cd21d7c0(0000) GS:ffff8883ef280000(0000) knlGS:0000000000000000 [ 42.672191] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 42.[ CR02: ;32m00007f3237366208 CR3: 0 OK 79e001 CR4: 0000000000770ef0 [ 42.672214] PKRU: 55555554 [ 42.672218] Call Trace: [ 42.672223] [ 42.672226] ? die_addr+0x41/0xa0 [ 42.672238] ? exc_general_protection+0x14c/0x230 [ 42.672250] ? asm_exc_general_protection+0x26/0x30 [ 42.672260] ? fuse_get_req+0x77a/0x990 [fuse] [ 42.672281] ? fuse_get_req+0x36b/0x990 [fuse] [ 42.672300] ? kasan_unpoison+0x27/0x60 [ 42.672310] ? __pfx_fuse_get_req+0x10/0x10 [fuse] [ 42.672327] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672333] ? alloc_pages_mpol_noprof+0x195/0x440 [ 42.672340] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672345] ? kasan_unpoison+0x27/0x60 [ 42.672350] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672355] ? __kasan_slab_alloc+0x4d/0x90 [ 42.672362] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672367] ? __kmalloc_cache_noprof+0x134/0x350 [ 42.672376] fuse_simple_background+0xe7/0x180 [fuse] [ 42.672406] cuse_channel_open+0x540/0x710 [cuse] [ 42.672415] misc_open+0x2a7/0x3a0 [ 42.672424] chrdev_open+0x1ef/0x5f0 [ 42.672432] ? __pfx_chrdev_open+0x10/0x10 [ 42.672439] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672443] ? security_file_open+0x3bb/0x720 [ 42.672451] do_dentry_open+0x43d/0x1200 [ 42.672459] ? __pfx_chrdev_open+0x10/0x10 [ 42.672468] vfs_open+0x79/0x340 [ 42.672475] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672482] do_open+0x68c/0x11e0 [ 42.672489] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672495] ? __pfx_do_open+0x10/0x10 [ 42.672501] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.672506] ? open_last_lookups+0x2a2/0x1370 [ 42.672515] path_openat+0x24f/0x640 [ 42.672522] ? __pfx_path_openat+0x10/0x10 [ 42.723972] ? stack_depot_save_flags+0x45/0x4b0 [ 42.724787] ? __fput+0x43c/0xa70 [ 42.725100] do_filp_open+0x1b3/0x3e0 [ 42.725710] ? poison_slab_object+0x10d/0x190 [ 42.726145] ? __kasan_slab_free+0x33/0x50 [ 42.726570] ? __pfx_do_filp_open+0x10/0x10 [ 42.726981] ? do_syscall_64+0x64/0x170 [ 42.727418] ? entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 42.728018] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.728505] ? do_raw_spin_lock+0x131/0x270 [ 42.728922] ? __pfx_do_raw_spin_lock+0x10/0x10 [ 42.729494] ? do_raw_spin_unlock+0x14c/0x1f0 [ 42.729992] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.730889] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.732178] ? alloc_fd+0x176/0x5e0 [ 42.732585] do_sys_openat2+0x122/0x160 [ 42.732929] ? __pfx_do_sys_openat2+0x10/0x10 [ 42.733448] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.734013] ? __pfx_map_id_up+0x10/0x10 [ 42.734482] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.735529] ? __memcg_slab_free_hook+0x292/0x500 [ 42.736131] __x64_sys_openat+0x123/0x1e0 [ 42.736526] ? __pfx___x64_sys_openat+0x10/0x10 [ 42.737369] ? __x64_sys_close+0x7c/0xd0 [ 42.737717] ? srso_alias_return_thunk+0x5/0xfbef5 [ 42.738192] ? syscall_trace_enter+0x11e/0x1b0 [ 42.738739] do_syscall_64+0x64/0x170 [ 42.739113] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 42.739638] RIP: 0033:0x7f28cd13e87b [ 42.740038] Code: 25 00 00 41 00 3d 00 00 41 00 74 4b 64 8b 04 25 18 00 00 00 85 c0 75 67 44 89 e2 48 89 ee bf 9c ff ff ff b8 01 01 00 00 0f 05 <48> 3d 00 f0 ff ff 0f 87 91 00 00 00 48 8b 54 24 28 64 48 2b 14 25 [ 42.741943] RSP: 002b:00007ffc992546c0 EFLAGS: 00000246 ORIG_RAX: 0000000000000101 [ 42.742951] RAX: ffffffffffffffda RBX: 00007f28cd44f1ee RCX: 00007f28cd13e87b [ 42.743660] RDX: 0000000000000002 RSI: 00007f28cd44f2fa RDI: 00000000ffffff9c [ 42.744518] RBP: 00007f28cd44f2fa R08: 0000000000000000 R09: 0000000000000001 [ 42.745211] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000002 [ 42.745920] R13: 00007f28cd44f2fa R14: 0000000000000000 R15: 0000000000000003 [ 42.746708] [ 42.746937] Modules linked in: cuse vfat fat ext4 mbcache jbd2 intel_rapl_msr intel_rapl_common kvm_amd ccp bochs drm_vram_helper kvm drm_ttm_helper ttm pcspkr i2c_piix4 drm_kms_helper i2c_smbus pvpanic_mmio pvpanic joydev sch_fq_codel drm fuse xfs nvme_tcp nvme_fabrics nvme_core sd_mod sg virtio_net net_failover virtio_scsi failover crct10dif_pclmul crc32_pclmul ata_generic pata_acpi ata_piix ghash_clmulni_intel virtio_pci sha512_ssse3 virtio_pci_legacy_dev sha256_ssse3 virtio_pci_modern_dev sha1_ssse3 libata serio_raw dm_multipath btrfs blake2b_generic xor zstd_compress raid6_pq sunrpc dm_mirror dm_region_hash dm_log dm_mod be2iscsi bnx2i cnic uio cxgb4i cxgb4 tls cxgb3i cxgb3 mdio libcxgbi libcxgb qla4xxx iscsi_boot_sysfs iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi qemu_fw_cfg aesni_intel crypto_simd cryptd [ 42.754333] ---[ end trace 0000000000000000 ]--- [ 42.756899] RIP: 0010:fuse_get_req+0x36b/0x990 [fuse] [ 42.757851] Code: 48 89 fa 48 c1 ea 03 80 3c 02 00 0f 85 8c 05 00 00 48 b8 00 00 00 00 00 fc ff df 48 8b 6d 08 48 8d 7d 58 48 89 fa 48 c1 ea 03 <80> 3c 02 00 0f 85 4d 05 00 00 f6 45 59 20 0f 85 06 03 00 00 48 83 [ 42.760334] RSP: 0018:ffffc900009a7730 EFLAGS: 00010212 [ 42.760940] RAX: dffffc0000000000 RBX: 1ffff92000134eed RCX: ffffffffc20dec9a [ 42.761697] RDX: 000000000000000b RSI: 0000000000000008 RDI: 0000000000000058 [ 42.763009] RBP: 0000000000000000 R08: 0000000000000001 R09: ffffed1022110172 [ 42.763920] R10: ffff888110880b97 R11: ffffc900009a737a R12: 0000000000000001 [ 42.764839] R13: ffff888110880b60 R14: ffff888110880b90 R15: ffff888169973840 [ 42.765716] FS: 00007f28cd21d7c0(0000) GS:ffff8883ef280000(0000) knlGS:0000000000000000 [ 42.766890] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 42.767828] CR2: 00007f3237366208 CR3: 000000012c79e001 CR4: 0000000000770ef0 [ 42.768730] PKRU: 55555554 [ 42.769022] Kernel panic - not syncing: Fatal exception [ 42.770758] Kernel Offset: 0x7200000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff) [ 42.771947] ---[ end Kernel panic - not syncing: Fatal exception ]--- It's obviously CUSE related callstack. For CUSE case, we don't have superblock and our checks for SB_I_NOIDMAP flag does not make any sense. Let's handle this case gracefully. Fixes: aa16880d9f13 ("fuse: add basic infrastructure to support idmappings") Link: https://lore.kernel.org/linux-next/87v7z586py.fsf@debian-BULLSEYE-live-builder-AMD64/ [1] Reported-by: Chandan Babu R Reported-by: syzbot+20c7e20cc8f5296dca12@syzkaller.appspotmail.com Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 5ca2ab48527e..3e7912c65bc6 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -148,7 +148,7 @@ static struct fuse_req *fuse_get_req(struct mnt_idmap *idmap, if (for_background) __set_bit(FR_BACKGROUND, &req->flags); - if ((fm->sb->s_iflags & SB_I_NOIDMAP) || idmap) { + if (!fm->sb || (fm->sb->s_iflags & SB_I_NOIDMAP) || idmap) { kuid_t idmapped_fsuid; kgid_t idmapped_fsgid; @@ -517,7 +517,7 @@ static void fuse_force_creds(struct fuse_req *req) { struct fuse_conn *fc = req->fm->fc; - if (req->fm->sb->s_iflags & SB_I_NOIDMAP) { + if (!req->fm->sb || req->fm->sb->s_iflags & SB_I_NOIDMAP) { req->in.h.uid = from_kuid_munged(fc->user_ns, current_fsuid()); req->in.h.gid = from_kgid_munged(fc->user_ns, current_fsgid()); } else { From 0c6793823d2b0eb079f4c6f54d9cdf6b2beec9d8 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Fri, 6 Sep 2024 16:34:51 +0200 Subject: [PATCH 099/107] fs/fuse: introduce and use fuse_simple_idmap_request() helper Let's convert all existing callers properly. No functional changes intended. Suggested-by: Christian Brauner Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dax.c | 4 ++-- fs/fuse/dev.c | 6 +++--- fs/fuse/dir.c | 26 +++++++++++++------------- fs/fuse/file.c | 32 ++++++++++++++++---------------- fs/fuse/fuse_i.h | 18 ++++++++++++++++-- fs/fuse/inode.c | 6 +++--- fs/fuse/ioctl.c | 2 +- fs/fuse/readdir.c | 4 ++-- fs/fuse/xattr.c | 8 ++++---- 9 files changed, 60 insertions(+), 46 deletions(-) diff --git a/fs/fuse/dax.c b/fs/fuse/dax.c index 6d8368d66dd4..12ef91d170bb 100644 --- a/fs/fuse/dax.c +++ b/fs/fuse/dax.c @@ -207,7 +207,7 @@ static int fuse_setup_one_mapping(struct inode *inode, unsigned long start_idx, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err < 0) return err; dmap->writable = writable; @@ -245,7 +245,7 @@ static int fuse_send_removemapping(struct inode *inode, args.in_args[0].value = inargp; args.in_args[1].size = inargp->count * sizeof(*remove_one); args.in_args[1].value = remove_one; - return fuse_simple_request(NULL, fm, &args); + return fuse_simple_request(fm, &args); } static int dmap_removemapping_list(struct inode *inode, unsigned int num, diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 3e7912c65bc6..317d2b30b21f 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -539,9 +539,9 @@ static void fuse_args_to_req(struct fuse_req *req, struct fuse_args *args) __set_bit(FR_ASYNC, &req->flags); } -ssize_t fuse_simple_request(struct mnt_idmap *idmap, - struct fuse_mount *fm, - struct fuse_args *args) +ssize_t __fuse_simple_request(struct mnt_idmap *idmap, + struct fuse_mount *fm, + struct fuse_args *args) { struct fuse_conn *fc = fm->fc; struct fuse_req *req; diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 99f9948bf68b..491e112819be 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -230,7 +230,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, unsigned int flags) parent = dget_parent(entry); fuse_lookup_init(fm->fc, &args, get_node_id(d_inode(parent)), &entry->d_name, &outarg); - ret = fuse_simple_request(NULL, fm, &args); + ret = fuse_simple_request(fm, &args); dput(parent); /* Zero nodeid is same as -ENOENT */ if (!ret && !outarg.nodeid) @@ -383,7 +383,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, const struct qstr *name attr_version = fuse_get_attr_version(fm->fc); fuse_lookup_init(fm->fc, &args, nodeid, name, outarg); - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); /* Zero nodeid is same as -ENOENT, but with valid timeout */ if (err || !outarg->nodeid) goto out_put_forget; @@ -677,7 +677,7 @@ static int fuse_create_open(struct mnt_idmap *idmap, struct inode *dir, if (err) goto out_free_ff; - err = fuse_simple_request(idmap, fm, &args); + err = fuse_simple_idmap_request(idmap, fm, &args); free_ext_value(&args); if (err) goto out_free_ff; @@ -809,7 +809,7 @@ static int create_new_entry(struct mnt_idmap *idmap, struct fuse_mount *fm, goto out_put_forget_req; } - err = fuse_simple_request(idmap, fm, args); + err = fuse_simple_idmap_request(idmap, fm, args); free_ext_value(args); if (err) goto out_put_forget_req; @@ -994,7 +994,7 @@ static int fuse_unlink(struct inode *dir, struct dentry *entry) args.in_numargs = 1; args.in_args[0].size = entry->d_name.len + 1; args.in_args[0].value = entry->d_name.name; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) { fuse_dir_changed(dir); fuse_entry_unlinked(entry); @@ -1017,7 +1017,7 @@ static int fuse_rmdir(struct inode *dir, struct dentry *entry) args.in_numargs = 1; args.in_args[0].size = entry->d_name.len + 1; args.in_args[0].value = entry->d_name.name; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) { fuse_dir_changed(dir); fuse_entry_unlinked(entry); @@ -1047,7 +1047,7 @@ static int fuse_rename_common(struct mnt_idmap *idmap, struct inode *olddir, str args.in_args[1].value = oldent->d_name.name; args.in_args[2].size = newent->d_name.len + 1; args.in_args[2].value = newent->d_name.name; - err = fuse_simple_request(idmap, fm, &args); + err = fuse_simple_idmap_request(idmap, fm, &args); if (!err) { /* ctime changes */ fuse_update_ctime(d_inode(oldent)); @@ -1222,7 +1222,7 @@ static int fuse_do_statx(struct mnt_idmap *idmap, struct inode *inode, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err) return err; @@ -1280,7 +1280,7 @@ static int fuse_do_getattr(struct mnt_idmap *idmap, struct inode *inode, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) { if (fuse_invalid_attr(&outarg.attr) || inode_wrong_type(inode, outarg.attr.mode)) { @@ -1492,7 +1492,7 @@ static int fuse_access(struct inode *inode, int mask) args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fm->fc->no_access = 1; err = 0; @@ -1604,7 +1604,7 @@ static int fuse_readlink_page(struct inode *inode, struct page *page) ap.args.page_zeroing = true; ap.args.out_numargs = 1; ap.args.out_args[0].size = desc.length; - res = fuse_simple_request(NULL, fm, &ap.args); + res = fuse_simple_request(fm, &ap.args); fuse_invalidate_atime(inode); @@ -1889,7 +1889,7 @@ int fuse_flush_times(struct inode *inode, struct fuse_file *ff) } fuse_setattr_fill(fm->fc, &args, inode, &inarg, &outarg); - return fuse_simple_request(NULL, fm, &args); + return fuse_simple_request(fm, &args); } /* @@ -2002,7 +2002,7 @@ int fuse_do_setattr(struct mnt_idmap *idmap, struct dentry *dentry, inarg.valid |= FATTR_KILL_SUIDGID; } fuse_setattr_fill(fc, &args, inode, &inarg, &outarg); - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err) { if (err == -EINTR) fuse_invalidate_attr(inode); diff --git a/fs/fuse/file.c b/fs/fuse/file.c index ca553d7a7c9e..b8afeca12487 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -48,7 +48,7 @@ static int fuse_send_open(struct fuse_mount *fm, u64 nodeid, args.out_args[0].size = sizeof(*outargp); args.out_args[0].value = outargp; - return fuse_simple_request(NULL, fm, &args); + return fuse_simple_request(fm, &args); } struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release) @@ -111,7 +111,7 @@ static void fuse_file_put(struct fuse_file *ff, bool sync) if (!args) { /* Do nothing when server does not implement 'open' */ } else if (sync) { - fuse_simple_request(NULL, ff->fm, args); + fuse_simple_request(ff->fm, args); fuse_release_end(ff->fm, args, 0); } else { args->end = fuse_release_end; @@ -539,7 +539,7 @@ static int fuse_flush(struct file *file, fl_owner_t id) args.in_args[0].value = &inarg; args.force = true; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fm->fc->no_flush = 1; err = 0; @@ -572,7 +572,7 @@ int fuse_fsync_common(struct file *file, loff_t start, loff_t end, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - return fuse_simple_request(NULL, fm, &args); + return fuse_simple_request(fm, &args); } static int fuse_fsync(struct file *file, loff_t start, loff_t end, @@ -814,7 +814,7 @@ static ssize_t fuse_send_read(struct fuse_io_args *ia, loff_t pos, size_t count, if (ia->io->async) return fuse_async_req_send(fm, ia, count); - return fuse_simple_request(NULL, fm, &ia->ap.args); + return fuse_simple_request(fm, &ia->ap.args); } static void fuse_read_update_size(struct inode *inode, loff_t size, @@ -878,7 +878,7 @@ static int fuse_do_readpage(struct file *file, struct page *page) desc.length--; fuse_read_args_fill(&ia, file, pos, desc.length, FUSE_READ); - res = fuse_simple_request(NULL, fm, &ia.ap.args); + res = fuse_simple_request(fm, &ia.ap.args); if (res < 0) return res; /* @@ -976,7 +976,7 @@ static void fuse_send_readpages(struct fuse_io_args *ia, struct file *file) if (!err) return; } else { - res = fuse_simple_request(NULL, fm, &ap->args); + res = fuse_simple_request(fm, &ap->args); err = res < 0 ? res : 0; } fuse_readpages_end(fm, &ap->args, err); @@ -1101,7 +1101,7 @@ static ssize_t fuse_send_write(struct fuse_io_args *ia, loff_t pos, if (ia->io->async) return fuse_async_req_send(fm, ia, count); - err = fuse_simple_request(NULL, fm, &ia->ap.args); + err = fuse_simple_request(fm, &ia->ap.args); if (!err && ia->write.out.size > count) err = -EIO; @@ -1147,7 +1147,7 @@ static ssize_t fuse_send_write_pages(struct fuse_io_args *ia, if (fm->fc->handle_killpriv_v2 && !capable(CAP_FSETID)) ia->write.in.write_flags |= FUSE_WRITE_KILL_SUIDGID; - err = fuse_simple_request(NULL, fm, &ap->args); + err = fuse_simple_request(fm, &ap->args); if (!err && ia->write.out.size > count) err = -EIO; @@ -2662,7 +2662,7 @@ static int fuse_getlk(struct file *file, struct file_lock *fl) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) err = convert_fuse_file_lock(fm->fc, &outarg.lk, fl); @@ -2686,7 +2686,7 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) } fuse_lk_fill(&args, file, fl, opcode, pid_nr, flock, &inarg); - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); /* locking is restartable */ if (err == -EINTR) @@ -2760,7 +2760,7 @@ static sector_t fuse_bmap(struct address_space *mapping, sector_t block) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) fm->fc->no_bmap = 1; @@ -2792,7 +2792,7 @@ static loff_t fuse_lseek(struct file *file, loff_t offset, int whence) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err) { if (err == -ENOSYS) { fm->fc->no_lseek = 1; @@ -2925,7 +2925,7 @@ __poll_t fuse_file_poll(struct file *file, poll_table *wait) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) return demangle_poll(outarg.revents); @@ -3147,7 +3147,7 @@ static long fuse_file_fallocate(struct file *file, int mode, loff_t offset, args.in_numargs = 1; args.in_args[0].size = sizeof(inarg); args.in_args[0].value = &inarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fm->fc->no_fallocate = 1; err = -EOPNOTSUPP; @@ -3259,7 +3259,7 @@ static ssize_t __fuse_copy_file_range(struct file *file_in, loff_t pos_in, args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fc->no_copy_file_range = 1; err = -EOPNOTSUPP; diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index 7dd0661aeab6..b2c7834f21b5 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -1147,8 +1147,22 @@ void __exit fuse_ctl_cleanup(void); /** * Simple request sending that does request allocation and freeing */ -ssize_t fuse_simple_request(struct mnt_idmap *idmap, struct fuse_mount *fm, - struct fuse_args *args); +ssize_t __fuse_simple_request(struct mnt_idmap *idmap, + struct fuse_mount *fm, + struct fuse_args *args); + +static inline ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args) +{ + return __fuse_simple_request(NULL, fm, args); +} + +static inline ssize_t fuse_simple_idmap_request(struct mnt_idmap *idmap, + struct fuse_mount *fm, + struct fuse_args *args) +{ + return __fuse_simple_request(idmap, fm, args); +} + int fuse_simple_background(struct fuse_mount *fm, struct fuse_args *args, gfp_t gfp_flags); diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index d7edb3fb829f..fd3321e29a3e 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -586,7 +586,7 @@ static void fuse_send_destroy(struct fuse_mount *fm) args.opcode = FUSE_DESTROY; args.force = true; args.nocreds = true; - fuse_simple_request(NULL, fm, &args); + fuse_simple_request(fm, &args); } } @@ -624,7 +624,7 @@ static int fuse_statfs(struct dentry *dentry, struct kstatfs *buf) args.out_numargs = 1; args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (!err) convert_fuse_statfs(buf, &outarg.st); return err; @@ -713,7 +713,7 @@ static int fuse_sync_fs(struct super_block *sb, int wait) args.nodeid = get_node_id(sb->s_root->d_inode); args.out_numargs = 0; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fc->sync_fs = 0; err = 0; diff --git a/fs/fuse/ioctl.c b/fs/fuse/ioctl.c index b40dd931167d..572ce8a82ceb 100644 --- a/fs/fuse/ioctl.c +++ b/fs/fuse/ioctl.c @@ -18,7 +18,7 @@ static ssize_t fuse_send_ioctl(struct fuse_mount *fm, struct fuse_args *args, args->out_args[0].size = sizeof(*outarg); args->out_args[0].value = outarg; - ret = fuse_simple_request(NULL, fm, args); + ret = fuse_simple_request(fm, args); /* Translate ENOSYS, which shouldn't be returned from fs */ if (ret == -ENOSYS) diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c index e8a093289421..0377b6dc24c8 100644 --- a/fs/fuse/readdir.c +++ b/fs/fuse/readdir.c @@ -279,7 +279,7 @@ static void fuse_force_forget(struct file *file, u64 nodeid) args.force = true; args.noreply = true; - fuse_simple_request(NULL, fm, &args); + fuse_simple_request(fm, &args); /* ignore errors */ } @@ -358,7 +358,7 @@ static int fuse_readdir_uncached(struct file *file, struct dir_context *ctx) FUSE_READDIR); } locked = fuse_lock_inode(inode); - res = fuse_simple_request(NULL, fm, &ap->args); + res = fuse_simple_request(fm, &ap->args); fuse_unlock_inode(inode, locked); if (res >= 0) { if (!res) { diff --git a/fs/fuse/xattr.c b/fs/fuse/xattr.c index 0a9b60de3668..9f568d345c51 100644 --- a/fs/fuse/xattr.c +++ b/fs/fuse/xattr.c @@ -37,7 +37,7 @@ int fuse_setxattr(struct inode *inode, const char *name, const void *value, args.in_args[1].value = name; args.in_args[2].size = size; args.in_args[2].value = value; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fm->fc->no_setxattr = 1; err = -EOPNOTSUPP; @@ -79,7 +79,7 @@ ssize_t fuse_getxattr(struct inode *inode, const char *name, void *value, args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; } - ret = fuse_simple_request(NULL, fm, &args); + ret = fuse_simple_request(fm, &args); if (!ret && !size) ret = min_t(size_t, outarg.size, XATTR_SIZE_MAX); if (ret == -ENOSYS) { @@ -141,7 +141,7 @@ ssize_t fuse_listxattr(struct dentry *entry, char *list, size_t size) args.out_args[0].size = sizeof(outarg); args.out_args[0].value = &outarg; } - ret = fuse_simple_request(NULL, fm, &args); + ret = fuse_simple_request(fm, &args); if (!ret && !size) ret = min_t(size_t, outarg.size, XATTR_LIST_MAX); if (ret > 0 && size) @@ -167,7 +167,7 @@ int fuse_removexattr(struct inode *inode, const char *name) args.in_numargs = 1; args.in_args[0].size = strlen(name) + 1; args.in_args[0].value = name; - err = fuse_simple_request(NULL, fm, &args); + err = fuse_simple_request(fm, &args); if (err == -ENOSYS) { fm->fc->no_removexattr = 1; err = -EOPNOTSUPP; From ffcdc4c628e1a30489da10dd78358e89c823b341 Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Fri, 6 Sep 2024 16:34:52 +0200 Subject: [PATCH 100/107] fs/mnt_idmapping: introduce an invalid_mnt_idmap Link: https://lore.kernel.org/linux-fsdevel/20240904-baugrube-erhoben-b3c1c49a2645@brauner/ Suggested-by: Christian Brauner Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/mnt_idmapping.c | 22 ++++++++++++++++++++-- include/linux/mnt_idmapping.h | 1 + 2 files changed, 21 insertions(+), 2 deletions(-) diff --git a/fs/mnt_idmapping.c b/fs/mnt_idmapping.c index 3c60f1eaca61..cbca6500848e 100644 --- a/fs/mnt_idmapping.c +++ b/fs/mnt_idmapping.c @@ -32,6 +32,15 @@ struct mnt_idmap nop_mnt_idmap = { }; EXPORT_SYMBOL_GPL(nop_mnt_idmap); +/* + * Carries the invalid idmapping of a full 0-4294967295 {g,u}id range. + * This means that all {g,u}ids are mapped to INVALID_VFS{G,U}ID. + */ +struct mnt_idmap invalid_mnt_idmap = { + .count = REFCOUNT_INIT(1), +}; +EXPORT_SYMBOL_GPL(invalid_mnt_idmap); + /** * initial_idmapping - check whether this is the initial mapping * @ns: idmapping to check @@ -75,6 +84,8 @@ vfsuid_t make_vfsuid(struct mnt_idmap *idmap, if (idmap == &nop_mnt_idmap) return VFSUIDT_INIT(kuid); + if (idmap == &invalid_mnt_idmap) + return INVALID_VFSUID; if (initial_idmapping(fs_userns)) uid = __kuid_val(kuid); else @@ -112,6 +123,8 @@ vfsgid_t make_vfsgid(struct mnt_idmap *idmap, if (idmap == &nop_mnt_idmap) return VFSGIDT_INIT(kgid); + if (idmap == &invalid_mnt_idmap) + return INVALID_VFSGID; if (initial_idmapping(fs_userns)) gid = __kgid_val(kgid); else @@ -140,6 +153,8 @@ kuid_t from_vfsuid(struct mnt_idmap *idmap, if (idmap == &nop_mnt_idmap) return AS_KUIDT(vfsuid); + if (idmap == &invalid_mnt_idmap) + return INVALID_UID; uid = map_id_up(&idmap->uid_map, __vfsuid_val(vfsuid)); if (uid == (uid_t)-1) return INVALID_UID; @@ -167,6 +182,8 @@ kgid_t from_vfsgid(struct mnt_idmap *idmap, if (idmap == &nop_mnt_idmap) return AS_KGIDT(vfsgid); + if (idmap == &invalid_mnt_idmap) + return INVALID_GID; gid = map_id_up(&idmap->gid_map, __vfsgid_val(vfsgid)); if (gid == (gid_t)-1) return INVALID_GID; @@ -296,7 +313,7 @@ struct mnt_idmap *alloc_mnt_idmap(struct user_namespace *mnt_userns) */ struct mnt_idmap *mnt_idmap_get(struct mnt_idmap *idmap) { - if (idmap != &nop_mnt_idmap) + if (idmap != &nop_mnt_idmap && idmap != &invalid_mnt_idmap) refcount_inc(&idmap->count); return idmap; @@ -312,7 +329,8 @@ EXPORT_SYMBOL_GPL(mnt_idmap_get); */ void mnt_idmap_put(struct mnt_idmap *idmap) { - if (idmap != &nop_mnt_idmap && refcount_dec_and_test(&idmap->count)) + if (idmap != &nop_mnt_idmap && idmap != &invalid_mnt_idmap && + refcount_dec_and_test(&idmap->count)) free_mnt_idmap(idmap); } EXPORT_SYMBOL_GPL(mnt_idmap_put); diff --git a/include/linux/mnt_idmapping.h b/include/linux/mnt_idmapping.h index cd4d5c8781f5..b1b219bc3422 100644 --- a/include/linux/mnt_idmapping.h +++ b/include/linux/mnt_idmapping.h @@ -9,6 +9,7 @@ struct mnt_idmap; struct user_namespace; extern struct mnt_idmap nop_mnt_idmap; +extern struct mnt_idmap invalid_mnt_idmap; extern struct user_namespace init_user_ns; typedef struct { From 106e4593ed1b9925ca732a74f490e4f52ea4e65c Mon Sep 17 00:00:00 2001 From: Alexander Mikhalitsyn Date: Fri, 6 Sep 2024 16:34:53 +0200 Subject: [PATCH 101/107] fs/fuse: convert to use invalid_mnt_idmap We should convert fs/fuse code to use a newly introduced invalid_mnt_idmap instead of passing a NULL as idmap pointer. Suggested-by: Christian Brauner Signed-off-by: Alexander Mikhalitsyn Reviewed-by: Christian Brauner Signed-off-by: Miklos Szeredi --- fs/fuse/dev.c | 47 +++++++++++++++++++++++------------------------ fs/fuse/dir.c | 6 +++--- fs/fuse/fuse_i.h | 2 +- 3 files changed, 27 insertions(+), 28 deletions(-) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 317d2b30b21f..cdf925e0b317 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -114,7 +114,11 @@ static struct fuse_req *fuse_get_req(struct mnt_idmap *idmap, { struct fuse_conn *fc = fm->fc; struct fuse_req *req; + bool no_idmap = !fm->sb || (fm->sb->s_iflags & SB_I_NOIDMAP); + kuid_t fsuid; + kgid_t fsgid; int err; + atomic_inc(&fc->num_waiting); if (fuse_block_alloc(fc, for_background)) { @@ -148,29 +152,24 @@ static struct fuse_req *fuse_get_req(struct mnt_idmap *idmap, if (for_background) __set_bit(FR_BACKGROUND, &req->flags); - if (!fm->sb || (fm->sb->s_iflags & SB_I_NOIDMAP) || idmap) { - kuid_t idmapped_fsuid; - kgid_t idmapped_fsgid; + /* + * Keep the old behavior when idmappings support was not + * declared by a FUSE server. + * + * For those FUSE servers who support idmapped mounts, + * we send UID/GID only along with "inode creation" + * fuse requests, otherwise idmap == &invalid_mnt_idmap and + * req->in.h.{u,g}id will be equal to FUSE_INVALID_UIDGID. + */ + fsuid = no_idmap ? current_fsuid() : mapped_fsuid(idmap, fc->user_ns); + fsgid = no_idmap ? current_fsgid() : mapped_fsgid(idmap, fc->user_ns); + req->in.h.uid = from_kuid(fc->user_ns, fsuid); + req->in.h.gid = from_kgid(fc->user_ns, fsgid); - /* - * Note, that when - * (fm->sb->s_iflags & SB_I_NOIDMAP) is true, then - * (idmap == &nop_mnt_idmap) is always true and therefore, - * mapped_fsuid(idmap, fc->user_ns) == current_fsuid(). - */ - idmapped_fsuid = idmap ? mapped_fsuid(idmap, fc->user_ns) : current_fsuid(); - idmapped_fsgid = idmap ? mapped_fsgid(idmap, fc->user_ns) : current_fsgid(); - req->in.h.uid = from_kuid(fc->user_ns, idmapped_fsuid); - req->in.h.gid = from_kgid(fc->user_ns, idmapped_fsgid); - - if (unlikely(req->in.h.uid == ((uid_t)-1) || - req->in.h.gid == ((gid_t)-1))) { - fuse_put_request(req); - return ERR_PTR(-EOVERFLOW); - } - } else { - req->in.h.uid = FUSE_INVALID_UIDGID; - req->in.h.gid = FUSE_INVALID_UIDGID; + if (no_idmap && unlikely(req->in.h.uid == ((uid_t)-1) || + req->in.h.gid == ((gid_t)-1))) { + fuse_put_request(req); + return ERR_PTR(-EOVERFLOW); } return req; @@ -619,7 +618,7 @@ int fuse_simple_background(struct fuse_mount *fm, struct fuse_args *args, __set_bit(FR_BACKGROUND, &req->flags); } else { WARN_ON(args->nocreds); - req = fuse_get_req(NULL, fm, true); + req = fuse_get_req(&invalid_mnt_idmap, fm, true); if (IS_ERR(req)) return PTR_ERR(req); } @@ -641,7 +640,7 @@ static int fuse_simple_notify_reply(struct fuse_mount *fm, struct fuse_req *req; struct fuse_iqueue *fiq = &fm->fc->iq; - req = fuse_get_req(NULL, fm, false); + req = fuse_get_req(&invalid_mnt_idmap, fm, false); if (IS_ERR(req)) return PTR_ERR(req); diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 491e112819be..54104dd48af7 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -1093,7 +1093,7 @@ static int fuse_rename2(struct mnt_idmap *idmap, struct inode *olddir, if (fc->no_rename2 || fc->minor < 23) return -EINVAL; - err = fuse_rename_common((flags & RENAME_WHITEOUT) ? idmap : NULL, + err = fuse_rename_common((flags & RENAME_WHITEOUT) ? idmap : &invalid_mnt_idmap, olddir, oldent, newdir, newent, flags, FUSE_RENAME2, sizeof(struct fuse_rename2_in)); @@ -1102,7 +1102,7 @@ static int fuse_rename2(struct mnt_idmap *idmap, struct inode *olddir, err = -EINVAL; } } else { - err = fuse_rename_common(NULL, olddir, oldent, newdir, newent, 0, + err = fuse_rename_common(&invalid_mnt_idmap, olddir, oldent, newdir, newent, 0, FUSE_RENAME, sizeof(struct fuse_rename_in)); } @@ -1127,7 +1127,7 @@ static int fuse_link(struct dentry *entry, struct inode *newdir, args.in_args[0].value = &inarg; args.in_args[1].size = newent->d_name.len + 1; args.in_args[1].value = newent->d_name.name; - err = create_new_entry(NULL, fm, &args, newdir, newent, inode->i_mode); + err = create_new_entry(&invalid_mnt_idmap, fm, &args, newdir, newent, inode->i_mode); if (!err) fuse_update_ctime_in_cache(inode); else if (err == -EINTR) diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index b2c7834f21b5..e6cc3d552b13 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -1153,7 +1153,7 @@ ssize_t __fuse_simple_request(struct mnt_idmap *idmap, static inline ssize_t fuse_simple_request(struct fuse_mount *fm, struct fuse_args *args) { - return __fuse_simple_request(NULL, fm, args); + return __fuse_simple_request(&invalid_mnt_idmap, fm, args); } static inline ssize_t fuse_simple_idmap_request(struct mnt_idmap *idmap, From 6630ea49103c3d45461e29b0f6eb0ce750aeb8f5 Mon Sep 17 00:00:00 2001 From: Yuezhang Mo Date: Mon, 23 Sep 2024 21:37:32 +0900 Subject: [PATCH 102/107] exfat: move extend valid_size into ->page_mkwrite() It is not a good way to extend valid_size to the end of the mmap area by writing zeros in mmap. Because after calling mmap, no data may be written, or only a small amount of data may be written to the head of the mmap area. This commit moves extending valid_size to exfat_page_mkwrite(). In exfat_page_mkwrite() only extend valid_size to the starting position of new data writing, which reduces unnecessary writing of zeros. If the block is not mapped and is marked as new after being mapped for writing, block_write_begin() will zero the page cache corresponding to the block, so there is no need to call zero_user_segment() in exfat_file_zeroed_range(). And after moving extending valid_size to exfat_page_mkwrite(), the data written by mmap will be copied to the page cache but the page cache may be not mapped to the disk. Calling zero_user_segment() will cause the data written by mmap to be cleared. So this commit removes calling zero_user_segment() from exfat_file_zeroed_range() and renames exfat_file_zeroed_range() to exfat_extend_valid_size(). Signed-off-by: Yuezhang Mo Signed-off-by: Namjae Jeon --- fs/exfat/file.c | 70 +++++++++++++++++++++++++++++++------------------ 1 file changed, 45 insertions(+), 25 deletions(-) diff --git a/fs/exfat/file.c b/fs/exfat/file.c index a945ede3bf97..a25d7eb789f4 100644 --- a/fs/exfat/file.c +++ b/fs/exfat/file.c @@ -532,32 +532,32 @@ int exfat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync) return blkdev_issue_flush(inode->i_sb->s_bdev); } -static int exfat_file_zeroed_range(struct file *file, loff_t start, loff_t end) +static int exfat_extend_valid_size(struct file *file, loff_t new_valid_size) { int err; + loff_t pos; struct inode *inode = file_inode(file); + struct exfat_inode_info *ei = EXFAT_I(inode); struct address_space *mapping = inode->i_mapping; const struct address_space_operations *ops = mapping->a_ops; - while (start < end) { - u32 zerofrom, len; + pos = ei->valid_size; + while (pos < new_valid_size) { + u32 len; struct folio *folio; - zerofrom = start & (PAGE_SIZE - 1); - len = PAGE_SIZE - zerofrom; - if (start + len > end) - len = end - start; + len = PAGE_SIZE - (pos & (PAGE_SIZE - 1)); + if (pos + len > new_valid_size) + len = new_valid_size - pos; - err = ops->write_begin(file, mapping, start, len, &folio, NULL); + err = ops->write_begin(file, mapping, pos, len, &folio, NULL); if (err) goto out; - folio_zero_range(folio, offset_in_folio(folio, start), len); - - err = ops->write_end(file, mapping, start, len, len, folio, NULL); + err = ops->write_end(file, mapping, pos, len, len, folio, NULL); if (err < 0) goto out; - start += len; + pos += len; balance_dirty_pages_ratelimited(mapping); cond_resched(); @@ -585,7 +585,7 @@ static ssize_t exfat_file_write_iter(struct kiocb *iocb, struct iov_iter *iter) goto unlock; if (pos > valid_size) { - ret = exfat_file_zeroed_range(file, valid_size, pos); + ret = exfat_extend_valid_size(file, pos); if (ret < 0 && ret != -ENOSPC) { exfat_err(inode->i_sb, "write: fail to zero from %llu to %llu(%zd)", @@ -619,26 +619,46 @@ unlock: return ret; } -static int exfat_file_mmap(struct file *file, struct vm_area_struct *vma) +static vm_fault_t exfat_page_mkwrite(struct vm_fault *vmf) { - int ret; + int err; + struct vm_area_struct *vma = vmf->vma; + struct file *file = vma->vm_file; struct inode *inode = file_inode(file); struct exfat_inode_info *ei = EXFAT_I(inode); - loff_t start = ((loff_t)vma->vm_pgoff << PAGE_SHIFT); - loff_t end = min_t(loff_t, i_size_read(inode), + loff_t start, end; + + if (!inode_trylock(inode)) + return VM_FAULT_RETRY; + + start = ((loff_t)vma->vm_pgoff << PAGE_SHIFT); + end = min_t(loff_t, i_size_read(inode), start + vma->vm_end - vma->vm_start); - if ((vma->vm_flags & VM_WRITE) && ei->valid_size < end) { - ret = exfat_file_zeroed_range(file, ei->valid_size, end); - if (ret < 0) { - exfat_err(inode->i_sb, - "mmap: fail to zero from %llu to %llu(%d)", - start, end, ret); - return ret; + if (ei->valid_size < end) { + err = exfat_extend_valid_size(file, end); + if (err < 0) { + inode_unlock(inode); + return vmf_fs_error(err); } } - return generic_file_mmap(file, vma); + inode_unlock(inode); + + return filemap_page_mkwrite(vmf); +} + +static const struct vm_operations_struct exfat_file_vm_ops = { + .fault = filemap_fault, + .map_pages = filemap_map_pages, + .page_mkwrite = exfat_page_mkwrite, +}; + +static int exfat_file_mmap(struct file *file, struct vm_area_struct *vma) +{ + file_accessed(file); + vma->vm_ops = &exfat_file_vm_ops; + return 0; } const struct file_operations exfat_file_operations = { From c290fe508eee36df1640c3cb35dc8f89e073c8a8 Mon Sep 17 00:00:00 2001 From: Daniel Yang Date: Mon, 16 Sep 2024 16:05:06 -0700 Subject: [PATCH 103/107] exfat: resolve memory leak from exfat_create_upcase_table() If exfat_load_upcase_table reaches end and returns -EINVAL, allocated memory doesn't get freed and while exfat_load_default_upcase_table allocates more memory, leading to a memory leak. Here's link to syzkaller crash report illustrating this issue: https://syzkaller.appspot.com/text?tag=CrashReport&x=1406c201980000 Reported-by: syzbot+e1c69cadec0f1a078e3d@syzkaller.appspotmail.com Fixes: a13d1a4de3b0 ("exfat: move freeing sbi, upcase table and dropping nls into rcu-delayed helper") Cc: stable@vger.kernel.org Signed-off-by: Daniel Yang Signed-off-by: Namjae Jeon --- fs/exfat/nls.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/fs/exfat/nls.c b/fs/exfat/nls.c index afdf13c34ff5..1ac011088ce7 100644 --- a/fs/exfat/nls.c +++ b/fs/exfat/nls.c @@ -779,8 +779,11 @@ int exfat_create_upcase_table(struct super_block *sb) le32_to_cpu(ep->dentry.upcase.checksum)); brelse(bh); - if (ret && ret != -EIO) + if (ret && ret != -EIO) { + /* free memory from exfat_load_upcase_table call */ + exfat_free_upcase_table(sbi); goto load_default; + } /* load successfully */ return ret; From cb7d85014fb1ca3387f7ff5f6067337b3d7f3c5a Mon Sep 17 00:00:00 2001 From: Yuezhang Mo Date: Sat, 14 Sep 2024 09:42:01 +0800 Subject: [PATCH 104/107] MAINTAINERS: exfat: add myself as reviewer I have been contributing to exfat for sometime and I would like to help with code reviews as well. Signed-off-by: Yuezhang Mo Acked-by: Sungjong Seo Signed-off-by: Namjae Jeon --- MAINTAINERS | 1 + 1 file changed, 1 insertion(+) diff --git a/MAINTAINERS b/MAINTAINERS index 5432239f3c5d..508ade6eee78 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -8415,6 +8415,7 @@ N: binfmt EXFAT FILE SYSTEM M: Namjae Jeon M: Sungjong Seo +R: Yuezhang Mo L: linux-fsdevel@vger.kernel.org S: Maintained T: git git://git.kernel.org/pub/scm/linux/kernel/git/linkinjeon/exfat.git From ae87b9c2dc9800e6ab52febd09341140599ff8e3 Mon Sep 17 00:00:00 2001 From: Jaegeuk Kim Date: Sat, 14 Sep 2024 21:44:10 +0000 Subject: [PATCH 105/107] f2fs: allow F2FS_IPU_NOCACHE for pinned file This patch allows f2fs to submit bios of in-place writes on pinned file. Reviewed-by: Daeho Jeong Reviewed-by: Chao Yu Signed-off-by: Jaegeuk Kim --- fs/f2fs/sysfs.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index d18931b7b094..c56e8c873935 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -792,7 +792,8 @@ out: if (!strcmp(a->attr.name, "ipu_policy")) { if (t >= BIT(F2FS_IPU_MAX)) return -EINVAL; - if (t && f2fs_lfs_mode(sbi)) + /* allow F2FS_IPU_NOCACHE only for IPU in the pinned file */ + if (f2fs_lfs_mode(sbi) && (t & ~BIT(F2FS_IPU_NOCACHE))) return -EINVAL; SM_I(sbi)->ipu_policy = (unsigned int)t; return count; From fcd2d9e1fdcd7cada612f2e8737fb13a2bce7d0e Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Tue, 24 Sep 2024 10:47:23 +0200 Subject: [PATCH 106/107] fuse: clear FR_PENDING if abort is detected when sending request The (!fiq->connected) check was moved into the queuing method resulting in the following: Fixes: 5de8acb41c86 ("fuse: cleanup request queuing towards virtiofs") Reported-by: Lai, Yi Closes: https://lore.kernel.org/all/ZvFEAM6JfrBKsOU0@ly-workstation/ Signed-off-by: Miklos Szeredi --- fs/fuse/dev.c | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index cdf925e0b317..53c4569d85a4 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -295,6 +295,7 @@ static void fuse_dev_queue_req(struct fuse_iqueue *fiq, struct fuse_req *req) } else { spin_unlock(&fiq->lock); req->out.h.error = -ENOTCONN; + clear_bit(FR_PENDING, &req->flags); fuse_request_end(req); } } From 2f3d8ff457982f4055fe8f7bf19d3821ba22c376 Mon Sep 17 00:00:00 2001 From: yangyun Date: Sat, 14 Sep 2024 16:51:31 +0800 Subject: [PATCH 107/107] fuse: use exclusive lock when FUSE_I_CACHE_IO_MODE is set This may be a typo. The comment has said shared locks are not allowed when this bit is set. If using shared lock, the wait in `fuse_file_cached_io_open` may be forever. Fixes: 205c1d802683 ("fuse: allow parallel dio writes with FUSE_DIRECT_IO_ALLOW_MMAP") CC: stable@vger.kernel.org # v6.9 Signed-off-by: yangyun Reviewed-by: Bernd Schubert Signed-off-by: Miklos Szeredi --- fs/fuse/file.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index b8afeca12487..1b5cd46c8225 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -1345,7 +1345,7 @@ static bool fuse_dio_wr_exclusive_lock(struct kiocb *iocb, struct iov_iter *from /* shared locks are not allowed with parallel page cache IO */ if (test_bit(FUSE_I_CACHE_IO_MODE, &fi->state)) - return false; + return true; /* Parallel dio beyond EOF is not supported, at least for now. */ if (fuse_io_past_eof(iocb, from))