From 2d63e6a3d97132449451c2f66fe24a2dc4e2938f Mon Sep 17 00:00:00 2001 From: Greg Kroah-Hartman Date: Wed, 21 Sep 2022 11:13:41 +0200 Subject: [PATCH 001/115] scripts: coccicheck: use "grep -E" instead of "egrep" The latest version of grep claims that egrep is now obsolete so the build now contains warnings that look like: egrep: warning: egrep is obsolescent; using grep -E fix this up by moving the vdso Makefile to use "grep -E" instead. Cc: Julia Lawall Cc: Nicolas Palix Cc: cocci@inria.fr Signed-off-by: Greg Kroah-Hartman --- scripts/coccicheck | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/coccicheck b/scripts/coccicheck index caba0bff6da7..2956fce8fa4f 100755 --- a/scripts/coccicheck +++ b/scripts/coccicheck @@ -47,7 +47,7 @@ FLAGS="--very-quiet" # inspected there. # # --profile will not output if --very-quiet is used, so avoid it. -echo $SPFLAGS | egrep -e "--profile|--show-trying" 2>&1 > /dev/null +echo $SPFLAGS | grep -E -e "--profile|--show-trying" 2>&1 > /dev/null if [ $? -eq 0 ]; then FLAGS="--quiet" fi From 6e0149a55379d6b91948cf6e94e700eb7dbaa518 Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Sun, 27 Nov 2022 13:59:25 +0100 Subject: [PATCH 002/115] 9p/fs: Remove unneeded idr.h #include The 9p fs does not use IDR or IDA functionalities. So there is no point in including . Remove it. Link: https://lkml.kernel.org/r/3d1e0ed9714eaee7e18d9f5b0b4bfa49b00b286d.1669553950.git.christophe.jaillet@wanadoo.fr Signed-off-by: Christophe JAILLET Reviewed-by: Christian Schoenebeck [Dominique: reword subject] Signed-off-by: Dominique Martinet --- fs/9p/fid.c | 1 - fs/9p/v9fs.c | 1 - fs/9p/vfs_addr.c | 1 - fs/9p/vfs_dentry.c | 1 - fs/9p/vfs_dir.c | 1 - fs/9p/vfs_file.c | 1 - fs/9p/vfs_inode.c | 1 - fs/9p/vfs_inode_dotl.c | 1 - fs/9p/vfs_super.c | 1 - 9 files changed, 9 deletions(-) diff --git a/fs/9p/fid.c b/fs/9p/fid.c index 23cf9b2fbfe4..805151114e96 100644 --- a/fs/9p/fid.c +++ b/fs/9p/fid.c @@ -11,7 +11,6 @@ #include #include #include -#include #include #include diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c index 0129de2ea31a..3a9c4517265f 100644 --- a/fs/9p/v9fs.c +++ b/fs/9p/v9fs.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_addr.c b/fs/9p/vfs_addr.c index 47b9a1122f34..93373486ab04 100644 --- a/fs/9p/vfs_addr.c +++ b/fs/9p/vfs_addr.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_dentry.c b/fs/9p/vfs_dentry.c index f89f01734587..65fa2df5e49b 100644 --- a/fs/9p/vfs_dentry.c +++ b/fs/9p/vfs_dentry.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c index 000fbaae9b18..1675a196c2ba 100644 --- a/fs/9p/vfs_dir.c +++ b/fs/9p/vfs_dir.c @@ -14,7 +14,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c index aec43ba83799..b740017634ef 100644 --- a/fs/9p/vfs_file.c +++ b/fs/9p/vfs_file.c @@ -18,7 +18,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 4d1a4a8d9277..27a04a226d97 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -17,7 +17,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index 5cfa4b4f070f..8696e8899c27 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include #include diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 2d9ee073d12c..266c4693e20c 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -15,7 +15,6 @@ #include #include #include -#include #include #include #include From 31fff92c9cadbca4cd294bcdff285ca3fc9bad7c Mon Sep 17 00:00:00 2001 From: Christophe JAILLET Date: Sun, 27 Nov 2022 15:46:45 +0100 Subject: [PATCH 003/115] 9p/net: Remove unneeded idr.h #include The 9p net files don't use IDR or IDA functionalities. So there is no point in including . Remove it. Link: https://lkml.kernel.org/r/9e386018601d7e4a9e5d7da8fc3e9555ebb25c87.1669560387.git.christophe.jaillet@wanadoo.fr Signed-off-by: Christophe JAILLET Reviewed-by: Christian Schoenebeck Signed-off-by: Dominique Martinet --- net/9p/trans_fd.c | 1 - net/9p/trans_rdma.c | 1 - net/9p/trans_virtio.c | 1 - 3 files changed, 3 deletions(-) diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c index eeea0a6a75b6..06ec9f7d3318 100644 --- a/net/9p/trans_fd.c +++ b/net/9p/trans_fd.c @@ -20,7 +20,6 @@ #include #include #include -#include #include #include #include diff --git a/net/9p/trans_rdma.c b/net/9p/trans_rdma.c index 6ff706760676..33a9ac6f2d55 100644 --- a/net/9p/trans_rdma.c +++ b/net/9p/trans_rdma.c @@ -21,7 +21,6 @@ #include #include #include -#include #include #include #include diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index e757f0601304..19bccfa0d593 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -22,7 +22,6 @@ #include #include #include -#include #include #include #include From 26273ade77f54716e30dfd40ac6e85ceb54ac0f9 Mon Sep 17 00:00:00 2001 From: Schspa Shi Date: Thu, 1 Dec 2022 11:33:10 +0800 Subject: [PATCH 004/115] 9p: set req refcount to zero to avoid uninitialized usage When a new request is allocated, the refcount will be zero if it is reused, but if the request is newly allocated from slab, it is not fully initialized before being added to idr. If the p9_read_work got a response before the refcount initiated. It will use a uninitialized req, which will result in a bad request data struct. Here is the logs from syzbot. Corrupted memory at 0xffff88807eade00b [ 0xff 0x07 0x00 0x00 0x00 0x00 0x00 0x00 . . . . . . . . ] (in kfence-#110): p9_fcall_fini net/9p/client.c:248 [inline] p9_req_put net/9p/client.c:396 [inline] p9_req_put+0x208/0x250 net/9p/client.c:390 p9_client_walk+0x247/0x540 net/9p/client.c:1165 clone_fid fs/9p/fid.h:21 [inline] v9fs_fid_xattr_set+0xe4/0x2b0 fs/9p/xattr.c:118 v9fs_xattr_set fs/9p/xattr.c:100 [inline] v9fs_xattr_handler_set+0x6f/0x120 fs/9p/xattr.c:159 __vfs_setxattr+0x119/0x180 fs/xattr.c:182 __vfs_setxattr_noperm+0x129/0x5f0 fs/xattr.c:216 __vfs_setxattr_locked+0x1d3/0x260 fs/xattr.c:277 vfs_setxattr+0x143/0x340 fs/xattr.c:309 setxattr+0x146/0x160 fs/xattr.c:617 path_setxattr+0x197/0x1c0 fs/xattr.c:636 __do_sys_setxattr fs/xattr.c:652 [inline] __se_sys_setxattr fs/xattr.c:648 [inline] __ia32_sys_setxattr+0xc0/0x160 fs/xattr.c:648 do_syscall_32_irqs_on arch/x86/entry/common.c:112 [inline] __do_fast_syscall_32+0x65/0xf0 arch/x86/entry/common.c:178 do_fast_syscall_32+0x33/0x70 arch/x86/entry/common.c:203 entry_SYSENTER_compat_after_hwframe+0x70/0x82 Below is a similar scenario, the scenario in the syzbot log looks more complicated than this one, but this patch can fix it. T21124 p9_read_work ======================== second trans ================================= p9_client_walk p9_client_rpc p9_client_prepare_req p9_tag_alloc req = kmem_cache_alloc(p9_req_cache, GFP_NOFS); tag = idr_alloc << preempted >> req->tc.tag = tag; /* req->[refcount/tag] == uninitialized */ m->rreq = p9_tag_lookup(m->client, m->rc.tag); /* increments uninitalized refcount */ refcount_set(&req->refcount, 2); /* cb drops one ref */ p9_client_cb(req) /* reader thread drops its ref: request is incorrectly freed */ p9_req_put(req) /* use after free and ref underflow */ p9_req_put(req) To fix it, we can initialize the refcount to zero before add to idr. Link: https://lkml.kernel.org/r/20221201033310.18589-1-schspa@gmail.com Cc: stable@vger.kernel.org # 6.0+ due to 6cda12864cb0 ("9p: Drop kref usage") Fixes: 728356dedeff ("9p: Add refcount to p9_req_t") Reported-by: syzbot+8f1060e2aaf8ca55220b@syzkaller.appspotmail.com Signed-off-by: Schspa Shi Reviewed-by: Christian Schoenebeck Signed-off-by: Dominique Martinet --- net/9p/client.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/net/9p/client.c b/net/9p/client.c index aaa37b07e30a..b554f8357f96 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -297,6 +297,11 @@ p9_tag_alloc(struct p9_client *c, int8_t type, uint t_size, uint r_size, p9pdu_reset(&req->rc); req->t_err = 0; req->status = REQ_STATUS_ALLOC; + /* refcount needs to be set to 0 before inserting into the idr + * so p9_tag_lookup does not accept a request that is not fully + * initialized. refcount_set to 2 below will mark request ready. + */ + refcount_set(&req->refcount, 0); init_waitqueue_head(&req->wq); INIT_LIST_HEAD(&req->req_list); From f15e006b831384aaec4b4f13265c0dff88ef09dd Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Tue, 22 Nov 2022 09:06:56 +0900 Subject: [PATCH 005/115] 9p/xen: do not memcpy header into req->rc while 'h' is packed and can be assumed to match the request payload, req->rc is a struct p9_fcall which is not packed and that memcpy could be wrong. Fix this by copying each fields individually instead. Reported-by: Christian Schoenebeck Reviewed-by: Christian Schoenebeck Suggested-by: Stefano Stabellini Reviewed-by: Stefano Stabellini Link: https://lkml.kernel.org/r/alpine.DEB.2.22.394.2211211454540.1049131@ubuntu-linux-20-04-desktop Link: https://lkml.kernel.org/r/20221122001025.119121-1-asmadeus@codewreck.org Signed-off-by: Dominique Martinet --- net/9p/trans_xen.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/net/9p/trans_xen.c b/net/9p/trans_xen.c index aaa5fd364691..de2d2ca8819a 100644 --- a/net/9p/trans_xen.c +++ b/net/9p/trans_xen.c @@ -216,7 +216,9 @@ static void p9_xen_response(struct work_struct *work) goto recv_error; } - memcpy(&req->rc, &h, sizeof(h)); + req->rc.size = h.size; + req->rc.id = h.id; + req->rc.tag = h.tag; req->rc.offset = 0; masked_cons = xen_9pfs_mask(cons, XEN_9PFS_RING_SIZE(ring)); From 97d73d978271ade27fc751ad606f23c1c4c43678 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:26 +0100 Subject: [PATCH 006/115] ALSA: hda: Allow for compress stream to hdac_ext_stream assignment Currently only PCM streams can enlist hdac_stream for their data transfer. Add cstream field to hdac_ext_stream to expose possibility of compress stream assignment in place of PCM one. Limited to HOST-type only as there no other users on the horizon. Signed-off-by: Cezary Rojewski Acked-by: Takashi Iwai Link: https://lore.kernel.org/r/20221202152841.672536-2-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- include/sound/hdaudio_ext.h | 2 ++ sound/hda/ext/hdac_ext_stream.c | 41 +++++++++++++++++++++++++++++++++ 2 files changed, 43 insertions(+) diff --git a/include/sound/hdaudio_ext.h b/include/sound/hdaudio_ext.h index 68ab89211de2..511211f4a2b6 100644 --- a/include/sound/hdaudio_ext.h +++ b/include/sound/hdaudio_ext.h @@ -75,6 +75,8 @@ struct hdac_ext_stream *snd_hdac_ext_stream_assign(struct hdac_bus *bus, struct snd_pcm_substream *substream, int type); void snd_hdac_ext_stream_release(struct hdac_ext_stream *hext_stream, int type); +struct hdac_ext_stream *snd_hdac_ext_cstream_assign(struct hdac_bus *bus, + struct snd_compr_stream *cstream); void snd_hdac_ext_stream_decouple_locked(struct hdac_bus *bus, struct hdac_ext_stream *hext_stream, bool decouple); void snd_hdac_ext_stream_decouple(struct hdac_bus *bus, diff --git a/sound/hda/ext/hdac_ext_stream.c b/sound/hda/ext/hdac_ext_stream.c index 2a071a09224d..11b7119cc47e 100644 --- a/sound/hda/ext/hdac_ext_stream.c +++ b/sound/hda/ext/hdac_ext_stream.c @@ -14,6 +14,7 @@ #include #include #include +#include /** * snd_hdac_ext_stream_init - initialize each stream (aka device) @@ -367,3 +368,43 @@ void snd_hdac_ext_stream_release(struct hdac_ext_stream *hext_stream, int type) } EXPORT_SYMBOL_GPL(snd_hdac_ext_stream_release); + +/** + * snd_hdac_ext_cstream_assign - assign a host stream for compress + * @bus: HD-audio core bus + * @cstream: Compress stream to assign + * + * Assign an unused host stream for the given compress stream. + * If no stream is free, NULL is returned. Stream is decoupled + * before assignment. + */ +struct hdac_ext_stream *snd_hdac_ext_cstream_assign(struct hdac_bus *bus, + struct snd_compr_stream *cstream) +{ + struct hdac_ext_stream *res = NULL; + struct hdac_stream *hstream; + + spin_lock_irq(&bus->reg_lock); + list_for_each_entry(hstream, &bus->stream_list, list) { + struct hdac_ext_stream *hext_stream = stream_to_hdac_ext_stream(hstream); + + if (hstream->direction != cstream->direction) + continue; + + if (!hstream->opened) { + res = hext_stream; + break; + } + } + + if (res) { + snd_hdac_ext_stream_decouple_locked(bus, res, true); + res->hstream.opened = 1; + res->hstream.running = 0; + res->hstream.cstream = cstream; + } + spin_unlock_irq(&bus->reg_lock); + + return res; +} +EXPORT_SYMBOL_GPL(snd_hdac_ext_cstream_assign); From f6b1254664a0a15c8bbe0a17b2c86840aa38d3d7 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:27 +0100 Subject: [PATCH 007/115] ALSA: hda: Prepare for compress stream support Before introducing compress specific changes, adjust BDL and parameters setup functions so these are not tightly coupled with PCM streams. Signed-off-by: Cezary Rojewski Acked-by: Takashi Iwai Link: https://lore.kernel.org/r/20221202152841.672536-3-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/hda/hdac_stream.c | 32 ++++++++++++++++---------------- 1 file changed, 16 insertions(+), 16 deletions(-) diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c index 3b250ee7f6a7..8a12c6347914 100644 --- a/sound/hda/hdac_stream.c +++ b/sound/hda/hdac_stream.c @@ -487,11 +487,15 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) { struct hdac_bus *bus = azx_dev->bus; struct snd_pcm_substream *substream = azx_dev->substream; - struct snd_pcm_runtime *runtime = substream->runtime; + struct snd_pcm_runtime *runtime; + struct snd_dma_buffer *dmab; __le32 *bdl; int i, ofs, periods, period_bytes; int pos_adj, pos_align; + runtime = substream->runtime; + dmab = snd_pcm_get_dma_buf(substream); + /* reset BDL address */ snd_hdac_stream_writel(azx_dev, SD_BDLPL, 0); snd_hdac_stream_writel(azx_dev, SD_BDLPU, 0); @@ -505,7 +509,7 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) azx_dev->frags = 0; pos_adj = bus->bdl_pos_adj; - if (!azx_dev->no_period_wakeup && pos_adj > 0) { + if (runtime && !azx_dev->no_period_wakeup && pos_adj > 0) { pos_align = pos_adj; pos_adj = DIV_ROUND_UP(pos_adj * runtime->rate, 48000); if (!pos_adj) @@ -518,8 +522,7 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) pos_adj); pos_adj = 0; } else { - ofs = setup_bdle(bus, snd_pcm_get_dma_buf(substream), - azx_dev, + ofs = setup_bdle(bus, dmab, azx_dev, &bdl, ofs, pos_adj, true); if (ofs < 0) goto error; @@ -529,13 +532,11 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) for (i = 0; i < periods; i++) { if (i == periods - 1 && pos_adj) - ofs = setup_bdle(bus, snd_pcm_get_dma_buf(substream), - azx_dev, &bdl, ofs, - period_bytes - pos_adj, 0); + ofs = setup_bdle(bus, dmab, azx_dev, + &bdl, ofs, period_bytes - pos_adj, 0); else - ofs = setup_bdle(bus, snd_pcm_get_dma_buf(substream), - azx_dev, &bdl, ofs, - period_bytes, + ofs = setup_bdle(bus, dmab, azx_dev, + &bdl, ofs, period_bytes, !azx_dev->no_period_wakeup); if (ofs < 0) goto error; @@ -560,26 +561,25 @@ EXPORT_SYMBOL_GPL(snd_hdac_stream_setup_periods); int snd_hdac_stream_set_params(struct hdac_stream *azx_dev, unsigned int format_val) { - - unsigned int bufsize, period_bytes; struct snd_pcm_substream *substream = azx_dev->substream; - struct snd_pcm_runtime *runtime; + unsigned int bufsize, period_bytes; + unsigned int no_period_wakeup; int err; if (!substream) return -EINVAL; - runtime = substream->runtime; bufsize = snd_pcm_lib_buffer_bytes(substream); period_bytes = snd_pcm_lib_period_bytes(substream); + no_period_wakeup = substream->runtime->no_period_wakeup; if (bufsize != azx_dev->bufsize || period_bytes != azx_dev->period_bytes || format_val != azx_dev->format_val || - runtime->no_period_wakeup != azx_dev->no_period_wakeup) { + no_period_wakeup != azx_dev->no_period_wakeup) { azx_dev->bufsize = bufsize; azx_dev->period_bytes = period_bytes; azx_dev->format_val = format_val; - azx_dev->no_period_wakeup = runtime->no_period_wakeup; + azx_dev->no_period_wakeup = no_period_wakeup; err = snd_hdac_stream_setup_periods(azx_dev); if (err < 0) return err; From 3e9582267e3a06bfd9622dbd2304a8cfac977b43 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:28 +0100 Subject: [PATCH 008/115] ALSA: hda: Interrupt servicing and BDL setup for compress streams Account for compress streams when receiving and servicing buffer completed interrupts. In case of compress stream enlisting hdac_stream for data transfer, setup BDL entries much like it is the case for PCM streams. Signed-off-by: Divya Prakash Signed-off-by: Cezary Rojewski Acked-by: Takashi Iwai Link: https://lore.kernel.org/r/20221202152841.672536-4-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/hda/hdac_controller.c | 4 ++-- sound/hda/hdac_stream.c | 27 ++++++++++++++++++++------- 2 files changed, 22 insertions(+), 9 deletions(-) diff --git a/sound/hda/hdac_controller.c b/sound/hda/hdac_controller.c index 9a60bfdb39ba..3c7af6558249 100644 --- a/sound/hda/hdac_controller.c +++ b/sound/hda/hdac_controller.c @@ -578,8 +578,8 @@ int snd_hdac_bus_handle_stream_irq(struct hdac_bus *bus, unsigned int status, sd_status = snd_hdac_stream_readb(azx_dev, SD_STS); snd_hdac_stream_writeb(azx_dev, SD_STS, SD_INT_MASK); handled |= 1 << azx_dev->index; - if (!azx_dev->substream || !azx_dev->running || - !(sd_status & SD_INT_COMPLETE)) + if ((!azx_dev->substream && !azx_dev->cstream) || + !azx_dev->running || !(sd_status & SD_INT_COMPLETE)) continue; if (ack) ack(bus, azx_dev); diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c index 8a12c6347914..8f625402505f 100644 --- a/sound/hda/hdac_stream.c +++ b/sound/hda/hdac_stream.c @@ -7,6 +7,7 @@ #include #include #include +#include #include #include #include @@ -487,14 +488,19 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) { struct hdac_bus *bus = azx_dev->bus; struct snd_pcm_substream *substream = azx_dev->substream; - struct snd_pcm_runtime *runtime; + struct snd_compr_stream *cstream = azx_dev->cstream; + struct snd_pcm_runtime *runtime = NULL; struct snd_dma_buffer *dmab; __le32 *bdl; int i, ofs, periods, period_bytes; int pos_adj, pos_align; - runtime = substream->runtime; - dmab = snd_pcm_get_dma_buf(substream); + if (substream) { + runtime = substream->runtime; + dmab = snd_pcm_get_dma_buf(substream); + } else if (cstream) { + dmab = snd_pcm_get_dma_buf(cstream); + } /* reset BDL address */ snd_hdac_stream_writel(azx_dev, SD_BDLPL, 0); @@ -562,15 +568,22 @@ int snd_hdac_stream_set_params(struct hdac_stream *azx_dev, unsigned int format_val) { struct snd_pcm_substream *substream = azx_dev->substream; + struct snd_compr_stream *cstream = azx_dev->cstream; unsigned int bufsize, period_bytes; unsigned int no_period_wakeup; int err; - if (!substream) + if (substream) { + bufsize = snd_pcm_lib_buffer_bytes(substream); + period_bytes = snd_pcm_lib_period_bytes(substream); + no_period_wakeup = substream->runtime->no_period_wakeup; + } else if (cstream) { + bufsize = cstream->runtime->buffer_size; + period_bytes = cstream->runtime->fragment_size; + no_period_wakeup = 0; + } else { return -EINVAL; - bufsize = snd_pcm_lib_buffer_bytes(substream); - period_bytes = snd_pcm_lib_period_bytes(substream); - no_period_wakeup = substream->runtime->no_period_wakeup; + } if (bufsize != azx_dev->bufsize || period_bytes != azx_dev->period_bytes || From bb03099bf2253fcd1a4d57e6f5ee4e8000911e77 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:29 +0100 Subject: [PATCH 009/115] ASoC: Intel: avs: Introduce avs_log_buffer_status_locked() Simplify locking of firmware log gathering by providing single location for such purpose. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-5-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/apl.c | 5 +---- sound/soc/intel/avs/avs.h | 12 ++++++++++++ sound/soc/intel/avs/ipc.c | 2 +- sound/soc/intel/avs/skl.c | 7 +------ 4 files changed, 15 insertions(+), 11 deletions(-) diff --git a/sound/soc/intel/avs/apl.c b/sound/soc/intel/avs/apl.c index 7c8ce98eda9d..821d5a9ad25f 100644 --- a/sound/soc/intel/avs/apl.c +++ b/sound/soc/intel/avs/apl.c @@ -50,7 +50,6 @@ static int apl_enable_logs(struct avs_dev *adev, enum avs_log_enable enable, u32 static int apl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) { struct apl_log_buffer_layout layout; - unsigned long flags; void __iomem *addr, *buf; addr = avs_log_buffer_addr(adev, msg->log.core); @@ -59,7 +58,6 @@ static int apl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg memcpy_fromio(&layout, addr, sizeof(layout)); - spin_lock_irqsave(&adev->dbg.trace_lock, flags); if (!kfifo_initialized(&adev->dbg.trace_fifo)) /* consume the logs regardless of consumer presence */ goto update_read_ptr; @@ -78,7 +76,6 @@ static int apl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg wake_up(&adev->dbg.trace_waitq); update_read_ptr: - spin_unlock_irqrestore(&adev->dbg.trace_lock, flags); writel(layout.write_ptr, addr); return 0; } @@ -140,7 +137,7 @@ static int apl_coredump(struct avs_dev *adev, union avs_notify_msg *msg) * gathered before dumping stack */ lbs_msg.log.core = msg->ext.coredump.core_id; - avs_dsp_op(adev, log_buffer_status, &lbs_msg); + avs_log_buffer_status_locked(adev, &lbs_msg); } pos = dump + AVS_FW_REGS_SIZE; diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index 8d05b27608fe..1c89af6240d2 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -344,6 +344,18 @@ unsigned int __kfifo_fromio_locked(struct kfifo *fifo, const void __iomem *src, (avs_sram_addr(adev, AVS_DEBUG_WINDOW) + __offset); \ }) +static inline int avs_log_buffer_status_locked(struct avs_dev *adev, union avs_notify_msg *msg) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&adev->dbg.trace_lock, flags); + ret = avs_dsp_op(adev, log_buffer_status, msg); + spin_unlock_irqrestore(&adev->dbg.trace_lock, flags); + + return ret; +} + struct apl_log_buffer_layout { u32 read_ptr; u32 write_ptr; diff --git a/sound/soc/intel/avs/ipc.c b/sound/soc/intel/avs/ipc.c index af8a260093f4..bdf013c3dd12 100644 --- a/sound/soc/intel/avs/ipc.c +++ b/sound/soc/intel/avs/ipc.c @@ -266,7 +266,7 @@ static void avs_dsp_process_notification(struct avs_dev *adev, u64 header) break; case AVS_NOTIFY_LOG_BUFFER_STATUS: - avs_dsp_op(adev, log_buffer_status, &msg); + avs_log_buffer_status_locked(adev, &msg); break; case AVS_NOTIFY_EXCEPTION_CAUGHT: diff --git a/sound/soc/intel/avs/skl.c b/sound/soc/intel/avs/skl.c index dc98b5cf900f..ff690e99d960 100644 --- a/sound/soc/intel/avs/skl.c +++ b/sound/soc/intel/avs/skl.c @@ -55,15 +55,11 @@ int skl_log_buffer_offset(struct avs_dev *adev, u32 core) static int skl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) { - unsigned long flags; void __iomem *buf; u16 size, write, offset; - spin_lock_irqsave(&adev->dbg.trace_lock, flags); - if (!kfifo_initialized(&adev->dbg.trace_fifo)) { - spin_unlock_irqrestore(&adev->dbg.trace_lock, flags); + if (!kfifo_initialized(&adev->dbg.trace_fifo)) return 0; - } size = avs_log_buffer_size(adev) / 2; write = readl(avs_sram_addr(adev, AVS_FW_REGS_WINDOW) + FW_REGS_DBG_LOG_WP(msg->log.core)); @@ -74,7 +70,6 @@ skl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) buf = avs_log_buffer_addr(adev, msg->log.core) + offset; __kfifo_fromio_locked(&adev->dbg.trace_fifo, buf, size, &adev->dbg.fifo_lock); wake_up(&adev->dbg.trace_waitq); - spin_unlock_irqrestore(&adev->dbg.trace_lock, flags); return 0; } From 58029b7734ec84738aeb8fb391e625832bb6b0a6 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:30 +0100 Subject: [PATCH 010/115] ASoC: Intel: avs: Drop fifo_lock Log gathering is already locked, thanks to ->trace_lock. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-6-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/apl.c | 9 ++++----- sound/soc/intel/avs/avs.h | 4 +--- sound/soc/intel/avs/skl.c | 2 +- sound/soc/intel/avs/utils.c | 6 +----- 4 files changed, 7 insertions(+), 14 deletions(-) diff --git a/sound/soc/intel/avs/apl.c b/sound/soc/intel/avs/apl.c index 821d5a9ad25f..66672ffd95df 100644 --- a/sound/soc/intel/avs/apl.c +++ b/sound/soc/intel/avs/apl.c @@ -65,13 +65,12 @@ static int apl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg buf = apl_log_payload_addr(addr); if (layout.read_ptr > layout.write_ptr) { - __kfifo_fromio_locked(&adev->dbg.trace_fifo, buf + layout.read_ptr, - apl_log_payload_size(adev) - layout.read_ptr, - &adev->dbg.fifo_lock); + __kfifo_fromio(&adev->dbg.trace_fifo, buf + layout.read_ptr, + apl_log_payload_size(adev) - layout.read_ptr); layout.read_ptr = 0; } - __kfifo_fromio_locked(&adev->dbg.trace_fifo, buf + layout.read_ptr, - layout.write_ptr - layout.read_ptr, &adev->dbg.fifo_lock); + __kfifo_fromio(&adev->dbg.trace_fifo, buf + layout.read_ptr, + layout.write_ptr - layout.read_ptr); wake_up(&adev->dbg.trace_waitq); diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index 1c89af6240d2..957151ecf39a 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -95,7 +95,6 @@ struct avs_fw_entry { struct avs_debug { struct kfifo trace_fifo; - spinlock_t fifo_lock; /* serialize I/O for trace_fifo */ spinlock_t trace_lock; /* serialize debug window I/O between each LOG_BUFFER_STATUS */ wait_queue_head_t trace_waitq; u32 aging_timer_period; @@ -331,8 +330,7 @@ void avs_unregister_all_boards(struct avs_dev *adev); /* Firmware tracing helpers */ -unsigned int __kfifo_fromio_locked(struct kfifo *fifo, const void __iomem *src, unsigned int len, - spinlock_t *lock); +unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len); #define avs_log_buffer_size(adev) \ ((adev)->fw_cfg.trace_log_bytes / (adev)->hw_cfg.dsp_cores) diff --git a/sound/soc/intel/avs/skl.c b/sound/soc/intel/avs/skl.c index ff690e99d960..936cd44eb73e 100644 --- a/sound/soc/intel/avs/skl.c +++ b/sound/soc/intel/avs/skl.c @@ -68,7 +68,7 @@ skl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) /* Address is guaranteed to exist in SRAM2. */ buf = avs_log_buffer_addr(adev, msg->log.core) + offset; - __kfifo_fromio_locked(&adev->dbg.trace_fifo, buf, size, &adev->dbg.fifo_lock); + __kfifo_fromio(&adev->dbg.trace_fifo, buf, size); wake_up(&adev->dbg.trace_waitq); return 0; diff --git a/sound/soc/intel/avs/utils.c b/sound/soc/intel/avs/utils.c index 13611dee9787..75ad434d7dfb 100644 --- a/sound/soc/intel/avs/utils.c +++ b/sound/soc/intel/avs/utils.c @@ -301,14 +301,11 @@ void avs_release_firmwares(struct avs_dev *adev) } } -unsigned int __kfifo_fromio_locked(struct kfifo *fifo, const void __iomem *src, unsigned int len, - spinlock_t *lock) +unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len) { struct __kfifo *__fifo = &fifo->kfifo; - unsigned long flags; unsigned int l, off; - spin_lock_irqsave(lock, flags); len = min(len, kfifo_avail(fifo)); off = __fifo->in & __fifo->mask; l = min(len, kfifo_size(fifo) - off); @@ -318,7 +315,6 @@ unsigned int __kfifo_fromio_locked(struct kfifo *fifo, const void __iomem *src, /* Make sure data copied from SRAM is visible to all CPUs. */ smp_mb(); __fifo->in += len; - spin_unlock_irqrestore(lock, flags); return len; } From 9e3c15beb8976771f95ba30b3da8bd35dc7188ac Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:31 +0100 Subject: [PATCH 011/115] ASoC: Intel: avs: Introduce debug-context aware helpers Debug-related fields and log-dumping are useful when debugfs is enabled. Define them under CONFIG_DEBUG_FS and provide stubs when the config is disabled so that the code that makes use of these needs not to be complicated unnecessarily. Members that are duplicated by this patch will be removed by the follow up changes. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-7-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/Makefile | 4 ++++ sound/soc/intel/avs/avs.h | 29 +++++++++++++++++++++++++++++ sound/soc/intel/avs/debugfs.c | 28 ++++++++++++++++++++++++++++ 3 files changed, 61 insertions(+) create mode 100644 sound/soc/intel/avs/debugfs.c diff --git a/sound/soc/intel/avs/Makefile b/sound/soc/intel/avs/Makefile index 919212825f21..a211a0b7b4a8 100644 --- a/sound/soc/intel/avs/Makefile +++ b/sound/soc/intel/avs/Makefile @@ -9,6 +9,10 @@ snd-soc-avs-objs += trace.o # tell define_trace.h where to find the trace header CFLAGS_trace.o := -I$(src) +ifneq ($(CONFIG_DEBUG_FS),) +snd-soc-avs-objs += debugfs.o +endif + obj-$(CONFIG_SND_SOC_INTEL_AVS) += snd-soc-avs.o # Machine support diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index 957151ecf39a..3687d03f87d4 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -9,6 +9,7 @@ #ifndef __SOUND_SOC_INTEL_AVS_H #define __SOUND_SOC_INTEL_AVS_H +#include #include #include #include @@ -146,6 +147,14 @@ struct avs_dev { struct mutex path_mutex; struct avs_debug dbg; + spinlock_t trace_lock; /* serialize debug window I/O between each LOG_BUFFER_STATUS */ +#ifdef CONFIG_DEBUG_FS + struct kfifo trace_fifo; + wait_queue_head_t trace_waitq; + u32 aging_timer_period; + u32 fifo_full_timer_period; + u32 logged_resources; /* context dependent: core or library */ +#endif }; /* from hda_bus to avs_dev */ @@ -366,4 +375,24 @@ struct apl_log_buffer_layout { #define apl_log_payload_addr(addr) \ (addr + sizeof(struct apl_log_buffer_layout)) +#ifdef CONFIG_DEBUG_FS +bool avs_logging_fw(struct avs_dev *adev); +void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int len); +void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len); +#else +static inline bool avs_logging_fw(struct avs_dev *adev) +{ + return false; +} + +static inline void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int len) +{ +} + +static inline void +avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len) +{ +} +#endif + #endif /* __SOUND_SOC_INTEL_AVS_H */ diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c new file mode 100644 index 000000000000..ac3889e21542 --- /dev/null +++ b/sound/soc/intel/avs/debugfs.c @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0-only +// +// Copyright(c) 2021-2022 Intel Corporation. All rights reserved. +// +// Authors: Cezary Rojewski +// Amadeusz Slawinski +// + +#include +#include +#include +#include "avs.h" + +bool avs_logging_fw(struct avs_dev *adev) +{ + return kfifo_initialized(&adev->trace_fifo); +} + +void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int len) +{ + __kfifo_fromio(&adev->trace_fifo, src, len); +} + +void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len) +{ + avs_dump_fw_log(adev, src, len); + wake_up(&adev->trace_waitq); +} From b3eefa5d8dbfe5286c3308fa706fc9c45b38fe19 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:32 +0100 Subject: [PATCH 012/115] ASoC: Intel: avs: Make enable_logs() dependent on DEBUG_FS Without debug filesystem present, this code is redundant. Operations: log_buffer_status and log_buffer_offset are left as is as EXCEPTION_CAUGHT and even unexpected LOG_BUFFER_STATUS notifications may occur without user ever touching debugfs. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-8-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/apl.c | 7 ++++--- sound/soc/intel/avs/avs.h | 5 +++++ sound/soc/intel/avs/messages.c | 36 ++++++++++++++++++---------------- sound/soc/intel/avs/skl.c | 7 ++++--- 4 files changed, 32 insertions(+), 23 deletions(-) diff --git a/sound/soc/intel/avs/apl.c b/sound/soc/intel/avs/apl.c index 66672ffd95df..beef308c9428 100644 --- a/sound/soc/intel/avs/apl.c +++ b/sound/soc/intel/avs/apl.c @@ -13,8 +13,9 @@ #include "path.h" #include "topology.h" -static int apl_enable_logs(struct avs_dev *adev, enum avs_log_enable enable, u32 aging_period, - u32 fifo_full_period, unsigned long resource_mask, u32 *priorities) +static int __maybe_unused +apl_enable_logs(struct avs_dev *adev, enum avs_log_enable enable, u32 aging_period, + u32 fifo_full_period, unsigned long resource_mask, u32 *priorities) { struct apl_log_state_info *info; u32 size, num_cores = adev->hw_cfg.dsp_cores; @@ -239,10 +240,10 @@ const struct avs_dsp_ops apl_dsp_ops = { .load_basefw = avs_hda_load_basefw, .load_lib = avs_hda_load_library, .transfer_mods = avs_hda_transfer_modules, - .enable_logs = apl_enable_logs, .log_buffer_offset = skl_log_buffer_offset, .log_buffer_status = apl_log_buffer_status, .coredump = apl_coredump, .d0ix_toggle = apl_d0ix_toggle, .set_d0ix = apl_set_d0ix, + AVS_SET_ENABLE_LOGS_OP(apl) }; diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index 3687d03f87d4..f8f11d8b5936 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -376,10 +376,15 @@ struct apl_log_buffer_layout { (addr + sizeof(struct apl_log_buffer_layout)) #ifdef CONFIG_DEBUG_FS +#define AVS_SET_ENABLE_LOGS_OP(name) \ + .enable_logs = name##_enable_logs + bool avs_logging_fw(struct avs_dev *adev); void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int len); void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len); #else +#define AVS_SET_ENABLE_LOGS_OP(name) + static inline bool avs_logging_fw(struct avs_dev *adev) { return false; diff --git a/sound/soc/intel/avs/messages.c b/sound/soc/intel/avs/messages.c index 6b0fecbf07c3..f734d49e42be 100644 --- a/sound/soc/intel/avs/messages.c +++ b/sound/soc/intel/avs/messages.c @@ -685,6 +685,24 @@ int avs_ipc_get_modules_info(struct avs_dev *adev, struct avs_mods_info **info) return 0; } +int avs_ipc_copier_set_sink_format(struct avs_dev *adev, u16 module_id, + u8 instance_id, u32 sink_id, + const struct avs_audio_format *src_fmt, + const struct avs_audio_format *sink_fmt) +{ + struct avs_copier_sink_format cpr_fmt; + + cpr_fmt.sink_id = sink_id; + /* Firmware expects driver to resend copier's input format. */ + cpr_fmt.src_fmt = *src_fmt; + cpr_fmt.sink_fmt = *sink_fmt; + + return avs_ipc_set_large_config(adev, module_id, instance_id, + AVS_COPIER_SET_SINK_FORMAT, + (u8 *)&cpr_fmt, sizeof(cpr_fmt)); +} + +#ifdef CONFIG_DEBUG_FS int avs_ipc_set_enable_logs(struct avs_dev *adev, u8 *log_info, size_t size) { return avs_ipc_set_large_config(adev, AVS_BASEFW_MOD_ID, AVS_BASEFW_INST_ID, @@ -704,20 +722,4 @@ int avs_ipc_set_system_time(struct avs_dev *adev) return avs_ipc_set_large_config(adev, AVS_BASEFW_MOD_ID, AVS_BASEFW_INST_ID, AVS_BASEFW_SYSTEM_TIME, (u8 *)&sys_time, sizeof(sys_time)); } - -int avs_ipc_copier_set_sink_format(struct avs_dev *adev, u16 module_id, - u8 instance_id, u32 sink_id, - const struct avs_audio_format *src_fmt, - const struct avs_audio_format *sink_fmt) -{ - struct avs_copier_sink_format cpr_fmt; - - cpr_fmt.sink_id = sink_id; - /* Firmware expects driver to resend copier's input format. */ - cpr_fmt.src_fmt = *src_fmt; - cpr_fmt.sink_fmt = *sink_fmt; - - return avs_ipc_set_large_config(adev, module_id, instance_id, - AVS_COPIER_SET_SINK_FORMAT, - (u8 *)&cpr_fmt, sizeof(cpr_fmt)); -} +#endif diff --git a/sound/soc/intel/avs/skl.c b/sound/soc/intel/avs/skl.c index 936cd44eb73e..c5edb0b0df14 100644 --- a/sound/soc/intel/avs/skl.c +++ b/sound/soc/intel/avs/skl.c @@ -12,8 +12,9 @@ #include "avs.h" #include "messages.h" -static int skl_enable_logs(struct avs_dev *adev, enum avs_log_enable enable, u32 aging_period, - u32 fifo_full_period, unsigned long resource_mask, u32 *priorities) +static int __maybe_unused +skl_enable_logs(struct avs_dev *adev, enum avs_log_enable enable, u32 aging_period, + u32 fifo_full_period, unsigned long resource_mask, u32 *priorities) { struct skl_log_state_info *info; u32 size, num_cores = adev->hw_cfg.dsp_cores; @@ -111,10 +112,10 @@ const struct avs_dsp_ops skl_dsp_ops = { .load_basefw = avs_cldma_load_basefw, .load_lib = avs_cldma_load_library, .transfer_mods = avs_cldma_transfer_modules, - .enable_logs = skl_enable_logs, .log_buffer_offset = skl_log_buffer_offset, .log_buffer_status = skl_log_buffer_status, .coredump = skl_coredump, .d0ix_toggle = skl_d0ix_toggle, .set_d0ix = skl_set_d0ix, + AVS_SET_ENABLE_LOGS_OP(skl) }; From f7de161fc8d5e1ebac3c361a37b1d748e7086330 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:33 +0100 Subject: [PATCH 013/115] ASoC: Intel: avs: Drop usage of debug members in non-debug code Switch to debug-context aware wrappers instead of accessing debug members directly allowing for readable separation of debug and non-debug related code. Duplicates are removed along the way. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-9-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/apl.c | 11 ++++------- sound/soc/intel/avs/avs.h | 16 ++-------------- sound/soc/intel/avs/debugfs.c | 18 ++++++++++++++++++ sound/soc/intel/avs/skl.c | 5 ++--- sound/soc/intel/avs/utils.c | 18 ------------------ 5 files changed, 26 insertions(+), 42 deletions(-) diff --git a/sound/soc/intel/avs/apl.c b/sound/soc/intel/avs/apl.c index beef308c9428..02683dce277a 100644 --- a/sound/soc/intel/avs/apl.c +++ b/sound/soc/intel/avs/apl.c @@ -59,21 +59,18 @@ static int apl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg memcpy_fromio(&layout, addr, sizeof(layout)); - if (!kfifo_initialized(&adev->dbg.trace_fifo)) + if (!avs_logging_fw(adev)) /* consume the logs regardless of consumer presence */ goto update_read_ptr; buf = apl_log_payload_addr(addr); if (layout.read_ptr > layout.write_ptr) { - __kfifo_fromio(&adev->dbg.trace_fifo, buf + layout.read_ptr, - apl_log_payload_size(adev) - layout.read_ptr); + avs_dump_fw_log(adev, buf + layout.read_ptr, + apl_log_payload_size(adev) - layout.read_ptr); layout.read_ptr = 0; } - __kfifo_fromio(&adev->dbg.trace_fifo, buf + layout.read_ptr, - layout.write_ptr - layout.read_ptr); - - wake_up(&adev->dbg.trace_waitq); + avs_dump_fw_log_wakeup(adev, buf + layout.read_ptr, layout.write_ptr - layout.read_ptr); update_read_ptr: writel(layout.write_ptr, addr); diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index f8f11d8b5936..7a9fb27d3845 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -94,15 +94,6 @@ struct avs_fw_entry { struct list_head node; }; -struct avs_debug { - struct kfifo trace_fifo; - spinlock_t trace_lock; /* serialize debug window I/O between each LOG_BUFFER_STATUS */ - wait_queue_head_t trace_waitq; - u32 aging_timer_period; - u32 fifo_full_timer_period; - u32 logged_resources; /* context dependent: core or library */ -}; - /* * struct avs_dev - Intel HD-Audio driver data * @@ -146,7 +137,6 @@ struct avs_dev { spinlock_t path_list_lock; struct mutex path_mutex; - struct avs_debug dbg; spinlock_t trace_lock; /* serialize debug window I/O between each LOG_BUFFER_STATUS */ #ifdef CONFIG_DEBUG_FS struct kfifo trace_fifo; @@ -339,8 +329,6 @@ void avs_unregister_all_boards(struct avs_dev *adev); /* Firmware tracing helpers */ -unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len); - #define avs_log_buffer_size(adev) \ ((adev)->fw_cfg.trace_log_bytes / (adev)->hw_cfg.dsp_cores) @@ -356,9 +344,9 @@ static inline int avs_log_buffer_status_locked(struct avs_dev *adev, union avs_n unsigned long flags; int ret; - spin_lock_irqsave(&adev->dbg.trace_lock, flags); + spin_lock_irqsave(&adev->trace_lock, flags); ret = avs_dsp_op(adev, log_buffer_status, msg); - spin_unlock_irqrestore(&adev->dbg.trace_lock, flags); + spin_unlock_irqrestore(&adev->trace_lock, flags); return ret; } diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c index ac3889e21542..78705bcb09fb 100644 --- a/sound/soc/intel/avs/debugfs.c +++ b/sound/soc/intel/avs/debugfs.c @@ -11,6 +11,24 @@ #include #include "avs.h" +static unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len) +{ + struct __kfifo *__fifo = &fifo->kfifo; + unsigned int l, off; + + len = min(len, kfifo_avail(fifo)); + off = __fifo->in & __fifo->mask; + l = min(len, kfifo_size(fifo) - off); + + memcpy_fromio(__fifo->data + off, src, l); + memcpy_fromio(__fifo->data, src + l, len - l); + /* Make sure data copied from SRAM is visible to all CPUs. */ + smp_mb(); + __fifo->in += len; + + return len; +} + bool avs_logging_fw(struct avs_dev *adev) { return kfifo_initialized(&adev->trace_fifo); diff --git a/sound/soc/intel/avs/skl.c b/sound/soc/intel/avs/skl.c index c5edb0b0df14..6bb8bbc70442 100644 --- a/sound/soc/intel/avs/skl.c +++ b/sound/soc/intel/avs/skl.c @@ -59,7 +59,7 @@ skl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) void __iomem *buf; u16 size, write, offset; - if (!kfifo_initialized(&adev->dbg.trace_fifo)) + if (!avs_logging_fw(adev)) return 0; size = avs_log_buffer_size(adev) / 2; @@ -69,8 +69,7 @@ skl_log_buffer_status(struct avs_dev *adev, union avs_notify_msg *msg) /* Address is guaranteed to exist in SRAM2. */ buf = avs_log_buffer_addr(adev, msg->log.core) + offset; - __kfifo_fromio(&adev->dbg.trace_fifo, buf, size); - wake_up(&adev->dbg.trace_waitq); + avs_dump_fw_log_wakeup(adev, buf, size); return 0; } diff --git a/sound/soc/intel/avs/utils.c b/sound/soc/intel/avs/utils.c index 75ad434d7dfb..82416b86662d 100644 --- a/sound/soc/intel/avs/utils.c +++ b/sound/soc/intel/avs/utils.c @@ -300,21 +300,3 @@ void avs_release_firmwares(struct avs_dev *adev) kfree(entry); } } - -unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len) -{ - struct __kfifo *__fifo = &fifo->kfifo; - unsigned int l, off; - - len = min(len, kfifo_avail(fifo)); - off = __fifo->in & __fifo->mask; - l = min(len, kfifo_size(fifo) - off); - - memcpy_fromio(__fifo->data + off, src, l); - memcpy_fromio(__fifo->data, src + l, len - l); - /* Make sure data copied from SRAM is visible to all CPUs. */ - smp_mb(); - __fifo->in += len; - - return len; -} From dab8d000e25c3e91154efca287434a4f78ab65d2 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:34 +0100 Subject: [PATCH 014/115] ASoC: Intel: avs: Add data probing requests Data probing is a cAVS firmware functionality that allows for data extraction and injection directly from or to DMA stream. To support it, new functions and types are added. These facilitate communication with the firmware. Total of eight IPCs: - probe module initialization and cleanup - addition and removal of probe points - addition and removal of injection DMAs - dumping list of currently connected probe points or enlisted DMAs Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-10-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/Makefile | 2 +- sound/soc/intel/avs/messages.c | 78 ++++++++++++++++++++++++++++++++++ sound/soc/intel/avs/messages.h | 53 +++++++++++++++++++++++ sound/soc/intel/avs/probes.c | 46 ++++++++++++++++++++ 4 files changed, 178 insertions(+), 1 deletion(-) create mode 100644 sound/soc/intel/avs/probes.c diff --git a/sound/soc/intel/avs/Makefile b/sound/soc/intel/avs/Makefile index a211a0b7b4a8..1c6924a1ebca 100644 --- a/sound/soc/intel/avs/Makefile +++ b/sound/soc/intel/avs/Makefile @@ -10,7 +10,7 @@ snd-soc-avs-objs += trace.o CFLAGS_trace.o := -I$(src) ifneq ($(CONFIG_DEBUG_FS),) -snd-soc-avs-objs += debugfs.o +snd-soc-avs-objs += probes.o debugfs.o endif obj-$(CONFIG_SND_SOC_INTEL_AVS) += snd-soc-avs.o diff --git a/sound/soc/intel/avs/messages.c b/sound/soc/intel/avs/messages.c index f734d49e42be..e11ae4246416 100644 --- a/sound/soc/intel/avs/messages.c +++ b/sound/soc/intel/avs/messages.c @@ -722,4 +722,82 @@ int avs_ipc_set_system_time(struct avs_dev *adev) return avs_ipc_set_large_config(adev, AVS_BASEFW_MOD_ID, AVS_BASEFW_INST_ID, AVS_BASEFW_SYSTEM_TIME, (u8 *)&sys_time, sizeof(sys_time)); } + +int avs_ipc_probe_get_dma(struct avs_dev *adev, struct avs_probe_dma **dmas, size_t *num_dmas) +{ + size_t payload_size; + u32 module_id; + u8 *payload; + int ret; + + module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + ret = avs_ipc_get_large_config(adev, module_id, AVS_PROBE_INST_ID, AVS_PROBE_INJECTION_DMA, + NULL, 0, &payload, &payload_size); + if (ret) + return ret; + + *dmas = (struct avs_probe_dma *)payload; + *num_dmas = payload_size / sizeof(**dmas); + + return 0; +} + +int avs_ipc_probe_attach_dma(struct avs_dev *adev, struct avs_probe_dma *dmas, size_t num_dmas) +{ + u32 module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + return avs_ipc_set_large_config(adev, module_id, AVS_PROBE_INST_ID, AVS_PROBE_INJECTION_DMA, + (u8 *)dmas, array_size(sizeof(*dmas), num_dmas)); +} + +int avs_ipc_probe_detach_dma(struct avs_dev *adev, union avs_connector_node_id *node_ids, + size_t num_node_ids) +{ + u32 module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + return avs_ipc_set_large_config(adev, module_id, AVS_PROBE_INST_ID, + AVS_PROBE_INJECTION_DMA_DETACH, (u8 *)node_ids, + array_size(sizeof(*node_ids), num_node_ids)); +} + +int avs_ipc_probe_get_points(struct avs_dev *adev, struct avs_probe_point_desc **descs, + size_t *num_descs) +{ + size_t payload_size; + u32 module_id; + u8 *payload; + int ret; + + module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + ret = avs_ipc_get_large_config(adev, module_id, AVS_PROBE_INST_ID, AVS_PROBE_POINTS, NULL, + 0, &payload, &payload_size); + if (ret) + return ret; + + *descs = (struct avs_probe_point_desc *)payload; + *num_descs = payload_size / sizeof(**descs); + + return 0; +} + +int avs_ipc_probe_connect_points(struct avs_dev *adev, struct avs_probe_point_desc *descs, + size_t num_descs) +{ + u32 module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + return avs_ipc_set_large_config(adev, module_id, AVS_PROBE_INST_ID, AVS_PROBE_POINTS, + (u8 *)descs, array_size(sizeof(*descs), num_descs)); +} + +int avs_ipc_probe_disconnect_points(struct avs_dev *adev, union avs_probe_point_id *ids, + size_t num_ids) +{ + u32 module_id = avs_get_module_id(adev, &AVS_PROBE_MOD_UUID); + + return avs_ipc_set_large_config(adev, module_id, AVS_PROBE_INST_ID, + AVS_PROBE_POINTS_DISCONNECT, (u8 *)ids, + array_size(sizeof(*ids), num_ids)); +} #endif diff --git a/sound/soc/intel/avs/messages.h b/sound/soc/intel/avs/messages.h index 02b3b7a74783..9dd835527e02 100644 --- a/sound/soc/intel/avs/messages.h +++ b/sound/soc/intel/avs/messages.h @@ -802,4 +802,57 @@ int avs_ipc_copier_set_sink_format(struct avs_dev *adev, u16 module_id, const struct avs_audio_format *src_fmt, const struct avs_audio_format *sink_fmt); +#define AVS_PROBE_INST_ID 0 + +enum avs_probe_runtime_param { + AVS_PROBE_INJECTION_DMA = 1, + AVS_PROBE_INJECTION_DMA_DETACH, + AVS_PROBE_POINTS, + AVS_PROBE_POINTS_DISCONNECT, +}; + +struct avs_probe_dma { + union avs_connector_node_id node_id; + u32 dma_buffer_size; +} __packed; + +enum avs_probe_type { + AVS_PROBE_TYPE_INPUT = 0, + AVS_PROBE_TYPE_OUTPUT, + AVS_PROBE_TYPE_INTERNAL +}; + +union avs_probe_point_id { + u32 value; + struct { + u32 module_id:16; + u32 instance_id:8; + u32 type:2; + u32 index:6; + } id; +} __packed; + +enum avs_connection_purpose { + AVS_CONNECTION_PURPOSE_EXTRACT = 0, + AVS_CONNECTION_PURPOSE_INJECT, + AVS_CONNECTION_PURPOSE_INJECT_REEXTRACT, +}; + +struct avs_probe_point_desc { + union avs_probe_point_id id; + u32 purpose; + union avs_connector_node_id node_id; +} __packed; + +int avs_ipc_probe_get_dma(struct avs_dev *adev, struct avs_probe_dma **dmas, size_t *num_dmas); +int avs_ipc_probe_attach_dma(struct avs_dev *adev, struct avs_probe_dma *dmas, size_t num_dmas); +int avs_ipc_probe_detach_dma(struct avs_dev *adev, union avs_connector_node_id *node_ids, + size_t num_node_ids); +int avs_ipc_probe_get_points(struct avs_dev *adev, struct avs_probe_point_desc **descs, + size_t *num_descs); +int avs_ipc_probe_connect_points(struct avs_dev *adev, struct avs_probe_point_desc *descs, + size_t num_descs); +int avs_ipc_probe_disconnect_points(struct avs_dev *adev, union avs_probe_point_id *ids, + size_t num_ids); + #endif /* __SOUND_SOC_INTEL_AVS_MSGS_H */ diff --git a/sound/soc/intel/avs/probes.c b/sound/soc/intel/avs/probes.c new file mode 100644 index 000000000000..339bad6fec22 --- /dev/null +++ b/sound/soc/intel/avs/probes.c @@ -0,0 +1,46 @@ +// SPDX-License-Identifier: GPL-2.0-only +// +// Copyright(c) 2021-2022 Intel Corporation. All rights reserved. +// +// Authors: Cezary Rojewski +// Amadeusz Slawinski +// + +#include "avs.h" +#include "messages.h" + +__maybe_unused +static int avs_dsp_init_probe(struct avs_dev *adev, union avs_connector_node_id node_id, + size_t buffer_size) + +{ + struct avs_probe_cfg cfg = {{0}}; + struct avs_module_entry mentry; + u16 dummy; + + avs_get_module_entry(adev, &AVS_PROBE_MOD_UUID, &mentry); + + /* + * Probe module uses no cycles, audio data format and input and output + * frame sizes are unused. It is also not owned by any pipeline. + */ + cfg.base.ibs = 1; + /* BSS module descriptor is always segment of index=2. */ + cfg.base.is_pages = mentry.segments[2].flags.length; + cfg.gtw_cfg.node_id = node_id; + cfg.gtw_cfg.dma_buffer_size = buffer_size; + + return avs_dsp_init_module(adev, mentry.module_id, INVALID_PIPELINE_ID, 0, 0, &cfg, + sizeof(cfg), &dummy); +} + +__maybe_unused +static void avs_dsp_delete_probe(struct avs_dev *adev) +{ + struct avs_module_entry mentry; + + avs_get_module_entry(adev, &AVS_PROBE_MOD_UUID, &mentry); + + /* There is only ever one probe module instance. */ + avs_dsp_delete_module(adev, mentry.module_id, 0, INVALID_PIPELINE_ID, 0); +} From 700462f55493c6831ad71b209eaebe310dcf11fd Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:35 +0100 Subject: [PATCH 015/115] ASoC: Intel: avs: Probe compress operations Add compress operations handlers for data extraction through probes. A single HDAudio stream is enlisted for said purpose. Operations follow same protocol as for standard PCM streaming on HOST side. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-11-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/Kconfig | 1 + sound/soc/intel/avs/avs.h | 3 + sound/soc/intel/avs/probes.c | 224 ++++++++++++++++++++++++++++++++++- 3 files changed, 225 insertions(+), 3 deletions(-) diff --git a/sound/soc/intel/Kconfig b/sound/soc/intel/Kconfig index ac799de4f7fd..4b9e498e3303 100644 --- a/sound/soc/intel/Kconfig +++ b/sound/soc/intel/Kconfig @@ -217,6 +217,7 @@ config SND_SOC_INTEL_AVS select SND_SOC_ACPI if ACPI select SND_SOC_TOPOLOGY select SND_SOC_HDA + select SND_SOC_COMPRESS if DEBUG_FS select SND_HDA_EXT_CORE select SND_HDA_DSP_LOADER select SND_INTEL_DSP_CONFIG diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index 7a9fb27d3845..e5e7c72eb511 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -144,6 +144,9 @@ struct avs_dev { u32 aging_timer_period; u32 fifo_full_timer_period; u32 logged_resources; /* context dependent: core or library */ + /* probes */ + struct hdac_ext_stream *extractor; + unsigned int num_probe_streams; #endif }; diff --git a/sound/soc/intel/avs/probes.c b/sound/soc/intel/avs/probes.c index 339bad6fec22..e90284ec8500 100644 --- a/sound/soc/intel/avs/probes.c +++ b/sound/soc/intel/avs/probes.c @@ -6,13 +6,15 @@ // Amadeusz Slawinski // +#include +#include +#include +#include #include "avs.h" #include "messages.h" -__maybe_unused static int avs_dsp_init_probe(struct avs_dev *adev, union avs_connector_node_id node_id, size_t buffer_size) - { struct avs_probe_cfg cfg = {{0}}; struct avs_module_entry mentry; @@ -34,7 +36,6 @@ static int avs_dsp_init_probe(struct avs_dev *adev, union avs_connector_node_id sizeof(cfg), &dummy); } -__maybe_unused static void avs_dsp_delete_probe(struct avs_dev *adev) { struct avs_module_entry mentry; @@ -44,3 +45,220 @@ static void avs_dsp_delete_probe(struct avs_dev *adev) /* There is only ever one probe module instance. */ avs_dsp_delete_module(adev, mentry.module_id, 0, INVALID_PIPELINE_ID, 0); } + +static inline struct hdac_ext_stream *avs_compr_get_host_stream(struct snd_compr_stream *cstream) +{ + return cstream->runtime->private_data; +} + +static int avs_probe_compr_open(struct snd_compr_stream *cstream, struct snd_soc_dai *dai) +{ + struct avs_dev *adev = to_avs_dev(dai->dev); + struct hdac_bus *bus = &adev->base.core; + struct hdac_ext_stream *host_stream; + + if (adev->extractor) { + dev_err(dai->dev, "Cannot open more than one extractor stream\n"); + return -EEXIST; + } + + host_stream = snd_hdac_ext_cstream_assign(bus, cstream); + if (!host_stream) { + dev_err(dai->dev, "Failed to assign HDAudio stream for extraction\n"); + return -EBUSY; + } + + adev->extractor = host_stream; + hdac_stream(host_stream)->curr_pos = 0; + cstream->runtime->private_data = host_stream; + + return 0; +} + +static int avs_probe_compr_free(struct snd_compr_stream *cstream, struct snd_soc_dai *dai) +{ + struct hdac_ext_stream *host_stream = avs_compr_get_host_stream(cstream); + struct avs_dev *adev = to_avs_dev(dai->dev); + struct avs_probe_point_desc *desc; + /* Extractor node identifier. */ + unsigned int vindex = INVALID_NODE_ID.vindex; + size_t num_desc; + int i, ret; + + /* Disconnect all probe points. */ + ret = avs_ipc_probe_get_points(adev, &desc, &num_desc); + if (ret) { + dev_err(dai->dev, "get probe points failed: %d\n", ret); + ret = AVS_IPC_RET(ret); + goto exit; + } + + for (i = 0; i < num_desc; i++) + if (desc[i].node_id.vindex == vindex) + avs_ipc_probe_disconnect_points(adev, &desc[i].id, 1); + kfree(desc); + +exit: + if (adev->num_probe_streams) { + adev->num_probe_streams--; + if (!adev->num_probe_streams) { + avs_dsp_delete_probe(adev); + avs_dsp_enable_d0ix(adev); + } + } + + snd_hdac_stream_cleanup(hdac_stream(host_stream)); + hdac_stream(host_stream)->prepared = 0; + snd_hdac_ext_stream_release(host_stream, HDAC_EXT_STREAM_TYPE_HOST); + + snd_compr_free_pages(cstream); + adev->extractor = NULL; + + return ret; +} + +static int avs_probe_compr_set_params(struct snd_compr_stream *cstream, + struct snd_compr_params *params, struct snd_soc_dai *dai) +{ + struct hdac_ext_stream *host_stream = avs_compr_get_host_stream(cstream); + struct snd_compr_runtime *rtd = cstream->runtime; + struct avs_dev *adev = to_avs_dev(dai->dev); + /* compr params do not store bit depth, default to S32_LE. */ + snd_pcm_format_t format = SNDRV_PCM_FORMAT_S32_LE; + unsigned int format_val; + int bps, ret; + + hdac_stream(host_stream)->bufsize = 0; + hdac_stream(host_stream)->period_bytes = 0; + hdac_stream(host_stream)->format_val = 0; + cstream->dma_buffer.dev.type = SNDRV_DMA_TYPE_DEV_SG; + cstream->dma_buffer.dev.dev = adev->dev; + + ret = snd_compr_malloc_pages(cstream, rtd->buffer_size); + if (ret < 0) + return ret; + bps = snd_pcm_format_physical_width(format); + if (bps < 0) + return bps; + format_val = snd_hdac_calc_stream_format(params->codec.sample_rate, params->codec.ch_out, + format, bps, 0); + ret = snd_hdac_stream_set_params(hdac_stream(host_stream), format_val); + if (ret < 0) + return ret; + ret = snd_hdac_stream_setup(hdac_stream(host_stream)); + if (ret < 0) + return ret; + + hdac_stream(host_stream)->prepared = 1; + + if (!adev->num_probe_streams) { + union avs_connector_node_id node_id; + + /* D0ix not allowed during probing. */ + ret = avs_dsp_disable_d0ix(adev); + if (ret) + return ret; + + node_id.vindex = hdac_stream(host_stream)->stream_tag - 1; + node_id.dma_type = AVS_DMA_HDA_HOST_INPUT; + + ret = avs_dsp_init_probe(adev, node_id, rtd->dma_bytes); + if (ret < 0) { + dev_err(dai->dev, "probe init failed: %d\n", ret); + avs_dsp_enable_d0ix(adev); + return ret; + } + } + + adev->num_probe_streams++; + return 0; +} + +static int avs_probe_compr_trigger(struct snd_compr_stream *cstream, int cmd, + struct snd_soc_dai *dai) +{ + struct hdac_ext_stream *host_stream = avs_compr_get_host_stream(cstream); + struct avs_dev *adev = to_avs_dev(dai->dev); + struct hdac_bus *bus = &adev->base.core; + unsigned long cookie; + + if (!hdac_stream(host_stream)->prepared) + return -EPIPE; + + switch (cmd) { + case SNDRV_PCM_TRIGGER_START: + case SNDRV_PCM_TRIGGER_PAUSE_RELEASE: + case SNDRV_PCM_TRIGGER_RESUME: + spin_lock_irqsave(&bus->reg_lock, cookie); + snd_hdac_stream_start(hdac_stream(host_stream), true); + spin_unlock_irqrestore(&bus->reg_lock, cookie); + break; + + case SNDRV_PCM_TRIGGER_PAUSE_PUSH: + case SNDRV_PCM_TRIGGER_SUSPEND: + case SNDRV_PCM_TRIGGER_STOP: + spin_lock_irqsave(&bus->reg_lock, cookie); + snd_hdac_stream_stop(hdac_stream(host_stream)); + spin_unlock_irqrestore(&bus->reg_lock, cookie); + break; + + default: + return -EINVAL; + } + + return 0; +} + +static int avs_probe_compr_pointer(struct snd_compr_stream *cstream, + struct snd_compr_tstamp *tstamp, struct snd_soc_dai *dai) +{ + struct hdac_ext_stream *host_stream = avs_compr_get_host_stream(cstream); + struct snd_soc_pcm_stream *pstream; + + pstream = &dai->driver->capture; + tstamp->copied_total = hdac_stream(host_stream)->curr_pos; + tstamp->sampling_rate = snd_pcm_rate_bit_to_rate(pstream->rates); + + return 0; +} + +static int avs_probe_compr_copy(struct snd_soc_component *comp, struct snd_compr_stream *cstream, + char __user *buf, size_t count) +{ + struct snd_compr_runtime *rtd = cstream->runtime; + unsigned int offset, n; + void *ptr; + int ret; + + if (count > rtd->buffer_size) + count = rtd->buffer_size; + + div_u64_rem(rtd->total_bytes_transferred, rtd->buffer_size, &offset); + ptr = rtd->dma_area + offset; + n = rtd->buffer_size - offset; + + if (count < n) { + ret = copy_to_user(buf, ptr, count); + } else { + ret = copy_to_user(buf, ptr, n); + ret += copy_to_user(buf + n, rtd->dma_area, count - n); + } + + if (ret) + return count - ret; + return count; +} + +__maybe_unused +static const struct snd_soc_cdai_ops avs_probe_dai_ops = { + .startup = avs_probe_compr_open, + .shutdown = avs_probe_compr_free, + .set_params = avs_probe_compr_set_params, + .trigger = avs_probe_compr_trigger, + .pointer = avs_probe_compr_pointer, +}; + +__maybe_unused +static const struct snd_compress_ops avs_probe_compress_ops = { + .copy = avs_probe_compr_copy, +}; From ed914a2a45a45e7d8f900ae8997ca4573792afcc Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:36 +0100 Subject: [PATCH 016/115] ASoC: Intel: avs: Data probing soc-component Define stub component for data probing. Stub as most operations from standard PCM case do not apply here. Specific bits are CPU DAIs and compress_ops. FE DAIs can link against these new CPU DAI to create new compress devices. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-12-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/avs.h | 10 +++++++ sound/soc/intel/avs/pcm.c | 6 ++-- sound/soc/intel/avs/probes.c | 53 ++++++++++++++++++++++++++++++++++-- 3 files changed, 64 insertions(+), 5 deletions(-) diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index e5e7c72eb511..e19d8d89455d 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -322,6 +322,9 @@ struct avs_soc_component { extern const struct snd_soc_dai_ops avs_dai_fe_ops; +int avs_soc_component_register(struct device *dev, const char *name, + const struct snd_soc_component_driver *drv, + struct snd_soc_dai_driver *cpu_dais, int num_cpu_dais); int avs_dmic_platform_register(struct avs_dev *adev, const char *name); int avs_i2s_platform_register(struct avs_dev *adev, const char *name, unsigned long port_mask, unsigned long *tdms); @@ -373,6 +376,8 @@ struct apl_log_buffer_layout { bool avs_logging_fw(struct avs_dev *adev); void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int len); void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len); + +int avs_probe_platform_register(struct avs_dev *adev, const char *name); #else #define AVS_SET_ENABLE_LOGS_OP(name) @@ -389,6 +394,11 @@ static inline void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len) { } + +static inline int avs_probe_platform_register(struct avs_dev *adev, const char *name) +{ + return 0; +} #endif #endif /* __SOUND_SOC_INTEL_AVS_H */ diff --git a/sound/soc/intel/avs/pcm.c b/sound/soc/intel/avs/pcm.c index 70d687fa9923..f930c5e86a84 100644 --- a/sound/soc/intel/avs/pcm.c +++ b/sound/soc/intel/avs/pcm.c @@ -1126,9 +1126,9 @@ static const struct snd_soc_component_driver avs_component_driver = { .topology_name_prefix = "intel/avs", }; -static int avs_soc_component_register(struct device *dev, const char *name, - const struct snd_soc_component_driver *drv, - struct snd_soc_dai_driver *cpu_dais, int num_cpu_dais) +int avs_soc_component_register(struct device *dev, const char *name, + const struct snd_soc_component_driver *drv, + struct snd_soc_dai_driver *cpu_dais, int num_cpu_dais) { struct avs_soc_component *acomp; int ret; diff --git a/sound/soc/intel/avs/probes.c b/sound/soc/intel/avs/probes.c index e90284ec8500..29d63f2a9616 100644 --- a/sound/soc/intel/avs/probes.c +++ b/sound/soc/intel/avs/probes.c @@ -249,7 +249,6 @@ static int avs_probe_compr_copy(struct snd_soc_component *comp, struct snd_compr return count; } -__maybe_unused static const struct snd_soc_cdai_ops avs_probe_dai_ops = { .startup = avs_probe_compr_open, .shutdown = avs_probe_compr_free, @@ -258,7 +257,57 @@ static const struct snd_soc_cdai_ops avs_probe_dai_ops = { .pointer = avs_probe_compr_pointer, }; -__maybe_unused static const struct snd_compress_ops avs_probe_compress_ops = { .copy = avs_probe_compr_copy, }; + +static struct snd_soc_dai_driver probe_cpu_dais[] = { +{ + .name = "Probe Extraction CPU DAI", + .compress_new = snd_soc_new_compress, + .cops = &avs_probe_dai_ops, + .capture = { + .stream_name = "Probe Extraction", + .channels_min = 1, + .channels_max = 8, + .rates = SNDRV_PCM_RATE_48000, + .rate_min = 48000, + .rate_max = 48000, + }, +}, +}; + +static int avs_probe_component_probe(struct snd_soc_component *component) +{ + struct avs_soc_component *acomp = to_avs_soc_component(component); + struct avs_dev *adev = to_avs_dev(component->dev); + + mutex_lock(&adev->comp_list_mutex); + list_add_tail(&acomp->node, &adev->comp_list); + mutex_unlock(&adev->comp_list_mutex); + return 0; +} + +static void avs_probe_component_remove(struct snd_soc_component *component) +{ + struct avs_soc_component *acomp = to_avs_soc_component(component); + struct avs_dev *adev = to_avs_dev(component->dev); + + mutex_lock(&adev->comp_list_mutex); + list_del(&acomp->node); + mutex_unlock(&adev->comp_list_mutex); +} + +static const struct snd_soc_component_driver avs_probe_component_driver = { + .name = "avs-probe-compr", + .probe = avs_probe_component_probe, + .remove = avs_probe_component_remove, + .compress_ops = &avs_probe_compress_ops, + .module_get_upon_open = 1, /* increment refcount when a stream is opened */ +}; + +int avs_probe_platform_register(struct avs_dev *adev, const char *name) +{ + return avs_soc_component_register(adev->dev, name, &avs_probe_component_driver, + probe_cpu_dais, ARRAY_SIZE(probe_cpu_dais)); +} From e17527e167ae5bd71fc9cb67da4e73bbb050e6f7 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:37 +0100 Subject: [PATCH 017/115] ASoC: Intel: avs: Add probe machine board Stub machine board driver with no custom DAPM routes and single FE DAI link for userspace to interact with. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-13-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/board_selection.c | 33 ++++++++++++++ sound/soc/intel/avs/boards/Kconfig | 8 ++++ sound/soc/intel/avs/boards/Makefile | 2 + sound/soc/intel/avs/boards/probe.c | 64 +++++++++++++++++++++++++++ 4 files changed, 107 insertions(+) create mode 100644 sound/soc/intel/avs/boards/probe.c diff --git a/sound/soc/intel/avs/board_selection.c b/sound/soc/intel/avs/board_selection.c index 02cc1ce8f5f5..b2823c2107f7 100644 --- a/sound/soc/intel/avs/board_selection.c +++ b/sound/soc/intel/avs/board_selection.c @@ -291,6 +291,33 @@ static void board_pdev_unregister(void *data) platform_device_unregister(data); } +static int __maybe_unused avs_register_probe_board(struct avs_dev *adev) +{ + struct platform_device *board; + struct snd_soc_acpi_mach mach = {{0}}; + int ret; + + ret = avs_probe_platform_register(adev, "probe-platform"); + if (ret < 0) + return ret; + + mach.mach_params.platform = "probe-platform"; + + board = platform_device_register_data(NULL, "avs_probe_mb", PLATFORM_DEVID_NONE, + (const void *)&mach, sizeof(mach)); + if (IS_ERR(board)) { + dev_err(adev->dev, "probe board register failed\n"); + return PTR_ERR(board); + } + + ret = devm_add_action(adev->dev, board_pdev_unregister, board); + if (ret < 0) { + platform_device_unregister(board); + return ret; + } + return 0; +} + static int avs_register_dmic_board(struct avs_dev *adev) { struct platform_device *codec, *board; @@ -500,6 +527,12 @@ int avs_register_all_boards(struct avs_dev *adev) { int ret; +#ifdef CONFIG_DEBUG_FS + ret = avs_register_probe_board(adev); + if (ret < 0) + dev_warn(adev->dev, "enumerate PROBE endpoints failed: %d\n", ret); +#endif + ret = avs_register_dmic_board(adev); if (ret < 0) dev_warn(adev->dev, "enumerate DMIC endpoints failed: %d\n", diff --git a/sound/soc/intel/avs/boards/Kconfig b/sound/soc/intel/avs/boards/Kconfig index 9bd40fdd9028..e4c230efe8d7 100644 --- a/sound/soc/intel/avs/boards/Kconfig +++ b/sound/soc/intel/avs/boards/Kconfig @@ -77,6 +77,14 @@ config SND_SOC_INTEL_AVS_MACH_NAU8825 Say Y or m if you have such a device. This is a recommended option. If unsure select "N". +config SND_SOC_INTEL_AVS_MACH_PROBE + tristate "Probing (data) board" + depends on DEBUG_FS + select SND_HWDEP + help + This adds support for data probing board which can be used to + gather data from runtime stream over compress operations. + config SND_SOC_INTEL_AVS_MACH_RT274 tristate "rt274 in I2S mode" depends on I2C diff --git a/sound/soc/intel/avs/boards/Makefile b/sound/soc/intel/avs/boards/Makefile index 4d70b8d09ce5..b81343420370 100644 --- a/sound/soc/intel/avs/boards/Makefile +++ b/sound/soc/intel/avs/boards/Makefile @@ -8,6 +8,7 @@ snd-soc-avs-max98927-objs := max98927.o snd-soc-avs-max98357a-objs := max98357a.o snd-soc-avs-max98373-objs := max98373.o snd-soc-avs-nau8825-objs := nau8825.o +snd-soc-avs-probe-objs := probe.o snd-soc-avs-rt274-objs := rt274.o snd-soc-avs-rt286-objs := rt286.o snd-soc-avs-rt298-objs := rt298.o @@ -22,6 +23,7 @@ obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_MAX98927) += snd-soc-avs-max98927.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_MAX98357A) += snd-soc-avs-max98357a.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_MAX98373) += snd-soc-avs-max98373.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_NAU8825) += snd-soc-avs-nau8825.o +obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_PROBE) += snd-soc-avs-probe.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_RT274) += snd-soc-avs-rt274.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_RT286) += snd-soc-avs-rt286.o obj-$(CONFIG_SND_SOC_INTEL_AVS_MACH_RT298) += snd-soc-avs-rt298.o diff --git a/sound/soc/intel/avs/boards/probe.c b/sound/soc/intel/avs/boards/probe.c new file mode 100644 index 000000000000..411acaee74f9 --- /dev/null +++ b/sound/soc/intel/avs/boards/probe.c @@ -0,0 +1,64 @@ +// SPDX-License-Identifier: GPL-2.0-only +// +// Copyright(c) 2021-2022 Intel Corporation. All rights reserved. +// +// Authors: Cezary Rojewski +// Amadeusz Slawinski +// + +#include +#include +#include +#include + +SND_SOC_DAILINK_DEF(dummy, DAILINK_COMP_ARRAY(COMP_DUMMY())); +SND_SOC_DAILINK_DEF(probe_cp, DAILINK_COMP_ARRAY(COMP_CPU("Probe Extraction CPU DAI"))); +SND_SOC_DAILINK_DEF(platform, DAILINK_COMP_ARRAY(COMP_PLATFORM("probe-platform"))); + +static struct snd_soc_dai_link probe_mb_dai_links[] = { + { + .name = "Compress Probe Capture", + .nonatomic = 1, + SND_SOC_DAILINK_REG(probe_cp, dummy, platform), + }, +}; + +static int avs_probe_mb_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct snd_soc_acpi_mach *mach; + struct snd_soc_card *card; + int ret; + + mach = dev_get_platdata(dev); + + card = devm_kzalloc(dev, sizeof(*card), GFP_KERNEL); + if (!card) + return -ENOMEM; + + card->name = "avs_probe_mb"; + card->dev = dev; + card->owner = THIS_MODULE; + card->dai_link = probe_mb_dai_links; + card->num_links = ARRAY_SIZE(probe_mb_dai_links); + card->fully_routed = true; + + ret = snd_soc_fixup_dai_links_platform_name(card, mach->mach_params.platform); + if (ret) + return ret; + + return devm_snd_soc_register_card(dev, card); +} + +static struct platform_driver avs_probe_mb_driver = { + .probe = avs_probe_mb_probe, + .driver = { + .name = "avs_probe_mb", + .pm = &snd_soc_pm_ops, + }, +}; + +module_platform_driver(avs_probe_mb_driver); + +MODULE_LICENSE("GPL"); +MODULE_ALIAS("platform:avs_probe_mb"); From 5a565ba23abe478f3d4c3b0c8798bcb5215b82f5 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:38 +0100 Subject: [PATCH 018/115] ASoC: Intel: avs: Probing and firmware tracing over debugfs Define debugfs subdirectory delegated for IPC communication with DSP. Input format: uint,uint,(...) which are later translated into DWORDS sequence and further into instances of struct of interest given the IPC type. For Extractor probes, following have been enabled: - PROBE_POINT_ADD (echo <..> probe_points) - PROBE_POINT_REMOVE (echo <..> probe_points_remove) - PROBE_POINT_INFO (cat probe_points) Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-14-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/avs.h | 7 + sound/soc/intel/avs/core.c | 2 + sound/soc/intel/avs/debugfs.c | 326 ++++++++++++++++++++++++++++++++++ 3 files changed, 335 insertions(+) diff --git a/sound/soc/intel/avs/avs.h b/sound/soc/intel/avs/avs.h index e19d8d89455d..d7fccdcb9c16 100644 --- a/sound/soc/intel/avs/avs.h +++ b/sound/soc/intel/avs/avs.h @@ -144,6 +144,7 @@ struct avs_dev { u32 aging_timer_period; u32 fifo_full_timer_period; u32 logged_resources; /* context dependent: core or library */ + struct dentry *debugfs_root; /* probes */ struct hdac_ext_stream *extractor; unsigned int num_probe_streams; @@ -378,6 +379,9 @@ void avs_dump_fw_log(struct avs_dev *adev, const void __iomem *src, unsigned int void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsigned int len); int avs_probe_platform_register(struct avs_dev *adev, const char *name); + +void avs_debugfs_init(struct avs_dev *adev); +void avs_debugfs_exit(struct avs_dev *adev); #else #define AVS_SET_ENABLE_LOGS_OP(name) @@ -399,6 +403,9 @@ static inline int avs_probe_platform_register(struct avs_dev *adev, const char * { return 0; } + +static inline void avs_debugfs_init(struct avs_dev *adev) { } +static inline void avs_debugfs_exit(struct avs_dev *adev) { } #endif #endif /* __SOUND_SOC_INTEL_AVS_H */ diff --git a/sound/soc/intel/avs/core.c b/sound/soc/intel/avs/core.c index f7bc06404dbc..2ca24273c491 100644 --- a/sound/soc/intel/avs/core.c +++ b/sound/soc/intel/avs/core.c @@ -214,6 +214,7 @@ static void avs_hda_probe_work(struct work_struct *work) adev->nhlt = intel_nhlt_init(adev->dev); if (!adev->nhlt) dev_info(bus->dev, "platform has no NHLT\n"); + avs_debugfs_init(adev); avs_register_all_boards(adev); @@ -491,6 +492,7 @@ static void avs_pci_remove(struct pci_dev *pci) avs_unregister_all_boards(adev); + avs_debugfs_exit(adev); if (adev->nhlt) intel_nhlt_free(adev->nhlt); diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c index 78705bcb09fb..678572ee6b9d 100644 --- a/sound/soc/intel/avs/debugfs.c +++ b/sound/soc/intel/avs/debugfs.c @@ -9,7 +9,10 @@ #include #include #include +#include +#include #include "avs.h" +#include "messages.h" static unsigned int __kfifo_fromio(struct kfifo *fifo, const void __iomem *src, unsigned int len) { @@ -44,3 +47,326 @@ void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsig avs_dump_fw_log(adev, src, len); wake_up(&adev->trace_waitq); } + +static ssize_t probe_points_read(struct file *file, char __user *to, size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + struct avs_probe_point_desc *desc; + size_t num_desc, len = 0; + char *buf; + int i, ret; + + /* Prevent chaining, send and dump IPC value just once. */ + if (*ppos) + return 0; + + buf = kzalloc(PAGE_SIZE, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + ret = avs_ipc_probe_get_points(adev, &desc, &num_desc); + if (ret) { + ret = AVS_IPC_RET(ret); + goto exit; + } + + for (i = 0; i < num_desc; i++) { + ret = snprintf(buf + len, PAGE_SIZE - len, + "Id: %#010x Purpose: %d Node id: %#x\n", + desc[i].id.value, desc[i].purpose, desc[i].node_id.val); + if (ret < 0) + goto free_desc; + len += ret; + } + + ret = simple_read_from_buffer(to, count, ppos, buf, len); +free_desc: + kfree(desc); +exit: + kfree(buf); + return ret; +} + +static ssize_t probe_points_write(struct file *file, const char __user *from, size_t count, + loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + struct avs_probe_point_desc *desc; + u32 *array, num_elems; + size_t bytes; + int ret; + + ret = parse_int_array_user(from, count, (int **)&array); + if (ret < 0) + return ret; + + num_elems = *array; + bytes = sizeof(*array) * num_elems; + if (bytes % sizeof(*desc)) { + ret = -EINVAL; + goto exit; + } + + desc = (struct avs_probe_point_desc *)&array[1]; + ret = avs_ipc_probe_connect_points(adev, desc, bytes / sizeof(*desc)); + if (ret) + ret = AVS_IPC_RET(ret); + else + ret = count; +exit: + kfree(array); + return ret; +} + +static const struct file_operations probe_points_fops = { + .open = simple_open, + .read = probe_points_read, + .write = probe_points_write, + .llseek = no_llseek, +}; + +static ssize_t probe_points_disconnect_write(struct file *file, const char __user *from, + size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + union avs_probe_point_id *id; + u32 *array, num_elems; + size_t bytes; + int ret; + + ret = parse_int_array_user(from, count, (int **)&array); + if (ret < 0) + return ret; + + num_elems = *array; + bytes = sizeof(*array) * num_elems; + if (bytes % sizeof(*id)) { + ret = -EINVAL; + goto exit; + } + + id = (union avs_probe_point_id *)&array[1]; + ret = avs_ipc_probe_disconnect_points(adev, id, bytes / sizeof(*id)); + if (ret) + ret = AVS_IPC_RET(ret); + else + ret = count; +exit: + kfree(array); + return ret; +} + +static const struct file_operations probe_points_disconnect_fops = { + .open = simple_open, + .write = probe_points_disconnect_write, + .llseek = default_llseek, +}; + +static ssize_t strace_read(struct file *file, char __user *to, size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + struct kfifo *fifo = &adev->trace_fifo; + unsigned int copied; + + if (kfifo_is_empty(fifo)) { + DEFINE_WAIT(wait); + + prepare_to_wait(&adev->trace_waitq, &wait, TASK_INTERRUPTIBLE); + if (!signal_pending(current)) + schedule(); + finish_wait(&adev->trace_waitq, &wait); + } + + if (kfifo_to_user(fifo, to, count, &copied)) + return -EFAULT; + *ppos += copied; + return copied; +} + +static int strace_open(struct inode *inode, struct file *file) +{ + struct avs_dev *adev = inode->i_private; + int ret; + + if (kfifo_initialized(&adev->trace_fifo)) + return -EBUSY; + + ret = kfifo_alloc(&adev->trace_fifo, PAGE_SIZE, GFP_KERNEL); + if (ret < 0) + return ret; + + file->private_data = adev; + return 0; +} + +static int strace_release(struct inode *inode, struct file *file) +{ + struct avs_dev *adev = file->private_data; + unsigned long flags; + + spin_lock_irqsave(&adev->trace_lock, flags); + kfifo_free(&adev->trace_fifo); + spin_unlock_irqrestore(&adev->trace_lock, flags); + + return 0; +} + +static const struct file_operations strace_fops = { + .llseek = default_llseek, + .read = strace_read, + .open = strace_open, + .release = strace_release, +}; + +#define DISABLE_TIMERS UINT_MAX + +static int enable_logs(struct avs_dev *adev, u32 resource_mask, u32 *priorities) +{ + int ret; + + /* Logging demands D0i0 state from DSP. */ + if (!adev->logged_resources) { + pm_runtime_get_sync(adev->dev); + + ret = avs_dsp_disable_d0ix(adev); + if (ret) + goto err_d0ix; + } + + ret = avs_ipc_set_system_time(adev); + if (ret && ret != AVS_IPC_NOT_SUPPORTED) { + ret = AVS_IPC_RET(ret); + goto err_ipc; + } + + ret = avs_dsp_op(adev, enable_logs, AVS_LOG_ENABLE, adev->aging_timer_period, + adev->fifo_full_timer_period, resource_mask, priorities); + if (ret) + goto err_ipc; + + adev->logged_resources |= resource_mask; + return 0; + +err_ipc: + if (!adev->logged_resources) { + avs_dsp_enable_d0ix(adev); +err_d0ix: + pm_runtime_mark_last_busy(adev->dev); + pm_runtime_put_autosuspend(adev->dev); + } + + return ret; +} + +static int disable_logs(struct avs_dev *adev, u32 resource_mask) +{ + int ret; + + /* Check if there's anything to do. */ + if (!adev->logged_resources) + return 0; + + ret = avs_dsp_op(adev, enable_logs, AVS_LOG_DISABLE, DISABLE_TIMERS, DISABLE_TIMERS, + resource_mask, NULL); + + /* + * If IPC fails causing recovery, logged_resources is already zero + * so unsetting bits is still safe. + */ + adev->logged_resources &= ~resource_mask; + + /* If that's the last resource, allow for D3. */ + if (!adev->logged_resources) { + avs_dsp_enable_d0ix(adev); + pm_runtime_mark_last_busy(adev->dev); + pm_runtime_put_autosuspend(adev->dev); + } + + return ret; +} + +static ssize_t trace_control_read(struct file *file, char __user *to, size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + char buf[64]; + int len; + + len = snprintf(buf, sizeof(buf), "0x%08x\n", adev->logged_resources); + + return simple_read_from_buffer(to, count, ppos, buf, len); +} + +static ssize_t trace_control_write(struct file *file, const char __user *from, size_t count, + loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + u32 *array, num_elems; + u32 resource_mask; + int ret; + + ret = parse_int_array_user(from, count, (int **)&array); + if (ret < 0) + return ret; + + num_elems = *array; + resource_mask = array[1]; + + /* + * Disable if just resource mask is provided - no log priority flags. + * + * Enable input format: mask, prio1, .., prioN + * Where 'N' equals number of bits set in the 'mask'. + */ + if (num_elems == 1) { + ret = disable_logs(adev, resource_mask); + } else { + if (num_elems != (hweight_long(resource_mask) + 1)) { + ret = -EINVAL; + goto free_array; + } + + ret = enable_logs(adev, resource_mask, &array[2]); + } + + if (!ret) + ret = count; +free_array: + kfree(array); + return ret; +} + +static const struct file_operations trace_control_fops = { + .llseek = default_llseek, + .read = trace_control_read, + .write = trace_control_write, + .open = simple_open, +}; + +void avs_debugfs_init(struct avs_dev *adev) +{ + init_waitqueue_head(&adev->trace_waitq); + spin_lock_init(&adev->trace_lock); + + adev->debugfs_root = debugfs_create_dir("avs", snd_soc_debugfs_root); + + /* Initialize timer periods with recommended defaults. */ + adev->aging_timer_period = 10; + adev->fifo_full_timer_period = 10; + + debugfs_create_file("strace", 0444, adev->debugfs_root, adev, &strace_fops); + debugfs_create_file("trace_control", 0644, adev->debugfs_root, adev, &trace_control_fops); + + debugfs_create_u32("trace_aging_period", 0644, adev->debugfs_root, + &adev->aging_timer_period); + debugfs_create_u32("trace_fifo_full_period", 0644, adev->debugfs_root, + &adev->fifo_full_timer_period); + + debugfs_create_file("probe_points", 0644, adev->debugfs_root, adev, &probe_points_fops); + debugfs_create_file("probe_points_disconnect", 0200, adev->debugfs_root, adev, + &probe_points_disconnect_fops); +} + +void avs_debugfs_exit(struct avs_dev *adev) +{ + debugfs_remove_recursive(adev->debugfs_root); +} From 34d27c71707c4ed615105376e0f3907d99b1b271 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:39 +0100 Subject: [PATCH 019/115] ASoC: Intel: avs: Gather remaining logs on strace_release() When user closes the tracer, some logs may still remain in the tail of the buffer as firmware sends LOG_BUFFER_STATUS notification only when certain threshold of data is reached. Add whatever is left to already gathered logs so no information is lost. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-15-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/debugfs.c | 16 +++++++++++++++- 1 file changed, 15 insertions(+), 1 deletion(-) diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c index 678572ee6b9d..e7b0b99824aa 100644 --- a/sound/soc/intel/avs/debugfs.c +++ b/sound/soc/intel/avs/debugfs.c @@ -201,11 +201,25 @@ static int strace_open(struct inode *inode, struct file *file) static int strace_release(struct inode *inode, struct file *file) { + union avs_notify_msg msg = AVS_NOTIFICATION(LOG_BUFFER_STATUS); struct avs_dev *adev = file->private_data; - unsigned long flags; + unsigned long resource_mask; + unsigned long flags, i; + u32 num_cores; + + resource_mask = adev->logged_resources; + num_cores = adev->hw_cfg.dsp_cores; spin_lock_irqsave(&adev->trace_lock, flags); + + /* Gather any remaining logs. */ + for_each_set_bit(i, &resource_mask, num_cores) { + msg.log.core = i; + avs_dsp_op(adev, log_buffer_status, &msg); + } + kfifo_free(&adev->trace_fifo); + spin_unlock_irqrestore(&adev->trace_lock, flags); return 0; From 870f6e5abba95ac78e750b61cf8f3f15be96796f Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:40 +0100 Subject: [PATCH 020/115] ASoC: Intel: avs: Allow for dumping FW_REGS area SRAM0 window begins with a block of memory, usually of size PAGE_SIZE, dedicated to the base firmware registers. When debugging firmware, it is desirable to be able to dump them at will. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-16-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/debugfs.c | 24 ++++++++++++++++++++++++ 1 file changed, 24 insertions(+) diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c index e7b0b99824aa..e9042d4328c4 100644 --- a/sound/soc/intel/avs/debugfs.c +++ b/sound/soc/intel/avs/debugfs.c @@ -48,6 +48,29 @@ void avs_dump_fw_log_wakeup(struct avs_dev *adev, const void __iomem *src, unsig wake_up(&adev->trace_waitq); } +static ssize_t fw_regs_read(struct file *file, char __user *to, size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + char *buf; + int ret; + + buf = kzalloc(AVS_FW_REGS_SIZE, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + memcpy_fromio(buf, avs_sram_addr(adev, AVS_FW_REGS_WINDOW), AVS_FW_REGS_SIZE); + + ret = simple_read_from_buffer(to, count, ppos, buf, AVS_FW_REGS_SIZE); + kfree(buf); + return ret; +} + +static const struct file_operations fw_regs_fops = { + .open = simple_open, + .read = fw_regs_read, + .llseek = no_llseek, +}; + static ssize_t probe_points_read(struct file *file, char __user *to, size_t count, loff_t *ppos) { struct avs_dev *adev = file->private_data; @@ -369,6 +392,7 @@ void avs_debugfs_init(struct avs_dev *adev) debugfs_create_file("strace", 0444, adev->debugfs_root, adev, &strace_fops); debugfs_create_file("trace_control", 0644, adev->debugfs_root, adev, &trace_control_fops); + debugfs_create_file("fw_regs", 0444, adev->debugfs_root, adev, &fw_regs_fops); debugfs_create_u32("trace_aging_period", 0644, adev->debugfs_root, &adev->aging_timer_period); From 85ac9c8c8eed76e8a320a9e017c6d36e2a52745b Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Fri, 2 Dec 2022 16:28:41 +0100 Subject: [PATCH 021/115] ASoC: Intel: avs: Allow for dumping debug window snapshot Add new read-only debugfs entry which dumps entire content of the SRAM window 2 i.e.: the debug window. Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221202152841.672536-17-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/avs/debugfs.c | 26 ++++++++++++++++++++++++++ sound/soc/intel/avs/registers.h | 3 ++- 2 files changed, 28 insertions(+), 1 deletion(-) diff --git a/sound/soc/intel/avs/debugfs.c b/sound/soc/intel/avs/debugfs.c index e9042d4328c4..bdd388ec01ea 100644 --- a/sound/soc/intel/avs/debugfs.c +++ b/sound/soc/intel/avs/debugfs.c @@ -71,6 +71,31 @@ static const struct file_operations fw_regs_fops = { .llseek = no_llseek, }; +static ssize_t debug_window_read(struct file *file, char __user *to, size_t count, loff_t *ppos) +{ + struct avs_dev *adev = file->private_data; + size_t size; + char *buf; + int ret; + + size = adev->hw_cfg.dsp_cores * AVS_WINDOW_CHUNK_SIZE; + buf = kzalloc(size, GFP_KERNEL); + if (!buf) + return -ENOMEM; + + memcpy_fromio(buf, avs_sram_addr(adev, AVS_DEBUG_WINDOW), size); + + ret = simple_read_from_buffer(to, count, ppos, buf, size); + kfree(buf); + return ret; +} + +static const struct file_operations debug_window_fops = { + .open = simple_open, + .read = debug_window_read, + .llseek = no_llseek, +}; + static ssize_t probe_points_read(struct file *file, char __user *to, size_t count, loff_t *ppos) { struct avs_dev *adev = file->private_data; @@ -393,6 +418,7 @@ void avs_debugfs_init(struct avs_dev *adev) debugfs_create_file("strace", 0444, adev->debugfs_root, adev, &strace_fops); debugfs_create_file("trace_control", 0644, adev->debugfs_root, adev, &trace_control_fops); debugfs_create_file("fw_regs", 0444, adev->debugfs_root, adev, &fw_regs_fops); + debugfs_create_file("debug_window", 0444, adev->debugfs_root, adev, &debug_window_fops); debugfs_create_u32("trace_aging_period", 0644, adev->debugfs_root, &adev->aging_timer_period); diff --git a/sound/soc/intel/avs/registers.h b/sound/soc/intel/avs/registers.h index 95be86148cf3..2b464e466ed5 100644 --- a/sound/soc/intel/avs/registers.h +++ b/sound/soc/intel/avs/registers.h @@ -59,7 +59,8 @@ #define AVS_FW_REG_STATUS(adev) (AVS_FW_REG_BASE(adev) + 0x0) #define AVS_FW_REG_ERROR_CODE(adev) (AVS_FW_REG_BASE(adev) + 0x4) -#define AVS_FW_REGS_SIZE PAGE_SIZE +#define AVS_WINDOW_CHUNK_SIZE PAGE_SIZE +#define AVS_FW_REGS_SIZE AVS_WINDOW_CHUNK_SIZE #define AVS_FW_REGS_WINDOW 0 /* DSP -> HOST communication window */ #define AVS_UPLINK_WINDOW AVS_FW_REGS_WINDOW From 1b41beaa7a58467505ec3023af8aad74f878b888 Mon Sep 17 00:00:00 2001 From: Yang Yingliang Date: Mon, 5 Dec 2022 22:37:21 +0800 Subject: [PATCH 022/115] ASoC: sof_es8336: fix possible use-after-free in sof_es8336_remove() sof_es8336_remove() calls cancel_delayed_work(). However, that function does not wait until the work function finishes. This means that the callback function may still be running after the driver's remove function has finished, which would result in a use-after-free. Fix by calling cancel_delayed_work_sync(), which ensures that the work is properly cancelled, no longer running, and unable to re-schedule itself. Fixes: 89cdb224f2ab ("ASoC: sof_es8336: reduce pop noise on speaker") Signed-off-by: Yang Yingliang Acked-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221205143721.3988988-1-yangyingliang@huawei.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/sof_es8336.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/intel/boards/sof_es8336.c b/sound/soc/intel/boards/sof_es8336.c index 70713e4b07dc..773e5d1d87d4 100644 --- a/sound/soc/intel/boards/sof_es8336.c +++ b/sound/soc/intel/boards/sof_es8336.c @@ -783,7 +783,7 @@ static int sof_es8336_remove(struct platform_device *pdev) struct snd_soc_card *card = platform_get_drvdata(pdev); struct sof_es8336_private *priv = snd_soc_card_get_drvdata(card); - cancel_delayed_work(&priv->pcm_pop_work); + cancel_delayed_work_sync(&priv->pcm_pop_work); gpiod_put(priv->gpio_speakers); device_remove_software_node(priv->codec_dev); put_device(priv->codec_dev); From 8e4c2eee1e15c1206c26f6b28b05fe9711a427c6 Mon Sep 17 00:00:00 2001 From: Christian Schoenebeck Date: Tue, 22 Nov 2022 20:20:22 +0100 Subject: [PATCH 023/115] net/9p: distinguish zero-copy requests Add boolean `zc` member to struct p9_fcall to distinguish zero-copy messages (not using the linear `sdata` buffer for message payload) from regular messages (which do copy message payload to `sdata` before being further processed). This new member is appended to end of structure to avoid inserting huge padding in generated layout. Link: https://lkml.kernel.org/r/8f2a5c12a446c3b544da64e0b1550e1fb2d6f972.1669144861.git.linux_oss@crudebyte.com Signed-off-by: Christian Schoenebeck Tested-by: Stefano Stabellini Signed-off-by: Dominique Martinet --- include/net/9p/9p.h | 2 ++ net/9p/client.c | 6 ++++++ 2 files changed, 8 insertions(+) diff --git a/include/net/9p/9p.h b/include/net/9p/9p.h index 13abe013af21..429adf6be29c 100644 --- a/include/net/9p/9p.h +++ b/include/net/9p/9p.h @@ -531,6 +531,7 @@ struct p9_rstatfs { * @offset: used by marshalling routines to track current position in buffer * @capacity: used by marshalling routines to track total malloc'd capacity * @sdata: payload + * @zc: whether zero-copy is used * * &p9_fcall represents the structure for all 9P RPC * transactions. Requests are packaged into fcalls, and reponses @@ -549,6 +550,7 @@ struct p9_fcall { struct kmem_cache *cache; u8 *sdata; + bool zc; }; int p9_errstr2errno(char *errstr, int len); diff --git a/net/9p/client.c b/net/9p/client.c index b554f8357f96..a2b4a965a5a9 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -685,6 +685,9 @@ p9_client_rpc(struct p9_client *c, int8_t type, const char *fmt, ...) if (IS_ERR(req)) return req; + req->tc.zc = false; + req->rc.zc = false; + if (signal_pending(current)) { sigpending = 1; clear_thread_flag(TIF_SIGPENDING); @@ -783,6 +786,9 @@ static struct p9_req_t *p9_client_zc_rpc(struct p9_client *c, int8_t type, if (IS_ERR(req)) return req; + req->tc.zc = true; + req->rc.zc = true; + if (signal_pending(current)) { sigpending = 1; clear_thread_flag(TIF_SIGPENDING); From a31b3cffbd8e5d032dcb267bf94ee48d71c1a28b Mon Sep 17 00:00:00 2001 From: Christian Schoenebeck Date: Tue, 22 Nov 2022 20:20:29 +0100 Subject: [PATCH 024/115] net/9p: fix response size check in p9_check_errors() Since commit 60ece0833b6c ("net/9p: allocate appropriate reduced message buffers") it is no longer appropriate to check server's response size against msize. Check against the previously allocated buffer capacity instead. - Omit this size check entirely for zero-copy messages, as those always allocate 4k (P9_ZC_HDR_SZ) linear buffers which are not used for actual payload and can be much bigger than 4k. - Replace p9_debug() by pr_err() to make sure this message is always printed in case this error is triggered. - Add 9p message type to error message to ease investigation. Link: https://lkml.kernel.org/r/e0edec84b1c80119ae937ce854b4f5f6dbe2d08c.1669144861.git.linux_oss@crudebyte.com Signed-off-by: Christian Schoenebeck Tested-by: Stefano Stabellini Reported-by: kernel test robot Signed-off-by: Dominique Martinet --- net/9p/client.c | 7 +++---- 1 file changed, 3 insertions(+), 4 deletions(-) diff --git a/net/9p/client.c b/net/9p/client.c index a2b4a965a5a9..7b2a997662d9 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -519,10 +519,9 @@ static int p9_check_errors(struct p9_client *c, struct p9_req_t *req) int ecode; err = p9_parse_header(&req->rc, NULL, &type, NULL, 0); - if (req->rc.size >= c->msize) { - p9_debug(P9_DEBUG_ERROR, - "requested packet size too big: %d\n", - req->rc.size); + if (req->rc.size > req->rc.capacity && !req->rc.zc) { + pr_err("requested packet size too big: %d does not fit %zu (type=%d)\n", + req->rc.size, req->rc.capacity, req->rc.id); return -EIO; } /* dump the response from server From 31b573946ea55e1ea0e08ae8e83bcf879b30f83a Mon Sep 17 00:00:00 2001 From: Jiao Zhou Date: Tue, 6 Dec 2022 13:53:11 -0500 Subject: [PATCH 025/115] ALSA: hda/hdmi: Add HP Device 0x8711 to force connect list HDMI audio is not working on the HP EliteDesk 800 G6 because the pin is unconnected. This issue can be resolved by using the 'hdajackretask' tool to override the unconnected pin to force it to connect. Signed-off-by: Jiao Zhou Cc: Link: https://lore.kernel.org/r/20221206185311.3669950-1-jiaozhou@google.com Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_hdmi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 21edf7a619f0..7a40ddfd695a 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1975,6 +1975,7 @@ static int hdmi_add_cvt(struct hda_codec *codec, hda_nid_t cvt_nid) static const struct snd_pci_quirk force_connect_list[] = { SND_PCI_QUIRK(0x103c, 0x870f, "HP", 1), SND_PCI_QUIRK(0x103c, 0x871a, "HP", 1), + SND_PCI_QUIRK(0x103c, 0x8711, "HP", 1), SND_PCI_QUIRK(0x1462, 0xec94, "MS-7C94", 1), SND_PCI_QUIRK(0x8086, 0x2081, "Intel NUC 10", 1), {} From 696b66ac26ef953aed5783ef26a252ec8f207013 Mon Sep 17 00:00:00 2001 From: wangdicheng Date: Wed, 7 Dec 2022 16:20:36 +0800 Subject: [PATCH 026/115] ALSA: usb-audio: add the quirk for KT0206 device Add relevant information to the quirks-table.h file. The test passes and the sound source file plays normally. Signed-off-by: wangdicheng Cc: Link: https://lore.kernel.org/r/SG2PR02MB587849631CB96809CF90DBED8A1A9@SG2PR02MB5878.apcprd02.prod.outlook.com Signed-off-by: Takashi Iwai --- sound/usb/quirks-table.h | 2 ++ 1 file changed, 2 insertions(+) diff --git a/sound/usb/quirks-table.h b/sound/usb/quirks-table.h index 874fcf245747..271884e35003 100644 --- a/sound/usb/quirks-table.h +++ b/sound/usb/quirks-table.h @@ -76,6 +76,8 @@ { USB_DEVICE_VENDOR_SPEC(0x041e, 0x3f0a) }, /* E-Mu 0204 USB */ { USB_DEVICE_VENDOR_SPEC(0x041e, 0x3f19) }, +/* Ktmicro Usb_audio device */ +{ USB_DEVICE_VENDOR_SPEC(0x31b2, 0x0011) }, /* * Creative Technology, Ltd Live! Cam Sync HD [VF0770] From 4bf5bf54476dffe60e6b6d8d539f67309ff599e2 Mon Sep 17 00:00:00 2001 From: Edward Pacman Date: Wed, 7 Dec 2022 21:32:18 +0800 Subject: [PATCH 027/115] ALSA: hda/realtek: Add quirk for Lenovo TianYi510Pro-14IOB Lenovo TianYi510Pro-14IOB (17aa:3742) require quirk for enabling headset-mic Signed-off-by: Edward Pacman Cc: Link: https://bugzilla.kernel.org/show_bug.cgi?id=216756 Link: https://lore.kernel.org/r/20221207133218.18989-1-edward@edward-p.xyz Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_realtek.c | 27 +++++++++++++++++++++++++++ 1 file changed, 27 insertions(+) diff --git a/sound/pci/hda/patch_realtek.c b/sound/pci/hda/patch_realtek.c index 5f51f8fc7fdc..e443d88f627f 100644 --- a/sound/pci/hda/patch_realtek.c +++ b/sound/pci/hda/patch_realtek.c @@ -10999,6 +10999,17 @@ static void alc897_fixup_lenovo_headset_mic(struct hda_codec *codec, } } +static void alc897_fixup_lenovo_headset_mode(struct hda_codec *codec, + const struct hda_fixup *fix, int action) +{ + struct alc_spec *spec = codec->spec; + + if (action == HDA_FIXUP_ACT_PRE_PROBE) { + spec->parse_flags |= HDA_PINCFG_HEADSET_MIC; + spec->gen.hp_automute_hook = alc897_hp_automute_hook; + } +} + static const struct coef_fw alc668_coefs[] = { WRITE_COEF(0x01, 0xbebe), WRITE_COEF(0x02, 0xaaaa), WRITE_COEF(0x03, 0x0), WRITE_COEF(0x04, 0x0180), WRITE_COEF(0x06, 0x0), WRITE_COEF(0x07, 0x0f80), @@ -11082,6 +11093,8 @@ enum { ALC897_FIXUP_LENOVO_HEADSET_MIC, ALC897_FIXUP_HEADSET_MIC_PIN, ALC897_FIXUP_HP_HSMIC_VERB, + ALC897_FIXUP_LENOVO_HEADSET_MODE, + ALC897_FIXUP_HEADSET_MIC_PIN2, }; static const struct hda_fixup alc662_fixups[] = { @@ -11508,6 +11521,19 @@ static const struct hda_fixup alc662_fixups[] = { { } }, }, + [ALC897_FIXUP_LENOVO_HEADSET_MODE] = { + .type = HDA_FIXUP_FUNC, + .v.func = alc897_fixup_lenovo_headset_mode, + }, + [ALC897_FIXUP_HEADSET_MIC_PIN2] = { + .type = HDA_FIXUP_PINS, + .v.pins = (const struct hda_pintbl[]) { + { 0x1a, 0x01a11140 }, /* use as headset mic, without its own jack detect */ + { } + }, + .chained = true, + .chain_id = ALC897_FIXUP_LENOVO_HEADSET_MODE + }, }; static const struct snd_pci_quirk alc662_fixup_tbl[] = { @@ -11560,6 +11586,7 @@ static const struct snd_pci_quirk alc662_fixup_tbl[] = { SND_PCI_QUIRK(0x17aa, 0x32cb, "Lenovo ThinkCentre M70", ALC897_FIXUP_HEADSET_MIC_PIN), SND_PCI_QUIRK(0x17aa, 0x32cf, "Lenovo ThinkCentre M950", ALC897_FIXUP_HEADSET_MIC_PIN), SND_PCI_QUIRK(0x17aa, 0x32f7, "Lenovo ThinkCentre M90", ALC897_FIXUP_HEADSET_MIC_PIN), + SND_PCI_QUIRK(0x17aa, 0x3742, "Lenovo TianYi510Pro-14IOB", ALC897_FIXUP_HEADSET_MIC_PIN2), SND_PCI_QUIRK(0x17aa, 0x38af, "Lenovo Ideapad Y550P", ALC662_FIXUP_IDEAPAD), SND_PCI_QUIRK(0x17aa, 0x3a0d, "Lenovo Ideapad Y550", ALC662_FIXUP_IDEAPAD), SND_PCI_QUIRK(0x1849, 0x5892, "ASRock B150M", ALC892_FIXUP_ASROCK_MOBO), From 72d9a541d7f186f0ec97c71ba7e477dd9bf4155f Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:25 +0100 Subject: [PATCH 028/115] ASoC: Intel: Skylake: Update pipe_config_idx before filling BE params Without updating the index before BE copier config is filled with hardware parameters, outdated parameters are used instead. Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-2-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl-topology.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c index e06eac592da1..fc3d719d93e1 100644 --- a/sound/soc/intel/skylake/skl-topology.c +++ b/sound/soc/intel/skylake/skl-topology.c @@ -1837,16 +1837,24 @@ static int skl_tplg_be_fill_pipe_params(struct snd_soc_dai *dai, { struct nhlt_specific_cfg *cfg; struct skl_pipe *pipe = mconfig->pipe; + struct skl_pipe_params save = *pipe->p_params; struct skl_pipe_fmt *pipe_fmt; struct skl_dev *skl = get_skl_ctx(dai->dev); int link_type = skl_tplg_be_link_type(mconfig->dev_type); u8 dev_type = skl_tplg_be_dev_type(mconfig->dev_type); + int ret; skl_tplg_fill_dma_id(mconfig, params); if (link_type == NHLT_LINK_HDA) return 0; + *pipe->p_params = *params; + ret = skl_tplg_get_pipe_config(skl, mconfig); + if (ret) + goto err; + + dev_dbg(skl->dev, "%s using pipe config: %d\n", __func__, pipe->pipe_config_idx); if (pipe->direction == SNDRV_PCM_STREAM_PLAYBACK) pipe_fmt = &pipe->configs[pipe->pipe_config_idx].out_fmt; else @@ -1865,10 +1873,15 @@ static int skl_tplg_be_fill_pipe_params(struct snd_soc_dai *dai, dev_err(dai->dev, "Blob NULL for id:%d type:%d dirn:%d ch:%d, freq:%d, fmt:%d\n", mconfig->vbus_id, link_type, params->stream, params->ch, params->s_freq, params->s_fmt); - return -EINVAL; + ret = -EINVAL; + goto err; } return 0; + +err: + *pipe->p_params = save; + return ret; } static int skl_tplg_be_set_src_pipe_params(struct snd_soc_dai *dai, From b0d16e54e7559f2055123ea7b1d9ff1bb808ebad Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:26 +0100 Subject: [PATCH 029/115] ASoC: Intel: Skylake: Remove skl_tplg_is_multi_fmt() Rather than forcing userspace to select proper format with enumerable kcontrols, select it ourselves based on provided hw_params. Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-3-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl-topology.c | 40 -------------------------- 1 file changed, 40 deletions(-) diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c index fc3d719d93e1..f144acae1b44 100644 --- a/sound/soc/intel/skylake/skl-topology.c +++ b/sound/soc/intel/skylake/skl-topology.c @@ -582,38 +582,6 @@ static int skl_tplg_unload_pipe_modules(struct skl_dev *skl, return ret; } -static bool skl_tplg_is_multi_fmt(struct skl_dev *skl, struct skl_pipe *pipe) -{ - struct skl_pipe_fmt *cur_fmt; - struct skl_pipe_fmt *next_fmt; - int i; - - if (pipe->nr_cfgs <= 1) - return false; - - if (pipe->conn_type != SKL_PIPE_CONN_TYPE_FE) - return true; - - for (i = 0; i < pipe->nr_cfgs - 1; i++) { - if (pipe->direction == SNDRV_PCM_STREAM_PLAYBACK) { - cur_fmt = &pipe->configs[i].out_fmt; - next_fmt = &pipe->configs[i + 1].out_fmt; - } else { - cur_fmt = &pipe->configs[i].in_fmt; - next_fmt = &pipe->configs[i + 1].in_fmt; - } - - if (!CHECK_HW_PARAMS(cur_fmt->channels, cur_fmt->freq, - cur_fmt->bps, - next_fmt->channels, - next_fmt->freq, - next_fmt->bps)) - return true; - } - - return false; -} - /* * Here, we select pipe format based on the pipe type and pipe * direction to determine the current config index for the pipeline. @@ -636,14 +604,6 @@ skl_tplg_get_pipe_config(struct skl_dev *skl, struct skl_module_cfg *mconfig) return 0; } - if (skl_tplg_is_multi_fmt(skl, pipe)) { - pipe->cur_config_idx = pipe->pipe_config_idx; - pipe->memory_pages = pconfig->mem_pages; - dev_dbg(skl->dev, "found pipe config idx:%d\n", - pipe->cur_config_idx); - return 0; - } - if (pipe->conn_type == SKL_PIPE_CONN_TYPE_NONE || pipe->nr_cfgs == 1) { dev_dbg(skl->dev, "No conn_type or just 1 pathcfg, taking 0th for %d\n", pipe->ppl_id); From 75ab3c00769009e32e5cf51c8b503de4f73114e4 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:27 +0100 Subject: [PATCH 030/115] ASoC: Intel: Skylake: Drop pipe_config_idx Field ->pipe_config_idx duplicates the job of ->cur_config_idx so remove it. Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-4-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl-topology.c | 10 +++++----- sound/soc/intel/skylake/skl-topology.h | 1 - 2 files changed, 5 insertions(+), 6 deletions(-) diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c index f144acae1b44..567a3b661ce4 100644 --- a/sound/soc/intel/skylake/skl-topology.c +++ b/sound/soc/intel/skylake/skl-topology.c @@ -1351,9 +1351,9 @@ static int skl_tplg_multi_config_set_get(struct snd_kcontrol *kcontrol, return -EIO; if (is_set) - pipe->pipe_config_idx = ucontrol->value.enumerated.item[0]; + pipe->cur_config_idx = ucontrol->value.enumerated.item[0]; else - ucontrol->value.enumerated.item[0] = pipe->pipe_config_idx; + ucontrol->value.enumerated.item[0] = pipe->cur_config_idx; return 0; } @@ -1814,11 +1814,11 @@ static int skl_tplg_be_fill_pipe_params(struct snd_soc_dai *dai, if (ret) goto err; - dev_dbg(skl->dev, "%s using pipe config: %d\n", __func__, pipe->pipe_config_idx); + dev_dbg(skl->dev, "%s using pipe config: %d\n", __func__, pipe->cur_config_idx); if (pipe->direction == SNDRV_PCM_STREAM_PLAYBACK) - pipe_fmt = &pipe->configs[pipe->pipe_config_idx].out_fmt; + pipe_fmt = &pipe->configs[pipe->cur_config_idx].out_fmt; else - pipe_fmt = &pipe->configs[pipe->pipe_config_idx].in_fmt; + pipe_fmt = &pipe->configs[pipe->cur_config_idx].in_fmt; /* update the blob based on virtual bus_id*/ cfg = intel_nhlt_get_endpoint_blob(dai->dev, skl->nhlt, diff --git a/sound/soc/intel/skylake/skl-topology.h b/sound/soc/intel/skylake/skl-topology.h index 017ac0ef324d..6db0fd7bad49 100644 --- a/sound/soc/intel/skylake/skl-topology.h +++ b/sound/soc/intel/skylake/skl-topology.h @@ -324,7 +324,6 @@ struct skl_pipe { struct skl_path_config configs[SKL_MAX_PATH_CONFIGS]; struct list_head w_list; bool passthru; - u32 pipe_config_idx; }; enum skl_module_state { From 4ac587f3578c5ca490e4df55af6403f5474eb2f0 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:28 +0100 Subject: [PATCH 031/115] ASoC: Intel: Skylake: Introduce single place for pipe-config selection Provide a single location for pipe config selection where all fields that have to be updated whenever ->pipe_config_idx changes can be updated accordingly. Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-5-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl-topology.c | 18 ++++++++++-------- 1 file changed, 10 insertions(+), 8 deletions(-) diff --git a/sound/soc/intel/skylake/skl-topology.c b/sound/soc/intel/skylake/skl-topology.c index 567a3b661ce4..b20643b83401 100644 --- a/sound/soc/intel/skylake/skl-topology.c +++ b/sound/soc/intel/skylake/skl-topology.c @@ -582,6 +582,12 @@ static int skl_tplg_unload_pipe_modules(struct skl_dev *skl, return ret; } +static void skl_tplg_set_pipe_config_idx(struct skl_pipe *pipe, int idx) +{ + pipe->cur_config_idx = idx; + pipe->memory_pages = pipe->configs[idx].mem_pages; +} + /* * Here, we select pipe format based on the pipe type and pipe * direction to determine the current config index for the pipeline. @@ -600,16 +606,14 @@ skl_tplg_get_pipe_config(struct skl_dev *skl, struct skl_module_cfg *mconfig) int i; if (pipe->nr_cfgs == 0) { - pipe->cur_config_idx = 0; + skl_tplg_set_pipe_config_idx(pipe, 0); return 0; } if (pipe->conn_type == SKL_PIPE_CONN_TYPE_NONE || pipe->nr_cfgs == 1) { dev_dbg(skl->dev, "No conn_type or just 1 pathcfg, taking 0th for %d\n", pipe->ppl_id); - pipe->cur_config_idx = 0; - pipe->memory_pages = pconfig->mem_pages; - + skl_tplg_set_pipe_config_idx(pipe, 0); return 0; } @@ -628,10 +632,8 @@ skl_tplg_get_pipe_config(struct skl_dev *skl, struct skl_module_cfg *mconfig) if (CHECK_HW_PARAMS(params->ch, params->s_freq, params->s_fmt, fmt->channels, fmt->freq, fmt->bps)) { - pipe->cur_config_idx = i; - pipe->memory_pages = pconfig->mem_pages; + skl_tplg_set_pipe_config_idx(pipe, i); dev_dbg(skl->dev, "Using pipe config: %d\n", i); - return 0; } } @@ -1351,7 +1353,7 @@ static int skl_tplg_multi_config_set_get(struct snd_kcontrol *kcontrol, return -EIO; if (is_set) - pipe->cur_config_idx = ucontrol->value.enumerated.item[0]; + skl_tplg_set_pipe_config_idx(pipe, ucontrol->value.enumerated.item[0]); else ucontrol->value.enumerated.item[0] = pipe->cur_config_idx; From 171107237246d66bce04f3769d33648f896b4ce3 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:29 +0100 Subject: [PATCH 032/115] ASoC: Intel: Skylake: Fix driver hang during shutdown AudioDSP cores and HDAudio links need to be turned off on shutdown to ensure no communication or data transfer occurs during the procedure. Fixes: c5a76a246989 ("ASoC: Intel: Skylake: Add shutdown callback") Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-6-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/sound/soc/intel/skylake/skl.c b/sound/soc/intel/skylake/skl.c index 9bd9f9866898..998bd0232cf1 100644 --- a/sound/soc/intel/skylake/skl.c +++ b/sound/soc/intel/skylake/skl.c @@ -1107,7 +1107,10 @@ static void skl_shutdown(struct pci_dev *pci) if (!skl->init_done) return; - snd_hdac_stop_streams_and_chip(bus); + snd_hdac_stop_streams(bus); + snd_hdac_ext_bus_link_power_down_all(bus); + skl_dsp_sleep(skl->dsp); + list_for_each_entry(s, &bus->stream_list, list) { stream = stream_to_hdac_ext_stream(s); snd_hdac_ext_stream_decouple(bus, stream, false); From 451d85c46cf719a09a052510d4d4cd920103163a Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Mon, 5 Dec 2022 09:53:30 +0100 Subject: [PATCH 033/115] ASoC: Intel: Skylake: Use SG allocation for SKL-based firmware load Resign from ->alloc_dma_buf() and use snd_dma_alloc_pages() directly. For data i.e.: base firmware binary transfer, make use of SG allocation to better adapt to memory-limited environment. For BDL descriptor, given its small size this is not required. Signed-off-by: Cezary Rojewski Tested-by: Lukasz Majczak Link: https://lore.kernel.org/r/20221205085330.857665-7-cezary.rojewski@intel.com Signed-off-by: Mark Brown --- sound/soc/intel/skylake/skl-sst-cldma.c | 27 +++++++++++++++---------- 1 file changed, 16 insertions(+), 11 deletions(-) diff --git a/sound/soc/intel/skylake/skl-sst-cldma.c b/sound/soc/intel/skylake/skl-sst-cldma.c index b91f7a652a2b..b0204ea00f07 100644 --- a/sound/soc/intel/skylake/skl-sst-cldma.c +++ b/sound/soc/intel/skylake/skl-sst-cldma.c @@ -11,6 +11,7 @@ #include #include #include +#include #include "../common/sst-dsp.h" #include "../common/sst-dsp-priv.h" @@ -79,21 +80,25 @@ static void skl_cldma_setup_bdle(struct sst_dsp *ctx, __le32 **bdlp, int size, int with_ioc) { __le32 *bdl = *bdlp; + int remaining = ctx->cl_dev.bufsize; + int offset = 0; ctx->cl_dev.frags = 0; - while (size > 0) { - phys_addr_t addr = virt_to_phys(dmab_data->area + - (ctx->cl_dev.frags * ctx->cl_dev.bufsize)); + while (remaining > 0) { + phys_addr_t addr; + int chunk; + addr = snd_sgbuf_get_addr(dmab_data, offset); bdl[0] = cpu_to_le32(lower_32_bits(addr)); bdl[1] = cpu_to_le32(upper_32_bits(addr)); + chunk = snd_sgbuf_get_chunk_size(dmab_data, offset, size); + bdl[2] = cpu_to_le32(chunk); - bdl[2] = cpu_to_le32(ctx->cl_dev.bufsize); - - size -= ctx->cl_dev.bufsize; - bdl[3] = (size || !with_ioc) ? 0 : cpu_to_le32(0x01); + remaining -= chunk; + bdl[3] = (remaining > 0) ? 0 : cpu_to_le32(0x01); bdl += 4; + offset += chunk; ctx->cl_dev.frags++; } } @@ -338,15 +343,15 @@ int skl_cldma_prepare(struct sst_dsp *ctx) ctx->cl_dev.ops.cl_stop_dma = skl_cldma_stop; /* Allocate buffer*/ - ret = ctx->dsp_ops.alloc_dma_buf(ctx->dev, - &ctx->cl_dev.dmab_data, ctx->cl_dev.bufsize); + ret = snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV_SG, ctx->dev, ctx->cl_dev.bufsize, + &ctx->cl_dev.dmab_data); if (ret < 0) { dev_err(ctx->dev, "Alloc buffer for base fw failed: %x\n", ret); return ret; } + /* Setup Code loader BDL */ - ret = ctx->dsp_ops.alloc_dma_buf(ctx->dev, - &ctx->cl_dev.dmab_bdl, PAGE_SIZE); + ret = snd_dma_alloc_pages(SNDRV_DMA_TYPE_DEV, ctx->dev, BDL_SIZE, &ctx->cl_dev.dmab_bdl); if (ret < 0) { dev_err(ctx->dev, "Alloc buffer for blde failed: %x\n", ret); ctx->dsp_ops.free_dma_buf(ctx->dev, &ctx->cl_dev.dmab_data); From c0660fce5e0672b9fcffaae02184d58c8ed2aec1 Mon Sep 17 00:00:00 2001 From: Brent Lu Date: Tue, 6 Dec 2022 15:25:05 -0600 Subject: [PATCH 034/115] ASoC: Intel: sof_rt5682: add jsl_rt5682 board config This configuration supports JSL boards which implement ALC5682I-VD/VS on SSP0 port. Reviewed-by: Bard Liao Reviewed-by: Ranjani Sridharan Signed-off-by: Brent Lu Signed-off-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221206212507.359993-2-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/sof_rt5682.c | 6 ++++++ sound/soc/intel/common/soc-acpi-intel-jsl-match.c | 5 +++++ 2 files changed, 11 insertions(+) diff --git a/sound/soc/intel/boards/sof_rt5682.c b/sound/soc/intel/boards/sof_rt5682.c index 4a2f91249b10..2eabc4b0fafa 100644 --- a/sound/soc/intel/boards/sof_rt5682.c +++ b/sound/soc/intel/boards/sof_rt5682.c @@ -1104,6 +1104,12 @@ static const struct platform_device_id board_ids[] = { SOF_RT5682_SSP_AMP(1) | SOF_RT5682_NUM_HDMIDEV(4)), }, + { + .name = "jsl_rt5682", + .driver_data = (kernel_ulong_t)(SOF_RT5682_MCLK_EN | + SOF_RT5682_MCLK_24MHZ | + SOF_RT5682_SSP_CODEC(0)), + }, { } }; MODULE_DEVICE_TABLE(platform, board_ids); diff --git a/sound/soc/intel/common/soc-acpi-intel-jsl-match.c b/sound/soc/intel/common/soc-acpi-intel-jsl-match.c index b95c4b2cda94..f5c7e1bbded0 100644 --- a/sound/soc/intel/common/soc-acpi-intel-jsl-match.c +++ b/sound/soc/intel/common/soc-acpi-intel-jsl-match.c @@ -78,6 +78,11 @@ struct snd_soc_acpi_mach snd_soc_acpi_intel_jsl_machines[] = { .quirk_data = &mx98360a_spk, .sof_tplg_filename = "sof-jsl-rt5682-mx98360a.tplg", }, + { + .comp_ids = &rt5682_rt5682s_hp, + .drv_name = "jsl_rt5682", + .sof_tplg_filename = "sof-jsl-rt5682.tplg", + }, { .id = "10134242", .drv_name = "jsl_cs4242_mx98360a", From 5c10da436ebd93f9bfa244ea933773d14b566499 Mon Sep 17 00:00:00 2001 From: Gongjun Song Date: Tue, 6 Dec 2022 15:25:06 -0600 Subject: [PATCH 035/115] ASoC: Intel: sof_sdw: use common helpers for all Realtek amps sof_sdw_rt1308.c/sof_sdw_rt1316.c/sof_sdw_rt1318.c handle amp in basically the same way, optimized and merged into one file. Reviewed-by: Ranjani Sridharan Reviewed-by: Bard Liao Signed-off-by: Gongjun Song Signed-off-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221206212507.359993-3-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/Makefile | 3 +- sound/soc/intel/boards/sof_sdw.c | 10 +- sound/soc/intel/boards/sof_sdw_common.h | 20 +- sound/soc/intel/boards/sof_sdw_rt1316.c | 239 ------------------ sound/soc/intel/boards/sof_sdw_rt1318.c | 120 --------- .../{sof_sdw_rt1308.c => sof_sdw_rt_amp.c} | 139 +++++++--- 6 files changed, 120 insertions(+), 411 deletions(-) delete mode 100644 sound/soc/intel/boards/sof_sdw_rt1316.c delete mode 100644 sound/soc/intel/boards/sof_sdw_rt1318.c rename sound/soc/intel/boards/{sof_sdw_rt1308.c => sof_sdw_rt_amp.c} (59%) diff --git a/sound/soc/intel/boards/Makefile b/sound/soc/intel/boards/Makefile index 7e1a4ff77ac3..d1fd7a2b32db 100644 --- a/sound/soc/intel/boards/Makefile +++ b/sound/soc/intel/boards/Makefile @@ -37,8 +37,7 @@ snd-soc-sof_da7219_max98373-objs := sof_da7219_max98373.o snd-soc-ehl-rt5660-objs := ehl_rt5660.o snd-soc-sof-ssp-amp-objs := sof_ssp_amp.o snd-soc-sof-sdw-objs += sof_sdw.o \ - sof_sdw_max98373.o sof_sdw_rt1308.o \ - sof_sdw_rt1316.o sof_sdw_rt1318.o \ + sof_sdw_max98373.o sof_sdw_rt_amp.o \ sof_sdw_rt5682.o sof_sdw_rt700.o \ sof_sdw_rt711.o sof_sdw_rt711_sdca.o \ sof_sdw_rt715.o sof_sdw_rt715_sdca.o \ diff --git a/sound/soc/intel/boards/sof_sdw.c b/sound/soc/intel/boards/sof_sdw.c index b58c7b35599d..d2ed807abde9 100644 --- a/sound/soc/intel/boards/sof_sdw.c +++ b/sound/soc/intel/boards/sof_sdw.c @@ -550,23 +550,23 @@ static struct sof_sdw_codec_info codec_info_list[] = { .direction = {true, false}, .dai_name = "rt1308-aif", .ops = &sof_sdw_rt1308_i2s_ops, - .init = sof_sdw_rt1308_init, - .exit = sof_sdw_rt1308_exit, + .init = sof_sdw_rt_amp_init, + .exit = sof_sdw_rt_amp_exit, .codec_type = SOF_SDW_CODEC_TYPE_AMP, }, { .part_id = 0x1316, .direction = {true, true}, .dai_name = "rt1316-aif", - .init = sof_sdw_rt1316_init, - .exit = sof_sdw_rt1316_exit, + .init = sof_sdw_rt_amp_init, + .exit = sof_sdw_rt_amp_exit, .codec_type = SOF_SDW_CODEC_TYPE_AMP, }, { .part_id = 0x1318, .direction = {true, true}, .dai_name = "rt1318-aif", - .init = sof_sdw_rt1318_init, + .init = sof_sdw_rt_amp_init, .codec_type = SOF_SDW_CODEC_TYPE_AMP, }, { diff --git a/sound/soc/intel/boards/sof_sdw_common.h b/sound/soc/intel/boards/sof_sdw_common.h index 54a50f7da4da..350010b0e5f4 100644 --- a/sound/soc/intel/boards/sof_sdw_common.h +++ b/sound/soc/intel/boards/sof_sdw_common.h @@ -125,30 +125,18 @@ int sof_sdw_rt700_init(struct snd_soc_card *card, struct sof_sdw_codec_info *info, bool playback); -/* RT1308 support */ +/* RT1308 I2S support */ extern struct snd_soc_ops sof_sdw_rt1308_i2s_ops; -int sof_sdw_rt1308_init(struct snd_soc_card *card, +/* generic amp support */ +int sof_sdw_rt_amp_init(struct snd_soc_card *card, const struct snd_soc_acpi_link_adr *link, struct snd_soc_dai_link *dai_links, struct sof_sdw_codec_info *info, bool playback); -int sof_sdw_rt1308_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link); +int sof_sdw_rt_amp_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link); /* RT1316 support */ -int sof_sdw_rt1316_init(struct snd_soc_card *card, - const struct snd_soc_acpi_link_adr *link, - struct snd_soc_dai_link *dai_links, - struct sof_sdw_codec_info *info, - bool playback); -int sof_sdw_rt1316_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link); - -/* RT1318 support */ -int sof_sdw_rt1318_init(struct snd_soc_card *card, - const struct snd_soc_acpi_link_adr *link, - struct snd_soc_dai_link *dai_links, - struct sof_sdw_codec_info *info, - bool playback); /* RT715 support */ int sof_sdw_rt715_init(struct snd_soc_card *card, diff --git a/sound/soc/intel/boards/sof_sdw_rt1316.c b/sound/soc/intel/boards/sof_sdw_rt1316.c deleted file mode 100644 index f6bbea0d3810..000000000000 --- a/sound/soc/intel/boards/sof_sdw_rt1316.c +++ /dev/null @@ -1,239 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-only -// Copyright (c) 2020 Intel Corporation - -/* - * sof_sdw_rt1316 - Helpers to handle RT1316 from generic machine driver - */ - -#include -#include -#include -#include -#include -#include -#include -#include -#include -#include "sof_sdw_common.h" -#include "sof_sdw_amp_coeff_tables.h" - -struct rt1316_platform_data { - const unsigned char *bq_params; - const unsigned int bq_params_cnt; -}; - -static const struct rt1316_platform_data dell_0b00_platform_data = { - .bq_params = dell_0b00_bq_params, - .bq_params_cnt = ARRAY_SIZE(dell_0b00_bq_params), -}; - -static const struct dmi_system_id dmi_platform_data[] = { - /* AlderLake devices */ - { - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), - DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0B00") - }, - .driver_data = (void *)&dell_0b00_platform_data, - }, - { - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), - DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0B01") - }, - .driver_data = (void *)&dell_0b00_platform_data, - }, - { - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), - DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0AFF") - }, - .driver_data = (void *)&dell_0b00_platform_data, - }, - { - .matches = { - DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), - DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0AFE") - }, - .driver_data = (void *)&dell_0b00_platform_data, - }, -}; - -static int rt1316_add_device_props(struct device *sdw_dev) -{ - struct property_entry props[3] = {}; - struct fwnode_handle *fwnode; - const struct dmi_system_id *dmi_data; - const struct rt1316_platform_data *pdata; - unsigned char params[RT1316_MAX_BQ_REG]; - int ret; - - dmi_data = dmi_first_match(dmi_platform_data); - if (!dmi_data) - return 0; - - pdata = dmi_data->driver_data; - memcpy(¶ms, pdata->bq_params, sizeof(unsigned char) * pdata->bq_params_cnt); - - props[0] = PROPERTY_ENTRY_U8_ARRAY("realtek,bq-params", params); - props[1] = PROPERTY_ENTRY_U32("realtek,bq-params-cnt", pdata->bq_params_cnt); - - fwnode = fwnode_create_software_node(props, NULL); - if (IS_ERR(fwnode)) - return PTR_ERR(fwnode); - - ret = device_add_software_node(sdw_dev, to_software_node(fwnode)); - - fwnode_handle_put(fwnode); - - return ret; -} - -static const struct snd_soc_dapm_widget rt1316_widgets[] = { - SND_SOC_DAPM_SPK("Speaker", NULL), -}; - -/* - * dapm routes for rt1316 will be registered dynamically according - * to the number of rt1316 used. The first two entries will be registered - * for one codec case, and the last two entries are also registered - * if two 1316s are used. - */ -static const struct snd_soc_dapm_route rt1316_map[] = { - { "Speaker", NULL, "rt1316-1 SPOL" }, - { "Speaker", NULL, "rt1316-1 SPOR" }, - { "Speaker", NULL, "rt1316-2 SPOL" }, - { "Speaker", NULL, "rt1316-2 SPOR" }, -}; - -static const struct snd_kcontrol_new rt1316_controls[] = { - SOC_DAPM_PIN_SWITCH("Speaker"), -}; - -static int first_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - struct snd_soc_card *card = rtd->card; - int ret; - - card->components = devm_kasprintf(card->dev, GFP_KERNEL, - "%s spk:rt1316", - card->components); - if (!card->components) - return -ENOMEM; - - ret = snd_soc_add_card_controls(card, rt1316_controls, - ARRAY_SIZE(rt1316_controls)); - if (ret) { - dev_err(card->dev, "rt1316 controls addition failed: %d\n", ret); - return ret; - } - - ret = snd_soc_dapm_new_controls(&card->dapm, rt1316_widgets, - ARRAY_SIZE(rt1316_widgets)); - if (ret) { - dev_err(card->dev, "rt1316 widgets addition failed: %d\n", ret); - return ret; - } - - ret = snd_soc_dapm_add_routes(&card->dapm, rt1316_map, 2); - if (ret) - dev_err(rtd->dev, "failed to add first SPK map: %d\n", ret); - - return ret; -} - -static int second_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - struct snd_soc_card *card = rtd->card; - int ret; - - ret = snd_soc_dapm_add_routes(&card->dapm, rt1316_map + 2, 2); - if (ret) - dev_err(rtd->dev, "failed to add second SPK map: %d\n", ret); - - return ret; -} - -static int all_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - int ret; - - ret = first_spk_init(rtd); - if (ret) - return ret; - - return second_spk_init(rtd); -} - -int sof_sdw_rt1316_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link) -{ - struct mc_private *ctx = snd_soc_card_get_drvdata(card); - - if (ctx->amp_dev1) { - device_remove_software_node(ctx->amp_dev1); - put_device(ctx->amp_dev1); - } - - if (ctx->amp_dev2) { - device_remove_software_node(ctx->amp_dev2); - put_device(ctx->amp_dev2); - } - - return 0; -} - -int sof_sdw_rt1316_init(struct snd_soc_card *card, - const struct snd_soc_acpi_link_adr *link, - struct snd_soc_dai_link *dai_links, - struct sof_sdw_codec_info *info, - bool playback) -{ - struct mc_private *ctx = snd_soc_card_get_drvdata(card); - struct device *sdw_dev1, *sdw_dev2; - int ret; - - /* Count amp number and do init on playback link only. */ - if (!playback) - return 0; - - info->amp_num++; - if (info->amp_num == 1) - dai_links->init = first_spk_init; - - if (info->amp_num == 2) { - sdw_dev1 = bus_find_device_by_name(&sdw_bus_type, NULL, dai_links->codecs[0].name); - if (!sdw_dev1) - return -EPROBE_DEFER; - - ret = rt1316_add_device_props(sdw_dev1); - if (ret < 0) { - put_device(sdw_dev1); - return ret; - } - ctx->amp_dev1 = sdw_dev1; - - sdw_dev2 = bus_find_device_by_name(&sdw_bus_type, NULL, dai_links->codecs[1].name); - if (!sdw_dev2) - return -EPROBE_DEFER; - - ret = rt1316_add_device_props(sdw_dev2); - if (ret < 0) { - put_device(sdw_dev2); - return ret; - } - ctx->amp_dev2 = sdw_dev2; - - /* - * if two 1316s are in one dai link, the init function - * in this dai link will be first set for the first speaker, - * and it should be reset to initialize all speakers when - * the second speaker is found. - */ - if (dai_links->init) - dai_links->init = all_spk_init; - else - dai_links->init = second_spk_init; - } - - return 0; -} diff --git a/sound/soc/intel/boards/sof_sdw_rt1318.c b/sound/soc/intel/boards/sof_sdw_rt1318.c deleted file mode 100644 index dbee4bf5c814..000000000000 --- a/sound/soc/intel/boards/sof_sdw_rt1318.c +++ /dev/null @@ -1,120 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0-only -// Copyright (c) 2022 Intel Corporation - -/* - * sof_sdw_rt1318 - Helpers to handle RT1318 from generic machine driver - */ - -#include -#include -#include -#include -#include -#include -#include "sof_sdw_common.h" - -static const struct snd_soc_dapm_widget rt1318_widgets[] = { - SND_SOC_DAPM_SPK("Speaker", NULL), -}; - -/* - * dapm routes for rt1318 will be registered dynamically according - * to the number of rt1318 used. The first two entries will be registered - * for one codec case, and the last two entries are also registered - * if two 1318s are used. - */ -static const struct snd_soc_dapm_route rt1318_map[] = { - { "Speaker", NULL, "rt1318-1 SPOL" }, - { "Speaker", NULL, "rt1318-1 SPOR" }, - { "Speaker", NULL, "rt1318-2 SPOL" }, - { "Speaker", NULL, "rt1318-2 SPOR" }, -}; - -static const struct snd_kcontrol_new rt1318_controls[] = { - SOC_DAPM_PIN_SWITCH("Speaker"), -}; - -static int first_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - struct snd_soc_card *card = rtd->card; - int ret; - - card->components = devm_kasprintf(card->dev, GFP_KERNEL, - "%s spk:rt1318", - card->components); - if (!card->components) - return -ENOMEM; - - ret = snd_soc_add_card_controls(card, rt1318_controls, - ARRAY_SIZE(rt1318_controls)); - if (ret) { - dev_err(card->dev, "rt1318 controls addition failed: %d\n", ret); - return ret; - } - - ret = snd_soc_dapm_new_controls(&card->dapm, rt1318_widgets, - ARRAY_SIZE(rt1318_widgets)); - if (ret) { - dev_err(card->dev, "rt1318 widgets addition failed: %d\n", ret); - return ret; - } - - ret = snd_soc_dapm_add_routes(&card->dapm, rt1318_map, 2); - if (ret) - dev_err(rtd->dev, "failed to add first SPK map: %d\n", ret); - - return ret; -} - -static int second_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - struct snd_soc_card *card = rtd->card; - int ret; - - ret = snd_soc_dapm_add_routes(&card->dapm, rt1318_map + 2, 2); - if (ret) - dev_err(rtd->dev, "failed to add second SPK map: %d\n", ret); - - return ret; -} - -static int all_spk_init(struct snd_soc_pcm_runtime *rtd) -{ - int ret; - - ret = first_spk_init(rtd); - if (ret) - return ret; - - return second_spk_init(rtd); -} - -int sof_sdw_rt1318_init(struct snd_soc_card *card, - const struct snd_soc_acpi_link_adr *link, - struct snd_soc_dai_link *dai_links, - struct sof_sdw_codec_info *info, - bool playback) -{ - /* Count amp number and do init on playback link only. */ - if (!playback) - return 0; - - info->amp_num++; - if (info->amp_num == 1) - dai_links->init = first_spk_init; - - if (info->amp_num == 2) { - /* - * if two 1318s are in one dai link, the init function - * in this dai link will be first set for the first speaker, - * and it should be reset to initialize all speakers when - * the second speaker is found. - */ - if (dai_links->init) - dai_links->init = all_spk_init; - else - dai_links->init = second_spk_init; - } - - return 0; -} diff --git a/sound/soc/intel/boards/sof_sdw_rt1308.c b/sound/soc/intel/boards/sof_sdw_rt_amp.c similarity index 59% rename from sound/soc/intel/boards/sof_sdw_rt1308.c rename to sound/soc/intel/boards/sof_sdw_rt_amp.c index a19b055b9c6f..26bf9e0dd3d2 100644 --- a/sound/soc/intel/boards/sof_sdw_rt1308.c +++ b/sound/soc/intel/boards/sof_sdw_rt_amp.c @@ -1,8 +1,8 @@ // SPDX-License-Identifier: GPL-2.0-only -// Copyright (c) 2020 Intel Corporation +// Copyright (c) 2022 Intel Corporation /* - * sof_sdw_rt1308 - Helpers to handle RT1308 from generic machine driver + * sof_sdw_rt_amp - Helpers to handle RT1308/RT1316/RT1318 from generic machine driver */ #include @@ -18,16 +18,26 @@ #include "sof_sdw_amp_coeff_tables.h" #include "../../codecs/rt1308.h" -struct rt1308_platform_data { +#define CODEC_NAME_SIZE 7 + +/* choose a larger value to resolve compatibility issues */ +#define RT_AMP_MAX_BQ_REG RT1316_MAX_BQ_REG + +struct rt_amp_platform_data { const unsigned char *bq_params; const unsigned int bq_params_cnt; }; -static const struct rt1308_platform_data dell_0a5d_platform_data = { +static const struct rt_amp_platform_data dell_0a5d_platform_data = { .bq_params = dell_0a5d_bq_params, .bq_params_cnt = ARRAY_SIZE(dell_0a5d_bq_params), }; +static const struct rt_amp_platform_data dell_0b00_platform_data = { + .bq_params = dell_0b00_bq_params, + .bq_params_cnt = ARRAY_SIZE(dell_0b00_bq_params), +}; + static const struct dmi_system_id dmi_platform_data[] = { /* CometLake devices */ { @@ -59,15 +69,45 @@ static const struct dmi_system_id dmi_platform_data[] = { }, .driver_data = (void *)&dell_0a5d_platform_data, }, + /* AlderLake devices */ + { + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), + DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0B00") + }, + .driver_data = (void *)&dell_0b00_platform_data, + }, + { + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), + DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0B01") + }, + .driver_data = (void *)&dell_0b00_platform_data, + }, + { + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), + DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0AFF") + }, + .driver_data = (void *)&dell_0b00_platform_data, + }, + { + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc"), + DMI_EXACT_MATCH(DMI_PRODUCT_SKU, "0AFE") + }, + .driver_data = (void *)&dell_0b00_platform_data, + }, + {}, }; -static int rt1308_add_device_props(struct device *sdw_dev) +static int rt_amp_add_device_props(struct device *sdw_dev) { struct property_entry props[3] = {}; struct fwnode_handle *fwnode; const struct dmi_system_id *dmi_data; - const struct rt1308_platform_data *pdata; - unsigned char params[RT1308_MAX_BQ_REG]; + const struct rt_amp_platform_data *pdata; + unsigned char params[RT_AMP_MAX_BQ_REG]; int ret; dmi_data = dmi_first_match(dmi_platform_data); @@ -91,15 +131,19 @@ static int rt1308_add_device_props(struct device *sdw_dev) return ret; } -static const struct snd_soc_dapm_widget rt1308_widgets[] = { +static const struct snd_kcontrol_new rt_amp_controls[] = { + SOC_DAPM_PIN_SWITCH("Speaker"), +}; + +static const struct snd_soc_dapm_widget rt_amp_widgets[] = { SND_SOC_DAPM_SPK("Speaker", NULL), }; /* - * dapm routes for rt1308 will be registered dynamically according - * to the number of rt1308 used. The first two entries will be registered - * for one codec case, and the last two entries are also registered - * if two 1308s are used. + * dapm routes for rt1308/rt1316/rt1318 will be registered dynamically + * according to the number of rt1308/rt1316/rt1318 used. The first two + * entries will be registered for one codec case, and the last two entries + * are also registered if two 1308s/1316s/1318s are used. */ static const struct snd_soc_dapm_route rt1308_map[] = { { "Speaker", NULL, "rt1308-1 SPOL" }, @@ -108,36 +152,69 @@ static const struct snd_soc_dapm_route rt1308_map[] = { { "Speaker", NULL, "rt1308-2 SPOR" }, }; -static const struct snd_kcontrol_new rt1308_controls[] = { - SOC_DAPM_PIN_SWITCH("Speaker"), +static const struct snd_soc_dapm_route rt1316_map[] = { + { "Speaker", NULL, "rt1316-1 SPOL" }, + { "Speaker", NULL, "rt1316-1 SPOR" }, + { "Speaker", NULL, "rt1316-2 SPOL" }, + { "Speaker", NULL, "rt1316-2 SPOR" }, }; +static const struct snd_soc_dapm_route rt1318_map[] = { + { "Speaker", NULL, "rt1318-1 SPOL" }, + { "Speaker", NULL, "rt1318-1 SPOR" }, + { "Speaker", NULL, "rt1318-2 SPOL" }, + { "Speaker", NULL, "rt1318-2 SPOR" }, +}; + +static const struct snd_soc_dapm_route *get_codec_name_and_route(struct snd_soc_pcm_runtime *rtd, + char *codec_name) +{ + const char *dai_name; + + dai_name = rtd->dai_link->codecs->dai_name; + + /* get the codec name */ + snprintf(codec_name, CODEC_NAME_SIZE, "%s", dai_name); + + /* choose the right codec's map */ + if (strcmp(codec_name, "rt1308") == 0) + return rt1308_map; + else if (strcmp(codec_name, "rt1316") == 0) + return rt1316_map; + else + return rt1318_map; +} + static int first_spk_init(struct snd_soc_pcm_runtime *rtd) { struct snd_soc_card *card = rtd->card; + const struct snd_soc_dapm_route *rt_amp_map; + char codec_name[CODEC_NAME_SIZE]; int ret; + rt_amp_map = get_codec_name_and_route(rtd, codec_name); + card->components = devm_kasprintf(card->dev, GFP_KERNEL, - "%s spk:rt1308", - card->components); + "%s spk:%s", + card->components, codec_name); if (!card->components) return -ENOMEM; - ret = snd_soc_add_card_controls(card, rt1308_controls, - ARRAY_SIZE(rt1308_controls)); + ret = snd_soc_add_card_controls(card, rt_amp_controls, + ARRAY_SIZE(rt_amp_controls)); if (ret) { - dev_err(card->dev, "rt1308 controls addition failed: %d\n", ret); + dev_err(card->dev, "%s controls addition failed: %d\n", codec_name, ret); return ret; } - ret = snd_soc_dapm_new_controls(&card->dapm, rt1308_widgets, - ARRAY_SIZE(rt1308_widgets)); + ret = snd_soc_dapm_new_controls(&card->dapm, rt_amp_widgets, + ARRAY_SIZE(rt_amp_widgets)); if (ret) { - dev_err(card->dev, "rt1308 widgets addition failed: %d\n", ret); + dev_err(card->dev, "%s widgets addition failed: %d\n", codec_name, ret); return ret; } - ret = snd_soc_dapm_add_routes(&card->dapm, rt1308_map, 2); + ret = snd_soc_dapm_add_routes(&card->dapm, rt_amp_map, 2); if (ret) dev_err(rtd->dev, "failed to add first SPK map: %d\n", ret); @@ -147,9 +224,13 @@ static int first_spk_init(struct snd_soc_pcm_runtime *rtd) static int second_spk_init(struct snd_soc_pcm_runtime *rtd) { struct snd_soc_card *card = rtd->card; + const struct snd_soc_dapm_route *rt_amp_map; + char codec_name[CODEC_NAME_SIZE]; int ret; - ret = snd_soc_dapm_add_routes(&card->dapm, rt1308_map + 2, 2); + rt_amp_map = get_codec_name_and_route(rtd, codec_name); + + ret = snd_soc_dapm_add_routes(&card->dapm, rt_amp_map + 2, 2); if (ret) dev_err(rtd->dev, "failed to add second SPK map: %d\n", ret); @@ -204,7 +285,7 @@ struct snd_soc_ops sof_sdw_rt1308_i2s_ops = { .hw_params = rt1308_i2s_hw_params, }; -int sof_sdw_rt1308_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link) +int sof_sdw_rt_amp_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_link) { struct mc_private *ctx = snd_soc_card_get_drvdata(card); @@ -221,7 +302,7 @@ int sof_sdw_rt1308_exit(struct snd_soc_card *card, struct snd_soc_dai_link *dai_ return 0; } -int sof_sdw_rt1308_init(struct snd_soc_card *card, +int sof_sdw_rt_amp_init(struct snd_soc_card *card, const struct snd_soc_acpi_link_adr *link, struct snd_soc_dai_link *dai_links, struct sof_sdw_codec_info *info, @@ -244,7 +325,7 @@ int sof_sdw_rt1308_init(struct snd_soc_card *card, if (!sdw_dev1) return -EPROBE_DEFER; - ret = rt1308_add_device_props(sdw_dev1); + ret = rt_amp_add_device_props(sdw_dev1); if (ret < 0) { put_device(sdw_dev1); return ret; @@ -255,7 +336,7 @@ int sof_sdw_rt1308_init(struct snd_soc_card *card, if (!sdw_dev2) return -EPROBE_DEFER; - ret = rt1308_add_device_props(sdw_dev2); + ret = rt_amp_add_device_props(sdw_dev2); if (ret < 0) { put_device(sdw_dev2); return ret; @@ -263,7 +344,7 @@ int sof_sdw_rt1308_init(struct snd_soc_card *card, ctx->amp_dev2 = sdw_dev2; /* - * if two 1308s are in one dai link, the init function + * if two amps are in one dai link, the init function * in this dai link will be first set for the first speaker, * and it should be reset to initialize all speakers when * the second speaker is found. From 47d2b66fec133cb27da3a551334686e465d19469 Mon Sep 17 00:00:00 2001 From: Bard Liao Date: Tue, 6 Dec 2022 15:25:07 -0600 Subject: [PATCH 036/115] ASoC: Intel: sof_realtek_common: set ret = 0 as initial value 'ret' will not be initialized if dai_fmt is not DSP_A or DSP_B. Reviewed-by: Ranjani Sridharan Signed-off-by: Bard Liao Signed-off-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221206212507.359993-4-pierre-louis.bossart@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/sof_realtek_common.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/sound/soc/intel/boards/sof_realtek_common.c b/sound/soc/intel/boards/sof_realtek_common.c index ff2851fc8930..6c12ca92f371 100644 --- a/sound/soc/intel/boards/sof_realtek_common.c +++ b/sound/soc/intel/boards/sof_realtek_common.c @@ -267,7 +267,8 @@ static int rt1015_hw_params(struct snd_pcm_substream *substream, struct snd_soc_pcm_runtime *rtd = asoc_substream_to_rtd(substream); struct snd_soc_dai_link *dai_link = rtd->dai_link; struct snd_soc_dai *codec_dai; - int i, clk_freq, ret; + int i, clk_freq; + int ret = 0; clk_freq = sof_dai_get_bclk(rtd); From 3327d721114c109ba0575f86f8fda3b525404054 Mon Sep 17 00:00:00 2001 From: Wang Yufen Date: Mon, 5 Dec 2022 18:04:24 +0800 Subject: [PATCH 037/115] ASoC: mediatek: mt8173-rt5650-rt5514: fix refcount leak in mt8173_rt5650_rt5514_dev_probe() The node returned by of_parse_phandle() with refcount incremented, of_node_put() needs be called when finish using it. So add it in the error path in mt8173_rt5650_rt5514_dev_probe(). Fixes: 0d1d7a664288 ("ASoC: mediatek: Refine mt8173 driver and change config option") Signed-off-by: Wang Yufen Link: https://lore.kernel.org/r/1670234664-24246-1-git-send-email-wangyufen@huawei.com Signed-off-by: Mark Brown --- sound/soc/mediatek/mt8173/mt8173-rt5650-rt5514.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/sound/soc/mediatek/mt8173/mt8173-rt5650-rt5514.c b/sound/soc/mediatek/mt8173/mt8173-rt5650-rt5514.c index 12f40c81b101..f803f121659d 100644 --- a/sound/soc/mediatek/mt8173/mt8173-rt5650-rt5514.c +++ b/sound/soc/mediatek/mt8173/mt8173-rt5650-rt5514.c @@ -200,14 +200,16 @@ static int mt8173_rt5650_rt5514_dev_probe(struct platform_device *pdev) if (!mt8173_rt5650_rt5514_dais[DAI_LINK_CODEC_I2S].codecs[0].of_node) { dev_err(&pdev->dev, "Property 'audio-codec' missing or invalid\n"); - return -EINVAL; + ret = -EINVAL; + goto out; } mt8173_rt5650_rt5514_dais[DAI_LINK_CODEC_I2S].codecs[1].of_node = of_parse_phandle(pdev->dev.of_node, "mediatek,audio-codec", 1); if (!mt8173_rt5650_rt5514_dais[DAI_LINK_CODEC_I2S].codecs[1].of_node) { dev_err(&pdev->dev, "Property 'audio-codec' missing or invalid\n"); - return -EINVAL; + ret = -EINVAL; + goto out; } mt8173_rt5650_rt5514_codec_conf[0].dlc.of_node = mt8173_rt5650_rt5514_dais[DAI_LINK_CODEC_I2S].codecs[1].of_node; @@ -216,6 +218,7 @@ static int mt8173_rt5650_rt5514_dev_probe(struct platform_device *pdev) ret = devm_snd_soc_register_card(&pdev->dev, card); +out: of_node_put(platform_node); return ret; } From 8ab2d12c726f0fde0692fa5d81d8019b3dcd62d0 Mon Sep 17 00:00:00 2001 From: Wang Yufen Date: Mon, 5 Dec 2022 16:15:27 +0800 Subject: [PATCH 038/115] ASoC: audio-graph-card: fix refcount leak of cpu_ep in __graph_for_each_link() The of_get_next_child() returns a node with refcount incremented, and decrements the refcount of prev. So in the error path of the while loop, of_node_put() needs be called for cpu_ep. Fixes: fce9b90c1ab7 ("ASoC: audio-graph-card: cleanup DAI link loop method - step2") Signed-off-by: Wang Yufen Acked-by: Kuninori Morimoto Link: https://lore.kernel.org/r/1670228127-13835-1-git-send-email-wangyufen@huawei.com Signed-off-by: Mark Brown --- sound/soc/generic/audio-graph-card.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/sound/soc/generic/audio-graph-card.c b/sound/soc/generic/audio-graph-card.c index fe7cf972d44c..5daa824a4ffc 100644 --- a/sound/soc/generic/audio-graph-card.c +++ b/sound/soc/generic/audio-graph-card.c @@ -485,8 +485,10 @@ static int __graph_for_each_link(struct asoc_simple_priv *priv, of_node_put(codec_ep); of_node_put(codec_port); - if (ret < 0) + if (ret < 0) { + of_node_put(cpu_ep); return ret; + } codec_port_old = codec_port; } From ef0a098efb36660326c133af9b5a04a96a00e3ca Mon Sep 17 00:00:00 2001 From: Wang Jingjin Date: Mon, 5 Dec 2022 11:28:02 +0800 Subject: [PATCH 039/115] ASoC: rockchip: pdm: Add missing clk_disable_unprepare() in rockchip_pdm_runtime_resume() The clk_disable_unprepare() should be called in the error handling of rockchip_pdm_runtime_resume(). Fixes: fc05a5b22253 ("ASoC: rockchip: add support for pdm controller") Signed-off-by: Wang Jingjin Link: https://lore.kernel.org/r/20221205032802.2422983-1-wangjingjin1@huawei.com Signed-off-by: Mark Brown --- sound/soc/rockchip/rockchip_pdm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/rockchip/rockchip_pdm.c b/sound/soc/rockchip/rockchip_pdm.c index a7549f827235..5b1e47bdc376 100644 --- a/sound/soc/rockchip/rockchip_pdm.c +++ b/sound/soc/rockchip/rockchip_pdm.c @@ -431,6 +431,7 @@ static int rockchip_pdm_runtime_resume(struct device *dev) ret = clk_prepare_enable(pdm->hclk); if (ret) { + clk_disable_unprepare(pdm->clk); dev_err(pdm->dev, "hclock enable failed %d\n", ret); return ret; } From a39bc7cf8e284653fb6fd9d897f269f4ac80cf52 Mon Sep 17 00:00:00 2001 From: ye xingchen Date: Mon, 5 Dec 2022 19:43:47 +0800 Subject: [PATCH 040/115] ASoC: imx-audmux: use sysfs_emit() to instead of scnprintf() Follow the advice of the Documentation/filesystems/sysfs.rst and show() should only use sysfs_emit() or sysfs_emit_at() when formatting the value to be returned to user space. Signed-off-by: ye xingchen Link: https://lore.kernel.org/r/202212051943476482106@zte.com.cn Signed-off-by: Mark Brown --- sound/soc/fsl/imx-audmux.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/sound/soc/fsl/imx-audmux.c b/sound/soc/fsl/imx-audmux.c index 50b71e5d4589..582f1e2431ee 100644 --- a/sound/soc/fsl/imx-audmux.c +++ b/sound/soc/fsl/imx-audmux.c @@ -75,8 +75,7 @@ static ssize_t audmux_read_file(struct file *file, char __user *user_buf, if (!buf) return -ENOMEM; - ret = scnprintf(buf, PAGE_SIZE, "PDCR: %08x\nPTCR: %08x\n", - pdcr, ptcr); + ret = sysfs_emit(buf, "PDCR: %08x\nPTCR: %08x\n", pdcr, ptcr); if (ptcr & IMX_AUDMUX_V2_PTCR_TFSDIR) ret += scnprintf(buf + ret, PAGE_SIZE - ret, From 38eef3be38ab895959c442702864212cc3beb96c Mon Sep 17 00:00:00 2001 From: Wang Yufen Date: Mon, 5 Dec 2022 17:56:28 +0800 Subject: [PATCH 041/115] ASoC: mediatek: mt8183: fix refcount leak in mt8183_mt6358_ts3a227_max98357_dev_probe() The node returned by of_parse_phandle() with refcount incremented, of_node_put() needs be called when finish using it. So add it in the error path in mt8183_mt6358_ts3a227_max98357_dev_probe(). Fixes: 11c0269017b2 ("ASoC: Mediatek: MT8183: Add machine driver with TS3A227") Signed-off-by: Wang Yufen Link: https://lore.kernel.org/r/1670234188-23596-1-git-send-email-wangyufen@huawei.com Signed-off-by: Mark Brown --- .../mt8183/mt8183-mt6358-ts3a227-max98357.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/sound/soc/mediatek/mt8183/mt8183-mt6358-ts3a227-max98357.c b/sound/soc/mediatek/mt8183/mt8183-mt6358-ts3a227-max98357.c index 8fb473543cf9..ce9aedde7e1e 100644 --- a/sound/soc/mediatek/mt8183/mt8183-mt6358-ts3a227-max98357.c +++ b/sound/soc/mediatek/mt8183/mt8183-mt6358-ts3a227-max98357.c @@ -677,8 +677,10 @@ mt8183_mt6358_ts3a227_max98357_dev_probe(struct platform_device *pdev) } card = (struct snd_soc_card *)of_device_get_match_data(&pdev->dev); - if (!card) + if (!card) { + of_node_put(platform_node); return -EINVAL; + } card->dev = &pdev->dev; ec_codec = of_parse_phandle(pdev->dev.of_node, "mediatek,ec-codec", 0); @@ -767,8 +769,10 @@ mt8183_mt6358_ts3a227_max98357_dev_probe(struct platform_device *pdev) } priv = devm_kzalloc(&pdev->dev, sizeof(*priv), GFP_KERNEL); - if (!priv) - return -ENOMEM; + if (!priv) { + ret = -ENOMEM; + goto out; + } snd_soc_card_set_drvdata(card, priv); @@ -776,7 +780,8 @@ mt8183_mt6358_ts3a227_max98357_dev_probe(struct platform_device *pdev) if (IS_ERR(priv->pinctrl)) { dev_err(&pdev->dev, "%s devm_pinctrl_get failed\n", __func__); - return PTR_ERR(priv->pinctrl); + ret = PTR_ERR(priv->pinctrl); + goto out; } for (i = 0; i < PIN_STATE_MAX; i++) { @@ -809,6 +814,7 @@ mt8183_mt6358_ts3a227_max98357_dev_probe(struct platform_device *pdev) ret = devm_snd_soc_register_card(&pdev->dev, card); +out: of_node_put(platform_node); of_node_put(ec_codec); of_node_put(hdmi_codec); From 81ed7d9de18768fe0cb3d74a7a163a8c082e1346 Mon Sep 17 00:00:00 2001 From: Jiapeng Chong Date: Mon, 5 Dec 2022 15:35:07 +0800 Subject: [PATCH 042/115] ASoC: codecs: wcd-clsh: Remove the unused function The function wcd_clsh_set_buck_mode() is defined in the wcd-clsh-v2.c file, but not called elsewhere, so remove this unused function. sound/soc/codecs/wcd-clsh-v2.c:133:20: warning: unused function 'wcd_clsh_enable_status'. Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3348 Reported-by: Abaci Robot Signed-off-by: Jiapeng Chong Link: https://lore.kernel.org/r/20221205073507.36071-1-jiapeng.chong@linux.alibaba.com Signed-off-by: Mark Brown --- sound/soc/codecs/wcd-clsh-v2.c | 6 ------ 1 file changed, 6 deletions(-) diff --git a/sound/soc/codecs/wcd-clsh-v2.c b/sound/soc/codecs/wcd-clsh-v2.c index 4c7ebc7fb400..a75db27e5205 100644 --- a/sound/soc/codecs/wcd-clsh-v2.c +++ b/sound/soc/codecs/wcd-clsh-v2.c @@ -130,12 +130,6 @@ static inline void wcd_enable_clsh_block(struct wcd_clsh_ctrl *ctrl, ctrl->clsh_users = 0; } -static inline bool wcd_clsh_enable_status(struct snd_soc_component *comp) -{ - return snd_soc_component_read(comp, WCD9XXX_A_CDC_CLSH_CRC) & - WCD9XXX_A_CDC_CLSH_CRC_CLK_EN_MASK; -} - static inline void wcd_clsh_set_buck_mode(struct snd_soc_component *comp, int mode) { From 1da681e52853f0abfbfff8c69833d31e538ff9c0 Mon Sep 17 00:00:00 2001 From: Chancel Liu Date: Mon, 28 Nov 2022 14:09:50 +0800 Subject: [PATCH 043/115] ASoC: soc-pcm.c: Clear DAIs parameters after stream_active is updated DAIs parameters should be cleared if there's no active stream. Before, we implemented it in soc_pcm_hw_free() by detecting stream_active. If the running stream is the last active stream, we're going to clear parameters. However it will cause DAIs parameters never be cleared if there're more than one stream. For example, we have stream1 and stream2 about to stop. stream2 executes soc_pcm_hw_free() before stream1 executes soc_pcm_close(). At the moment, stream2 should clear DAIs parameters. Since stream_active is not yet updated by stream1 in soc_pcm_close(), stream2 will not clear DAIs parameters. In result both stream1 and stream2 don't clear the parameters. This patch moves DAIs parameters cleanup after stream_active is updated. Signed-off-by: Chancel Liu Link: https://lore.kernel.org/r/20221128060950.3540845-1-chancel.liu@nxp.com Signed-off-by: Mark Brown --- sound/soc/soc-pcm.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) diff --git a/sound/soc/soc-pcm.c b/sound/soc/soc-pcm.c index a7810c78ffa1..579a44d81d9a 100644 --- a/sound/soc/soc-pcm.c +++ b/sound/soc/soc-pcm.c @@ -709,8 +709,17 @@ static int soc_pcm_clean(struct snd_soc_pcm_runtime *rtd, snd_soc_dpcm_mutex_assert_held(rtd); - if (!rollback) + if (!rollback) { snd_soc_runtime_deactivate(rtd, substream->stream); + /* clear the corresponding DAIs parameters when going to be inactive */ + for_each_rtd_dais(rtd, i, dai) { + if (snd_soc_dai_active(dai) == 0) + soc_pcm_set_dai_params(dai, NULL); + + if (snd_soc_dai_stream_active(dai, substream->stream) == 0) + snd_soc_dai_digital_mute(dai, 1, substream->stream); + } + } for_each_rtd_dais(rtd, i, dai) snd_soc_dai_shutdown(dai, substream, rollback); @@ -940,15 +949,6 @@ static int soc_pcm_hw_clean(struct snd_soc_pcm_runtime *rtd, snd_soc_dpcm_mutex_assert_held(rtd); - /* clear the corresponding DAIs parameters when going to be inactive */ - for_each_rtd_dais(rtd, i, dai) { - if (snd_soc_dai_active(dai) == 1) - soc_pcm_set_dai_params(dai, NULL); - - if (snd_soc_dai_stream_active(dai, substream->stream) == 1) - snd_soc_dai_digital_mute(dai, 1, substream->stream); - } - /* run the stream event */ snd_soc_dapm_stream_stop(rtd, substream->stream); From e85b1f5a9769ac30f4d2f6fb1cdcd9570c38e0c1 Mon Sep 17 00:00:00 2001 From: Marek Vasut Date: Thu, 8 Dec 2022 04:53:54 +0100 Subject: [PATCH 044/115] ASoC: dt-bindings: fsl-sai: Reinstate i.MX93 SAI compatible string The ASoC: dt-bindings: fsl-sai: Fix mx6ul and mx7d compatible strings dropped i.MX93 SAI compatible string, reinstate it. Fixes: 81b6c043e7ba ("ASoC: dt-bindings: fsl-sai: Fix mx6ul and mx7d compatible strings") Signed-off-by: Marek Vasut Link: https://lore.kernel.org/r/20221208035354.255438-1-marex@denx.de Signed-off-by: Mark Brown --- Documentation/devicetree/bindings/sound/fsl,sai.yaml | 1 + 1 file changed, 1 insertion(+) diff --git a/Documentation/devicetree/bindings/sound/fsl,sai.yaml b/Documentation/devicetree/bindings/sound/fsl,sai.yaml index 5b28d2d51327..7e56337d8edc 100644 --- a/Documentation/devicetree/bindings/sound/fsl,sai.yaml +++ b/Documentation/devicetree/bindings/sound/fsl,sai.yaml @@ -38,6 +38,7 @@ properties: - fsl,imx8mq-sai - fsl,imx8qm-sai - fsl,imx8ulp-sai + - fsl,imx93-sai - fsl,vf610-sai reg: From 084ca216931ab9313e6fb862c2ec9ec5e0702cd5 Mon Sep 17 00:00:00 2001 From: Cezary Rojewski Date: Thu, 8 Dec 2022 15:26:35 +0100 Subject: [PATCH 045/115] ALSA: hda: Error out if invalid stream is being setup Scenario when snd_hdac_stream_setup_periods() receives an instance of struct hdac_stream with neither ->substream nor ->cstream initialized is invalid. Simultaneously addresses "uninitialized symbol 'dmab'" error reported by Smatch. Fixes: 3e9582267e3a ("ALSA: hda: Interrupt servicing and BDL setup for compress streams") Signed-off-by: Cezary Rojewski Link: https://lore.kernel.org/r/20221208142635.1514944-1-cezary.rojewski@intel.com Signed-off-by: Takashi Iwai --- sound/hda/hdac_stream.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c index 8f625402505f..547adbc22590 100644 --- a/sound/hda/hdac_stream.c +++ b/sound/hda/hdac_stream.c @@ -500,6 +500,9 @@ int snd_hdac_stream_setup_periods(struct hdac_stream *azx_dev) dmab = snd_pcm_get_dma_buf(substream); } else if (cstream) { dmab = snd_pcm_get_dma_buf(cstream); + } else { + WARN(1, "No substream or cstream assigned\n"); + return -EINVAL; } /* reset BDL address */ From ada261b690ecd5c2f55f0c51bdf11d852a4561a6 Mon Sep 17 00:00:00 2001 From: Kai Vehmanen Date: Fri, 9 Dec 2022 12:18:20 +0200 Subject: [PATCH 046/115] ALSA: hda/hdmi: fix i915 silent stream programming flow The i915 display codec may not successfully transition to normal audio streaming mode, if the stream id is programmed while codec is actively transmitting data. This can happen when silent stream is enabled in KAE mode. Fix the issue by implementing a i915 specific programming flow, where the silent streaming is temporarily stopped, a small delay is applied to ensure display codec becomes idle, and then proceed with reprogramming the stream ID. Fixes: 15175a4f2bbb ("ALSA: hda/hdmi: add keep-alive support for ADL-P and DG2") Link: https://gitlab.freedesktop.org/drm/intel/-/issues/7353 Signed-off-by: Kai Vehmanen Reviewed-by: Pierre-Louis Bossart Tested-by: Rodrigo Vivi Link: https://lore.kernel.org/r/20221209101822.3893675-2-kai.vehmanen@linux.intel.com Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_hdmi.c | 28 ++++++++++++++++++++++++++-- 1 file changed, 26 insertions(+), 2 deletions(-) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 7a40ddfd695a..48bb23745ed3 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -2879,9 +2879,33 @@ static int i915_hsw_setup_stream(struct hda_codec *codec, hda_nid_t cvt_nid, hda_nid_t pin_nid, int dev_id, u32 stream_tag, int format) { + struct hdmi_spec *spec = codec->spec; + int pin_idx = pin_id_to_pin_index(codec, pin_nid, dev_id); + struct hdmi_spec_per_pin *per_pin; + int res; + + if (pin_idx < 0) + per_pin = NULL; + else + per_pin = get_pin(spec, pin_idx); + haswell_verify_D0(codec, cvt_nid, pin_nid); - return hdmi_setup_stream(codec, cvt_nid, pin_nid, dev_id, - stream_tag, format); + + if (spec->silent_stream_type == SILENT_STREAM_KAE && per_pin && per_pin->silent_stream) { + silent_stream_set_kae(codec, per_pin, false); + /* wait for pending transfers in codec to clear */ + usleep_range(100, 200); + } + + res = hdmi_setup_stream(codec, cvt_nid, pin_nid, dev_id, + stream_tag, format); + + if (spec->silent_stream_type == SILENT_STREAM_KAE && per_pin && per_pin->silent_stream) { + usleep_range(100, 200); + silent_stream_set_kae(codec, per_pin, true); + } + + return res; } /* pin_cvt_fixup ops override for HSW+ and VLV+ */ From b17e7ea041d8b565063632501ca4597afd105102 Mon Sep 17 00:00:00 2001 From: Kai Vehmanen Date: Fri, 9 Dec 2022 12:18:21 +0200 Subject: [PATCH 047/115] ALSA: hda/hdmi: set default audio parameters for KAE silent-stream If the stream-id is zero, the keep-alive (KAE) will only ensure clock is generated, but no audio samples are sent over display link. This happens before first real audio stream is played out to a newly connected receiver. Reuse the code in silent_stream_enable() to set up stream parameters to sane defaults values, also when using the newer keep-alive flow. Fixes: 15175a4f2bbb ("ALSA: hda/hdmi: add keep-alive support for ADL-P and DG2") Signed-off-by: Kai Vehmanen Reviewed-by: Pierre-Louis Bossart Tested-by: Rodrigo Vivi Link: https://lore.kernel.org/r/20221209101822.3893675-3-kai.vehmanen@linux.intel.com Signed-off-by: Takashi Iwai --- sound/pci/hda/patch_hdmi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 48bb23745ed3..f8e6ff7f8820 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -1738,6 +1738,7 @@ static void silent_stream_enable(struct hda_codec *codec, switch (spec->silent_stream_type) { case SILENT_STREAM_KAE: + silent_stream_enable_i915(codec, per_pin); silent_stream_set_kae(codec, per_pin, true); break; case SILENT_STREAM_I915: From ee0b089d660021792e4ab4dda191b097ce1e964f Mon Sep 17 00:00:00 2001 From: Kai Vehmanen Date: Fri, 9 Dec 2022 12:18:22 +0200 Subject: [PATCH 048/115] ALSA: hda/hdmi: fix stream-id config keep-alive for rt suspend When the new style KAE keep-alive implementation is used on compatible Intel hardware, the clocks are maintained when codec is in D3. The generic code in hda_cleanup_all_streams() can however interfere with generation of audio samples in this mode, by setting the stream and channel ids to zero. To get full benefit of the keepalive, set the new no_stream_clean_at_suspend quirk bit on affected Intel hardware. When this bit is set, stream cleanup is skipped in hda_call_codec_suspend(). Special handling is needed for the case when system goes to suspend. The stream id programming can be lost in this case. This will also cause codec->cvt_setups to be out of sync. Handle this by implementing custom suspend/resume handlers. If keep-alive is active for any converter, set the quirk flags no_stream_clean_at_suspend and forced_resume. Upon resume, keepalive programming is restored if needed. Fixes: 15175a4f2bbb ("ALSA: hda/hdmi: add keep-alive support for ADL-P and DG2") Signed-off-by: Kai Vehmanen Reviewed-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221209101822.3893675-4-kai.vehmanen@linux.intel.com Signed-off-by: Takashi Iwai --- include/sound/hda_codec.h | 1 + sound/pci/hda/hda_codec.c | 3 +- sound/pci/hda/patch_hdmi.c | 90 +++++++++++++++++++++++++++++++++++++- 3 files changed, 92 insertions(+), 2 deletions(-) diff --git a/include/sound/hda_codec.h b/include/sound/hda_codec.h index 25ec8c181688..eba23daf2c29 100644 --- a/include/sound/hda_codec.h +++ b/include/sound/hda_codec.h @@ -258,6 +258,7 @@ struct hda_codec { unsigned int link_down_at_suspend:1; /* link down at runtime suspend */ unsigned int relaxed_resume:1; /* don't resume forcibly for jack */ unsigned int forced_resume:1; /* forced resume for jack */ + unsigned int no_stream_clean_at_suspend:1; /* do not clean streams at suspend */ #ifdef CONFIG_PM unsigned long power_on_acct; diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index b4d1e658c556..edd653ece70d 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -2886,7 +2886,8 @@ static unsigned int hda_call_codec_suspend(struct hda_codec *codec) snd_hdac_enter_pm(&codec->core); if (codec->patch_ops.suspend) codec->patch_ops.suspend(codec); - hda_cleanup_all_streams(codec); + if (!codec->no_stream_clean_at_suspend) + hda_cleanup_all_streams(codec); state = hda_set_power_state(codec, AC_PWRST_D3); update_power_acct(codec, true); snd_hdac_leave_pm(&codec->core); diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index f8e6ff7f8820..8015e4471267 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -2926,6 +2926,88 @@ static void i915_pin_cvt_fixup(struct hda_codec *codec, } } +#ifdef CONFIG_PM +static int i915_adlp_hdmi_suspend(struct hda_codec *codec) +{ + struct hdmi_spec *spec = codec->spec; + bool silent_streams = false; + int pin_idx, res; + + res = generic_hdmi_suspend(codec); + + for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + struct hdmi_spec_per_pin *per_pin = get_pin(spec, pin_idx); + + if (per_pin->silent_stream) { + silent_streams = true; + break; + } + } + + if (silent_streams && spec->silent_stream_type == SILENT_STREAM_KAE) { + /* + * stream-id should remain programmed when codec goes + * to runtime suspend + */ + codec->no_stream_clean_at_suspend = 1; + + /* + * the system might go to S3, in which case keep-alive + * must be reprogrammed upon resume + */ + codec->forced_resume = 1; + + codec_dbg(codec, "HDMI: KAE active at suspend\n"); + } else { + codec->no_stream_clean_at_suspend = 0; + codec->forced_resume = 0; + } + + return res; +} + +static int i915_adlp_hdmi_resume(struct hda_codec *codec) +{ + struct hdmi_spec *spec = codec->spec; + int pin_idx, res; + + res = generic_hdmi_resume(codec); + + /* KAE not programmed at suspend, nothing to do here */ + if (!codec->no_stream_clean_at_suspend) + return res; + + for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { + struct hdmi_spec_per_pin *per_pin = get_pin(spec, pin_idx); + + /* + * If system was in suspend with monitor connected, + * the codec setting may have been lost. Re-enable + * keep-alive. + */ + if (per_pin->silent_stream) { + unsigned int param; + + param = snd_hda_codec_read(codec, per_pin->cvt_nid, 0, + AC_VERB_GET_CONV, 0); + if (!param) { + codec_dbg(codec, "HDMI: KAE: restore stream id\n"); + silent_stream_enable_i915(codec, per_pin); + } + + param = snd_hda_codec_read(codec, per_pin->cvt_nid, 0, + AC_VERB_GET_DIGI_CONVERT_1, 0); + if (!(param & (AC_DIG3_KAE << 16))) { + codec_dbg(codec, "HDMI: KAE: restore DIG3_KAE\n"); + silent_stream_set_kae(codec, per_pin, true); + } + } + } + + return res; +} +#endif + /* precondition and allocation for Intel codecs */ static int alloc_intel_hdmi(struct hda_codec *codec) { @@ -3056,8 +3138,14 @@ static int patch_i915_adlp_hdmi(struct hda_codec *codec) if (!res) { spec = codec->spec; - if (spec->silent_stream_type) + if (spec->silent_stream_type) { spec->silent_stream_type = SILENT_STREAM_KAE; + +#ifdef CONFIG_PM + codec->patch_ops.resume = i915_adlp_hdmi_resume; + codec->patch_ops.suspend = i915_adlp_hdmi_suspend; +#endif + } } return res; From 2c1da39008fee00596ed33baeacaffa0dc62df25 Mon Sep 17 00:00:00 2001 From: Matt Roper Date: Mon, 28 Nov 2022 15:30:10 -0800 Subject: [PATCH 049/115] drm/i915/gt: Correct kerneldoc for intel_gt_mcr_wait_for_reg() The kerneldoc function name was not updated when this function was converted to a non-fw form. Fixes: 41f425adbce9 ("drm/i915/gt: Manage uncore->lock while waiting on MCR register") Reported-by: kernel test robot Signed-off-by: Matt Roper Reviewed-by: Balasubramani Vivekanandan Link: https://patchwork.freedesktop.org/patch/msgid/20221128233014.4000136-2-matthew.d.roper@intel.com (cherry picked from commit 03b713d029bd17a1ed426590609af79843db95e2) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_gt_mcr.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/intel_gt_mcr.c b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c index d9a8ff9e5e57..ea86c1ab5dc5 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_mcr.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c @@ -702,7 +702,7 @@ void intel_gt_mcr_get_ss_steering(struct intel_gt *gt, unsigned int dss, } /** - * intel_gt_mcr_wait_for_reg_fw - wait until MCR register matches expected state + * intel_gt_mcr_wait_for_reg - wait until MCR register matches expected state * @gt: GT structure * @reg: the register to read * @mask: mask to apply to register value From 449a0ef584d42ed24b7432c899863eaabe2583b5 Mon Sep 17 00:00:00 2001 From: Miaoqian Lin Date: Wed, 7 Dec 2022 15:29:09 +0400 Subject: [PATCH 050/115] drm/i915: Fix documentation for intel_uncore_forcewake_put__locked intel_uncore_forcewake_put__locked() is used to release a reference. Fixes: a6111f7b6604 ("drm/i915: Reduce locking in execlist command submission") Signed-off-by: Miaoqian Lin Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20221207112909.2655251-1-linmq006@gmail.com (cherry picked from commit 955f4d7176eb154db587ae162ec2b392dc8d5f27) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/intel_uncore.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 8006a6c61466..614013745fca 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -824,9 +824,9 @@ void intel_uncore_forcewake_flush(struct intel_uncore *uncore, } /** - * intel_uncore_forcewake_put__locked - grab forcewake domain references + * intel_uncore_forcewake_put__locked - release forcewake domain references * @uncore: the intel_uncore structure - * @fw_domains: forcewake domains to get reference on + * @fw_domains: forcewake domains to put references * * See intel_uncore_forcewake_put(). This variant places the onus * on the caller to explicitly handle the dev_priv->uncore.lock spinlock. From d4d4c6fbae3837623708594a7499f40673fb0692 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Wed, 23 Nov 2022 15:53:42 -0800 Subject: [PATCH 051/115] drm/i915/perf: Do not parse context image for HSW An earlier commit introduced a mechanism to parse the context image to find the OA context control offset. This resulted in an NPD on haswell when gem_context was passed into i915_perf_open_ioctl params. Haswell does not support logical ring contexts, so ensure that the context image is parsed only for platforms with logical ring contexts and also validate lrc_reg_state. v2: Fix build failure v3: Fix checkpatch error Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7432 Fixes: a5c3a3cbf029 ("drm/i915/perf: Determine gen12 oa ctx offset at runtime") Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Ashutosh Dixit Signed-off-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20221123235342.713068-1-umesh.nerlige.ramappa@intel.com (cherry picked from commit 95c713d722017b26e301303713d638e0b95b1f68) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/i915_perf.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 00e09bb18b13..125b6ca25a75 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1383,6 +1383,9 @@ static u32 oa_context_image_offset(struct intel_context *ce, u32 reg) u32 offset, len = (ce->engine->context_size - PAGE_SIZE) / 4; u32 *state = ce->lrc_reg_state; + if (drm_WARN_ON(&ce->engine->i915->drm, !state)) + return U32_MAX; + for (offset = 0; offset < len; ) { if (IS_MI_LRI_CMD(state[offset])) { /* @@ -1447,7 +1450,8 @@ static int oa_get_render_ctx_id(struct i915_perf_stream *stream) if (IS_ERR(ce)) return PTR_ERR(ce); - if (engine_supports_mi_query(stream->engine)) { + if (engine_supports_mi_query(stream->engine) && + HAS_LOGICAL_RING_CONTEXTS(stream->perf->i915)) { /* * We are enabling perf query here. If we don't find the context * offset here, just return an error. From 2165359b7ed4e0b93fc23f49ede38d76e91fffe1 Mon Sep 17 00:00:00 2001 From: Colin Ian King Date: Thu, 8 Dec 2022 12:18:54 +0000 Subject: [PATCH 052/115] drm/amd/display: Fix spelling mistake: "dram_clk_chanage" -> "dram_clk_change" There is a spelling mistake in the struct field dram_clk_chanage. Fix it. Signed-off-by: Colin Ian King Signed-off-by: Hamza Mahfooz Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c | 8 ++++---- drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 2 +- .../drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c | 4 ++-- drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c | 8 ++++---- drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubbub.c | 8 ++++---- drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c | 8 ++++---- drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h | 2 +- 7 files changed, 20 insertions(+), 20 deletions(-) diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c index 0f746bb4e500..d51f1ce02874 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hubbub.c @@ -55,7 +55,7 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A); s = &wm->sets[1]; s->wm_set = 1; @@ -65,7 +65,7 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B); s = &wm->sets[2]; s->wm_set = 2; @@ -75,7 +75,7 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C); s = &wm->sets[3]; s->wm_set = 3; @@ -85,7 +85,7 @@ void hubbub1_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); } void hubbub1_allow_self_refresh_control(struct hubbub *hubbub, bool allow) diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c index c8ec11839b4d..fe2023f18b7d 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c @@ -159,7 +159,7 @@ static void dcn10_log_hubbub_state(struct dc *dc, DTN_INFO_MICRO_SEC(s->pte_meta_urgent); DTN_INFO_MICRO_SEC(s->sr_enter); DTN_INFO_MICRO_SEC(s->sr_exit); - DTN_INFO_MICRO_SEC(s->dram_clk_chanage); + DTN_INFO_MICRO_SEC(s->dram_clk_change); DTN_INFO("\n"); } diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c index e8b6065fffad..a0f8e31d2adc 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c +++ b/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer_debug.c @@ -83,7 +83,7 @@ static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned i memset(&wm, 0, sizeof(struct dcn_hubbub_wm)); dc->res_pool->hubbub->funcs->wm_read_state(dc->res_pool->hubbub, &wm); - chars_printed = snprintf_count(pBuf, remaining_buffer, "wm_set_index,data_urgent,pte_meta_urgent,sr_enter,sr_exit,dram_clk_chanage\n"); + chars_printed = snprintf_count(pBuf, remaining_buffer, "wm_set_index,data_urgent,pte_meta_urgent,sr_enter,sr_exit,dram_clk_change\n"); remaining_buffer -= chars_printed; pBuf += chars_printed; @@ -98,7 +98,7 @@ static unsigned int dcn10_get_hubbub_state(struct dc *dc, char *pBuf, unsigned i (s->pte_meta_urgent * frac) / ref_clk_mhz / frac, (s->pte_meta_urgent * frac) / ref_clk_mhz % frac, (s->sr_enter * frac) / ref_clk_mhz / frac, (s->sr_enter * frac) / ref_clk_mhz % frac, (s->sr_exit * frac) / ref_clk_mhz / frac, (s->sr_exit * frac) / ref_clk_mhz % frac, - (s->dram_clk_chanage * frac) / ref_clk_mhz / frac, (s->dram_clk_chanage * frac) / ref_clk_mhz % frac); + (s->dram_clk_change * frac) / ref_clk_mhz / frac, (s->dram_clk_change * frac) / ref_clk_mhz % frac); remaining_buffer -= chars_printed; pBuf += chars_printed; } diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c index aacb1fb5c73e..24bd93219936 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubbub.c @@ -500,7 +500,7 @@ void hubbub2_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_A); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A); s = &wm->sets[1]; s->wm_set = 1; @@ -511,7 +511,7 @@ void hubbub2_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_B); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B); s = &wm->sets[2]; s->wm_set = 2; @@ -522,7 +522,7 @@ void hubbub2_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_C); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C); s = &wm->sets[3]; s->wm_set = 3; @@ -533,7 +533,7 @@ void hubbub2_wm_read_state(struct hubbub *hubbub, s->sr_enter = REG_READ(DCHUBBUB_ARB_ALLOW_SR_ENTER_WATERMARK_D); s->sr_exit = REG_READ(DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D); } - s->dram_clk_chanage = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); + s->dram_clk_change = REG_READ(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D); } void hubbub2_get_dchub_ref_freq(struct hubbub *hubbub, diff --git a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubbub.c index c5e200d09038..aeb0e0d9b70a 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn21/dcn21_hubbub.c @@ -635,7 +635,7 @@ void hubbub21_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A, &s->sr_exit); REG_GET(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A, - DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A, &s->dram_clk_chanage); + DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_A, &s->dram_clk_change); s = &wm->sets[1]; s->wm_set = 1; @@ -649,7 +649,7 @@ void hubbub21_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B, &s->sr_exit); REG_GET(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B, - DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B, &s->dram_clk_chanage); + DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_B, &s->dram_clk_change); s = &wm->sets[2]; s->wm_set = 2; @@ -663,7 +663,7 @@ void hubbub21_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C, &s->sr_exit); REG_GET(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C, - DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C, &s->dram_clk_chanage); + DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_C, &s->dram_clk_change); s = &wm->sets[3]; s->wm_set = 3; @@ -677,7 +677,7 @@ void hubbub21_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D, &s->sr_exit); REG_GET(DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D, - DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D, &s->dram_clk_chanage); + DCHUBBUB_ARB_ALLOW_DRAM_CLK_CHANGE_WATERMARK_D, &s->dram_clk_change); } static void hubbub21_apply_DEDCN21_147_wa(struct hubbub *hubbub) diff --git a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c index 5947c2cb0f30..9501403a48a9 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c +++ b/drivers/gpu/drm/amd/display/dc/dcn32/dcn32_hubbub.c @@ -865,7 +865,7 @@ static void hubbub32_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_A, &s->sr_exit); REG_GET(DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_A, - DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_A, &s->dram_clk_chanage); + DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_A, &s->dram_clk_change); REG_GET(DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_A, DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_A, &s->usr_retrain); @@ -885,7 +885,7 @@ static void hubbub32_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_B, &s->sr_exit); REG_GET(DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_B, - DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_B, &s->dram_clk_chanage); + DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_B, &s->dram_clk_change); REG_GET(DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_B, DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_B, &s->usr_retrain); @@ -905,7 +905,7 @@ static void hubbub32_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_C, &s->sr_exit); REG_GET(DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_C, - DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_C, &s->dram_clk_chanage); + DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_C, &s->dram_clk_change); REG_GET(DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_C, DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_C, &s->usr_retrain); @@ -925,7 +925,7 @@ static void hubbub32_wm_read_state(struct hubbub *hubbub, DCHUBBUB_ARB_ALLOW_SR_EXIT_WATERMARK_D, &s->sr_exit); REG_GET(DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_D, - DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_D, &s->dram_clk_chanage); + DCHUBBUB_ARB_UCLK_PSTATE_CHANGE_WATERMARK_D, &s->dram_clk_change); REG_GET(DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_D, DCHUBBUB_ARB_USR_RETRAINING_WATERMARK_D, &s->usr_retrain); diff --git a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h index f2e1fcb668fb..5b0265c0df61 100644 --- a/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h +++ b/drivers/gpu/drm/amd/display/dc/inc/hw/dchubbub.h @@ -46,7 +46,7 @@ struct dcn_hubbub_wm_set { uint32_t pte_meta_urgent; uint32_t sr_enter; uint32_t sr_exit; - uint32_t dram_clk_chanage; + uint32_t dram_clk_change; uint32_t usr_retrain; uint32_t fclk_pstate_change; }; From 81d0bcf9900932633d270d5bc4a54ff599c6ebdb Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Wed, 7 Dec 2022 11:08:53 -0500 Subject: [PATCH 053/115] drm/amdgpu: make display pinning more flexible (v2) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Only apply the static threshold for Stoney and Carrizo. This hardware has certain requirements that don't allow mixing of GTT and VRAM. Newer asics do not have these requirements so we should be able to be more flexible with where buffers end up. Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2270 Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2291 Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2255 Acked-by: Luben Tuikov Reviewed-by: Christian König Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 919bbea2e3ac..2df55cc7e07f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -1506,7 +1506,8 @@ u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo) uint32_t amdgpu_bo_get_preferred_domain(struct amdgpu_device *adev, uint32_t domain) { - if (domain == (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) { + if ((domain == (AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT)) && + ((adev->asic_type == CHIP_CARRIZO) || (adev->asic_type == CHIP_STONEY))) { domain = AMDGPU_GEM_DOMAIN_VRAM; if (adev->gmc.real_vram_size <= AMDGPU_SG_THRESHOLD) domain = AMDGPU_GEM_DOMAIN_GTT; From 1d4624cd72b912b2680c08d0be48338a1629a858 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Mon, 21 Nov 2022 15:52:19 -0500 Subject: [PATCH 054/115] drm/amdgpu: handle polaris10/11 overlap asics (v2) Some special polaris 10 chips overlap with the polaris11 DID range. Handle this properly in the driver. v2: use local flags for other function calls. Acked-by: Luben Tuikov Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 13 +++++++++++-- 1 file changed, 11 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 7383272c6a3a..b4f2d61ea0d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -2039,6 +2039,15 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, "See modparam exp_hw_support\n"); return -ENODEV; } + /* differentiate between P10 and P11 asics with the same DID */ + if (pdev->device == 0x67FF && + (pdev->revision == 0xE3 || + pdev->revision == 0xE7 || + pdev->revision == 0xF3 || + pdev->revision == 0xF7)) { + flags &= ~AMD_ASIC_MASK; + flags |= CHIP_POLARIS10; + } /* Due to hardware bugs, S/G Display on raven requires a 1:1 IOMMU mapping, * however, SME requires an indirect IOMMU mapping because the encryption @@ -2108,12 +2117,12 @@ static int amdgpu_pci_probe(struct pci_dev *pdev, pci_set_drvdata(pdev, ddev); - ret = amdgpu_driver_load_kms(adev, ent->driver_data); + ret = amdgpu_driver_load_kms(adev, flags); if (ret) goto err_pci; retry_init: - ret = drm_dev_register(ddev, ent->driver_data); + ret = drm_dev_register(ddev, flags); if (ret == -EAGAIN && ++retry <= 3) { DRM_INFO("retry init %d\n", retry); /* Don't request EX mode too frequently which is attacking */ From 2aa2a5ead0ee0a358bf80a2984a641d1bf2adc2a Mon Sep 17 00:00:00 2001 From: Kai Vehmanen Date: Fri, 9 Dec 2022 13:45:28 +0200 Subject: [PATCH 055/115] ASoC: SOF: Intel: pci-tgl: unblock S5 entry if DMA stop has failed" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit If system shutdown has not been completed cleanly, it is possible the DMA stream shutdown has not been done, or was not clean. If this is the case, Intel TGL/ADL HDA platforms may fail to shutdown cleanly due to pending HDA DMA transactions. To avoid this, detect this scenario in the shutdown callback, and perform an additional controller reset. This has been tested to unblock S5 entry if this condition is hit. Co-developed-by: Archana Patni Signed-off-by: Archana Patni Signed-off-by: Kai Vehmanen Reviewed-by: Pierre-Louis Bossart Reviewed-by: Péter Ujfalusi Reviewed-by: Ranjani Sridharan Link: https://lore.kernel.org/r/20221209114529.3909192-2-kai.vehmanen@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/sof/intel/hda-dsp.c | 72 +++++++++++++++++++++++++++++++++++ sound/soc/sof/intel/hda.h | 1 + sound/soc/sof/intel/tgl.c | 2 +- 3 files changed, 74 insertions(+), 1 deletion(-) diff --git a/sound/soc/sof/intel/hda-dsp.c b/sound/soc/sof/intel/hda-dsp.c index 5fa29df54b42..b4eacae8564c 100644 --- a/sound/soc/sof/intel/hda-dsp.c +++ b/sound/soc/sof/intel/hda-dsp.c @@ -878,6 +878,78 @@ int hda_dsp_suspend(struct snd_sof_dev *sdev, u32 target_state) return snd_sof_dsp_set_power_state(sdev, &target_dsp_state); } +static unsigned int hda_dsp_check_for_dma_streams(struct snd_sof_dev *sdev) +{ + struct hdac_bus *bus = sof_to_bus(sdev); + struct hdac_stream *s; + unsigned int active_streams = 0; + int sd_offset; + u32 val; + + list_for_each_entry(s, &bus->stream_list, list) { + sd_offset = SOF_STREAM_SD_OFFSET(s); + val = snd_sof_dsp_read(sdev, HDA_DSP_HDA_BAR, + sd_offset); + if (val & SOF_HDA_SD_CTL_DMA_START) + active_streams |= BIT(s->index); + } + + return active_streams; +} + +static int hda_dsp_s5_quirk(struct snd_sof_dev *sdev) +{ + int ret; + + /* + * Do not assume a certain timing between the prior + * suspend flow, and running of this quirk function. + * This is needed if the controller was just put + * to reset before calling this function. + */ + usleep_range(500, 1000); + + /* + * Take controller out of reset to flush DMA + * transactions. + */ + ret = hda_dsp_ctrl_link_reset(sdev, false); + if (ret < 0) + return ret; + + usleep_range(500, 1000); + + /* Restore state for shutdown, back to reset */ + ret = hda_dsp_ctrl_link_reset(sdev, true); + if (ret < 0) + return ret; + + return ret; +} + +int hda_dsp_shutdown_dma_flush(struct snd_sof_dev *sdev) +{ + unsigned int active_streams; + int ret, ret2; + + /* check if DMA cleanup has been successful */ + active_streams = hda_dsp_check_for_dma_streams(sdev); + + sdev->system_suspend_target = SOF_SUSPEND_S3; + ret = snd_sof_suspend(sdev->dev); + + if (active_streams) { + dev_warn(sdev->dev, + "There were active DSP streams (%#x) at shutdown, trying to recover\n", + active_streams); + ret2 = hda_dsp_s5_quirk(sdev); + if (ret2 < 0) + dev_err(sdev->dev, "shutdown recovery failed (%d)\n", ret2); + } + + return ret; +} + int hda_dsp_shutdown(struct snd_sof_dev *sdev) { sdev->system_suspend_target = SOF_SUSPEND_S3; diff --git a/sound/soc/sof/intel/hda.h b/sound/soc/sof/intel/hda.h index 022ce80968dd..caccaf8fba9c 100644 --- a/sound/soc/sof/intel/hda.h +++ b/sound/soc/sof/intel/hda.h @@ -592,6 +592,7 @@ int hda_dsp_resume(struct snd_sof_dev *sdev); int hda_dsp_runtime_suspend(struct snd_sof_dev *sdev); int hda_dsp_runtime_resume(struct snd_sof_dev *sdev); int hda_dsp_runtime_idle(struct snd_sof_dev *sdev); +int hda_dsp_shutdown_dma_flush(struct snd_sof_dev *sdev); int hda_dsp_shutdown(struct snd_sof_dev *sdev); int hda_dsp_set_hw_params_upon_resume(struct snd_sof_dev *sdev); void hda_dsp_dump(struct snd_sof_dev *sdev, u32 flags); diff --git a/sound/soc/sof/intel/tgl.c b/sound/soc/sof/intel/tgl.c index 30f2f49ee149..58ac3a46e6a7 100644 --- a/sound/soc/sof/intel/tgl.c +++ b/sound/soc/sof/intel/tgl.c @@ -60,7 +60,7 @@ int sof_tgl_ops_init(struct snd_sof_dev *sdev) memcpy(&sof_tgl_ops, &sof_hda_common_ops, sizeof(struct snd_sof_dsp_ops)); /* probe/remove/shutdown */ - sof_tgl_ops.shutdown = hda_dsp_shutdown; + sof_tgl_ops.shutdown = hda_dsp_shutdown_dma_flush; if (sdev->pdata->ipc_type == SOF_IPC) { /* doorbell */ From 44fda61d2bcfb74a942df93959e083a4e8eff75f Mon Sep 17 00:00:00 2001 From: Kai Vehmanen Date: Fri, 9 Dec 2022 13:45:29 +0200 Subject: [PATCH 056/115] ASoC: SOF: Revert: "core: unregister clients and machine drivers in .shutdown" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The unregister machine drivers call is not safe to do when kexec is used. Kexec-lite gets blocked with following backtrace: [ 84.943749] Freezing user space processes ... (elapsed 0.111 seconds) done. [ 246.784446] INFO: task kexec-lite:5123 blocked for more than 122 seconds. [ 246.819035] Call Trace: [ 246.821782] [ 246.824186] __schedule+0x5f9/0x1263 [ 246.828231] schedule+0x87/0xc5 [ 246.831779] snd_card_disconnect_sync+0xb5/0x127 ... [ 246.889249] snd_sof_device_shutdown+0xb4/0x150 [ 246.899317] pci_device_shutdown+0x37/0x61 [ 246.903990] device_shutdown+0x14c/0x1d6 [ 246.908391] kernel_kexec+0x45/0xb9 This reverts commit 83bfc7e793b555291785136c3ae86abcdc046887. Reported-by: Ricardo Ribalda Cc: Ricardo Ribalda Signed-off-by: Kai Vehmanen Reviewed-by: Pierre-Louis Bossart Reviewed-by: Péter Ujfalusi Reviewed-by: Ranjani Sridharan Link: https://lore.kernel.org/r/20221209114529.3909192-3-kai.vehmanen@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/sof/core.c | 9 --------- 1 file changed, 9 deletions(-) diff --git a/sound/soc/sof/core.c b/sound/soc/sof/core.c index 3e6141d03770..625977a29d8a 100644 --- a/sound/soc/sof/core.c +++ b/sound/soc/sof/core.c @@ -475,19 +475,10 @@ EXPORT_SYMBOL(snd_sof_device_remove); int snd_sof_device_shutdown(struct device *dev) { struct snd_sof_dev *sdev = dev_get_drvdata(dev); - struct snd_sof_pdata *pdata = sdev->pdata; if (IS_ENABLED(CONFIG_SND_SOC_SOF_PROBE_WORK_QUEUE)) cancel_work_sync(&sdev->probe_work); - /* - * make sure clients and machine driver(s) are unregistered to force - * all userspace devices to be closed prior to the DSP shutdown sequence - */ - sof_unregister_clients(sdev); - - snd_sof_machine_unregister(sdev, pdata); - if (sdev->fw_state == SOF_FW_BOOT_COMPLETE) return snd_sof_shutdown(sdev); From 83f1b7f39af73b01edf098fe3141404670703281 Mon Sep 17 00:00:00 2001 From: YC Hung Date: Fri, 9 Dec 2022 11:10:53 +0800 Subject: [PATCH 057/115] ASoC: mediatek: mt8195: add sof be ops to check audio active In MT8195 SOF design, both DSP and audio driver would access audio registers. Before DSP accesses audio registers, audio power and clock should be enabled. DSP will hang up if DSP access audio register but audio power and clock are disabled. Therefore, we add audio pm runtime active checking before accessing audio registers in SOF BE's callback hw_params function to avoid this situation. Signed-off-by: YC Hung Reviewed-by: AngeloGioacchino Del Regno Acked-by: Curtis Malainey Link: https://lore.kernel.org/r/20221209031053.8444-1-yc.hung@mediatek.com Signed-off-by: Mark Brown --- sound/soc/mediatek/mt8195/mt8195-mt6359.c | 30 +++++++++++++++++++++++ 1 file changed, 30 insertions(+) diff --git a/sound/soc/mediatek/mt8195/mt8195-mt6359.c b/sound/soc/mediatek/mt8195/mt8195-mt6359.c index 61be66f47723..4682748d82be 100644 --- a/sound/soc/mediatek/mt8195/mt8195-mt6359.c +++ b/sound/soc/mediatek/mt8195/mt8195-mt6359.c @@ -633,6 +633,32 @@ static const struct snd_soc_ops mt8195_rt1011_etdm_ops = { .hw_params = mt8195_rt1011_etdm_hw_params, }; +static int mt8195_sof_be_hw_params(struct snd_pcm_substream *substream, + struct snd_pcm_hw_params *params) +{ + struct snd_soc_pcm_runtime *rtd = asoc_substream_to_rtd(substream); + struct snd_soc_component *cmpnt_afe = NULL; + struct snd_soc_pcm_runtime *runtime; + + /* find afe component */ + for_each_card_rtds(rtd->card, runtime) { + cmpnt_afe = snd_soc_rtdcom_lookup(runtime, AFE_PCM_NAME); + if (cmpnt_afe) + break; + } + + if (cmpnt_afe && !pm_runtime_active(cmpnt_afe->dev)) { + dev_err(rtd->dev, "afe pm runtime is not active!!\n"); + return -EINVAL; + } + + return 0; +} + +static const struct snd_soc_ops mt8195_sof_be_ops = { + .hw_params = mt8195_sof_be_hw_params, +}; + static int mt8195_rt1011_init(struct snd_soc_pcm_runtime *rtd) { struct snd_soc_card *card = rtd->card; @@ -1272,24 +1298,28 @@ static struct snd_soc_dai_link mt8195_mt6359_dai_links[] = { .name = "AFE_SOF_DL2", .no_pcm = 1, .dpcm_playback = 1, + .ops = &mt8195_sof_be_ops, SND_SOC_DAILINK_REG(AFE_SOF_DL2), }, [DAI_LINK_SOF_DL3_BE] = { .name = "AFE_SOF_DL3", .no_pcm = 1, .dpcm_playback = 1, + .ops = &mt8195_sof_be_ops, SND_SOC_DAILINK_REG(AFE_SOF_DL3), }, [DAI_LINK_SOF_UL4_BE] = { .name = "AFE_SOF_UL4", .no_pcm = 1, .dpcm_capture = 1, + .ops = &mt8195_sof_be_ops, SND_SOC_DAILINK_REG(AFE_SOF_UL4), }, [DAI_LINK_SOF_UL5_BE] = { .name = "AFE_SOF_UL5", .no_pcm = 1, .dpcm_capture = 1, + .ops = &mt8195_sof_be_ops, SND_SOC_DAILINK_REG(AFE_SOF_UL5), }, }; From 9529dc167ffcdfd201b9f0eda71015f174095f7e Mon Sep 17 00:00:00 2001 From: Marek Szyprowski Date: Fri, 9 Dec 2022 10:16:57 +0100 Subject: [PATCH 058/115] ASoC: wm8994: Fix potential deadlock Fix this by dropping wm8994->accdet_lock while calling cancel_delayed_work_sync(&wm8994->mic_work) in wm1811_jackdet_irq(). Fixes: c0cc3f166525 ("ASoC: wm8994: Allow a delay between jack insertion and microphone detect") Signed-off-by: Marek Szyprowski Acked-by: Charles Keepax Link: https://lore.kernel.org/r/20221209091657.1183-1-m.szyprowski@samsung.com Signed-off-by: Mark Brown --- sound/soc/codecs/wm8994.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/sound/soc/codecs/wm8994.c b/sound/soc/codecs/wm8994.c index d3cfd3788f2a..8fe9a75d1235 100644 --- a/sound/soc/codecs/wm8994.c +++ b/sound/soc/codecs/wm8994.c @@ -3853,7 +3853,12 @@ static irqreturn_t wm1811_jackdet_irq(int irq, void *data) } else { dev_dbg(component->dev, "Jack not detected\n"); + /* Release wm8994->accdet_lock to avoid deadlock: + * cancel_delayed_work_sync() takes wm8994->mic_work internal + * lock and wm1811_mic_work takes wm8994->accdet_lock */ + mutex_unlock(&wm8994->accdet_lock); cancel_delayed_work_sync(&wm8994->mic_work); + mutex_lock(&wm8994->accdet_lock); snd_soc_component_update_bits(component, WM8958_MICBIAS2, WM8958_MICB2_DISCH, WM8958_MICB2_DISCH); From 6d94d0090527b1763872275a7ccd44df7219b31e Mon Sep 17 00:00:00 2001 From: Wang Jingjin Date: Thu, 8 Dec 2022 14:39:00 +0800 Subject: [PATCH 059/115] ASoC: rockchip: spdif: Add missing clk_disable_unprepare() in rk_spdif_runtime_resume() rk_spdif_runtime_resume() may have called clk_prepare_enable() before return from failed branches, add missing clk_disable_unprepare() in this case. Fixes: f874b80e1571 ("ASoC: rockchip: Add rockchip SPDIF transceiver driver") Signed-off-by: Wang Jingjin Link: https://lore.kernel.org/r/20221208063900.4180790-1-wangjingjin1@huawei.com Signed-off-by: Mark Brown --- sound/soc/rockchip/rockchip_spdif.c | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/rockchip/rockchip_spdif.c b/sound/soc/rockchip/rockchip_spdif.c index 8bef572d3cbc..5b4f00457587 100644 --- a/sound/soc/rockchip/rockchip_spdif.c +++ b/sound/soc/rockchip/rockchip_spdif.c @@ -88,6 +88,7 @@ static int __maybe_unused rk_spdif_runtime_resume(struct device *dev) ret = clk_prepare_enable(spdif->hclk); if (ret) { + clk_disable_unprepare(spdif->mclk); dev_err(spdif->dev, "hclk clock enable failed %d\n", ret); return ret; } From 0612d748003ce7bcd0d67a8d270900fcdadb1009 Mon Sep 17 00:00:00 2001 From: Gongjun Song Date: Mon, 12 Dec 2022 16:55:27 +0800 Subject: [PATCH 060/115] ASoC: Intel: soc-acpi: update codec addr on 0C11/0C4F product The unique ID is determined by the ADR pin level of rt1318. ODM changed design, update codec addr to match new design. Fixes: 0050e3d3d43d ("ASoC: Intel: soc-acpi: add SKU 0C11 SoundWire configuration") Signed-off-by: Gongjun Song Signed-off-by: Bard Liao Link: https://lore.kernel.org/r/20221212085527.1886168-1-yung-chuan.liao@linux.intel.com Signed-off-by: Mark Brown --- sound/soc/intel/common/soc-acpi-intel-rpl-match.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/intel/common/soc-acpi-intel-rpl-match.c b/sound/soc/intel/common/soc-acpi-intel-rpl-match.c index 3c5229f41bb0..31b43116e3d8 100644 --- a/sound/soc/intel/common/soc-acpi-intel-rpl-match.c +++ b/sound/soc/intel/common/soc-acpi-intel-rpl-match.c @@ -112,7 +112,7 @@ static const struct snd_soc_acpi_adr_device rt1316_1_group2_adr[] = { static const struct snd_soc_acpi_adr_device rt1318_1_group1_adr[] = { { - .adr = 0x000131025D131801ull, + .adr = 0x000132025D131801ull, .num_endpoints = 1, .endpoints = &spk_l_endpoint, .name_prefix = "rt1318-1" From 1a4f69ef15ec29b213e2b086b2502644e8ef76ee Mon Sep 17 00:00:00 2001 From: Dominique Martinet Date: Mon, 5 Dec 2022 21:39:01 +0900 Subject: [PATCH 061/115] 9p/client: fix data race on req->status KCSAN reported a race between writing req->status in p9_client_cb and accessing it in p9_client_rpc's wait_event. Accesses to req itself is protected by the data barrier (writing req fields, write barrier, writing status // reading status, read barrier, reading other req fields), but status accesses themselves apparently also must be annotated properly with WRITE_ONCE/READ_ONCE when we access it without locks. Follows: - error paths writing status in various threads all can notify p9_client_rpc, so these all also need WRITE_ONCE - there's a similar read loop in trans_virtio for zc case that also needs READ_ONCE - other reads in trans_fd should be protected by the trans_fd lock and lists state machine, as corresponding writers all are within trans_fd and should be under the same lock. If KCSAN complains on them we likely will have something else to fix as well, so it's better to leave them unmarked and look again if required. Link: https://lkml.kernel.org/r/20221205124756.426350-1-asmadeus@codewreck.org Reported-by: Naresh Kamboju Suggested-by: Marco Elver Acked-by: Marco Elver Reviewed-by: Christian Schoenebeck Signed-off-by: Dominique Martinet --- net/9p/client.c | 15 ++++++++------- net/9p/trans_fd.c | 12 ++++++------ net/9p/trans_rdma.c | 4 ++-- net/9p/trans_virtio.c | 9 +++++---- net/9p/trans_xen.c | 4 ++-- 5 files changed, 23 insertions(+), 21 deletions(-) diff --git a/net/9p/client.c b/net/9p/client.c index 7b2a997662d9..fef6516a0639 100644 --- a/net/9p/client.c +++ b/net/9p/client.c @@ -443,7 +443,7 @@ void p9_client_cb(struct p9_client *c, struct p9_req_t *req, int status) * the status change is visible to another thread */ smp_wmb(); - req->status = status; + WRITE_ONCE(req->status, status); wake_up(&req->wq); p9_debug(P9_DEBUG_MUX, "wakeup: %d\n", req->tc.tag); @@ -604,7 +604,7 @@ static int p9_client_flush(struct p9_client *c, struct p9_req_t *oldreq) /* if we haven't received a response for oldreq, * remove it from the list */ - if (oldreq->status == REQ_STATUS_SENT) { + if (READ_ONCE(oldreq->status) == REQ_STATUS_SENT) { if (c->trans_mod->cancelled) c->trans_mod->cancelled(c, oldreq); } @@ -704,7 +704,8 @@ p9_client_rpc(struct p9_client *c, int8_t type, const char *fmt, ...) } again: /* Wait for the response */ - err = wait_event_killable(req->wq, req->status >= REQ_STATUS_RCVD); + err = wait_event_killable(req->wq, + READ_ONCE(req->status) >= REQ_STATUS_RCVD); /* Make sure our req is coherent with regard to updates in other * threads - echoes to wmb() in the callback @@ -718,7 +719,7 @@ again: goto again; } - if (req->status == REQ_STATUS_ERROR) { + if (READ_ONCE(req->status) == REQ_STATUS_ERROR) { p9_debug(P9_DEBUG_ERROR, "req_status error %d\n", req->t_err); err = req->t_err; } @@ -731,7 +732,7 @@ again: p9_client_flush(c, req); /* if we received the response anyway, don't signal error */ - if (req->status == REQ_STATUS_RCVD) + if (READ_ONCE(req->status) == REQ_STATUS_RCVD) err = 0; } recalc_sigpending: @@ -803,7 +804,7 @@ static struct p9_req_t *p9_client_zc_rpc(struct p9_client *c, int8_t type, if (err != -ERESTARTSYS) goto recalc_sigpending; } - if (req->status == REQ_STATUS_ERROR) { + if (READ_ONCE(req->status) == REQ_STATUS_ERROR) { p9_debug(P9_DEBUG_ERROR, "req_status error %d\n", req->t_err); err = req->t_err; } @@ -816,7 +817,7 @@ static struct p9_req_t *p9_client_zc_rpc(struct p9_client *c, int8_t type, p9_client_flush(c, req); /* if we received the response anyway, don't signal error */ - if (req->status == REQ_STATUS_RCVD) + if (READ_ONCE(req->status) == REQ_STATUS_RCVD) err = 0; } recalc_sigpending: diff --git a/net/9p/trans_fd.c b/net/9p/trans_fd.c index 06ec9f7d3318..f8899745571c 100644 --- a/net/9p/trans_fd.c +++ b/net/9p/trans_fd.c @@ -201,11 +201,11 @@ static void p9_conn_cancel(struct p9_conn *m, int err) list_for_each_entry_safe(req, rtmp, &m->req_list, req_list) { list_move(&req->req_list, &cancel_list); - req->status = REQ_STATUS_ERROR; + WRITE_ONCE(req->status, REQ_STATUS_ERROR); } list_for_each_entry_safe(req, rtmp, &m->unsent_req_list, req_list) { list_move(&req->req_list, &cancel_list); - req->status = REQ_STATUS_ERROR; + WRITE_ONCE(req->status, REQ_STATUS_ERROR); } spin_unlock(&m->req_lock); @@ -466,7 +466,7 @@ static void p9_write_work(struct work_struct *work) req = list_entry(m->unsent_req_list.next, struct p9_req_t, req_list); - req->status = REQ_STATUS_SENT; + WRITE_ONCE(req->status, REQ_STATUS_SENT); p9_debug(P9_DEBUG_TRANS, "move req %p\n", req); list_move_tail(&req->req_list, &m->req_list); @@ -675,7 +675,7 @@ static int p9_fd_request(struct p9_client *client, struct p9_req_t *req) return m->err; spin_lock(&m->req_lock); - req->status = REQ_STATUS_UNSENT; + WRITE_ONCE(req->status, REQ_STATUS_UNSENT); list_add_tail(&req->req_list, &m->unsent_req_list); spin_unlock(&m->req_lock); @@ -702,7 +702,7 @@ static int p9_fd_cancel(struct p9_client *client, struct p9_req_t *req) if (req->status == REQ_STATUS_UNSENT) { list_del(&req->req_list); - req->status = REQ_STATUS_FLSHD; + WRITE_ONCE(req->status, REQ_STATUS_FLSHD); p9_req_put(client, req); ret = 0; } @@ -731,7 +731,7 @@ static int p9_fd_cancelled(struct p9_client *client, struct p9_req_t *req) * remove it from the list. */ list_del(&req->req_list); - req->status = REQ_STATUS_FLSHD; + WRITE_ONCE(req->status, REQ_STATUS_FLSHD); spin_unlock(&m->req_lock); p9_req_put(client, req); diff --git a/net/9p/trans_rdma.c b/net/9p/trans_rdma.c index 33a9ac6f2d55..83f9100d46bf 100644 --- a/net/9p/trans_rdma.c +++ b/net/9p/trans_rdma.c @@ -506,7 +506,7 @@ dont_need_post_recv: * because doing if after could erase the REQ_STATUS_RCVD * status in case of a very fast reply. */ - req->status = REQ_STATUS_SENT; + WRITE_ONCE(req->status, REQ_STATUS_SENT); err = ib_post_send(rdma->qp, &wr, NULL); if (err) goto send_error; @@ -516,7 +516,7 @@ dont_need_post_recv: /* Handle errors that happened during or while preparing the send: */ send_error: - req->status = REQ_STATUS_ERROR; + WRITE_ONCE(req->status, REQ_STATUS_ERROR); kfree(c); p9_debug(P9_DEBUG_ERROR, "Error %d in rdma_request()\n", err); diff --git a/net/9p/trans_virtio.c b/net/9p/trans_virtio.c index 19bccfa0d593..3c27ffb781e3 100644 --- a/net/9p/trans_virtio.c +++ b/net/9p/trans_virtio.c @@ -262,7 +262,7 @@ p9_virtio_request(struct p9_client *client, struct p9_req_t *req) p9_debug(P9_DEBUG_TRANS, "9p debug: virtio request\n"); - req->status = REQ_STATUS_SENT; + WRITE_ONCE(req->status, REQ_STATUS_SENT); req_retry: spin_lock_irqsave(&chan->lock, flags); @@ -468,7 +468,7 @@ p9_virtio_zc_request(struct p9_client *client, struct p9_req_t *req, inlen = n; } } - req->status = REQ_STATUS_SENT; + WRITE_ONCE(req->status, REQ_STATUS_SENT); req_retry_pinned: spin_lock_irqsave(&chan->lock, flags); @@ -531,9 +531,10 @@ req_retry_pinned: spin_unlock_irqrestore(&chan->lock, flags); kicked = 1; p9_debug(P9_DEBUG_TRANS, "virtio request kicked\n"); - err = wait_event_killable(req->wq, req->status >= REQ_STATUS_RCVD); + err = wait_event_killable(req->wq, + READ_ONCE(req->status) >= REQ_STATUS_RCVD); // RERROR needs reply (== error string) in static data - if (req->status == REQ_STATUS_RCVD && + if (READ_ONCE(req->status) == REQ_STATUS_RCVD && unlikely(req->rc.sdata[4] == P9_RERROR)) handle_rerror(req, in_hdr_len, offs, in_pages); diff --git a/net/9p/trans_xen.c b/net/9p/trans_xen.c index de2d2ca8819a..9630b1275557 100644 --- a/net/9p/trans_xen.c +++ b/net/9p/trans_xen.c @@ -157,7 +157,7 @@ again: &masked_prod, masked_cons, XEN_9PFS_RING_SIZE(ring)); - p9_req->status = REQ_STATUS_SENT; + WRITE_ONCE(p9_req->status, REQ_STATUS_SENT); virt_wmb(); /* write ring before updating pointer */ prod += size; ring->intf->out_prod = prod; @@ -212,7 +212,7 @@ static void p9_xen_response(struct work_struct *work) dev_warn(&priv->dev->dev, "requested packet size too big: %d for tag %d with capacity %zd\n", h.size, h.tag, req->rc.capacity); - req->status = REQ_STATUS_ERROR; + WRITE_ONCE(req->status, REQ_STATUS_ERROR); goto recv_error; } From a1dec9d70b6ad97087b60b81d2492134a84208c6 Mon Sep 17 00:00:00 2001 From: Hans de Goede Date: Tue, 13 Dec 2022 13:32:46 +0100 Subject: [PATCH 062/115] ASoC: Intel: bytcr_rt5640: Add quirk for the Advantech MICA-071 tablet The Advantech MICA-071 tablet deviates from the defaults for a non CR Bay Trail based tablet in several ways: 1. It uses an analog MIC on IN3 rather then using DMIC1 2. It only has 1 speaker 3. It needs the OVCD current threshold to be set to 1500uA instead of the default 2000uA to reliable differentiate between headphones vs headsets Add a quirk with these settings for this tablet. Signed-off-by: Hans de Goede Acked-by: Pierre-Louis Bossart Link: https://lore.kernel.org/r/20221213123246.11226-1-hdegoede@redhat.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/bytcr_rt5640.c | 15 +++++++++++++++ 1 file changed, 15 insertions(+) diff --git a/sound/soc/intel/boards/bytcr_rt5640.c b/sound/soc/intel/boards/bytcr_rt5640.c index 6db07b2417ca..44e6a0f1e2d9 100644 --- a/sound/soc/intel/boards/bytcr_rt5640.c +++ b/sound/soc/intel/boards/bytcr_rt5640.c @@ -570,6 +570,21 @@ static const struct dmi_system_id byt_rt5640_quirk_table[] = { BYT_RT5640_SSP0_AIF1 | BYT_RT5640_MCLK_EN), }, + { + /* Advantech MICA-071 */ + .matches = { + DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Advantech"), + DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "MICA-071"), + }, + /* OVCD Th = 1500uA to reliable detect head-phones vs -set */ + .driver_data = (void *)(BYT_RT5640_IN3_MAP | + BYT_RT5640_JD_SRC_JD2_IN4N | + BYT_RT5640_OVCD_TH_1500UA | + BYT_RT5640_OVCD_SF_0P75 | + BYT_RT5640_MONO_SPEAKER | + BYT_RT5640_DIFF_MIC | + BYT_RT5640_MCLK_EN), + }, { .matches = { DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ARCHOS"), From 6c900dcc3f7331a67ed29739d74524e428d137fb Mon Sep 17 00:00:00 2001 From: Hans de Goede Date: Tue, 13 Dec 2022 13:33:19 +0100 Subject: [PATCH 063/115] ASoC: rt5670: Remove unbalanced pm_runtime_put() For some reason rt5670_i2c_probe() does a pm_runtime_put() at the end of a successful probe. But it has never done a pm_runtime_get() leading to the following error being logged into dmesg: rt5670 i2c-10EC5640:00: Runtime PM usage count underflow! Fix this by removing the unnecessary pm_runtime_put(). Fixes: 64e89e5f5548 ("ASoC: rt5670: Add runtime PM support") Signed-off-by: Hans de Goede Link: https://lore.kernel.org/r/20221213123319.11285-1-hdegoede@redhat.com Signed-off-by: Mark Brown --- sound/soc/codecs/rt5670.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/sound/soc/codecs/rt5670.c b/sound/soc/codecs/rt5670.c index ebac6caeb40a..a230f441559a 100644 --- a/sound/soc/codecs/rt5670.c +++ b/sound/soc/codecs/rt5670.c @@ -3311,8 +3311,6 @@ static int rt5670_i2c_probe(struct i2c_client *i2c) if (ret < 0) goto err; - pm_runtime_put(&i2c->dev); - return 0; err: pm_runtime_disable(&i2c->dev); From 7bd220f2ba9014b78f0304178103393554b8c4fe Mon Sep 17 00:00:00 2001 From: YC Hung Date: Tue, 13 Dec 2022 19:56:17 +0800 Subject: [PATCH 064/115] ASoC: SOF: mediatek: initialize panic_info to zero Coverity spotted that panic_info is not initialized to zero in mtk_adsp_dump. Using uninitialized value panic_info.linenum when calling snd_sof_get_status. Fix this coverity by initializing panic_info struct as zero. Signed-off-by: YC Hung Reviewed-by: Curtis Malainey Link: https://lore.kernel.org/r/20221213115617.25086-1-yc.hung@mediatek.com Signed-off-by: Mark Brown --- sound/soc/sof/mediatek/mtk-adsp-common.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/sof/mediatek/mtk-adsp-common.c b/sound/soc/sof/mediatek/mtk-adsp-common.c index 1e0769c668a7..de8dbe27cd0d 100644 --- a/sound/soc/sof/mediatek/mtk-adsp-common.c +++ b/sound/soc/sof/mediatek/mtk-adsp-common.c @@ -60,7 +60,7 @@ void mtk_adsp_dump(struct snd_sof_dev *sdev, u32 flags) { char *level = (flags & SOF_DBG_DUMP_OPTIONAL) ? KERN_DEBUG : KERN_ERR; struct sof_ipc_dsp_oops_xtensa xoops; - struct sof_ipc_panic_info panic_info; + struct sof_ipc_panic_info panic_info = {}; u32 stack[MTK_ADSP_STACK_DUMP_SIZE]; u32 status; From b2e9e6a9cb87ce4a82fb106ae16c94639835fd47 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Thu, 8 Dec 2022 00:52:19 +0200 Subject: [PATCH 065/115] drm/i915: Fix VLV/CHV HDMI/DP audio enable MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Despite what I claimed in commit c3c5dc1d9224 ("drm/i915/audio: Do the vblank waits") the vblank interrupts are in fact not enabled yet when we do the audio enable sequence on VLV/CHV (all other platforms are fine). Reorder the enable sequence on VLV/CHV to match that of the other platforms so that the audio enable happens after the pipe has been enabled. Fixes: c3c5dc1d9224 ("drm/i915/audio: Do the vblank waits") Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20221207225219.29060-1-ville.syrjala@linux.intel.com Reviewed-by: Jani Nikula (cherry picked from commit a467a243554a64b418c14d7531a3b18c03d53bff) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/g4x_dp.c | 4 ++-- drivers/gpu/drm/i915/display/g4x_hdmi.c | 25 +++++++++++++++++++------ 2 files changed, 21 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/display/g4x_dp.c b/drivers/gpu/drm/i915/display/g4x_dp.c index 3593938dcd87..24ef36ec2d3d 100644 --- a/drivers/gpu/drm/i915/display/g4x_dp.c +++ b/drivers/gpu/drm/i915/display/g4x_dp.c @@ -673,8 +673,6 @@ static void intel_enable_dp(struct intel_atomic_state *state, intel_dp_pcon_dsc_configure(intel_dp, pipe_config); intel_dp_start_link_train(intel_dp, pipe_config); intel_dp_stop_link_train(intel_dp, pipe_config); - - intel_audio_codec_enable(encoder, pipe_config, conn_state); } static void g4x_enable_dp(struct intel_atomic_state *state, @@ -683,6 +681,7 @@ static void g4x_enable_dp(struct intel_atomic_state *state, const struct drm_connector_state *conn_state) { intel_enable_dp(state, encoder, pipe_config, conn_state); + intel_audio_codec_enable(encoder, pipe_config, conn_state); intel_edp_backlight_on(pipe_config, conn_state); } @@ -691,6 +690,7 @@ static void vlv_enable_dp(struct intel_atomic_state *state, const struct intel_crtc_state *pipe_config, const struct drm_connector_state *conn_state) { + intel_audio_codec_enable(encoder, pipe_config, conn_state); intel_edp_backlight_on(pipe_config, conn_state); } diff --git a/drivers/gpu/drm/i915/display/g4x_hdmi.c b/drivers/gpu/drm/i915/display/g4x_hdmi.c index 121caeaa409b..c3580d96765c 100644 --- a/drivers/gpu/drm/i915/display/g4x_hdmi.c +++ b/drivers/gpu/drm/i915/display/g4x_hdmi.c @@ -157,10 +157,8 @@ static void intel_hdmi_get_config(struct intel_encoder *encoder, &pipe_config->infoframes.hdmi); } -static void g4x_enable_hdmi(struct intel_atomic_state *state, - struct intel_encoder *encoder, - const struct intel_crtc_state *pipe_config, - const struct drm_connector_state *conn_state) +static void g4x_hdmi_enable_port(struct intel_encoder *encoder, + const struct intel_crtc_state *pipe_config) { struct drm_device *dev = encoder->base.dev; struct drm_i915_private *dev_priv = to_i915(dev); @@ -175,6 +173,16 @@ static void g4x_enable_hdmi(struct intel_atomic_state *state, intel_de_write(dev_priv, intel_hdmi->hdmi_reg, temp); intel_de_posting_read(dev_priv, intel_hdmi->hdmi_reg); +} + +static void g4x_enable_hdmi(struct intel_atomic_state *state, + struct intel_encoder *encoder, + const struct intel_crtc_state *pipe_config, + const struct drm_connector_state *conn_state) +{ + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + + g4x_hdmi_enable_port(encoder, pipe_config); drm_WARN_ON(&dev_priv->drm, pipe_config->has_audio && !pipe_config->has_hdmi_sink); @@ -294,6 +302,11 @@ static void vlv_enable_hdmi(struct intel_atomic_state *state, const struct intel_crtc_state *pipe_config, const struct drm_connector_state *conn_state) { + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); + + drm_WARN_ON(&dev_priv->drm, pipe_config->has_audio && + !pipe_config->has_hdmi_sink); + intel_audio_codec_enable(encoder, pipe_config, conn_state); } static void intel_disable_hdmi(struct intel_atomic_state *state, @@ -415,7 +428,7 @@ static void vlv_hdmi_pre_enable(struct intel_atomic_state *state, pipe_config->has_infoframe, pipe_config, conn_state); - g4x_enable_hdmi(state, encoder, pipe_config, conn_state); + g4x_hdmi_enable_port(encoder, pipe_config); vlv_wait_port_ready(dev_priv, dig_port, 0x0); } @@ -492,7 +505,7 @@ static void chv_hdmi_pre_enable(struct intel_atomic_state *state, pipe_config->has_infoframe, pipe_config, conn_state); - g4x_enable_hdmi(state, encoder, pipe_config, conn_state); + g4x_hdmi_enable_port(encoder, pipe_config); vlv_wait_port_ready(dev_priv, dig_port, 0x0); From 3153eebb7a76e663ac76d6670dc113296de96622 Mon Sep 17 00:00:00 2001 From: Khaled Almahallawy Date: Wed, 23 Nov 2022 14:09:26 -0800 Subject: [PATCH 066/115] drm/i915/display: Don't disable DDI/Transcoder when setting phy test pattern Bspecs has updated recently to remove the restriction to disable DDI/Transcoder before setting PHY test pattern. This update is to address PHY compliance test failures observed on a port with LTTPR. The issue is that when Transc. is disabled, the main link signals fed to LTTPR will be dropped invalidating link training, which will affect the quality of the phy test pattern when the transcoder is enabled again. v2: Update commit message (Clint) v3: Add missing Signed-off in v2 v4: Update Bspec and commit message for pre-gen12 (Jani) Bspec: 50482, 7555 Fixes: 8cdf72711928 ("drm/i915/dp: Program vswing, pre-emphasis, test-pattern") Cc: Imre Deak Cc: Clint Taylor CC: Jani Nikula Tested-by: Khaled Almahallawy Reviewed-by: Clint Taylor Signed-off-by: Khaled Almahallawy Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20221123220926.170034-1-khaled.almahallawy@intel.com (cherry picked from commit be4a847652056b067d6dc6fe0fc024a9e2e987ca) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_dp.c | 59 ------------------------- 1 file changed, 59 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_dp.c b/drivers/gpu/drm/i915/display/intel_dp.c index 67089711d9e2..75070eb07d4b 100644 --- a/drivers/gpu/drm/i915/display/intel_dp.c +++ b/drivers/gpu/drm/i915/display/intel_dp.c @@ -3679,61 +3679,6 @@ static void intel_dp_phy_pattern_update(struct intel_dp *intel_dp, } } -static void -intel_dp_autotest_phy_ddi_disable(struct intel_dp *intel_dp, - const struct intel_crtc_state *crtc_state) -{ - struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_device *dev = dig_port->base.base.dev; - struct drm_i915_private *dev_priv = to_i915(dev); - struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc); - enum pipe pipe = crtc->pipe; - u32 trans_ddi_func_ctl_value, trans_conf_value, dp_tp_ctl_value; - - trans_ddi_func_ctl_value = intel_de_read(dev_priv, - TRANS_DDI_FUNC_CTL(pipe)); - trans_conf_value = intel_de_read(dev_priv, PIPECONF(pipe)); - dp_tp_ctl_value = intel_de_read(dev_priv, TGL_DP_TP_CTL(pipe)); - - trans_ddi_func_ctl_value &= ~(TRANS_DDI_FUNC_ENABLE | - TGL_TRANS_DDI_PORT_MASK); - trans_conf_value &= ~PIPECONF_ENABLE; - dp_tp_ctl_value &= ~DP_TP_CTL_ENABLE; - - intel_de_write(dev_priv, PIPECONF(pipe), trans_conf_value); - intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(pipe), - trans_ddi_func_ctl_value); - intel_de_write(dev_priv, TGL_DP_TP_CTL(pipe), dp_tp_ctl_value); -} - -static void -intel_dp_autotest_phy_ddi_enable(struct intel_dp *intel_dp, - const struct intel_crtc_state *crtc_state) -{ - struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); - struct drm_device *dev = dig_port->base.base.dev; - struct drm_i915_private *dev_priv = to_i915(dev); - enum port port = dig_port->base.port; - struct intel_crtc *crtc = to_intel_crtc(dig_port->base.base.crtc); - enum pipe pipe = crtc->pipe; - u32 trans_ddi_func_ctl_value, trans_conf_value, dp_tp_ctl_value; - - trans_ddi_func_ctl_value = intel_de_read(dev_priv, - TRANS_DDI_FUNC_CTL(pipe)); - trans_conf_value = intel_de_read(dev_priv, PIPECONF(pipe)); - dp_tp_ctl_value = intel_de_read(dev_priv, TGL_DP_TP_CTL(pipe)); - - trans_ddi_func_ctl_value |= TRANS_DDI_FUNC_ENABLE | - TGL_TRANS_DDI_SELECT_PORT(port); - trans_conf_value |= PIPECONF_ENABLE; - dp_tp_ctl_value |= DP_TP_CTL_ENABLE; - - intel_de_write(dev_priv, PIPECONF(pipe), trans_conf_value); - intel_de_write(dev_priv, TGL_DP_TP_CTL(pipe), dp_tp_ctl_value); - intel_de_write(dev_priv, TRANS_DDI_FUNC_CTL(pipe), - trans_ddi_func_ctl_value); -} - static void intel_dp_process_phy_request(struct intel_dp *intel_dp, const struct intel_crtc_state *crtc_state) { @@ -3752,14 +3697,10 @@ static void intel_dp_process_phy_request(struct intel_dp *intel_dp, intel_dp_get_adjust_train(intel_dp, crtc_state, DP_PHY_DPRX, link_status); - intel_dp_autotest_phy_ddi_disable(intel_dp, crtc_state); - intel_dp_set_signal_levels(intel_dp, crtc_state, DP_PHY_DPRX); intel_dp_phy_pattern_update(intel_dp, crtc_state); - intel_dp_autotest_phy_ddi_enable(intel_dp, crtc_state); - drm_dp_dpcd_write(&intel_dp->aux, DP_TRAINING_LANE0_SET, intel_dp->train_set, crtc_state->lane_count); From 31a2e6cbe8a4eb0d1650fff4b77872b744e14a62 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 2 Dec 2022 12:28:42 +0000 Subject: [PATCH 067/115] drm/i915/migrate: Account for the reserved_space If the ring is nearly full when calling into emit_pte(), we might incorrectly trample the reserved_space when constructing the packet to emit the PTEs. This then triggers the GEM_BUG_ON(rq->reserved_space > ring->space) when later submitting the request, since the request itself doesn't have enough space left in the ring to emit things like workarounds, breadcrumbs etc. v2: Fix the whitespace errors Testcase: igt@i915_selftests@live_emit_pte_full_ring Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/7535 Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/6889 Fixes: cf586021642d ("drm/i915/gt: Pipelined page migration") Signed-off-by: Chris Wilson Signed-off-by: Matthew Auld Cc: Andrzej Hajda Cc: Andi Shyti Cc: Nirmoy Das Cc: # v5.15+ Tested-by: Nirmoy Das Reviewed-by: Nirmoy Das Reviewed-by: Andrzej Hajda Reviewed-by: Andi Shyti Link: https://patchwork.freedesktop.org/patch/msgid/20221202122844.428006-1-matthew.auld@intel.com (cherry picked from commit 35168a6c4ed53db4f786858bac23b1474fd7d0dc) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_migrate.c | 16 ++++++++++++---- 1 file changed, 12 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c index b405a04135ca..b783f6f740c8 100644 --- a/drivers/gpu/drm/i915/gt/intel_migrate.c +++ b/drivers/gpu/drm/i915/gt/intel_migrate.c @@ -342,6 +342,16 @@ static int emit_no_arbitration(struct i915_request *rq) return 0; } +static int max_pte_pkt_size(struct i915_request *rq, int pkt) +{ + struct intel_ring *ring = rq->ring; + + pkt = min_t(int, pkt, (ring->space - rq->reserved_space) / sizeof(u32) + 5); + pkt = min_t(int, pkt, (ring->size - ring->emit) / sizeof(u32) + 5); + + return pkt; +} + static int emit_pte(struct i915_request *rq, struct sgt_dma *it, enum i915_cache_level cache_level, @@ -388,8 +398,7 @@ static int emit_pte(struct i915_request *rq, return PTR_ERR(cs); /* Pack as many PTE updates as possible into a single MI command */ - pkt = min_t(int, dword_length, ring->space / sizeof(u32) + 5); - pkt = min_t(int, pkt, (ring->size - ring->emit) / sizeof(u32) + 5); + pkt = max_pte_pkt_size(rq, dword_length); hdr = cs; *cs++ = MI_STORE_DATA_IMM | REG_BIT(21); /* as qword elements */ @@ -422,8 +431,7 @@ static int emit_pte(struct i915_request *rq, } } - pkt = min_t(int, dword_rem, ring->space / sizeof(u32) + 5); - pkt = min_t(int, pkt, (ring->size - ring->emit) / sizeof(u32) + 5); + pkt = max_pte_pkt_size(rq, dword_rem); hdr = cs; *cs++ = MI_STORE_DATA_IMM | REG_BIT(21); From 47ea20762bb7875a62e10433a3cd5d34e9133f47 Mon Sep 17 00:00:00 2001 From: Shikang Fan Date: Thu, 8 Dec 2022 19:53:14 +0800 Subject: [PATCH 068/115] drm/amdgpu: Add an extra evict_resource call during device_suspend. MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - evict_resource is taking too long causing sriov full access mode timeout. So, add an extra evict_resource in the beginning as an early evict. Signed-off-by: Shikang Fan Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index cfa411c12072..64660a41d53c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -4112,6 +4112,11 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) adev->in_suspend = true; + /* Evict the majority of BOs before grabbing the full access */ + r = amdgpu_device_evict_resources(adev); + if (r) + return r; + if (amdgpu_sriov_vf(adev)) { amdgpu_virt_fini_data_exchange(adev); r = amdgpu_virt_request_full_gpu(adev, false); From 0debed5b117d11e33cba52870c4dcb64f5911891 Mon Sep 17 00:00:00 2001 From: ChiYuan Huang Date: Wed, 14 Dec 2022 09:37:11 +0800 Subject: [PATCH 069/115] regulator: core: Fix resolve supply lookup issue From Marek's log, the previous change modify the parent of rdev. https://lore.kernel.org/all/58b92e75-f373-dae7-7031-8abd465bb874@samsung.com/ In 'regulator_resolve_supply', it uses the parent DT node of rdev as the DT-lookup starting node. But the parent DT node may not exist. This will cause the NULL supply issue. This patch modify the parent of rdev back to the device that provides from 'regulator_config' in 'regulator_register'. Fixes: 8f3cbcd6b440 ("regulator: core: Use different devices for resource allocation and DT lookup") Reported-by: Marek Szyprowski Tested-by: Marek Szyprowski Signed-off-by: ChiYuan Huang Link: https://lore.kernel.org/r/1670981831-12583-1-git-send-email-u0084500@gmail.com Signed-off-by: Mark Brown --- drivers/regulator/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c index c84e8d78dc7e..35a7785c53f6 100644 --- a/drivers/regulator/core.c +++ b/drivers/regulator/core.c @@ -5539,7 +5539,7 @@ regulator_register(struct device *dev, /* register with sysfs */ rdev->dev.class = ®ulator_class; - rdev->dev.parent = dev; + rdev->dev.parent = config->dev; dev_set_name(&rdev->dev, "regulator.%lu", (unsigned long) atomic_inc_return(®ulator_no)); dev_set_drvdata(&rdev->dev, rdev); From 3b553e0041a65e499fa4e25ee146f01f4ec4e617 Mon Sep 17 00:00:00 2001 From: Christophe Leroy Date: Wed, 14 Dec 2022 14:41:33 +0100 Subject: [PATCH 070/115] spi: fsl_spi: Don't change speed while chipselect is active Commit c9bfcb315104 ("spi_mpc83xx: much improved driver") made modifications to the driver to not perform speed changes while chipselect is active. But those changes where lost with the convertion to tranfer_one. Previous implementation was allowing speed changes during message transfer when cs_change flag was set. At the time being, core SPI does not provide any feature to change speed while chipselect is off, so do not allow any speed change during message transfer, and perform the transfer setup in prepare_message in order to set correct speed while chipselect is still off. Reported-by: Herve Codina Fixes: 64ca1a034f00 ("spi: fsl_spi: Convert to transfer_one") Cc: stable@vger.kernel.org Signed-off-by: Christophe Leroy Tested-by: Herve Codina Reviewed-by: Herve Codina Link: https://lore.kernel.org/r/8aab84c51aa330cf91f4b43782a1c483e150a4e3.1671025244.git.christophe.leroy@csgroup.eu Signed-off-by: Mark Brown --- drivers/spi/spi-fsl-spi.c | 19 ++++++++++++++++--- 1 file changed, 16 insertions(+), 3 deletions(-) diff --git a/drivers/spi/spi-fsl-spi.c b/drivers/spi/spi-fsl-spi.c index 731624f157fc..93152144fd2e 100644 --- a/drivers/spi/spi-fsl-spi.c +++ b/drivers/spi/spi-fsl-spi.c @@ -333,13 +333,26 @@ static int fsl_spi_prepare_message(struct spi_controller *ctlr, { struct mpc8xxx_spi *mpc8xxx_spi = spi_controller_get_devdata(ctlr); struct spi_transfer *t; + struct spi_transfer *first; + + first = list_first_entry(&m->transfers, struct spi_transfer, + transfer_list); /* * In CPU mode, optimize large byte transfers to use larger * bits_per_word values to reduce number of interrupts taken. + * + * Some glitches can appear on the SPI clock when the mode changes. + * Check that there is no speed change during the transfer and set it up + * now to change the mode without having a chip-select asserted. */ - if (!(mpc8xxx_spi->flags & SPI_CPM_MODE)) { - list_for_each_entry(t, &m->transfers, transfer_list) { + list_for_each_entry(t, &m->transfers, transfer_list) { + if (t->speed_hz != first->speed_hz) { + dev_err(&m->spi->dev, + "speed_hz cannot change during message.\n"); + return -EINVAL; + } + if (!(mpc8xxx_spi->flags & SPI_CPM_MODE)) { if (t->len < 256 || t->bits_per_word != 8) continue; if ((t->len & 3) == 0) @@ -348,7 +361,7 @@ static int fsl_spi_prepare_message(struct spi_controller *ctlr, t->bits_per_word = 16; } } - return 0; + return fsl_spi_setup_transfer(m->spi, first); } static int fsl_spi_transfer_one(struct spi_controller *controller, From 9c3db58bf8f7d0007049f686ce8c419eed4325d1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Wed, 7 Dec 2022 08:47:30 +0100 Subject: [PATCH 071/115] drm/amdgpu: fixx NULL pointer deref in gmc_v9_0_get_vm_pte MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We not only need to make sure that we have a BO, but also that the BO has some backing store. Fixes: d1a372af1c3d ("drm/amdgpu: Set MTYPE in PTE based on BO flags") Signed-off-by: Christian König Reviewed-by: Felix Kuehling Reviewed-by: Alex Deucher Reviewed-by: Luben Tuikov Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index 50386eb2eec8..08d6cf79fb15 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -1185,6 +1185,8 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping *mapping, uint64_t *flags) { + struct amdgpu_bo *bo = mapping->bo_va->base.bo; + *flags &= ~AMDGPU_PTE_EXECUTABLE; *flags |= mapping->flags & AMDGPU_PTE_EXECUTABLE; @@ -1196,7 +1198,7 @@ static void gmc_v9_0_get_vm_pte(struct amdgpu_device *adev, *flags &= ~AMDGPU_PTE_VALID; } - if (mapping->bo_va->base.bo) + if (bo && bo->tbo.resource) gmc_v9_0_get_coherence_flags(adev, mapping->bo_va->base.bo, mapping, flags); } From 4d2ccd96ac25846749fc58691f5142a966e65b3a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Wed, 16 Nov 2022 15:45:36 +0100 Subject: [PATCH 072/115] drm/amdgpu: WARN when freeing kernel memory during suspend MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit When buffers are freed during suspend there is no guarantee that they can be re-allocated during resume. The PSP subsystem seems to be quite buggy regarding this, so add a WARN_ON() to point out those bugs. Signed-off-by: Christian König Reviewed-by: Alex Deucher Tested-by: Guilherme G. Piccoli Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 2df55cc7e07f..3393c1a6a0ff 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -422,6 +422,8 @@ void amdgpu_bo_free_kernel(struct amdgpu_bo **bo, u64 *gpu_addr, if (*bo == NULL) return; + WARN_ON(amdgpu_ttm_adev((*bo)->tbo.bdev)->in_suspend); + if (likely(amdgpu_bo_reserve(*bo, true) == 0)) { if (cpu_addr) amdgpu_bo_kunmap(*bo); From fe6872adb05e85bde38f2cdec01a0f4cfb826998 Mon Sep 17 00:00:00 2001 From: Yifan Zhang Date: Thu, 8 Dec 2022 11:55:15 +0800 Subject: [PATCH 073/115] drm/amd/display: Add DCN314 display SG Support Add display SG support for DCN 3.1.4. Signed-off-by: Yifan Zhang Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org --- drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 77277d90b6e2..50c783e19f5a 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -1503,6 +1503,7 @@ static int amdgpu_dm_init(struct amdgpu_device *adev) case IP_VERSION(3, 0, 1): case IP_VERSION(3, 1, 2): case IP_VERSION(3, 1, 3): + case IP_VERSION(3, 1, 4): case IP_VERSION(3, 1, 5): case IP_VERSION(3, 1, 6): init_data.flags.gpu_vm_support = true; From f95f51a4c3357eabf74fe14ab7daa5b5c0422b27 Mon Sep 17 00:00:00 2001 From: Felix Kuehling Date: Wed, 21 Apr 2021 21:09:54 -0400 Subject: [PATCH 074/115] drm/amdgpu: Add notifier lock for KFD userptrs Add a per-process MMU notifier lock for processing notifiers from userptrs. Use that lock to properly synchronize page table updates with MMU notifiers. Signed-off-by: Felix Kuehling Reviewed-by: Xiaogang Chen Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 13 +- .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 212 ++++++++++++------ drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c | 12 +- drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h | 3 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 17 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h | 6 + 6 files changed, 172 insertions(+), 91 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index f50e3ba4d7a5..589939631ed4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -29,6 +29,7 @@ #include #include #include +#include #include #include #include "amdgpu_sync.h" @@ -65,6 +66,7 @@ struct kgd_mem { struct mutex lock; struct amdgpu_bo *bo; struct dma_buf *dmabuf; + struct hmm_range *range; struct list_head attachments; /* protected by amdkfd_process_info.lock */ struct ttm_validate_buffer validate_list; @@ -75,7 +77,7 @@ struct kgd_mem { uint32_t alloc_flags; - atomic_t invalid; + uint32_t invalid; struct amdkfd_process_info *process_info; struct amdgpu_sync sync; @@ -131,7 +133,8 @@ struct amdkfd_process_info { struct amdgpu_amdkfd_fence *eviction_fence; /* MMU-notifier related fields */ - atomic_t evicted_bos; + struct mutex notifier_lock; + uint32_t evicted_bos; struct delayed_work restore_userptr_work; struct pid *pid; bool block_mmu_notifications; @@ -180,7 +183,8 @@ int kfd_debugfs_kfd_mem_limits(struct seq_file *m, void *data); bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm); struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f); int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo); -int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm); +int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni, + unsigned long cur_seq, struct kgd_mem *mem); #else static inline bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm) @@ -201,7 +205,8 @@ int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo) } static inline -int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, struct mm_struct *mm) +int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni, + unsigned long cur_seq, struct kgd_mem *mem) { return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 8782916e64a0..0a854bb8b47e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -964,7 +964,9 @@ static int init_user_pages(struct kgd_mem *mem, uint64_t user_addr, * later stage when it is scheduled by another ioctl called by * CRIU master process for the target pid for restore. */ - atomic_inc(&mem->invalid); + mutex_lock(&process_info->notifier_lock); + mem->invalid++; + mutex_unlock(&process_info->notifier_lock); mutex_unlock(&process_info->lock); return 0; } @@ -1301,6 +1303,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, return -ENOMEM; mutex_init(&info->lock); + mutex_init(&info->notifier_lock); INIT_LIST_HEAD(&info->vm_list_head); INIT_LIST_HEAD(&info->kfd_bo_list); INIT_LIST_HEAD(&info->userptr_valid_list); @@ -1317,7 +1320,6 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info, } info->pid = get_task_pid(current->group_leader, PIDTYPE_PID); - atomic_set(&info->evicted_bos, 0); INIT_DELAYED_WORK(&info->restore_userptr_work, amdgpu_amdkfd_restore_userptr_worker); @@ -1372,6 +1374,7 @@ reserve_pd_fail: put_pid(info->pid); create_evict_fence_fail: mutex_destroy(&info->lock); + mutex_destroy(&info->notifier_lock); kfree(info); } return ret; @@ -1496,6 +1499,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev, cancel_delayed_work_sync(&process_info->restore_userptr_work); put_pid(process_info->pid); mutex_destroy(&process_info->lock); + mutex_destroy(&process_info->notifier_lock); kfree(process_info); } } @@ -1548,7 +1552,9 @@ int amdgpu_amdkfd_criu_resume(void *p) mutex_lock(&pinfo->lock); pr_debug("scheduling work\n"); - atomic_inc(&pinfo->evicted_bos); + mutex_lock(&pinfo->notifier_lock); + pinfo->evicted_bos++; + mutex_unlock(&pinfo->notifier_lock); if (!READ_ONCE(pinfo->block_mmu_notifications)) { ret = -EINVAL; goto out_unlock; @@ -1773,8 +1779,13 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu( list_del(&bo_list_entry->head); mutex_unlock(&process_info->lock); - /* No more MMU notifiers */ - amdgpu_hmm_unregister(mem->bo); + /* Cleanup user pages and MMU notifiers */ + if (amdgpu_ttm_tt_get_usermm(mem->bo->tbo.ttm)) { + amdgpu_hmm_unregister(mem->bo); + mutex_lock(&process_info->notifier_lock); + amdgpu_ttm_tt_discard_user_pages(mem->bo->tbo.ttm, mem->range); + mutex_unlock(&process_info->notifier_lock); + } ret = reserve_bo_and_cond_vms(mem, NULL, BO_VM_ALL, &ctx); if (unlikely(ret)) @@ -1864,6 +1875,16 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu( */ mutex_lock(&mem->process_info->lock); + /* Lock notifier lock. If we find an invalid userptr BO, we can be + * sure that the MMU notifier is no longer running + * concurrently and the queues are actually stopped + */ + if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) { + mutex_lock(&mem->process_info->notifier_lock); + is_invalid_userptr = !!mem->invalid; + mutex_unlock(&mem->process_info->notifier_lock); + } + mutex_lock(&mem->lock); domain = mem->domain; @@ -2241,34 +2262,38 @@ int amdgpu_amdkfd_gpuvm_import_dmabuf(struct amdgpu_device *adev, * * Runs in MMU notifier, may be in RECLAIM_FS context. This means it * cannot do any memory allocations, and cannot take any locks that - * are held elsewhere while allocating memory. Therefore this is as - * simple as possible, using atomic counters. + * are held elsewhere while allocating memory. * * It doesn't do anything to the BO itself. The real work happens in * restore, where we get updated page addresses. This function only * ensures that GPU access to the BO is stopped. */ -int amdgpu_amdkfd_evict_userptr(struct kgd_mem *mem, - struct mm_struct *mm) +int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni, + unsigned long cur_seq, struct kgd_mem *mem) { struct amdkfd_process_info *process_info = mem->process_info; - int evicted_bos; int r = 0; - /* Do not process MMU notifications until stage-4 IOCTL is received */ + /* Do not process MMU notifications during CRIU restore until + * KFD_CRIU_OP_RESUME IOCTL is received + */ if (READ_ONCE(process_info->block_mmu_notifications)) return 0; - atomic_inc(&mem->invalid); - evicted_bos = atomic_inc_return(&process_info->evicted_bos); - if (evicted_bos == 1) { + mutex_lock(&process_info->notifier_lock); + mmu_interval_set_seq(mni, cur_seq); + + mem->invalid++; + if (++process_info->evicted_bos == 1) { /* First eviction, stop the queues */ - r = kgd2kfd_quiesce_mm(mm, KFD_QUEUE_EVICTION_TRIGGER_USERPTR); + r = kgd2kfd_quiesce_mm(mni->mm, + KFD_QUEUE_EVICTION_TRIGGER_USERPTR); if (r) pr_err("Failed to quiesce KFD\n"); schedule_delayed_work(&process_info->restore_userptr_work, msecs_to_jiffies(AMDGPU_USERPTR_RESTORE_DELAY_MS)); } + mutex_unlock(&process_info->notifier_lock); return r; } @@ -2285,54 +2310,58 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, struct kgd_mem *mem, *tmp_mem; struct amdgpu_bo *bo; struct ttm_operation_ctx ctx = { false, false }; - int invalid, ret; + uint32_t invalid; + int ret = 0; - /* Move all invalidated BOs to the userptr_inval_list and - * release their user pages by migration to the CPU domain - */ + mutex_lock(&process_info->notifier_lock); + + /* Move all invalidated BOs to the userptr_inval_list */ list_for_each_entry_safe(mem, tmp_mem, &process_info->userptr_valid_list, - validate_list.head) { - if (!atomic_read(&mem->invalid)) - continue; /* BO is still valid */ - - bo = mem->bo; - - if (amdgpu_bo_reserve(bo, true)) - return -EAGAIN; - amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); - ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); - amdgpu_bo_unreserve(bo); - if (ret) { - pr_err("%s: Failed to invalidate userptr BO\n", - __func__); - return -EAGAIN; - } - - list_move_tail(&mem->validate_list.head, - &process_info->userptr_inval_list); - } - - if (list_empty(&process_info->userptr_inval_list)) - return 0; /* All evicted userptr BOs were freed */ + validate_list.head) + if (mem->invalid) + list_move_tail(&mem->validate_list.head, + &process_info->userptr_inval_list); /* Go through userptr_inval_list and update any invalid user_pages */ list_for_each_entry(mem, &process_info->userptr_inval_list, validate_list.head) { - struct hmm_range *range; - - invalid = atomic_read(&mem->invalid); + invalid = mem->invalid; if (!invalid) /* BO hasn't been invalidated since the last - * revalidation attempt. Keep its BO list. + * revalidation attempt. Keep its page list. */ continue; bo = mem->bo; + amdgpu_ttm_tt_discard_user_pages(bo->tbo.ttm, mem->range); + mem->range = NULL; + + /* BO reservations and getting user pages (hmm_range_fault) + * must happen outside the notifier lock + */ + mutex_unlock(&process_info->notifier_lock); + + /* Move the BO to system (CPU) domain if necessary to unmap + * and free the SG table + */ + if (bo->tbo.resource->mem_type != TTM_PL_SYSTEM) { + if (amdgpu_bo_reserve(bo, true)) + return -EAGAIN; + amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); + ret = ttm_bo_validate(&bo->tbo, &bo->placement, &ctx); + amdgpu_bo_unreserve(bo); + if (ret) { + pr_err("%s: Failed to invalidate userptr BO\n", + __func__); + return -EAGAIN; + } + } + /* Get updated user pages */ ret = amdgpu_ttm_tt_get_user_pages(bo, bo->tbo.ttm->pages, - &range); + &mem->range); if (ret) { pr_debug("Failed %d to get user pages\n", ret); @@ -2345,30 +2374,32 @@ static int update_invalid_user_pages(struct amdkfd_process_info *process_info, */ if (ret != -EFAULT) return ret; - } else { - /* - * FIXME: Cannot ignore the return code, must hold - * notifier_lock - */ - amdgpu_ttm_tt_get_user_pages_done(bo->tbo.ttm, range); + ret = 0; } + mutex_lock(&process_info->notifier_lock); + /* Mark the BO as valid unless it was invalidated * again concurrently. */ - if (atomic_cmpxchg(&mem->invalid, invalid, 0) != invalid) - return -EAGAIN; + if (mem->invalid != invalid) { + ret = -EAGAIN; + goto unlock_out; + } + mem->invalid = 0; } - return 0; +unlock_out: + mutex_unlock(&process_info->notifier_lock); + + return ret; } /* Validate invalid userptr BOs * - * Validates BOs on the userptr_inval_list, and moves them back to the - * userptr_valid_list. Also updates GPUVM page tables with new page - * addresses and waits for the page table updates to complete. + * Validates BOs on the userptr_inval_list. Also updates GPUVM page tables + * with new page addresses and waits for the page table updates to complete. */ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) { @@ -2439,9 +2470,6 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) } } - list_move_tail(&mem->validate_list.head, - &process_info->userptr_valid_list); - /* Update mapping. If the BO was not validated * (because we couldn't get user pages), this will * clear the page table entries, which will result in @@ -2457,7 +2485,9 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info) if (ret) { pr_err("%s: update PTE failed\n", __func__); /* make sure this gets validated again */ - atomic_inc(&mem->invalid); + mutex_lock(&process_info->notifier_lock); + mem->invalid++; + mutex_unlock(&process_info->notifier_lock); goto unreserve_out; } } @@ -2477,6 +2507,36 @@ out_no_mem: return ret; } +/* Confirm that all user pages are valid while holding the notifier lock + * + * Moves valid BOs from the userptr_inval_list back to userptr_val_list. + */ +static int confirm_valid_user_pages_locked(struct amdkfd_process_info *process_info) +{ + struct kgd_mem *mem, *tmp_mem; + int ret = 0; + + list_for_each_entry_safe(mem, tmp_mem, + &process_info->userptr_inval_list, + validate_list.head) { + bool valid = amdgpu_ttm_tt_get_user_pages_done( + mem->bo->tbo.ttm, mem->range); + + mem->range = NULL; + if (!valid) { + WARN(!mem->invalid, "Invalid BO not marked invalid"); + ret = -EAGAIN; + continue; + } + WARN(mem->invalid, "Valid BO is marked invalid"); + + list_move_tail(&mem->validate_list.head, + &process_info->userptr_valid_list); + } + + return ret; +} + /* Worker callback to restore evicted userptr BOs * * Tries to update and validate all userptr BOs. If successful and no @@ -2491,9 +2551,11 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) restore_userptr_work); struct task_struct *usertask; struct mm_struct *mm; - int evicted_bos; + uint32_t evicted_bos; - evicted_bos = atomic_read(&process_info->evicted_bos); + mutex_lock(&process_info->notifier_lock); + evicted_bos = process_info->evicted_bos; + mutex_unlock(&process_info->notifier_lock); if (!evicted_bos) return; @@ -2516,9 +2578,6 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) * and we can just restart the queues. */ if (!list_empty(&process_info->userptr_inval_list)) { - if (atomic_read(&process_info->evicted_bos) != evicted_bos) - goto unlock_out; /* Concurrent eviction, try again */ - if (validate_invalid_user_pages(process_info)) goto unlock_out; } @@ -2527,10 +2586,17 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) * be a first eviction that calls quiesce_mm. The eviction * reference counting inside KFD will handle this case. */ - if (atomic_cmpxchg(&process_info->evicted_bos, evicted_bos, 0) != - evicted_bos) - goto unlock_out; - evicted_bos = 0; + mutex_lock(&process_info->notifier_lock); + if (process_info->evicted_bos != evicted_bos) + goto unlock_notifier_out; + + if (confirm_valid_user_pages_locked(process_info)) { + WARN(1, "User pages unexpectedly invalid"); + goto unlock_notifier_out; + } + + process_info->evicted_bos = evicted_bos = 0; + if (kgd2kfd_resume_mm(mm)) { pr_err("%s: Failed to resume KFD\n", __func__); /* No recovery from this failure. Probably the CP is @@ -2538,6 +2604,8 @@ static void amdgpu_amdkfd_restore_userptr_worker(struct work_struct *work) */ } +unlock_notifier_out: + mutex_unlock(&process_info->notifier_lock); unlock_out: mutex_unlock(&process_info->lock); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c index 65715cb395d8..2dadcfe43d03 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.c @@ -105,17 +105,11 @@ static bool amdgpu_hmm_invalidate_hsa(struct mmu_interval_notifier *mni, unsigned long cur_seq) { struct amdgpu_bo *bo = container_of(mni, struct amdgpu_bo, notifier); - struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); if (!mmu_notifier_range_blockable(range)) return false; - mutex_lock(&adev->notifier_lock); - - mmu_interval_set_seq(mni, cur_seq); - - amdgpu_amdkfd_evict_userptr(bo->kfd_bo, bo->notifier.mm); - mutex_unlock(&adev->notifier_lock); + amdgpu_amdkfd_evict_userptr(mni, cur_seq, bo->kfd_bo); return true; } @@ -244,9 +238,9 @@ out_free_range: return r; } -int amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range) +bool amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range) { - int r; + bool r; r = mmu_interval_read_retry(hmm_range->notifier, hmm_range->notifier_seq); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h index 13ed94d3b01b..e2edcd010ccc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_hmm.h @@ -29,12 +29,13 @@ #include #include #include +#include int amdgpu_hmm_range_get_pages(struct mmu_interval_notifier *notifier, uint64_t start, uint64_t npages, bool readonly, void *owner, struct page **pages, struct hmm_range **phmm_range); -int amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range); +bool amdgpu_hmm_range_get_pages_done(struct hmm_range *hmm_range); #if defined(CONFIG_HMM_MIRROR) int amdgpu_hmm_register(struct amdgpu_bo *bo, unsigned long addr); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index b4236572eae1..f0e4c7309438 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -695,8 +695,19 @@ out_unlock: return r; } +/* amdgpu_ttm_tt_discard_user_pages - Discard range and pfn array allocations + */ +void amdgpu_ttm_tt_discard_user_pages(struct ttm_tt *ttm, + struct hmm_range *range) +{ + struct amdgpu_ttm_tt *gtt = (void *)ttm; + + if (gtt && gtt->userptr && range) + amdgpu_hmm_range_get_pages_done(range); +} + /* - * amdgpu_ttm_tt_userptr_range_done - stop HMM track the CPU page table change + * amdgpu_ttm_tt_get_user_pages_done - stop HMM track the CPU page table change * Check if the pages backing this ttm range have been invalidated * * Returns: true if pages are still valid @@ -714,10 +725,6 @@ bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm, WARN_ONCE(!range->hmm_pfns, "No user pages to check\n"); - /* - * FIXME: Must always hold notifier_lock for this, and must - * not ignore the return code. - */ return !amdgpu_hmm_range_get_pages_done(range); } #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h index b4d8ba2789f3..e2cd5894afc9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.h @@ -159,6 +159,8 @@ uint64_t amdgpu_ttm_domain_start(struct amdgpu_device *adev, uint32_t type); #if IS_ENABLED(CONFIG_DRM_AMDGPU_USERPTR) int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, struct page **pages, struct hmm_range **range); +void amdgpu_ttm_tt_discard_user_pages(struct ttm_tt *ttm, + struct hmm_range *range); bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm, struct hmm_range *range); #else @@ -168,6 +170,10 @@ static inline int amdgpu_ttm_tt_get_user_pages(struct amdgpu_bo *bo, { return -EPERM; } +static inline void amdgpu_ttm_tt_discard_user_pages(struct ttm_tt *ttm, + struct hmm_range *range) +{ +} static inline bool amdgpu_ttm_tt_get_user_pages_done(struct ttm_tt *ttm, struct hmm_range *range) { From 592cd24a08763975c75be850a7d4e461bfd353bf Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 9 Dec 2022 16:05:12 +0800 Subject: [PATCH 075/115] drm/amd/pm: add missing SMU13.0.0 mm_dpm feature mapping Without this, the pp_dpm_vclk and pp_dpm_dclk outputs are not with correct data. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index 87d7c66e49ef..21d89c3302f1 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -210,6 +210,8 @@ static struct cmn2asic_mapping smu_v13_0_0_feature_mask_map[SMU_FEATURE_COUNT] = FEA_MAP(MEM_TEMP_READ), FEA_MAP(ATHUB_MMHUB_PG), FEA_MAP(SOC_PCC), + [SMU_FEATURE_DPM_VCLK_BIT] = {1, FEATURE_MM_DPM_BIT}, + [SMU_FEATURE_DPM_DCLK_BIT] = {1, FEATURE_MM_DPM_BIT}, }; static struct cmn2asic_mapping smu_v13_0_0_table_map[SMU_TABLE_COUNT] = { From e0607c10ebf551a654c3577fc74b4bf5533e1cea Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 9 Dec 2022 16:09:58 +0800 Subject: [PATCH 076/115] drm/amd/pm: add missing SMU13.0.7 mm_dpm feature mapping Without this, the pp_dpm_vclk and pp_dpm_dclk outputs are not with correct data. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index c3c9ef523e59..c270f94a1b86 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -189,6 +189,8 @@ static struct cmn2asic_mapping smu_v13_0_7_feature_mask_map[SMU_FEATURE_COUNT] = FEA_MAP(MEM_TEMP_READ), FEA_MAP(ATHUB_MMHUB_PG), FEA_MAP(SOC_PCC), + [SMU_FEATURE_DPM_VCLK_BIT] = {1, FEATURE_MM_DPM_BIT}, + [SMU_FEATURE_DPM_DCLK_BIT] = {1, FEATURE_MM_DPM_BIT}, }; static struct cmn2asic_mapping smu_v13_0_7_table_map[SMU_TABLE_COUNT] = { From 56b0989e2939811c11ed9c449ff84cf85878ffe3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 25 Nov 2022 16:04:25 +0100 Subject: [PATCH 077/115] drm/amdgpu: fix GDS/GWS/OA switch handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Bas pointed out that this isn't working as expected and could cause crashes. Fix the handling by storing the marker that a switch is needed inside the job instead. Reported-by: Bas Nieuwenhuizen Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c | 42 ++++++++++++++++---- drivers/gpu/drm/amd/amdgpu/amdgpu_job.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 52 +++++++++---------------- 3 files changed, 53 insertions(+), 42 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index 2a9a2593dc18..01878145a586 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -165,6 +165,26 @@ bool amdgpu_vmid_had_gpu_reset(struct amdgpu_device *adev, atomic_read(&adev->gpu_reset_counter); } +/* Check if we need to switch to another set of resources */ +static bool amdgpu_vmid_gds_switch_needed(struct amdgpu_vmid *id, + struct amdgpu_job *job) +{ + return id->gds_base != job->gds_base || + id->gds_size != job->gds_size || + id->gws_base != job->gws_base || + id->gws_size != job->gws_size || + id->oa_base != job->oa_base || + id->oa_size != job->oa_size; +} + +/* Check if the id is compatible with the job */ +static bool amdgpu_vmid_compatible(struct amdgpu_vmid *id, + struct amdgpu_job *job) +{ + return id->pd_gpu_addr == job->vm_pd_addr && + !amdgpu_vmid_gds_switch_needed(id, job); +} + /** * amdgpu_vmid_grab_idle - grab idle VMID * @@ -265,7 +285,7 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm, *id = vm->reserved_vmid[vmhub]; if ((*id)->owner != vm->immediate.fence_context || - (*id)->pd_gpu_addr != job->vm_pd_addr || + !amdgpu_vmid_compatible(*id, job) || (*id)->flushed_updates < updates || !(*id)->last_flush || ((*id)->last_flush->context != fence_context && @@ -294,7 +314,6 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm, if (r) return r; - (*id)->flushed_updates = updates; job->vm_needs_flush = needs_flush; return 0; } @@ -333,7 +352,7 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm, if ((*id)->owner != vm->immediate.fence_context) continue; - if ((*id)->pd_gpu_addr != job->vm_pd_addr) + if (!amdgpu_vmid_compatible(*id, job)) continue; if (!(*id)->last_flush || @@ -355,7 +374,6 @@ static int amdgpu_vmid_grab_used(struct amdgpu_vm *vm, if (r) return r; - (*id)->flushed_updates = updates; job->vm_needs_flush |= needs_flush; return 0; } @@ -408,22 +426,30 @@ int amdgpu_vmid_grab(struct amdgpu_vm *vm, struct amdgpu_ring *ring, if (r) goto error; - id->flushed_updates = amdgpu_vm_tlb_seq(vm); job->vm_needs_flush = true; } list_move_tail(&id->list, &id_mgr->ids_lru); } - id->pd_gpu_addr = job->vm_pd_addr; - id->owner = vm->immediate.fence_context; - + job->gds_switch_needed = amdgpu_vmid_gds_switch_needed(id, job); if (job->vm_needs_flush) { + id->flushed_updates = amdgpu_vm_tlb_seq(vm); dma_fence_put(id->last_flush); id->last_flush = NULL; } job->vmid = id - id_mgr->ids; job->pasid = vm->pasid; + + id->gds_base = job->gds_base; + id->gds_size = job->gds_size; + id->gws_base = job->gws_base; + id->gws_size = job->gws_size; + id->oa_base = job->oa_base; + id->oa_size = job->oa_size; + id->pd_gpu_addr = job->vm_pd_addr; + id->owner = vm->immediate.fence_context; + trace_amdgpu_vm_grab_id(vm, ring, job); error: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h index a372802ea4e0..02e85b040baf 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h @@ -53,6 +53,7 @@ struct amdgpu_job { uint32_t preamble_status; uint32_t preemption_status; bool vm_needs_flush; + bool gds_switch_needed; uint64_t vm_pd_addr; unsigned vmid; unsigned pasid; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index c05cff979004..245c66ea10e7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -484,25 +484,20 @@ bool amdgpu_vm_need_pipeline_sync(struct amdgpu_ring *ring, struct amdgpu_device *adev = ring->adev; unsigned vmhub = ring->funcs->vmhub; struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; - struct amdgpu_vmid *id; - bool gds_switch_needed; - bool vm_flush_needed = job->vm_needs_flush || ring->has_compute_vm_bug; if (job->vmid == 0) return false; - id = &id_mgr->ids[job->vmid]; - gds_switch_needed = ring->funcs->emit_gds_switch && ( - id->gds_base != job->gds_base || - id->gds_size != job->gds_size || - id->gws_base != job->gws_base || - id->gws_size != job->gws_size || - id->oa_base != job->oa_base || - id->oa_size != job->oa_size); - if (amdgpu_vmid_had_gpu_reset(adev, id)) + if (job->vm_needs_flush || ring->has_compute_vm_bug) return true; - return vm_flush_needed || gds_switch_needed; + if (ring->funcs->emit_gds_switch && job->gds_switch_needed) + return true; + + if (amdgpu_vmid_had_gpu_reset(adev, &id_mgr->ids[job->vmid])) + return true; + + return false; } /** @@ -524,13 +519,8 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, unsigned vmhub = ring->funcs->vmhub; struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; struct amdgpu_vmid *id = &id_mgr->ids[job->vmid]; - bool gds_switch_needed = ring->funcs->emit_gds_switch && ( - id->gds_base != job->gds_base || - id->gds_size != job->gds_size || - id->gws_base != job->gws_base || - id->gws_size != job->gws_size || - id->oa_base != job->oa_base || - id->oa_size != job->oa_size); + bool gds_switch_needed = ring->funcs->emit_gds_switch && + job->gds_switch_needed; bool vm_flush_needed = job->vm_needs_flush; struct dma_fence *fence = NULL; bool pasid_mapping_needed = false; @@ -577,6 +567,14 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, if (pasid_mapping_needed) amdgpu_gmc_emit_pasid_mapping(ring, job->vmid, job->pasid); + if (!ring->is_mes_queue && ring->funcs->emit_gds_switch && + gds_switch_needed) { + amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base, + job->gds_size, job->gws_base, + job->gws_size, job->oa_base, + job->oa_size); + } + if (vm_flush_needed || pasid_mapping_needed) { r = amdgpu_fence_emit(ring, &fence, NULL, 0); if (r) @@ -601,20 +599,6 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, } dma_fence_put(fence); - if (!ring->is_mes_queue && ring->funcs->emit_gds_switch && - gds_switch_needed) { - id->gds_base = job->gds_base; - id->gds_size = job->gds_size; - id->gws_base = job->gws_base; - id->gws_size = job->gws_size; - id->oa_base = job->oa_base; - id->oa_size = job->oa_size; - amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base, - job->gds_size, job->gws_base, - job->gws_size, job->oa_base, - job->oa_size); - } - if (ring->funcs->patch_cond_exec) amdgpu_ring_patch_cond_exec(ring, patch_offset); From 5f3c40e9e2460c42f5bf6c51b1e393d7159241c3 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 25 Nov 2022 16:42:45 +0100 Subject: [PATCH 078/115] drm/amdgpu: cleanup SPM support a bit MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This should probably not access job->vm and also emit the SPM switch under the conditional execute. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_job.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 9 +++++---- 3 files changed, 7 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index 01878145a586..6949dfec75d5 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -315,6 +315,7 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm, return r; job->vm_needs_flush = needs_flush; + job->spm_update_needed = true; return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h index 02e85b040baf..52f2e313ea17 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h @@ -54,6 +54,7 @@ struct amdgpu_job { uint32_t preemption_status; bool vm_needs_flush; bool gds_switch_needed; + bool spm_update_needed; uint64_t vm_pd_addr; unsigned vmid; unsigned pasid; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 245c66ea10e7..a05cce3f3170 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -519,22 +519,20 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, unsigned vmhub = ring->funcs->vmhub; struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; struct amdgpu_vmid *id = &id_mgr->ids[job->vmid]; + bool spm_update_needed = job->spm_update_needed; bool gds_switch_needed = ring->funcs->emit_gds_switch && job->gds_switch_needed; bool vm_flush_needed = job->vm_needs_flush; struct dma_fence *fence = NULL; bool pasid_mapping_needed = false; unsigned patch_offset = 0; - bool update_spm_vmid_needed = (job->vm && (job->vm->reserved_vmid[vmhub] != NULL)); int r; - if (update_spm_vmid_needed && adev->gfx.rlc.funcs->update_spm_vmid) - adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid); - if (amdgpu_vmid_had_gpu_reset(adev, id)) { gds_switch_needed = true; vm_flush_needed = true; pasid_mapping_needed = true; + spm_update_needed = true; } mutex_lock(&id_mgr->lock); @@ -567,6 +565,9 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, if (pasid_mapping_needed) amdgpu_gmc_emit_pasid_mapping(ring, job->vmid, job->pasid); + if (spm_update_needed && adev->gfx.rlc.funcs->update_spm_vmid) + adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid); + if (!ring->is_mes_queue && ring->funcs->emit_gds_switch && gds_switch_needed) { amdgpu_ring_emit_gds_switch(ring, job->vmid, job->gds_base, From 053499f7b45dc56758240615569b349fe9e2fc8d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 25 Nov 2022 16:45:09 +0100 Subject: [PATCH 079/115] drm/amdgpu: stop waiting for the VM during unreserve MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This is completely pointless since the VMID always stays allocated until the VM is idle. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 16 ---------------- 1 file changed, 16 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index a05cce3f3170..dc379dc22c77 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -2368,7 +2368,6 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) union drm_amdgpu_vm *args = data; struct amdgpu_device *adev = drm_to_adev(dev); struct amdgpu_fpriv *fpriv = filp->driver_priv; - long timeout = msecs_to_jiffies(2000); int r; switch (args->in.op) { @@ -2380,21 +2379,6 @@ int amdgpu_vm_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) return r; break; case AMDGPU_VM_OP_UNRESERVE_VMID: - if (amdgpu_sriov_runtime(adev)) - timeout = 8 * timeout; - - /* Wait vm idle to make sure the vmid set in SPM_VMID is - * not referenced anymore. - */ - r = amdgpu_bo_reserve(fpriv->vm.root.bo, true); - if (r) - return r; - - r = amdgpu_vm_wait_idle(&fpriv->vm, timeout); - if (r < 0) - return r; - - amdgpu_bo_unreserve(fpriv->vm.root.bo); amdgpu_vmid_free_reserved(adev, &fpriv->vm, AMDGPU_GFXHUB_0); break; default: From e44a0fe630c58b0a87d8281f5c1077a3479e5fce Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Fri, 25 Nov 2022 17:04:25 +0100 Subject: [PATCH 080/115] drm/amdgpu: rework reserved VMID handling MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Instead of reserving a VMID for a single process allow that many processes use the reserved ID. This allows for proper isolation between the processes. Signed-off-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c | 47 ++++++++++++------------- drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h | 3 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 6 +--- 3 files changed, 25 insertions(+), 31 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c index 6949dfec75d5..fcb711a11a5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c @@ -278,12 +278,13 @@ static int amdgpu_vmid_grab_reserved(struct amdgpu_vm *vm, { struct amdgpu_device *adev = ring->adev; unsigned vmhub = ring->funcs->vmhub; + struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; uint64_t fence_context = adev->fence_context + ring->idx; bool needs_flush = vm->use_cpu_for_update; uint64_t updates = amdgpu_vm_tlb_seq(vm); int r; - *id = vm->reserved_vmid[vmhub]; + *id = id_mgr->reserved; if ((*id)->owner != vm->immediate.fence_context || !amdgpu_vmid_compatible(*id, job) || (*id)->flushed_updates < updates || @@ -462,31 +463,27 @@ int amdgpu_vmid_alloc_reserved(struct amdgpu_device *adev, struct amdgpu_vm *vm, unsigned vmhub) { - struct amdgpu_vmid_mgr *id_mgr; - struct amdgpu_vmid *idle; - int r = 0; + struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; - id_mgr = &adev->vm_manager.id_mgr[vmhub]; mutex_lock(&id_mgr->lock); if (vm->reserved_vmid[vmhub]) goto unlock; - if (atomic_inc_return(&id_mgr->reserved_vmid_num) > - AMDGPU_VM_MAX_RESERVED_VMID) { - DRM_ERROR("Over limitation of reserved vmid\n"); - atomic_dec(&id_mgr->reserved_vmid_num); - r = -EINVAL; - goto unlock; - } - /* Select the first entry VMID */ - idle = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vmid, list); - list_del_init(&idle->list); - vm->reserved_vmid[vmhub] = idle; - mutex_unlock(&id_mgr->lock); - return 0; + ++id_mgr->reserved_use_count; + if (!id_mgr->reserved) { + struct amdgpu_vmid *id; + + id = list_first_entry(&id_mgr->ids_lru, struct amdgpu_vmid, + list); + /* Remove from normal round robin handling */ + list_del_init(&id->list); + id_mgr->reserved = id; + } + vm->reserved_vmid[vmhub] = true; + unlock: mutex_unlock(&id_mgr->lock); - return r; + return 0; } void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, @@ -496,12 +493,12 @@ void amdgpu_vmid_free_reserved(struct amdgpu_device *adev, struct amdgpu_vmid_mgr *id_mgr = &adev->vm_manager.id_mgr[vmhub]; mutex_lock(&id_mgr->lock); - if (vm->reserved_vmid[vmhub]) { - list_add(&vm->reserved_vmid[vmhub]->list, - &id_mgr->ids_lru); - vm->reserved_vmid[vmhub] = NULL; - atomic_dec(&id_mgr->reserved_vmid_num); + if (vm->reserved_vmid[vmhub] && + !--id_mgr->reserved_use_count) { + /* give the reserved ID back to normal round robin */ + list_add(&id_mgr->reserved->list, &id_mgr->ids_lru); } + vm->reserved_vmid[vmhub] = false; mutex_unlock(&id_mgr->lock); } @@ -568,7 +565,7 @@ void amdgpu_vmid_mgr_init(struct amdgpu_device *adev) mutex_init(&id_mgr->lock); INIT_LIST_HEAD(&id_mgr->ids_lru); - atomic_set(&id_mgr->reserved_vmid_num, 0); + id_mgr->reserved_use_count = 0; /* manage only VMIDs not used by KFD */ id_mgr->num_ids = adev->vm_manager.first_kfd_vmid; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h index 57efe61dceed..d1cc09b45da4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.h @@ -67,7 +67,8 @@ struct amdgpu_vmid_mgr { unsigned num_ids; struct list_head ids_lru; struct amdgpu_vmid ids[AMDGPU_NUM_VMID]; - atomic_t reserved_vmid_num; + struct amdgpu_vmid *reserved; + unsigned int reserved_use_count; }; int amdgpu_pasid_alloc(unsigned int bits); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h index 6546e786bf00..094bb4807303 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h @@ -119,9 +119,6 @@ struct amdgpu_bo_vm; /* Reserve 2MB at top/bottom of address space for kernel use */ #define AMDGPU_VA_RESERVED_SIZE (2ULL << 20) -/* max vmids dedicated for process */ -#define AMDGPU_VM_MAX_RESERVED_VMID 1 - /* See vm_update_mode */ #define AMDGPU_VM_USE_CPU_FOR_GFX (1 << 0) #define AMDGPU_VM_USE_CPU_FOR_COMPUTE (1 << 1) @@ -298,8 +295,7 @@ struct amdgpu_vm { struct dma_fence *last_unlocked; unsigned int pasid; - /* dedicated to vm */ - struct amdgpu_vmid *reserved_vmid[AMDGPU_MAX_VMHUBS]; + bool reserved_vmid[AMDGPU_MAX_VMHUBS]; /* Flag to indicate if VM tables are updated by CPU or GPU (SDMA) */ bool use_cpu_for_update; From 952d19190c6d482ec725f22e8bc8646bc0189d41 Mon Sep 17 00:00:00 2001 From: Matthew Auld Date: Mon, 12 Dec 2022 17:19:57 +0000 Subject: [PATCH 081/115] drm/i915/migrate: fix corner case in CCS aux copying MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In the case of lmem -> lmem transfers, which is currently only possible with small-bar systems, we need to ensure we copy the CCS aux state as-is, rather than nuke it. This should fix some nasty display corruption sometimes seen on DG2 small-bar systems, when also using DG2_RC_CCS_CC for the surface. Fixes: e3afc690188b ("drm/i915/display: consider DG2_RC_CCS_CC when migrating buffers") Signed-off-by: Matthew Auld Cc: Ville Syrjälä Cc: Nirmoy Das Cc: Andrzej Hajda Cc: Shuicheng Lin Reviewed-by: Nirmoy Das Link: https://patchwork.freedesktop.org/patch/msgid/20221212171958.82593-1-matthew.auld@intel.com (cherry picked from commit b29d26fbcb862526d5047caec82878be2eb75c0f) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_migrate.c | 37 +++++++++++++++++++------ 1 file changed, 29 insertions(+), 8 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c index b783f6f740c8..5fb74e71f27b 100644 --- a/drivers/gpu/drm/i915/gt/intel_migrate.c +++ b/drivers/gpu/drm/i915/gt/intel_migrate.c @@ -837,14 +837,35 @@ intel_context_migrate_copy(struct intel_context *ce, if (err) goto out_rq; - /* - * While we can't always restore/manage the CCS state, - * we still need to ensure we don't leak the CCS state - * from the previous user, so make sure we overwrite it - * with something. - */ - err = emit_copy_ccs(rq, dst_offset, INDIRECT_ACCESS, - dst_offset, DIRECT_ACCESS, len); + if (src_is_lmem) { + /* + * If the src is already in lmem, then we must + * be doing an lmem -> lmem transfer, and so + * should be safe to directly copy the CCS + * state. In this case we have either + * initialised the CCS aux state when first + * clearing the pages (since it is already + * allocated in lmem), or the user has + * potentially populated it, in which case we + * need to copy the CCS state as-is. + */ + err = emit_copy_ccs(rq, + dst_offset, INDIRECT_ACCESS, + src_offset, INDIRECT_ACCESS, + len); + } else { + /* + * While we can't always restore/manage the CCS + * state, we still need to ensure we don't leak + * the CCS state from the previous user, so make + * sure we overwrite it with something. + */ + err = emit_copy_ccs(rq, + dst_offset, INDIRECT_ACCESS, + dst_offset, DIRECT_ACCESS, + len); + } + if (err) goto out_rq; From ad0fca2dceeab8fdd8e1135f4b4ef2dc46c2ead9 Mon Sep 17 00:00:00 2001 From: Matthew Auld Date: Mon, 12 Dec 2022 17:19:58 +0000 Subject: [PATCH 082/115] drm/i915/ttm: consider CCS for backup objects MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit It seems we can have one or more framebuffers that are still pinned when suspending lmem, in such a case we end up creating a shmem backup object, instead of evicting the object directly, but this will skip copying the CCS aux state, since we don't allocate the extra storage for the CCS pages as part of the ttm_tt construction. Since we can already deal with pinned objects just fine, it doesn't seem too nasty to just extend to support dealing with the CCS aux state, if the object is a pinned framebuffer. This fixes display corruption (like in gnome-shell) seen on DG2 when returning from suspend. Fixes: da0595ae91da ("drm/i915/migrate: Evict and restore the flatccs capable lmem obj") Signed-off-by: Matthew Auld Cc: Ville Syrjälä Cc: Nirmoy Das Cc: Andrzej Hajda Cc: Shuicheng Lin Cc: # v5.19+ Tested-by: Nirmoy Das Reviewed-by: Nirmoy Das Link: https://patchwork.freedesktop.org/patch/msgid/20221212171958.82593-2-matthew.auld@intel.com (cherry picked from commit 95df9cc24bee8a09d39c62bcef4319b984814e18) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gem/i915_gem_object.c | 3 +++ .../gpu/drm/i915/gem/i915_gem_object_types.h | 10 ++++++---- drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c | 18 +++++++++++++++++- 3 files changed, 26 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c index 733696057761..1a0886b8aaa1 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c @@ -785,6 +785,9 @@ bool i915_gem_object_needs_ccs_pages(struct drm_i915_gem_object *obj) if (!HAS_FLAT_CCS(to_i915(obj->base.dev))) return false; + if (obj->flags & I915_BO_ALLOC_CCS_AUX) + return true; + for (i = 0; i < obj->mm.n_placements; i++) { /* Compression is not allowed for the objects with smem placement */ if (obj->mm.placements[i]->type == INTEL_MEMORY_SYSTEM) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h index d0d6772e6f36..ab4c2f90a564 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object_types.h +++ b/drivers/gpu/drm/i915/gem/i915_gem_object_types.h @@ -327,16 +327,18 @@ struct drm_i915_gem_object { * dealing with userspace objects the CPU fault handler is free to ignore this. */ #define I915_BO_ALLOC_GPU_ONLY BIT(6) +#define I915_BO_ALLOC_CCS_AUX BIT(7) #define I915_BO_ALLOC_FLAGS (I915_BO_ALLOC_CONTIGUOUS | \ I915_BO_ALLOC_VOLATILE | \ I915_BO_ALLOC_CPU_CLEAR | \ I915_BO_ALLOC_USER | \ I915_BO_ALLOC_PM_VOLATILE | \ I915_BO_ALLOC_PM_EARLY | \ - I915_BO_ALLOC_GPU_ONLY) -#define I915_BO_READONLY BIT(7) -#define I915_TILING_QUIRK_BIT 8 /* unknown swizzling; do not release! */ -#define I915_BO_PROTECTED BIT(9) + I915_BO_ALLOC_GPU_ONLY | \ + I915_BO_ALLOC_CCS_AUX) +#define I915_BO_READONLY BIT(8) +#define I915_TILING_QUIRK_BIT 9 /* unknown swizzling; do not release! */ +#define I915_BO_PROTECTED BIT(10) /** * @mem_flags - Mutable placement-related flags * diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c index 07e49f22f2de..7e67742bc65e 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm_pm.c @@ -50,6 +50,7 @@ static int i915_ttm_backup(struct i915_gem_apply_to_region *apply, container_of(bo->bdev, typeof(*i915), bdev); struct drm_i915_gem_object *backup; struct ttm_operation_ctx ctx = {}; + unsigned int flags; int err = 0; if (bo->resource->mem_type == I915_PL_SYSTEM || obj->ttm.backup) @@ -65,7 +66,22 @@ static int i915_ttm_backup(struct i915_gem_apply_to_region *apply, if (obj->flags & I915_BO_ALLOC_PM_VOLATILE) return 0; - backup = i915_gem_object_create_shmem(i915, obj->base.size); + /* + * It seems that we might have some framebuffers still pinned at this + * stage, but for such objects we might also need to deal with the CCS + * aux state. Make sure we force the save/restore of the CCS state, + * otherwise we might observe display corruption, when returning from + * suspend. + */ + flags = 0; + if (i915_gem_object_needs_ccs_pages(obj)) { + WARN_ON_ONCE(!i915_gem_object_is_framebuffer(obj)); + WARN_ON_ONCE(!pm_apply->allow_gpu); + + flags = I915_BO_ALLOC_CCS_AUX; + } + backup = i915_gem_object_create_region(i915->mm.regions[INTEL_REGION_SMEM], + obj->base.size, 0, flags); if (IS_ERR(backup)) return PTR_ERR(backup); From ff5870a76c2abda389650d3711cdddc031d12665 Mon Sep 17 00:00:00 2001 From: Moises Cardona Date: Wed, 14 Dec 2022 07:08:30 -0500 Subject: [PATCH 083/115] ASoC: Intel: Add HP Stream 8 to bytcr_rt5640.c The bytcr_rt5640.c file already supports the HP Stream 7. The HP Stream 8 is almost identical in terms of the hardware with the exception of it having stereo speakers, a SIM card slot and the obvious size difference. Signed-off-by: Moises Cardona Reviewed-by: Hans de Goede Link: https://lore.kernel.org/r/20221214120830.1572474-1-moisesmcardona@gmail.com Signed-off-by: Mark Brown --- sound/soc/intel/boards/bytcr_rt5640.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/sound/soc/intel/boards/bytcr_rt5640.c b/sound/soc/intel/boards/bytcr_rt5640.c index 44e6a0f1e2d9..4699ca79f3ea 100644 --- a/sound/soc/intel/boards/bytcr_rt5640.c +++ b/sound/soc/intel/boards/bytcr_rt5640.c @@ -811,6 +811,16 @@ static const struct dmi_system_id byt_rt5640_quirk_table[] = { BYT_RT5640_SSP0_AIF1 | BYT_RT5640_MCLK_EN), }, + { /* HP Stream 8 */ + .matches = { + DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Hewlett-Packard"), + DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "HP Stream 8 Tablet"), + }, + .driver_data = (void *)(BYTCR_INPUT_DEFAULTS | + BYT_RT5640_JD_NOT_INV | + BYT_RT5640_SSP0_AIF1 | + BYT_RT5640_MCLK_EN), + }, { /* I.T.Works TW891 */ .matches = { DMI_EXACT_MATCH(DMI_SYS_VENDOR, "To be filled by O.E.M."), From 28afcb0ad54c858d0f426b340e88e0277a375597 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Mon, 12 Dec 2022 12:04:42 -0500 Subject: [PATCH 084/115] drm/amdgpu: Check if fru_addr is not NULL (v2) Always check if fru_addr is not NULL. This commit also fixes a "smatch" warning. v2: Add a Fixes tag. Cc: Alex Deucher Cc: Dan Carpenter Cc: kernel test robot Cc: AMD Graphics Fixes: afbe5d1e4bd7c7 ("drm/amdgpu: Bug-fix: Reading I2C FRU data on newer ASICs") Signed-off-by: Luben Tuikov Reviewed-by: Kent Russell Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c index 2c38ac7bc643..4620c4712ce3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fru_eeprom.c @@ -64,7 +64,8 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr) sizeof(atom_ctx->vbios_version)) || strnstr(atom_ctx->vbios_version, "D163", sizeof(atom_ctx->vbios_version))) { - *fru_addr = FRU_EEPROM_MADDR_6; + if (fru_addr) + *fru_addr = FRU_EEPROM_MADDR_6; return true; } else { return false; @@ -83,7 +84,8 @@ static bool is_fru_eeprom_supported(struct amdgpu_device *adev, u32 *fru_addr) sizeof(atom_ctx->vbios_version))) { return false; } else { - *fru_addr = FRU_EEPROM_MADDR_6; + if (fru_addr) + *fru_addr = FRU_EEPROM_MADDR_6; return true; } } else { From 7554886daa31eacc8e7fac9e15bbce67d10b8f1f Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Sat, 10 Dec 2022 02:51:19 -0500 Subject: [PATCH 085/115] drm/amdgpu: Fix size validation for non-exclusive domains (v4) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix amdgpu_bo_validate_size() to check whether the TTM domain manager for the requested memory exists, else we get a kernel oops when dereferencing "man". v2: Make the patch standalone, i.e. not dependent on local patches. v3: Preserve old behaviour and just check that the manager pointer is not NULL. v4: Complain if GTT domain requested and it is uninitialized--most likely a bug. Cc: Alex Deucher Cc: Christian König Cc: AMD Graphics Signed-off-by: Luben Tuikov Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 19 ++++++++----------- 1 file changed, 8 insertions(+), 11 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 3393c1a6a0ff..26c6e9b2d460 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -448,27 +448,24 @@ static bool amdgpu_bo_validate_size(struct amdgpu_device *adev, /* * If GTT is part of requested domains the check must succeed to - * allow fall back to GTT + * allow fall back to GTT. */ if (domain & AMDGPU_GEM_DOMAIN_GTT) { man = ttm_manager_type(&adev->mman.bdev, TTM_PL_TT); - if (size < man->size) + if (man && size < man->size) return true; - else - goto fail; - } - - if (domain & AMDGPU_GEM_DOMAIN_VRAM) { + else if (!man) + WARN_ON_ONCE("GTT domain requested but GTT mem manager uninitialized"); + goto fail; + } else if (domain & AMDGPU_GEM_DOMAIN_VRAM) { man = ttm_manager_type(&adev->mman.bdev, TTM_PL_VRAM); - if (size < man->size) + if (man && size < man->size) return true; - else - goto fail; + goto fail; } - /* TODO add more domains checks, such as AMDGPU_GEM_DOMAIN_CPU */ return true; From 3273f11675ef11959d25a56df3279f712bcd41b7 Mon Sep 17 00:00:00 2001 From: Luben Tuikov Date: Wed, 14 Dec 2022 03:56:03 -0500 Subject: [PATCH 086/115] drm/amdgpu: Remove unnecessary domain argument MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Remove the "domain" argument to amdgpu_bo_create_kernel_at() since this function takes an "offset" argument which is the offset off of VRAM, and as such allocation always takes place in VRAM. Thus, the "domain" argument is unnecessary. Cc: Alex Deucher Cc: Christian König Cc: AMD Graphics Signed-off-by: Luben Tuikov Reviewed-by: Christian König Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 10 +++++----- drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 7 ------- drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c | 1 - 4 files changed, 6 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 26c6e9b2d460..0f6385941480 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -346,17 +346,16 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev, * @adev: amdgpu device object * @offset: offset of the BO * @size: size of the BO - * @domain: where to place it * @bo_ptr: used to initialize BOs in structures * @cpu_addr: optional CPU address mapping * - * Creates a kernel BO at a specific offset in the address space of the domain. + * Creates a kernel BO at a specific offset in VRAM. * * Returns: * 0 on success, negative error code otherwise. */ int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev, - uint64_t offset, uint64_t size, uint32_t domain, + uint64_t offset, uint64_t size, struct amdgpu_bo **bo_ptr, void **cpu_addr) { struct ttm_operation_ctx ctx = { false, false }; @@ -366,8 +365,9 @@ int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev, offset &= PAGE_MASK; size = ALIGN(size, PAGE_SIZE); - r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE, domain, bo_ptr, - NULL, cpu_addr); + r = amdgpu_bo_create_reserved(adev, size, PAGE_SIZE, + AMDGPU_GEM_DOMAIN_VRAM, bo_ptr, NULL, + cpu_addr); if (r) return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 147b79c10cbb..93207badf83f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -284,7 +284,7 @@ int amdgpu_bo_create_kernel(struct amdgpu_device *adev, u32 domain, struct amdgpu_bo **bo_ptr, u64 *gpu_addr, void **cpu_addr); int amdgpu_bo_create_kernel_at(struct amdgpu_device *adev, - uint64_t offset, uint64_t size, uint32_t domain, + uint64_t offset, uint64_t size, struct amdgpu_bo **bo_ptr, void **cpu_addr); int amdgpu_bo_create_user(struct amdgpu_device *adev, struct amdgpu_bo_param *bp, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index f0e4c7309438..55e0284b2bdd 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -1576,7 +1576,6 @@ static int amdgpu_ttm_fw_reserve_vram_init(struct amdgpu_device *adev) return amdgpu_bo_create_kernel_at(adev, adev->mman.fw_vram_usage_start_offset, adev->mman.fw_vram_usage_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.fw_vram_usage_reserved_bo, &adev->mman.fw_vram_usage_va); } @@ -1602,7 +1601,6 @@ static int amdgpu_ttm_drv_reserve_vram_init(struct amdgpu_device *adev) return amdgpu_bo_create_kernel_at(adev, adev->mman.drv_vram_usage_start_offset, adev->mman.drv_vram_usage_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.drv_vram_usage_reserved_bo, &adev->mman.drv_vram_usage_va); } @@ -1683,7 +1681,6 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev) ret = amdgpu_bo_create_kernel_at(adev, ctx->c2p_train_data_offset, ctx->train_data_size, - AMDGPU_GEM_DOMAIN_VRAM, &ctx->c2p_bo, NULL); if (ret) { @@ -1697,7 +1694,6 @@ static int amdgpu_ttm_reserve_tmr(struct amdgpu_device *adev) ret = amdgpu_bo_create_kernel_at(adev, adev->gmc.real_vram_size - adev->mman.discovery_tmr_size, adev->mman.discovery_tmr_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.discovery_memory, NULL); if (ret) { @@ -1798,21 +1794,18 @@ int amdgpu_ttm_init(struct amdgpu_device *adev) * avoid display artifacts while transitioning between pre-OS * and driver. */ r = amdgpu_bo_create_kernel_at(adev, 0, adev->mman.stolen_vga_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.stolen_vga_memory, NULL); if (r) return r; r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_vga_size, adev->mman.stolen_extended_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.stolen_extended_memory, NULL); if (r) return r; r = amdgpu_bo_create_kernel_at(adev, adev->mman.stolen_reserved_offset, adev->mman.stolen_reserved_size, - AMDGPU_GEM_DOMAIN_VRAM, &adev->mman.stolen_reserved_memory, NULL); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c index 15544f262ec1..2994b9db196f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c @@ -395,7 +395,6 @@ static void amdgpu_virt_ras_reserve_bps(struct amdgpu_device *adev) */ if (amdgpu_bo_create_kernel_at(adev, bp << AMDGPU_GPU_PAGE_SHIFT, AMDGPU_GPU_PAGE_SIZE, - AMDGPU_GEM_DOMAIN_VRAM, &bo, NULL)) DRM_DEBUG("RAS WARN: reserve vram for retired page %llx fail\n", bp); From 47722220660cfb935e27e62d385959ecc296cddb Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 12 Dec 2022 17:31:57 +0100 Subject: [PATCH 087/115] drm/amdgpu: revert "generally allow over-commit during BO allocation" MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This reverts commit f9d00a4a8dc8fff951c97b3213f90d6bc7a72175. This causes problem for KFD because when we overcommit we accidentially bind the BO to GTT for moving it into VRAM. We also need to make sure that this is done only as fallback after trying to evict first. Signed-off-by: Christian König Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 16 +++++++++++++--- drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 6 +++++- 2 files changed, 18 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 62e98f1ad770..a0780a4e3e61 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -113,7 +113,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, bp.resv = resv; bp.preferred_domain = initial_domain; bp.flags = flags; - bp.domain = initial_domain | AMDGPU_GEM_DOMAIN_CPU; + bp.domain = initial_domain; bp.bo_ptr_size = sizeof(struct amdgpu_bo); r = amdgpu_bo_create_user(adev, &bp, &ubo); @@ -332,10 +332,20 @@ int amdgpu_gem_create_ioctl(struct drm_device *dev, void *data, } initial_domain = (u32)(0xffffffff & args->in.domains); +retry: r = amdgpu_gem_object_create(adev, size, args->in.alignment, - initial_domain, flags, ttm_bo_type_device, - resv, &gobj); + initial_domain, + flags, ttm_bo_type_device, resv, &gobj); if (r && r != -ERESTARTSYS) { + if (flags & AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED) { + flags &= ~AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED; + goto retry; + } + + if (initial_domain == AMDGPU_GEM_DOMAIN_VRAM) { + initial_domain |= AMDGPU_GEM_DOMAIN_GTT; + goto retry; + } DRM_DEBUG("Failed to allocate GEM object (%llu, %d, %llu, %d)\n", size, initial_domain, args->in.alignment, r); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index 0f6385941480..4e684c2afc70 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -580,7 +580,11 @@ int amdgpu_bo_create(struct amdgpu_device *adev, bo->flags |= AMDGPU_GEM_CREATE_VRAM_WIPE_ON_RELEASE; bo->tbo.bdev = &adev->mman.bdev; - amdgpu_bo_placement_from_domain(bo, bp->domain); + if (bp->domain & (AMDGPU_GEM_DOMAIN_GWS | AMDGPU_GEM_DOMAIN_OA | + AMDGPU_GEM_DOMAIN_GDS)) + amdgpu_bo_placement_from_domain(bo, AMDGPU_GEM_DOMAIN_CPU); + else + amdgpu_bo_placement_from_domain(bo, bp->domain); if (bp->type == ttm_bo_type_kernel) bo->tbo.priority = 1; From 1a17e5b513ceebf21100027745b8731b4728edf7 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Fri, 9 Dec 2022 11:54:57 -0800 Subject: [PATCH 088/115] LoadPin: Ignore the "contents" argument of the LSM hooks LoadPin only enforces the read-only origin of kernel file reads. Whether or not it was a partial read isn't important. Remove the overly conservative checks so that things like partial firmware reads will succeed (i.e. reading a firmware header). Fixes: 2039bda1fa8d ("LSM: Add "contents" flag to kernel_read_file hook") Cc: Paul Moore Cc: James Morris Cc: "Serge E. Hallyn" Cc: linux-security-module@vger.kernel.org Signed-off-by: Kees Cook Acked-by: Serge Hallyn Tested-by: Ping-Ke Shih Link: https://lore.kernel.org/r/20221209195453.never.494-kees@kernel.org --- security/loadpin/loadpin.c | 30 ++++++++++++++++++------------ 1 file changed, 18 insertions(+), 12 deletions(-) diff --git a/security/loadpin/loadpin.c b/security/loadpin/loadpin.c index de41621f4998..110a5ab2b46b 100644 --- a/security/loadpin/loadpin.c +++ b/security/loadpin/loadpin.c @@ -122,21 +122,11 @@ static void loadpin_sb_free_security(struct super_block *mnt_sb) } } -static int loadpin_read_file(struct file *file, enum kernel_read_file_id id, - bool contents) +static int loadpin_check(struct file *file, enum kernel_read_file_id id) { struct super_block *load_root; const char *origin = kernel_read_file_id_str(id); - /* - * If we will not know that we'll be seeing the full contents - * then we cannot trust a load will be complete and unchanged - * off disk. Treat all contents=false hooks as if there were - * no associated file struct. - */ - if (!contents) - file = NULL; - /* If the file id is excluded, ignore the pinning. */ if ((unsigned int)id < ARRAY_SIZE(ignore_read_file_id) && ignore_read_file_id[id]) { @@ -192,9 +182,25 @@ static int loadpin_read_file(struct file *file, enum kernel_read_file_id id, return 0; } +static int loadpin_read_file(struct file *file, enum kernel_read_file_id id, + bool contents) +{ + /* + * LoadPin only cares about the _origin_ of a file, not its + * contents, so we can ignore the "are full contents available" + * argument here. + */ + return loadpin_check(file, id); +} + static int loadpin_load_data(enum kernel_load_data_id id, bool contents) { - return loadpin_read_file(NULL, (enum kernel_read_file_id) id, contents); + /* + * LoadPin only cares about the _origin_ of a file, not its + * contents, so a NULL file is passed, and we can ignore the + * state of "contents". + */ + return loadpin_check(NULL, (enum kernel_read_file_id) id); } static struct security_hook_list loadpin_hooks[] __lsm_ro_after_init = { From 00dd027f721e0458418f7750d8a5a664ed3e5994 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Wed, 14 Dec 2022 14:35:47 -0800 Subject: [PATCH 089/115] docs: Fix path paste-o for /sys/kernel/warn_count Running "make htmldocs" shows that "/sys/kernel/oops_count" was duplicated. This should have been "warn_count": Warning: /sys/kernel/oops_count is defined 2 times: ./Documentation/ABI/testing/sysfs-kernel-warn_count:0 ./Documentation/ABI/testing/sysfs-kernel-oops_count:0 Fix the typo. Reported-by: kernel test robot Link: https://lore.kernel.org/linux-doc/202212110529.A3Qav8aR-lkp@intel.com Fixes: 8b05aa263361 ("panic: Expose "warn_count" to sysfs") Cc: linux-hardening@vger.kernel.org Signed-off-by: Kees Cook --- Documentation/ABI/testing/sysfs-kernel-warn_count | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/Documentation/ABI/testing/sysfs-kernel-warn_count b/Documentation/ABI/testing/sysfs-kernel-warn_count index 08f083d2fd51..90a029813717 100644 --- a/Documentation/ABI/testing/sysfs-kernel-warn_count +++ b/Documentation/ABI/testing/sysfs-kernel-warn_count @@ -1,4 +1,4 @@ -What: /sys/kernel/oops_count +What: /sys/kernel/warn_count Date: November 2022 KernelVersion: 6.2.0 Contact: Linux Kernel Hardening List From f68022ae0aeb0803450e05abc0e984027c33ef1b Mon Sep 17 00:00:00 2001 From: Kristina Martsenko Date: Fri, 9 Dec 2022 17:34:41 +0000 Subject: [PATCH 090/115] lkdtm: cfi: Make PAC test work with GCC 7 and 8 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The CFI test uses the branch-protection=none compiler attribute to disable PAC return address protection on a function. While newer GCC versions support this attribute, older versions (GCC 7 and 8) instead supported the sign-return-address=none attribute, leading to a build failure when the test is built with older compilers. Fix it by checking which attribute is supported and using the correct one. Fixes: 2e53b877dc12 ("lkdtm: Add CFI_BACKWARD to test ROP mitigations") Reported-by: Daniel Díaz Signed-off-by: Kristina Martsenko Signed-off-by: Kees Cook Link: https://lore.kernel.org/all/CAEUSe78kDPxQmQqCWW-_9LCgJDFhAeMoVBFnX9QLx18Z4uT4VQ@mail.gmail.com/ --- drivers/misc/lkdtm/cfi.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/drivers/misc/lkdtm/cfi.c b/drivers/misc/lkdtm/cfi.c index 5245cf6013c9..fc28714ae3a6 100644 --- a/drivers/misc/lkdtm/cfi.c +++ b/drivers/misc/lkdtm/cfi.c @@ -54,7 +54,11 @@ static void lkdtm_CFI_FORWARD_PROTO(void) # ifdef CONFIG_ARM64_BTI_KERNEL # define __no_pac "branch-protection=bti" # else -# define __no_pac "branch-protection=none" +# ifdef CONFIG_CC_HAS_BRANCH_PROT_PAC_RET +# define __no_pac "branch-protection=none" +# else +# define __no_pac "sign-return-address=none" +# endif # endif # define __no_ret_protection __noscs __attribute__((__target__(__no_pac))) #else From d6a9fb87e9d18f3394a9845546bbe868efdccfd2 Mon Sep 17 00:00:00 2001 From: Nathan Chancellor Date: Wed, 14 Dec 2022 16:26:03 -0700 Subject: [PATCH 091/115] security: Restrict CONFIG_ZERO_CALL_USED_REGS to gcc or clang > 15.0.6 A bad bug in clang's implementation of -fzero-call-used-regs can result in NULL pointer dereferences (see the links above the check for more information). Restrict CONFIG_CC_HAS_ZERO_CALL_USED_REGS to either a supported GCC version or a clang newer than 15.0.6, which will catch both a theoretical 15.0.7 and the upcoming 16.0.0, which will both have the bug fixed. Cc: stable@vger.kernel.org # v5.15+ Signed-off-by: Nathan Chancellor Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20221214232602.4118147-1-nathan@kernel.org --- security/Kconfig.hardening | 3 +++ 1 file changed, 3 insertions(+) diff --git a/security/Kconfig.hardening b/security/Kconfig.hardening index d766b7d0ffd1..53baa95cb644 100644 --- a/security/Kconfig.hardening +++ b/security/Kconfig.hardening @@ -257,6 +257,9 @@ config INIT_ON_FREE_DEFAULT_ON config CC_HAS_ZERO_CALL_USED_REGS def_bool $(cc-option,-fzero-call-used-regs=used-gpr) + # https://github.com/ClangBuiltLinux/linux/issues/1766 + # https://github.com/llvm/llvm-project/issues/59242 + depends on !CC_IS_CLANG || CLANG_VERSION > 150006 config ZERO_CALL_USED_REGS bool "Enable register zeroing on function exit" From 76d62f24db07f22ccf9bc18ca793c27d4ebef721 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Wed, 14 Dec 2022 23:18:34 +0000 Subject: [PATCH 092/115] pstore: Switch pmsg_lock to an rt_mutex to avoid priority inversion Wei Wang reported seeing priority inversion caused latencies caused by contention on pmsg_lock, and suggested it be switched to a rt_mutex. I was initially hesitant this would help, as the tasks in that trace all seemed to be SCHED_NORMAL, so the benefit would be limited to only nice boosting. However, another similar issue was raised where the priority inversion was seen did involve a blocked RT task so it is clear this would be helpful in that case. Cc: Wei Wang Cc: Midas Chien Cc: Connor O'Brien Cc: Kees Cook Cc: Anton Vorontsov Cc: Colin Cross Cc: Tony Luck Cc: kernel-team@android.com Fixes: 9d5438f462ab ("pstore: Add pmsg - user-space accessible pstore object") Reported-by: Wei Wang Signed-off-by: John Stultz Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20221214231834.3711880-1-jstultz@google.com --- fs/pstore/pmsg.c | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/fs/pstore/pmsg.c b/fs/pstore/pmsg.c index d8542ec2f38c..18cf94b597e0 100644 --- a/fs/pstore/pmsg.c +++ b/fs/pstore/pmsg.c @@ -7,9 +7,10 @@ #include #include #include +#include #include "internal.h" -static DEFINE_MUTEX(pmsg_lock); +static DEFINE_RT_MUTEX(pmsg_lock); static ssize_t write_pmsg(struct file *file, const char __user *buf, size_t count, loff_t *ppos) @@ -28,9 +29,9 @@ static ssize_t write_pmsg(struct file *file, const char __user *buf, if (!access_ok(buf, count)) return -EFAULT; - mutex_lock(&pmsg_lock); + rt_mutex_lock(&pmsg_lock); ret = psinfo->write_user(&record, buf); - mutex_unlock(&pmsg_lock); + rt_mutex_unlock(&pmsg_lock); return ret ? ret : count; } From cb3543cff90a4448ed560ac86c98033ad5fecda9 Mon Sep 17 00:00:00 2001 From: Johan Hovold Date: Thu, 15 Dec 2022 11:46:46 +0100 Subject: [PATCH 093/115] regulator: core: fix deadlock on regulator enable When updating the operating mode as part of regulator enable, the caller has already locked the regulator tree and drms_uA_update() must not try to do the same in order not to trigger a deadlock. The lock inversion is reported by lockdep as: ====================================================== WARNING: possible circular locking dependency detected 6.1.0-next-20221215 #142 Not tainted ------------------------------------------------------ udevd/154 is trying to acquire lock: ffffc11f123d7e50 (regulator_list_mutex){+.+.}-{3:3}, at: regulator_lock_dependent+0x54/0x280 but task is already holding lock: ffff80000e4c36e8 (regulator_ww_class_acquire){+.+.}-{0:0}, at: regulator_enable+0x34/0x80 which lock already depends on the new lock. ... Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(regulator_ww_class_acquire); lock(regulator_list_mutex); lock(regulator_ww_class_acquire); lock(regulator_list_mutex); *** DEADLOCK *** just before probe of a Qualcomm UFS controller (occasionally) deadlocks when enabling one of its regulators. Fixes: 9243a195be7a ("regulator: core: Change voltage setting path") Fixes: f8702f9e4aa7 ("regulator: core: Use ww_mutex for regulators locking") Cc: stable@vger.kernel.org # 5.0 Signed-off-by: Johan Hovold Link: https://lore.kernel.org/r/20221215104646.19818-1-johan+linaro@kernel.org Signed-off-by: Mark Brown --- drivers/regulator/core.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/regulator/core.c b/drivers/regulator/core.c index 35a7785c53f6..c661be129818 100644 --- a/drivers/regulator/core.c +++ b/drivers/regulator/core.c @@ -1002,7 +1002,7 @@ static int drms_uA_update(struct regulator_dev *rdev) /* get input voltage */ input_uV = 0; if (rdev->supply) - input_uV = regulator_get_voltage(rdev->supply); + input_uV = regulator_get_voltage_rdev(rdev->supply->rdev); if (input_uV <= 0) input_uV = rdev->constraints->input_uV; From a12a383e59ce486abd719b6bda33c353a3b385e7 Mon Sep 17 00:00:00 2001 From: Krzysztof Kozlowski Date: Thu, 15 Dec 2022 14:43:37 +0100 Subject: [PATCH 094/115] ASoC: lochnagar: Fix unused lochnagar_of_match warning MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit lochnagar_of_match is used unconditionally, so COMPILE_TEST builds without OF warn: sound/soc/codecs/lochnagar-sc.c:247:34: error: ‘lochnagar_of_match’ defined but not used [-Werror=unused-const-variable=] Reported-by: kernel test robot Signed-off-by: Krzysztof Kozlowski Link: https://lore.kernel.org/r/20221215134337.77944-1-krzysztof.kozlowski@linaro.org Signed-off-by: Mark Brown --- sound/soc/codecs/lochnagar-sc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/soc/codecs/lochnagar-sc.c b/sound/soc/codecs/lochnagar-sc.c index 13fbd8830b09..5e0bd0d24ed3 100644 --- a/sound/soc/codecs/lochnagar-sc.c +++ b/sound/soc/codecs/lochnagar-sc.c @@ -253,7 +253,7 @@ MODULE_DEVICE_TABLE(of, lochnagar_of_match); static struct platform_driver lochnagar_sc_codec_driver = { .driver = { .name = "lochnagar-soundcard", - .of_match_table = of_match_ptr(lochnagar_of_match), + .of_match_table = lochnagar_of_match, }, .probe = lochnagar_sc_probe, From 1794f6a9535bb5234c2b747d1bc6dad03249245a Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 2 Dec 2022 13:56:35 +0800 Subject: [PATCH 095/115] drm/amd/pm: enable GPO dynamic control support for SMU13.0.0 To better support UMD pstate profilings, the GPO feature needs to be switched on/off accordingly. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h | 3 ++- drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h | 3 +++ drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c | 15 +++++++++++++++ .../gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 2 ++ 4 files changed, 22 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h index a4e3425b1027..4180c71d930f 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h @@ -241,7 +241,8 @@ __SMU_DUMMY_MAP(GetGfxOffEntryCount), \ __SMU_DUMMY_MAP(LogGfxOffResidency), \ __SMU_DUMMY_MAP(SetNumBadMemoryPagesRetired), \ - __SMU_DUMMY_MAP(SetBadMemoryPagesRetiredFlagsPerChannel), + __SMU_DUMMY_MAP(SetBadMemoryPagesRetiredFlagsPerChannel), \ + __SMU_DUMMY_MAP(AllowGpo), #undef __SMU_DUMMY_MAP #define __SMU_DUMMY_MAP(type) SMU_MSG_##type diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h index 865d6358918d..ea29ac6a80e6 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h @@ -272,6 +272,9 @@ int smu_v13_0_init_pptable_microcode(struct smu_context *smu); int smu_v13_0_run_btc(struct smu_context *smu); +int smu_v13_0_gpo_control(struct smu_context *smu, + bool enablement); + int smu_v13_0_deep_sleep_control(struct smu_context *smu, bool enablement); diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index f5e90e0a99df..e3a80ac987df 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -2180,6 +2180,21 @@ int smu_v13_0_run_btc(struct smu_context *smu) return res; } +int smu_v13_0_gpo_control(struct smu_context *smu, + bool enablement) +{ + int res; + + res = smu_cmn_send_smc_msg_with_param(smu, + SMU_MSG_AllowGpo, + enablement ? 1 : 0, + NULL); + if (res) + dev_err(smu->adev->dev, "SetGpoAllow %d failed!\n", enablement); + + return res; +} + int smu_v13_0_deep_sleep_control(struct smu_context *smu, bool enablement) { diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index 21d89c3302f1..cc66828c7a84 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -144,6 +144,7 @@ static struct cmn2asic_msg_mapping smu_v13_0_0_message_map[SMU_MSG_MAX_COUNT] = MSG_MAP(SetNumBadMemoryPagesRetired, PPSMC_MSG_SetNumBadMemoryPagesRetired, 0), MSG_MAP(SetBadMemoryPagesRetiredFlagsPerChannel, PPSMC_MSG_SetBadMemoryPagesRetiredFlagsPerChannel, 0), + MSG_MAP(AllowGpo, PPSMC_MSG_SetGpoAllow, 0), }; static struct cmn2asic_mapping smu_v13_0_0_clk_map[SMU_CLK_COUNT] = { @@ -1949,6 +1950,7 @@ static const struct pptable_funcs smu_v13_0_0_ppt_funcs = { .set_df_cstate = smu_v13_0_0_set_df_cstate, .send_hbm_bad_pages_num = smu_v13_0_0_smu_send_bad_mem_page_num, .send_hbm_bad_channel_flag = smu_v13_0_0_send_bad_mem_channel_flag, + .gpo_control = smu_v13_0_gpo_control, }; void smu_v13_0_0_set_ppt_funcs(struct smu_context *smu) From 62b9f835a6c60171845642afec4ce4b44865f10f Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 2 Dec 2022 14:03:45 +0800 Subject: [PATCH 096/115] drm/amd/pm: enable GPO dynamic control support for SMU13.0.7 To better support UMD pstate profilings, the GPO feature needs to be switched on/off accordingly. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index c270f94a1b86..ab1c004606be 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -123,6 +123,7 @@ static struct cmn2asic_msg_mapping smu_v13_0_7_message_map[SMU_MSG_MAX_COUNT] = MSG_MAP(SetMGpuFanBoostLimitRpm, PPSMC_MSG_SetMGpuFanBoostLimitRpm, 0), MSG_MAP(DFCstateControl, PPSMC_MSG_SetExternalClientDfCstateAllow, 0), MSG_MAP(ArmD3, PPSMC_MSG_ArmD3, 0), + MSG_MAP(AllowGpo, PPSMC_MSG_SetGpoAllow, 0), }; static struct cmn2asic_mapping smu_v13_0_7_clk_map[SMU_CLK_COUNT] = { @@ -1690,6 +1691,7 @@ static const struct pptable_funcs smu_v13_0_7_ppt_funcs = { .mode1_reset = smu_v13_0_mode1_reset, .set_mp1_state = smu_v13_0_7_set_mp1_state, .set_df_cstate = smu_v13_0_7_set_df_cstate, + .gpo_control = smu_v13_0_gpo_control, }; void smu_v13_0_7_set_ppt_funcs(struct smu_context *smu) From 32a7819ff8e25375c7515aaae5cfcb8c44a461b7 Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Mon, 5 Dec 2022 14:53:34 +0800 Subject: [PATCH 097/115] drm/amd/pm: correct SMU13.0.0 pstate profiling clock settings Correct the pstate standard/peak profiling mode clock settings for SMU13.0.0. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- .../drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 22 +++++++++++++------ 1 file changed, 15 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index cc66828c7a84..d689fcab963d 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -1307,9 +1307,17 @@ static int smu_v13_0_0_populate_umd_state_clk(struct smu_context *smu) &dpm_context->dpm_tables.fclk_table; struct smu_umd_pstate_table *pstate_table = &smu->pstate_table; + struct smu_table_context *table_context = &smu->smu_table; + PPTable_t *pptable = table_context->driver_pptable; + DriverReportedClocks_t driver_clocks = + pptable->SkuTable.DriverReportedClocks; pstate_table->gfxclk_pstate.min = gfx_table->min; - pstate_table->gfxclk_pstate.peak = gfx_table->max; + if (driver_clocks.GameClockAc && + (driver_clocks.GameClockAc < gfx_table->max)) + pstate_table->gfxclk_pstate.peak = driver_clocks.GameClockAc; + else + pstate_table->gfxclk_pstate.peak = gfx_table->max; pstate_table->uclk_pstate.min = mem_table->min; pstate_table->uclk_pstate.peak = mem_table->max; @@ -1326,12 +1334,12 @@ static int smu_v13_0_0_populate_umd_state_clk(struct smu_context *smu) pstate_table->fclk_pstate.min = fclk_table->min; pstate_table->fclk_pstate.peak = fclk_table->max; - /* - * For now, just use the mininum clock frequency. - * TODO: update them when the real pstate settings available - */ - pstate_table->gfxclk_pstate.standard = gfx_table->min; - pstate_table->uclk_pstate.standard = mem_table->min; + if (driver_clocks.BaseClockAc && + driver_clocks.BaseClockAc < gfx_table->max) + pstate_table->gfxclk_pstate.standard = driver_clocks.BaseClockAc; + else + pstate_table->gfxclk_pstate.standard = gfx_table->max; + pstate_table->uclk_pstate.standard = mem_table->max; pstate_table->socclk_pstate.standard = soc_table->min; pstate_table->vclk_pstate.standard = vclk_table->min; pstate_table->dclk_pstate.standard = dclk_table->min; From 7a18e089eff02f17eaee49fc18641f5d16a8284b Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Mon, 5 Dec 2022 15:33:31 +0800 Subject: [PATCH 098/115] drm/amd/pm: update SMU13.0.0 reported maximum shader clock Update the reported maximum shader clock to the value which can be guarded to be achieved on all cards. This is to align with Window setting. Signed-off-by: Evan Quan Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0.x --- .../drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 70 ++++++++++++++++++- 1 file changed, 69 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index d689fcab963d..713fb6ad39f6 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -543,6 +543,23 @@ static int smu_v13_0_0_set_default_dpm_table(struct smu_context *smu) dpm_table); if (ret) return ret; + + /* + * Update the reported maximum shader clock to the value + * which can be guarded to be achieved on all cards. This + * is aligned with Window setting. And considering that value + * might be not the peak frequency the card can achieve, it + * is normal some real-time clock frequency can overtake this + * labelled maximum clock frequency(for example in pp_dpm_sclk + * sysfs output). + */ + if (skutable->DriverReportedClocks.GameClockAc && + (dpm_table->dpm_levels[dpm_table->count - 1].value > + skutable->DriverReportedClocks.GameClockAc)) { + dpm_table->dpm_levels[dpm_table->count - 1].value = + skutable->DriverReportedClocks.GameClockAc; + dpm_table->max = skutable->DriverReportedClocks.GameClockAc; + } } else { dpm_table->count = 1; dpm_table->dpm_levels[0].value = smu->smu_table.boot_values.gfxclk / 100; @@ -805,6 +822,57 @@ static int smu_v13_0_0_get_smu_metrics_data(struct smu_context *smu, return ret; } +static int smu_v13_0_0_get_dpm_ultimate_freq(struct smu_context *smu, + enum smu_clk_type clk_type, + uint32_t *min, + uint32_t *max) +{ + struct smu_13_0_dpm_context *dpm_context = + smu->smu_dpm.dpm_context; + struct smu_13_0_dpm_table *dpm_table; + + switch (clk_type) { + case SMU_MCLK: + case SMU_UCLK: + /* uclk dpm table */ + dpm_table = &dpm_context->dpm_tables.uclk_table; + break; + case SMU_GFXCLK: + case SMU_SCLK: + /* gfxclk dpm table */ + dpm_table = &dpm_context->dpm_tables.gfx_table; + break; + case SMU_SOCCLK: + /* socclk dpm table */ + dpm_table = &dpm_context->dpm_tables.soc_table; + break; + case SMU_FCLK: + /* fclk dpm table */ + dpm_table = &dpm_context->dpm_tables.fclk_table; + break; + case SMU_VCLK: + case SMU_VCLK1: + /* vclk dpm table */ + dpm_table = &dpm_context->dpm_tables.vclk_table; + break; + case SMU_DCLK: + case SMU_DCLK1: + /* dclk dpm table */ + dpm_table = &dpm_context->dpm_tables.dclk_table; + break; + default: + dev_err(smu->adev->dev, "Unsupported clock type!\n"); + return -EINVAL; + } + + if (min) + *min = dpm_table->min; + if (max) + *max = dpm_table->max; + + return 0; +} + static int smu_v13_0_0_read_sensor(struct smu_context *smu, enum amd_pp_sensors sensor, void *data, @@ -1910,7 +1978,7 @@ static const struct pptable_funcs smu_v13_0_0_ppt_funcs = { .get_enabled_mask = smu_cmn_get_enabled_mask, .dpm_set_vcn_enable = smu_v13_0_set_vcn_enable, .dpm_set_jpeg_enable = smu_v13_0_set_jpeg_enable, - .get_dpm_ultimate_freq = smu_v13_0_get_dpm_ultimate_freq, + .get_dpm_ultimate_freq = smu_v13_0_0_get_dpm_ultimate_freq, .get_vbios_bootup_values = smu_v13_0_get_vbios_bootup_values, .read_sensor = smu_v13_0_0_read_sensor, .feature_is_enabled = smu_cmn_feature_is_enabled, From 7535b832c6399b5ebfc5b53af5c51dd915ee2538 Mon Sep 17 00:00:00 2001 From: Kees Cook Date: Fri, 16 Dec 2022 12:26:57 -0800 Subject: [PATCH 099/115] exit: Use READ_ONCE() for all oops/warn limit reads Use a temporary variable to take full advantage of READ_ONCE() behavior. Without this, the report (and even the test) might be out of sync with the initial test. Reported-by: Peter Zijlstra Link: https://lore.kernel.org/lkml/Y5x7GXeluFmZ8E0E@hirez.programming.kicks-ass.net Fixes: 9fc9e278a5c0 ("panic: Introduce warn_limit") Fixes: d4ccd54d28d3 ("exit: Put an upper limit on how often we can oops") Cc: "Eric W. Biederman" Cc: Jann Horn Cc: Arnd Bergmann Cc: Petr Mladek Cc: Andrew Morton Cc: Luis Chamberlain Cc: Marco Elver Cc: tangmeng Cc: Sebastian Andrzej Siewior Cc: Tiezhu Yang Signed-off-by: Kees Cook --- kernel/exit.c | 6 ++++-- kernel/panic.c | 7 +++++-- 2 files changed, 9 insertions(+), 4 deletions(-) diff --git a/kernel/exit.c b/kernel/exit.c index deffb8e4b1b2..15dc2ec80c46 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -931,6 +931,7 @@ void __noreturn make_task_dead(int signr) * Then do everything else. */ struct task_struct *tsk = current; + unsigned int limit; if (unlikely(in_interrupt())) panic("Aiee, killing interrupt handler!"); @@ -954,8 +955,9 @@ void __noreturn make_task_dead(int signr) * To make sure this can't happen, place an upper bound on how often the * kernel may oops without panic(). */ - if (atomic_inc_return(&oops_count) >= READ_ONCE(oops_limit) && oops_limit) - panic("Oopsed too often (kernel.oops_limit is %d)", oops_limit); + limit = READ_ONCE(oops_limit); + if (atomic_inc_return(&oops_count) >= limit && limit) + panic("Oopsed too often (kernel.oops_limit is %d)", limit); /* * We're taking recursive faults here in make_task_dead. Safest is to just diff --git a/kernel/panic.c b/kernel/panic.c index 54deb743b2d5..7834c9854e02 100644 --- a/kernel/panic.c +++ b/kernel/panic.c @@ -231,12 +231,15 @@ static void panic_print_sys_info(bool console_flush) void check_panic_on_warn(const char *origin) { + unsigned int limit; + if (panic_on_warn) panic("%s: panic_on_warn set ...\n", origin); - if (atomic_inc_return(&warn_count) >= READ_ONCE(warn_limit) && warn_limit) + limit = READ_ONCE(warn_limit); + if (atomic_inc_return(&warn_count) >= limit && limit) panic("%s: system warned too often (kernel.warn_limit is %d)", - origin, warn_limit); + origin, limit); } /** From 29d48b87db64b6697ddad007548e51d032081c59 Mon Sep 17 00:00:00 2001 From: Philip Yang Date: Wed, 14 Dec 2022 10:15:17 -0500 Subject: [PATCH 100/115] drm/amdkfd: Fix kfd_process_device_init_vm error handling Should only destroy the ib_mem and let process cleanup worker to free the outstanding BOs. Reset the pointer in pdd->qpd structure, to avoid NULL pointer access in process destroy worker. BUG: kernel NULL pointer dereference, address: 0000000000000010 Call Trace: amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel+0x46/0xb0 [amdgpu] kfd_process_device_destroy_cwsr_dgpu+0x40/0x70 [amdgpu] kfd_process_destroy_pdds+0x71/0x190 [amdgpu] kfd_process_wq_release+0x2a2/0x3b0 [amdgpu] process_one_work+0x2a1/0x600 worker_thread+0x39/0x3d0 Signed-off-by: Philip Yang Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdkfd/kfd_process.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index a26257171ab7..6caa9dd57ff1 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -689,13 +689,13 @@ void kfd_process_destroy_wq(void) } static void kfd_process_free_gpuvm(struct kgd_mem *mem, - struct kfd_process_device *pdd, void *kptr) + struct kfd_process_device *pdd, void **kptr) { struct kfd_dev *dev = pdd->dev; - if (kptr) { + if (kptr && *kptr) { amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem); - kptr = NULL; + *kptr = NULL; } amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(dev->adev, mem, pdd->drm_priv); @@ -795,7 +795,7 @@ static void kfd_process_device_destroy_ib_mem(struct kfd_process_device *pdd) if (!qpd->ib_kaddr || !qpd->ib_base) return; - kfd_process_free_gpuvm(qpd->ib_mem, pdd, qpd->ib_kaddr); + kfd_process_free_gpuvm(qpd->ib_mem, pdd, &qpd->ib_kaddr); } struct kfd_process *kfd_create_process(struct file *filep) @@ -1277,7 +1277,7 @@ static void kfd_process_device_destroy_cwsr_dgpu(struct kfd_process_device *pdd) if (!dev->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base) return; - kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, qpd->cwsr_kaddr); + kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, &qpd->cwsr_kaddr); } void kfd_process_set_trap_handler(struct qcm_process_device *qpd, @@ -1598,8 +1598,8 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd, return 0; err_init_cwsr: + kfd_process_device_destroy_ib_mem(pdd); err_reserve_ib_mem: - kfd_process_device_free_bos(pdd); pdd->drm_priv = NULL; return ret; From 1a799c4c190ea9f0e81028e3eb3037ed0ab17ff5 Mon Sep 17 00:00:00 2001 From: Philip Yang Date: Tue, 13 Dec 2022 00:50:03 -0500 Subject: [PATCH 101/115] drm/amdkfd: Fix double release compute pasid If kfd_process_device_init_vm returns failure after vm is converted to compute vm and vm->pasid set to compute pasid, KFD will not take pdd->drm_file reference. As a result, drm close file handler maybe called to release the compute pasid before KFD process destroy worker to release the same pasid and set vm->pasid to zero, this generates below WARNING backtrace and NULL pointer access. Add helper amdgpu_amdkfd_gpuvm_set_vm_pasid and call it at the last step of kfd_process_device_init_vm, to ensure vm pasid is the original pasid if acquiring vm failed or is the compute pasid with pdd->drm_file reference taken to avoid double release same pasid. amdgpu: Failed to create process VM object ida_free called for id=32770 which is not allocated. WARNING: CPU: 57 PID: 72542 at ../lib/idr.c:522 ida_free+0x96/0x140 RIP: 0010:ida_free+0x96/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 BUG: kernel NULL pointer dereference, address: 0000000000000000 RIP: 0010:ida_free+0x76/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 Signed-off-by: Philip Yang Reviewed-by: Felix Kuehling Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 4 +- .../gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 41 +++++++++++++------ drivers/gpu/drm/amd/amdkfd/kfd_process.c | 12 ++++-- 3 files changed, 41 insertions(+), 16 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h index 589939631ed4..0040deaf8a83 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h @@ -270,8 +270,10 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_ (&((struct amdgpu_fpriv *) \ ((struct drm_file *)(drm_priv))->driver_priv)->vm) +int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev, + struct file *filp, u32 pasid); int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, - struct file *filp, u32 pasid, + struct file *filp, void **process_info, struct dma_fence **ef); void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c index 0a854bb8b47e..b15091d8310d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c @@ -1429,8 +1429,36 @@ static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo) amdgpu_bo_unreserve(bo); } +int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev, + struct file *filp, u32 pasid) + +{ + struct amdgpu_fpriv *drv_priv; + struct amdgpu_vm *avm; + int ret; + + ret = amdgpu_file_to_fpriv(filp, &drv_priv); + if (ret) + return ret; + avm = &drv_priv->vm; + + /* Free the original amdgpu allocated pasid, + * will be replaced with kfd allocated pasid. + */ + if (avm->pasid) { + amdgpu_pasid_free(avm->pasid); + amdgpu_vm_set_pasid(adev, avm, 0); + } + + ret = amdgpu_vm_set_pasid(adev, avm, pasid); + if (ret) + return ret; + + return 0; +} + int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, - struct file *filp, u32 pasid, + struct file *filp, void **process_info, struct dma_fence **ef) { @@ -1447,22 +1475,11 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev, if (avm->process_info) return -EINVAL; - /* Free the original amdgpu allocated pasid, - * will be replaced with kfd allocated pasid. - */ - if (avm->pasid) { - amdgpu_pasid_free(avm->pasid); - amdgpu_vm_set_pasid(adev, avm, 0); - } - /* Convert VM into a compute VM */ ret = amdgpu_vm_make_compute(adev, avm); if (ret) return ret; - ret = amdgpu_vm_set_pasid(adev, avm, pasid); - if (ret) - return ret; /* Initialize KFD part of the VM and process info */ ret = init_kfd_vm(avm, process_info, ef); if (ret) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process.c b/drivers/gpu/drm/amd/amdkfd/kfd_process.c index 6caa9dd57ff1..51b1683ac5c1 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process.c @@ -1576,9 +1576,9 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd, p = pdd->process; dev = pdd->dev; - ret = amdgpu_amdkfd_gpuvm_acquire_process_vm( - dev->adev, drm_file, p->pasid, - &p->kgd_process_info, &p->ef); + ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(dev->adev, drm_file, + &p->kgd_process_info, + &p->ef); if (ret) { pr_err("Failed to create process VM object\n"); return ret; @@ -1593,10 +1593,16 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd, if (ret) goto err_init_cwsr; + ret = amdgpu_amdkfd_gpuvm_set_vm_pasid(dev->adev, drm_file, p->pasid); + if (ret) + goto err_set_pasid; + pdd->drm_file = drm_file; return 0; +err_set_pasid: + kfd_process_device_destroy_cwsr_dgpu(pdd); err_init_cwsr: kfd_process_device_destroy_ib_mem(pdd); err_reserve_ib_mem: From d118b18fb1da02b41df2da78cb2794b3638d89cd Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Thu, 15 Dec 2022 17:36:31 +0100 Subject: [PATCH 102/115] drm/amd/pm: avoid large variable on kernel stack The activity_monitor_external[] array is too big to fit on the kernel stack, resulting in this warning with clang: drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu13/smu_v13_0_7_ppt.c:1438:12: error: stack frame size (1040) exceeds limit (1024) in 'smu_v13_0_7_get_power_profile_mode' [-Werror,-Wframe-larger-than] Use dynamic allocation instead. It should also be possible to have single element here instead of the array, but this seems easier. v2: fix up argument to sizeof() (Alex) Fixes: 334682ae8151 ("drm/amd/pm: enable workload type change on smu_v13_0_7") Signed-off-by: Arnd Bergmann Signed-off-by: Alex Deucher --- .../drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c | 21 ++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index ab1c004606be..07e3dc18c8b8 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -1440,7 +1440,7 @@ static int smu_v13_0_7_get_power_limit(struct smu_context *smu, static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf) { - DpmActivityMonitorCoeffIntExternal_t activity_monitor_external[PP_SMC_POWER_PROFILE_COUNT]; + DpmActivityMonitorCoeffIntExternal_t *activity_monitor_external; uint32_t i, j, size = 0; int16_t workload_type = 0; int result = 0; @@ -1448,6 +1448,12 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf if (!buf) return -EINVAL; + activity_monitor_external = kcalloc(PP_SMC_POWER_PROFILE_COUNT, + sizeof(*activity_monitor_external), + GFP_KERNEL); + if (!activity_monitor_external) + return -ENOMEM; + size += sysfs_emit_at(buf, size, " "); for (i = 0; i <= PP_SMC_POWER_PROFILE_WINDOW3D; i++) size += sysfs_emit_at(buf, size, "%-14s%s", amdgpu_pp_profile_name[i], @@ -1460,15 +1466,17 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf workload_type = smu_cmn_to_asic_specific_index(smu, CMN2ASIC_MAPPING_WORKLOAD, i); - if (workload_type < 0) - return -EINVAL; + if (workload_type < 0) { + result = -EINVAL; + goto out; + } result = smu_cmn_update_table(smu, SMU_TABLE_ACTIVITY_MONITOR_COEFF, workload_type, (void *)(&activity_monitor_external[i]), false); if (result) { dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__); - return result; + goto out; } } @@ -1496,7 +1504,10 @@ do { \ PRINT_DPM_MONITOR(Fclk_BoosterFreq); #undef PRINT_DPM_MONITOR - return size; + result = size; +out: + kfree(activity_monitor_external); + return result; } static int smu_v13_0_7_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size) From afa6646b1c5d3affd541f76bd7476e4b835a9174 Mon Sep 17 00:00:00 2001 From: Alex Deucher Date: Fri, 16 Dec 2022 11:42:20 -0500 Subject: [PATCH 103/115] drm/amdgpu: skip MES for S0ix as well since it's part of GFX It's also part of gfxoff. Cc: stable@vger.kernel.org # 6.0, 6.1 Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 64660a41d53c..afe6af9c0138 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -3016,14 +3016,15 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev) continue; } - /* skip suspend of gfx and psp for S0ix + /* skip suspend of gfx/mes and psp for S0ix * gfx is in gfxoff state, so on resume it will exit gfxoff just * like at runtime. PSP is also part of the always on hardware * so no need to suspend it. */ if (adev->in_s0ix && (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP || - adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX)) + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX || + adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_MES)) continue; /* XXX handle errors */ From 272b981416f8be0180c4d8066f90635fa7c1c501 Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Thu, 15 Dec 2022 13:38:46 +0800 Subject: [PATCH 104/115] drm/amd/pm: bump SMU13.0.0 driver_if header to version 0x34 To fit the latest PMFW and suppress the warning emerged on driver loading. Signed-off-by: Evan Quan Acked-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0, 6.1 --- .../gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h | 2 +- drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h | 1 + drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c | 2 ++ 3 files changed, 4 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h index b76f0f7e4299..d6b964cf73bd 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h @@ -522,9 +522,9 @@ typedef enum { TEMP_HOTSPOT_M, TEMP_MEM, TEMP_VR_GFX, + TEMP_VR_SOC, TEMP_VR_MEM0, TEMP_VR_MEM1, - TEMP_VR_SOC, TEMP_VR_U, TEMP_LIQUID0, TEMP_LIQUID1, diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h index ea29ac6a80e6..e8c6febb8b64 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h @@ -28,6 +28,7 @@ #define SMU13_DRIVER_IF_VERSION_INV 0xFFFFFFFF #define SMU13_DRIVER_IF_VERSION_YELLOW_CARP 0x04 #define SMU13_DRIVER_IF_VERSION_ALDE 0x08 +#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0 0x34 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_4 0x07 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_5 0x04 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10 0x32 diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c index e3a80ac987df..e54b760b875b 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c @@ -290,6 +290,8 @@ int smu_v13_0_check_fw_version(struct smu_context *smu) smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_ALDE; break; case IP_VERSION(13, 0, 0): + smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0; + break; case IP_VERSION(13, 0, 10): smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10; break; From e73fc71e8f015d61f3adca7659cb209fd5117aa5 Mon Sep 17 00:00:00 2001 From: Evan Quan Date: Fri, 16 Dec 2022 17:04:24 +0800 Subject: [PATCH 105/115] drm/amd/pm: correct the fan speed retrieving in PWM for some SMU13 asics For SMU 13.0.0 and 13.0.7, the output from PMFW is in percent. Driver need to convert that into correct PMW(255) based. Signed-off-by: Evan Quan Acked-by: Alex Deucher Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0, 6.1 --- .../drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c | 17 ++++++++++++++--- .../drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c | 17 ++++++++++++++--- 2 files changed, 28 insertions(+), 6 deletions(-) diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c index 713fb6ad39f6..9643b21c636a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c @@ -1441,12 +1441,23 @@ out: static int smu_v13_0_0_get_fan_speed_pwm(struct smu_context *smu, uint32_t *speed) { + int ret; + if (!speed) return -EINVAL; - return smu_v13_0_0_get_smu_metrics_data(smu, - METRICS_CURR_FANPWM, - speed); + ret = smu_v13_0_0_get_smu_metrics_data(smu, + METRICS_CURR_FANPWM, + speed); + if (ret) { + dev_err(smu->adev->dev, "Failed to get fan speed(PWM)!"); + return ret; + } + + /* Convert the PMFW output which is in percent to pwm(255) based */ + *speed = MIN(*speed * 255 / 100, 255); + + return 0; } static int smu_v13_0_0_get_fan_speed_rpm(struct smu_context *smu, diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c index 07e3dc18c8b8..5c6c6ad011ca 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c @@ -1363,12 +1363,23 @@ static int smu_v13_0_7_populate_umd_state_clk(struct smu_context *smu) static int smu_v13_0_7_get_fan_speed_pwm(struct smu_context *smu, uint32_t *speed) { + int ret; + if (!speed) return -EINVAL; - return smu_v13_0_7_get_smu_metrics_data(smu, - METRICS_CURR_FANPWM, - speed); + ret = smu_v13_0_7_get_smu_metrics_data(smu, + METRICS_CURR_FANPWM, + speed); + if (ret) { + dev_err(smu->adev->dev, "Failed to get fan speed(PWM)!"); + return ret; + } + + /* Convert the PMFW output which is in percent to pwm(255) based */ + *speed = MIN(*speed * 255 / 100, 255); + + return 0; } static int smu_v13_0_7_get_fan_speed_rpm(struct smu_context *smu, From 8660495a9c5b9afeec4cc006b3b75178f0fb2f10 Mon Sep 17 00:00:00 2001 From: Tim Huang Date: Mon, 19 Dec 2022 18:32:32 +0800 Subject: [PATCH 106/115] drm/amdgpu: skip mes self test after s0i3 resume for MES IP v11.0 MES is part of gfxoff and MES suspend and resume are skipped for S0i3. But the mes_self_test call path is still in the amdgpu_device_ip_late_init. it's should also be skipped for s0ix as no hardware re-initialization happened. Besides, mes_self_test will free the BO that triggers a lot of warning messages while in the suspend state. [ 81.656085] WARNING: CPU: 2 PID: 1550 at drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:425 amdgpu_bo_free_kernel+0xfc/0x110 [amdgpu] [ 81.679435] Call Trace: [ 81.679726] [ 81.679981] amdgpu_mes_remove_hw_queue+0x17a/0x230 [amdgpu] [ 81.680857] amdgpu_mes_self_test+0x390/0x430 [amdgpu] [ 81.681665] mes_v11_0_late_init+0x37/0x50 [amdgpu] [ 81.682423] amdgpu_device_ip_late_init+0x53/0x280 [amdgpu] [ 81.683257] amdgpu_device_resume+0xae/0x2a0 [amdgpu] [ 81.684043] amdgpu_pmops_resume+0x37/0x70 [amdgpu] [ 81.684818] pci_pm_resume+0x5c/0xa0 [ 81.685247] ? pci_pm_thaw+0x90/0x90 [ 81.685658] dpm_run_callback+0x4e/0x160 [ 81.686110] device_resume+0xad/0x210 [ 81.686529] async_resume+0x1e/0x40 [ 81.686931] async_run_entry_fn+0x33/0x120 [ 81.687405] process_one_work+0x21d/0x3f0 [ 81.687869] worker_thread+0x4a/0x3c0 [ 81.688293] ? process_one_work+0x3f0/0x3f0 [ 81.688777] kthread+0xff/0x130 [ 81.689157] ? kthread_complete_and_exit+0x20/0x20 [ 81.689707] ret_from_fork+0x22/0x30 [ 81.690118] [ 81.690380] ---[ end trace 0000000000000000 ]--- v2: make the comment clean and use adev->in_s0ix instead of adev->suspend Signed-off-by: Tim Huang Reviewed-by: Mario Limonciello Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0, 6.1 --- drivers/gpu/drm/amd/amdgpu/mes_v11_0.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c index 5459366f49ff..970b066b37bb 100644 --- a/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/mes_v11_0.c @@ -1342,7 +1342,8 @@ static int mes_v11_0_late_init(void *handle) { struct amdgpu_device *adev = (struct amdgpu_device *)handle; - if (!amdgpu_in_reset(adev) && + /* it's only intended for use in mes_self_test case, not for s0ix and reset */ + if (!amdgpu_in_reset(adev) && !adev->in_s0ix && (adev->ip_versions[GC_HWIP][0] != IP_VERSION(11, 0, 3))) amdgpu_mes_self_test(adev); From db918321275d648294a3fb2bdefeb99e33f00467 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 20 Dec 2022 09:05:36 +0100 Subject: [PATCH 107/115] ALSA: memalloc: don't use GFP_COMP for non-coherent dma allocations While not quite as bogus as for the dma-coherent allocations that were fixed earlier, GFP_COMP for these allocations has no benefits for the dma-direct case, and can't be supported at all by dma dma-iommu backend which splits up allocations into smaller orders. Due to an oversight in ffcb75458460 that flag stopped being cleared for all dma allocations, but only got rejected for coherent ones. Start fixing this by not requesting __GFP_COMP in the sound code, which is the only place that did this. Fixes: ffcb75458460 ("dma-mapping: reject __GFP_COMP in dma_alloc_attrs") Reported-by: Mikhail Gavrilov Reported-by: Kai Vehmanen Signed-off-by: Christoph Hellwig Acked-by: Takashi Iwai Tested-by: Mikhail Gavrilov Tested-by: Kai Vehmanen --- sound/core/memalloc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/sound/core/memalloc.c b/sound/core/memalloc.c index 3cf5a87d69ea..81025f50a542 100644 --- a/sound/core/memalloc.c +++ b/sound/core/memalloc.c @@ -542,7 +542,7 @@ static void *snd_dma_noncontig_alloc(struct snd_dma_buffer *dmab, size_t size) void *p; sgt = dma_alloc_noncontiguous(dmab->dev.dev, size, dmab->dev.dir, - DEFAULT_GFP | __GFP_COMP, 0); + DEFAULT_GFP, 0); #ifdef CONFIG_SND_DMA_SGBUF if (!sgt && !get_dma_ops(dmab->dev.dev)) { if (dmab->dev.type == SNDRV_DMA_TYPE_DEV_WC_SG) @@ -820,7 +820,7 @@ static void *snd_dma_noncoherent_alloc(struct snd_dma_buffer *dmab, size_t size) void *p; p = dma_alloc_noncoherent(dmab->dev.dev, size, &dmab->addr, - dmab->dev.dir, DEFAULT_GFP | __GFP_COMP); + dmab->dev.dir, DEFAULT_GFP); if (p) dmab->dev.need_sync = dma_need_sync(dmab->dev.dev, dmab->addr); return p; From 3622b86f49f84e52fb41fee9eb55f9290613dfc3 Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Tue, 20 Dec 2022 09:08:12 +0100 Subject: [PATCH 108/115] dma-mapping: reject GFP_COMP for noncoherent allocations While not quite as bogus as for the dma-coherent allocations that were fixed earlier, GFP_COMP for these allocations has no benefits for the dma-direct case, and can't be supported at all by dma dma-iommu backend which splits up allocations into smaller orders. Due to an oversight in ffcb75458460 that flag stopped being cleared for all dma allocations, but only got rejected for coherent ones, so fix up these callers to not allow __GFP_COMP as well after the sound code has been fixed to not ask for it. Fixes: ffcb75458460 ("dma-mapping: reject __GFP_COMP in dma_alloc_attrs") Reported-by: Mikhail Gavrilov Reported-by: Kai Vehmanen Signed-off-by: Christoph Hellwig Acked-by: Takashi Iwai Tested-by: Mikhail Gavrilov Tested-by: Kai Vehmanen --- kernel/dma/mapping.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/kernel/dma/mapping.c b/kernel/dma/mapping.c index c026a5a5e046..68106e3791f6 100644 --- a/kernel/dma/mapping.c +++ b/kernel/dma/mapping.c @@ -560,6 +560,8 @@ static struct page *__dma_alloc_pages(struct device *dev, size_t size, return NULL; if (WARN_ON_ONCE(gfp & (__GFP_DMA | __GFP_DMA32 | __GFP_HIGHMEM))) return NULL; + if (WARN_ON_ONCE(gfp & __GFP_COMP)) + return NULL; size = PAGE_ALIGN(size); if (dma_alloc_direct(dev, ops)) @@ -645,6 +647,8 @@ struct sg_table *dma_alloc_noncontiguous(struct device *dev, size_t size, if (WARN_ON_ONCE(attrs & ~DMA_ATTR_ALLOC_SINGLE_PAGES)) return NULL; + if (WARN_ON_ONCE(gfp & __GFP_COMP)) + return NULL; if (ops && ops->alloc_noncontiguous) sgt = ops->alloc_noncontiguous(dev, size, dir, gfp, attrs); From e1d900df63adcb748905131dd6258e570e11aed1 Mon Sep 17 00:00:00 2001 From: Saleemkhan Jamadar Date: Tue, 20 Dec 2022 13:21:44 +0530 Subject: [PATCH 109/115] drm/amdgpu: enable VCN DPG for GC IP v11.0.4 Enable VCN Dynamic Power Gating control for GC IP v11.0.4. Signed-off-by: Saleemkhan Jamadar Reviewed-by: Veerabadhran Gopalakrishnan Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org # 6.0, 6.1 --- drivers/gpu/drm/amd/amdgpu/soc21.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/amd/amdgpu/soc21.c b/drivers/gpu/drm/amd/amdgpu/soc21.c index 7d5fdf450d0c..5562670b7b52 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc21.c +++ b/drivers/gpu/drm/amd/amdgpu/soc21.c @@ -666,6 +666,7 @@ static int soc21_common_early_init(void *handle) AMD_CG_SUPPORT_VCN_MGCG | AMD_CG_SUPPORT_JPEG_MGCG; adev->pg_flags = AMD_PG_SUPPORT_VCN | + AMD_PG_SUPPORT_VCN_DPG | AMD_PG_SUPPORT_GFX_PG | AMD_PG_SUPPORT_JPEG; adev->external_rev_id = adev->rev_id + 0x1; From c1c4a8b217213c1924eabf4f28385bbee9cc50c1 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Christian=20K=C3=B6nig?= Date: Mon, 19 Dec 2022 11:47:18 +0100 Subject: [PATCH 110/115] drm/amdgpu: grab extra fence reference for drm_sched_job_add_dependency MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit That function consumes the reference. Reviewed-by: Luben Tuikov Reported-by: Borislav Petkov (AMD) Tested-by: Borislav Petkov (AMD) Signed-off-by: Christian König Fixes: aab9cf7b6954 ("drm/amdgpu: use scheduler dependencies for VM updates") Signed-off-by: Alex Deucher --- drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c index 59cf64216fbb..535cd6569bcc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c @@ -238,8 +238,10 @@ static int amdgpu_vm_sdma_update(struct amdgpu_vm_update_params *p, /* Wait for PD/PT moves to be completed */ dma_resv_iter_begin(&cursor, bo->tbo.base.resv, DMA_RESV_USAGE_KERNEL); dma_resv_for_each_fence_unlocked(&cursor, fence) { + dma_fence_get(fence); r = drm_sched_job_add_dependency(&p->job->base, fence); if (r) { + dma_fence_put(fence); dma_resv_iter_end(&cursor); return r; } From a95e163a4bfa7780f64e589bbedc6bdeb7cf3839 Mon Sep 17 00:00:00 2001 From: Jiapeng Chong Date: Tue, 13 Dec 2022 14:13:55 +0800 Subject: [PATCH 111/115] ALSA: azt3328: Remove the unused function snd_azf3328_codec_outl() The function snd_azf3328_codec_outl is defined in the azt3328.c file, but not called elsewhere, so remove this unused function. sound/pci/azt3328.c:367:1: warning: unused function 'snd_azf3328_codec_outl'. Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3432 Reported-by: Abaci Robot Signed-off-by: Jiapeng Chong Link: https://lore.kernel.org/r/20221213061355.62856-1-jiapeng.chong@linux.alibaba.com Signed-off-by: Takashi Iwai --- sound/pci/azt3328.c | 9 --------- 1 file changed, 9 deletions(-) diff --git a/sound/pci/azt3328.c b/sound/pci/azt3328.c index 7f329dfc5404..0c6754bf9455 100644 --- a/sound/pci/azt3328.c +++ b/sound/pci/azt3328.c @@ -363,15 +363,6 @@ snd_azf3328_codec_inw(const struct snd_azf3328_codec_data *codec, unsigned reg) return inw(codec->io_base + reg); } -static inline void -snd_azf3328_codec_outl(const struct snd_azf3328_codec_data *codec, - unsigned reg, - u32 value -) -{ - outl(value, codec->io_base + reg); -} - static inline void snd_azf3328_codec_outl_multi(const struct snd_azf3328_codec_data *codec, unsigned reg, const void *buffer, int count From fd28941cff1cd9d8ffa59fe11eb64148e09b6ed6 Mon Sep 17 00:00:00 2001 From: Jaroslav Kysela Date: Thu, 15 Dec 2022 16:30:37 +0100 Subject: [PATCH 112/115] ALSA: usb-audio: Add new quirk FIXED_RATE for JBL Quantum810 Wireless It seems that the firmware is broken and does not accept the UAC_EP_CS_ATTR_SAMPLE_RATE URB. There is only one rate (48000Hz) available in the descriptors for the output endpoint. Create a new quirk QUIRK_FLAG_FIXED_RATE to skip the rate setup when only one rate is available (fixed). BugLink: https://bugzilla.kernel.org/show_bug.cgi?id=216798 Signed-off-by: Jaroslav Kysela Link: https://lore.kernel.org/r/20221215153037.1163786-1-perex@perex.cz Signed-off-by: Takashi Iwai --- sound/usb/card.h | 1 + sound/usb/endpoint.c | 16 ++++++++++------ sound/usb/endpoint.h | 3 ++- sound/usb/implicit.c | 6 +++++- sound/usb/implicit.h | 2 +- sound/usb/pcm.c | 36 +++++++++++++++++++++++++++++++++--- sound/usb/pcm.h | 2 ++ sound/usb/quirks.c | 2 ++ sound/usb/usbaudio.h | 4 ++++ 9 files changed, 60 insertions(+), 12 deletions(-) diff --git a/sound/usb/card.h b/sound/usb/card.h index 40061550105a..6ec95b2edf86 100644 --- a/sound/usb/card.h +++ b/sound/usb/card.h @@ -131,6 +131,7 @@ struct snd_usb_endpoint { bool lowlatency_playback; /* low-latency playback mode */ bool need_setup; /* (re-)need for hw_params? */ bool need_prepare; /* (re-)need for prepare? */ + bool fixed_rate; /* skip rate setup */ /* for hw constraints */ const struct audioformat *cur_audiofmt; diff --git a/sound/usb/endpoint.c b/sound/usb/endpoint.c index 4aaf0784940b..419302e2057e 100644 --- a/sound/usb/endpoint.c +++ b/sound/usb/endpoint.c @@ -769,7 +769,8 @@ struct snd_usb_endpoint * snd_usb_endpoint_open(struct snd_usb_audio *chip, const struct audioformat *fp, const struct snd_pcm_hw_params *params, - bool is_sync_ep) + bool is_sync_ep, + bool fixed_rate) { struct snd_usb_endpoint *ep; int ep_num = is_sync_ep ? fp->sync_ep : fp->endpoint; @@ -825,6 +826,7 @@ snd_usb_endpoint_open(struct snd_usb_audio *chip, ep->implicit_fb_sync = fp->implicit_fb; ep->need_setup = true; ep->need_prepare = true; + ep->fixed_rate = fixed_rate; usb_audio_dbg(chip, " channels=%d, rate=%d, format=%s, period_bytes=%d, periods=%d, implicit_fb=%d\n", ep->cur_channels, ep->cur_rate, @@ -1413,11 +1415,13 @@ static int init_sample_rate(struct snd_usb_audio *chip, if (clock && !clock->need_setup) return 0; - err = snd_usb_init_sample_rate(chip, ep->cur_audiofmt, rate); - if (err < 0) { - if (clock) - clock->rate = 0; /* reset rate */ - return err; + if (!ep->fixed_rate) { + err = snd_usb_init_sample_rate(chip, ep->cur_audiofmt, rate); + if (err < 0) { + if (clock) + clock->rate = 0; /* reset rate */ + return err; + } } if (clock) diff --git a/sound/usb/endpoint.h b/sound/usb/endpoint.h index e67ea28faa54..924f4351588c 100644 --- a/sound/usb/endpoint.h +++ b/sound/usb/endpoint.h @@ -14,7 +14,8 @@ struct snd_usb_endpoint * snd_usb_endpoint_open(struct snd_usb_audio *chip, const struct audioformat *fp, const struct snd_pcm_hw_params *params, - bool is_sync_ep); + bool is_sync_ep, + bool fixed_rate); void snd_usb_endpoint_close(struct snd_usb_audio *chip, struct snd_usb_endpoint *ep); int snd_usb_endpoint_set_params(struct snd_usb_audio *chip, diff --git a/sound/usb/implicit.c b/sound/usb/implicit.c index f3e8484b3d9c..41ac7185b42b 100644 --- a/sound/usb/implicit.c +++ b/sound/usb/implicit.c @@ -15,6 +15,7 @@ #include "usbaudio.h" #include "card.h" #include "helper.h" +#include "pcm.h" #include "implicit.h" enum { @@ -455,7 +456,8 @@ const struct audioformat * snd_usb_find_implicit_fb_sync_format(struct snd_usb_audio *chip, const struct audioformat *target, const struct snd_pcm_hw_params *params, - int stream) + int stream, + bool *fixed_rate) { struct snd_usb_substream *subs; const struct audioformat *fp, *sync_fmt = NULL; @@ -483,6 +485,8 @@ snd_usb_find_implicit_fb_sync_format(struct snd_usb_audio *chip, } } + if (fixed_rate) + *fixed_rate = snd_usb_pcm_has_fixed_rate(subs); return sync_fmt; } diff --git a/sound/usb/implicit.h b/sound/usb/implicit.h index ccb415a0ea86..7f1577b6c4d3 100644 --- a/sound/usb/implicit.h +++ b/sound/usb/implicit.h @@ -9,6 +9,6 @@ const struct audioformat * snd_usb_find_implicit_fb_sync_format(struct snd_usb_audio *chip, const struct audioformat *target, const struct snd_pcm_hw_params *params, - int stream); + int stream, bool *fixed_rate); #endif /* __USBAUDIO_IMPLICIT_H */ diff --git a/sound/usb/pcm.c b/sound/usb/pcm.c index 9557bd4d1bbc..99a66d0ef5b2 100644 --- a/sound/usb/pcm.c +++ b/sound/usb/pcm.c @@ -157,6 +157,31 @@ find_substream_format(struct snd_usb_substream *subs, true, subs); } +bool snd_usb_pcm_has_fixed_rate(struct snd_usb_substream *subs) +{ + const struct audioformat *fp; + struct snd_usb_audio *chip = subs->stream->chip; + int rate = -1; + + if (!(chip->quirk_flags & QUIRK_FLAG_FIXED_RATE)) + return false; + list_for_each_entry(fp, &subs->fmt_list, list) { + if (fp->rates & SNDRV_PCM_RATE_CONTINUOUS) + return false; + if (fp->nr_rates < 1) + continue; + if (fp->nr_rates > 1) + return false; + if (rate < 0) { + rate = fp->rate_table[0]; + continue; + } + if (rate != fp->rate_table[0]) + return false; + } + return true; +} + static int init_pitch_v1(struct snd_usb_audio *chip, int ep) { struct usb_device *dev = chip->dev; @@ -450,12 +475,14 @@ static int snd_usb_hw_params(struct snd_pcm_substream *substream, struct snd_usb_audio *chip = subs->stream->chip; const struct audioformat *fmt; const struct audioformat *sync_fmt; + bool fixed_rate, sync_fixed_rate; int ret; ret = snd_media_start_pipeline(subs); if (ret) return ret; + fixed_rate = snd_usb_pcm_has_fixed_rate(subs); fmt = find_substream_format(subs, hw_params); if (!fmt) { usb_audio_dbg(chip, @@ -469,7 +496,8 @@ static int snd_usb_hw_params(struct snd_pcm_substream *substream, if (fmt->implicit_fb) { sync_fmt = snd_usb_find_implicit_fb_sync_format(chip, fmt, hw_params, - !substream->stream); + !substream->stream, + &sync_fixed_rate); if (!sync_fmt) { usb_audio_dbg(chip, "cannot find sync format: ep=0x%x, iface=%d:%d, format=%s, rate=%d, channels=%d\n", @@ -482,6 +510,7 @@ static int snd_usb_hw_params(struct snd_pcm_substream *substream, } } else { sync_fmt = fmt; + sync_fixed_rate = fixed_rate; } ret = snd_usb_lock_shutdown(chip); @@ -499,7 +528,7 @@ static int snd_usb_hw_params(struct snd_pcm_substream *substream, close_endpoints(chip, subs); } - subs->data_endpoint = snd_usb_endpoint_open(chip, fmt, hw_params, false); + subs->data_endpoint = snd_usb_endpoint_open(chip, fmt, hw_params, false, fixed_rate); if (!subs->data_endpoint) { ret = -EINVAL; goto unlock; @@ -508,7 +537,8 @@ static int snd_usb_hw_params(struct snd_pcm_substream *substream, if (fmt->sync_ep) { subs->sync_endpoint = snd_usb_endpoint_open(chip, sync_fmt, hw_params, - fmt == sync_fmt); + fmt == sync_fmt, + sync_fixed_rate); if (!subs->sync_endpoint) { ret = -EINVAL; goto unlock; diff --git a/sound/usb/pcm.h b/sound/usb/pcm.h index 493a4e34d78d..388fe2ba346d 100644 --- a/sound/usb/pcm.h +++ b/sound/usb/pcm.h @@ -6,6 +6,8 @@ void snd_usb_set_pcm_ops(struct snd_pcm *pcm, int stream); int snd_usb_pcm_suspend(struct snd_usb_stream *as); int snd_usb_pcm_resume(struct snd_usb_stream *as); +bool snd_usb_pcm_has_fixed_rate(struct snd_usb_substream *as); + int snd_usb_init_pitch(struct snd_usb_audio *chip, const struct audioformat *fmt); void snd_usb_preallocate_buffer(struct snd_usb_substream *subs); diff --git a/sound/usb/quirks.c b/sound/usb/quirks.c index 58b37bfc885c..3d13fdf7590c 100644 --- a/sound/usb/quirks.c +++ b/sound/usb/quirks.c @@ -2152,6 +2152,8 @@ static const struct usb_audio_quirk_flags_table quirk_flags_table[] = { QUIRK_FLAG_GENERIC_IMPLICIT_FB), DEVICE_FLG(0x0525, 0xa4ad, /* Hamedal C20 usb camero */ QUIRK_FLAG_IFACE_SKIP_CLOSE), + DEVICE_FLG(0x0ecb, 0x2069, /* JBL Quantum810 Wireless */ + QUIRK_FLAG_FIXED_RATE), /* Vendor matches */ VENDOR_FLG(0x045e, /* MS Lifecam */ diff --git a/sound/usb/usbaudio.h b/sound/usb/usbaudio.h index 2aba508a4831..f5a8dca66457 100644 --- a/sound/usb/usbaudio.h +++ b/sound/usb/usbaudio.h @@ -175,6 +175,9 @@ extern bool snd_usb_skip_validation; * QUIRK_FLAG_FORCE_IFACE_RESET * Force an interface reset whenever stopping & restarting a stream * (e.g. after xrun) + * QUIRK_FLAG_FIXED_RATE + * Do not set PCM rate (frequency) when only one rate is available + * for the given endpoint. */ #define QUIRK_FLAG_GET_SAMPLE_RATE (1U << 0) @@ -198,5 +201,6 @@ extern bool snd_usb_skip_validation; #define QUIRK_FLAG_SKIP_IMPLICIT_FB (1U << 18) #define QUIRK_FLAG_IFACE_SKIP_CLOSE (1U << 19) #define QUIRK_FLAG_FORCE_IFACE_RESET (1U << 20) +#define QUIRK_FLAG_FIXED_RATE (1U << 21) #endif /* __USBAUDIO_H */ From cf8016408d880afe9c5dc495af40dc2932874e77 Mon Sep 17 00:00:00 2001 From: Sami Tolvanen Date: Thu, 22 Dec 2022 22:57:47 +0000 Subject: [PATCH 113/115] cfi: Fix CFI failure with KASAN When CFI_CLANG and KASAN are both enabled, LLVM doesn't generate a CFI type hash for asan.module_ctor functions in translation units where CFI is disabled, which leads to a CFI failure during boot when do_ctors calls the affected constructors: CFI failure at do_basic_setup+0x64/0x90 (target: asan.module_ctor+0x0/0x28; expected type: 0xa540670c) Specifically, this happens because CFI is disabled for kernel/cfi.c. There's no reason to keep CFI disabled here anymore, so fix the failure by not filtering out CC_FLAGS_CFI for the file. Note that https://reviews.llvm.org/rG3b14862f0a96 fixed the issue where LLVM didn't emit CFI type hashes for any sanitizer constructors, but now type hashes are emitted correctly for TUs that use CFI. Link: https://github.com/ClangBuiltLinux/linux/issues/1742 Fixes: 89245600941e ("cfi: Switch to -fsanitize=kcfi") Reported-by: Mark Rutland Signed-off-by: Sami Tolvanen Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20221222225747.3538676-1-samitolvanen@google.com --- kernel/Makefile | 3 --- 1 file changed, 3 deletions(-) diff --git a/kernel/Makefile b/kernel/Makefile index d754e0be1176..ebc692242b68 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -41,9 +41,6 @@ UBSAN_SANITIZE_kcov.o := n KMSAN_SANITIZE_kcov.o := n CFLAGS_kcov.o := $(call cc-option, -fno-conserve-stack) -fno-stack-protector -# Don't instrument error handlers -CFLAGS_REMOVE_cfi.o := $(CC_FLAGS_CFI) - obj-y += sched/ obj-y += locking/ obj-y += power/ From 2f4fec5943407318b9523f01ce1f5d668c028332 Mon Sep 17 00:00:00 2001 From: John Stultz Date: Wed, 21 Dec 2022 05:18:55 +0000 Subject: [PATCH 114/115] pstore: Make sure CONFIG_PSTORE_PMSG selects CONFIG_RT_MUTEXES In commit 76d62f24db07 ("pstore: Switch pmsg_lock to an rt_mutex to avoid priority inversion") I changed a lock to an rt_mutex. However, its possible that CONFIG_RT_MUTEXES is not enabled, which then results in a build failure, as the 0day bot detected: https://lore.kernel.org/linux-mm/202212211244.TwzWZD3H-lkp@intel.com/ Thus this patch changes CONFIG_PSTORE_PMSG to select CONFIG_RT_MUTEXES, which ensures the build will not fail. Cc: Wei Wang Cc: Midas Chien Cc: Connor O'Brien Cc: Kees Cook Cc: Anton Vorontsov Cc: Colin Cross Cc: Tony Luck Cc: kernel test robot Cc: kernel-team@android.com Fixes: 76d62f24db07 ("pstore: Switch pmsg_lock to an rt_mutex to avoid priority inversion") Reported-by: kernel test robot Signed-off-by: John Stultz Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20221221051855.15761-1-jstultz@google.com --- fs/pstore/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/fs/pstore/Kconfig b/fs/pstore/Kconfig index 8adabde685f1..c49d554cc9ae 100644 --- a/fs/pstore/Kconfig +++ b/fs/pstore/Kconfig @@ -126,6 +126,7 @@ config PSTORE_CONSOLE config PSTORE_PMSG bool "Log user space messages" depends on PSTORE + select RT_MUTEXES help When the option is enabled, pstore will export a character interface /dev/pmsg0 to log user space messages. On reboot From beca3e311a49cd3c55a056096531737d7afa4361 Mon Sep 17 00:00:00 2001 From: Luca Stefani Date: Thu, 22 Dec 2022 14:10:49 +0100 Subject: [PATCH 115/115] pstore: Properly assign mem_type property If mem-type is specified in the device tree it would end up overriding the record_size field instead of populating mem_type. As record_size is currently parsed after the improper assignment with default size 0 it continued to work as expected regardless of the value found in the device tree. Simply changing the target field of the struct is enough to get mem-type working as expected. Fixes: 9d843e8fafc7 ("pstore: Add mem_type property DT parsing support") Cc: stable@vger.kernel.org Signed-off-by: Luca Stefani Signed-off-by: Kees Cook Link: https://lore.kernel.org/r/20221222131049.286288-1-luca@osomprivacy.com --- fs/pstore/ram.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fs/pstore/ram.c b/fs/pstore/ram.c index 9a5052431fd3..ade66dbe5f39 100644 --- a/fs/pstore/ram.c +++ b/fs/pstore/ram.c @@ -680,7 +680,7 @@ static int ramoops_parse_dt(struct platform_device *pdev, field = value; \ } - parse_u32("mem-type", pdata->record_size, pdata->mem_type); + parse_u32("mem-type", pdata->mem_type, pdata->mem_type); parse_u32("record-size", pdata->record_size, 0); parse_u32("console-size", pdata->console_size, 0); parse_u32("ftrace-size", pdata->ftrace_size, 0);