The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 25dc194b34dd5919dd07b8873ee338182e15df9d Mon Sep 17 00:00:00 2001
From: Nicholas Kazlauskas <nicholas.kazlauskas(a)amd.com>
Date: Mon, 7 Jan 2019 12:41:46 -0500
Subject: [PATCH] drm: Block fb changes for async plane updates
The prepare_fb call always happens on new_plane_state.
The drm_atomic_helper_cleanup_planes checks to see if
plane state pointer has changed when deciding to call cleanup_fb on
either the new_plane_state or the old_plane_state.
For a non-async atomic commit the state pointer is swapped, so this
helper calls prepare_fb on the new_plane_state and cleanup_fb on the
old_plane_state. This makes sense, since we want to prepare the
framebuffer we are going to use and cleanup the the framebuffer we are
no longer using.
For the async atomic update helpers this differs. The async atomic
update helpers perform in-place updates on the existing state. They call
drm_atomic_helper_cleanup_planes but the state pointer is not swapped.
This means that prepare_fb is called on the new_plane_state and
cleanup_fb is called on the new_plane_state (not the old).
In the case where old_plane_state->fb == new_plane_state->fb then
there should be no behavioral difference between an async update
and a non-async commit. But there are issues that arise when
old_plane_state->fb != new_plane_state->fb.
The first is that the new_plane_state->fb is immediately cleaned up
after it has been prepared, so we're using a fb that we shouldn't
be.
The second occurs during a sequence of async atomic updates and
non-async regular atomic commits. Suppose there are two framebuffers
being interleaved in a double-buffering scenario, fb1 and fb2:
- Async update, oldfb = NULL, newfb = fb1, prepare fb1, cleanup fb1
- Async update, oldfb = fb1, newfb = fb2, prepare fb2, cleanup fb2
- Non-async commit, oldfb = fb2, newfb = fb1, prepare fb1, cleanup fb2
We call cleanup_fb on fb2 twice in this example scenario, and any
further use will result in use-after-free.
The simple fix to this problem is to block framebuffer changes
in the drm_atomic_helper_async_check function for now.
v2: Move check by itself, add a FIXME (Daniel)
Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch>
Cc: Harry Wentland <harry.wentland(a)amd.com>
Cc: Andrey Grodzovsky <andrey.grodzovsky(a)amd.com>
Cc: <stable(a)vger.kernel.org> # v4.14+
Fixes: fef9df8b5945 ("drm/atomic: initial support for asynchronous plane update")
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas(a)amd.com>
Acked-by: Andrey Grodzovsky <andrey.grodzovsky(a)amd.com>
Acked-by: Harry Wentland <harry.wentland(a)amd.com>
Reviewed-by: Daniel Vetter <daniel(a)ffwll.ch>
Signed-off-by: Harry Wentland <harry.wentland(a)amd.com>
Link: https://patchwork.freedesktop.org/patch/275364/
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index 54e2ae614dcc..f4290f6b0c38 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -1602,6 +1602,15 @@ int drm_atomic_helper_async_check(struct drm_device *dev,
old_plane_state->crtc != new_plane_state->crtc)
return -EINVAL;
+ /*
+ * FIXME: Since prepare_fb and cleanup_fb are always called on
+ * the new_plane_state for async updates we need to block framebuffer
+ * changes. This prevents use of a fb that's been cleaned up and
+ * double cleanups from occuring.
+ */
+ if (old_plane_state->fb != new_plane_state->fb)
+ return -EINVAL;
+
funcs = plane->helper_private;
if (!funcs->atomic_async_update)
return -EINVAL;
Hi Greg and Sasha,
Apologies if this has already been reported but our CI noticed a build
error on arm64 on 4.9.163 due to commit 1e8e52342060 ("net: hns: Fix
for missing of_node_put() after of_parse_phandle()").
https://travis-ci.com/ClangBuiltLinux/continuous-integration/jobs/185127507…
Commit d20dc1493db4 ("of: Support const and non-const use for
to_of_node()") will resolve this and it applies cleanly with a
three way merge. I did a local test and it doesn't appear to
introduce any other build regressions I've added the
author/reviewer/maintainer to see if they have any objections.
Thanks!
Nathan
Hi,
I'm facing a bug on the latest stable 4.19.29 driver i40e which had been fixed on later version (commit id cdec2141c24ef177d929765c5a6f95549c266fb3)
The patch applies on 4.9.29 with hunks but I've tested and it fixes the issue.
Regards,
Emeric Brun
Commit 5c27ff5db1491a947264d6d4e4cbe43ae6535bae upstream.
I have encountered an interrupt storm during the eMMC chip probing (and
the chip finally didn't get detected). It turned out that U-Boot left
the DMAC interrupts enabled while the Linux driver didn't use those.
The SDHI driver's interrupt handler somehow assumes that, even if an
SDIO interrupt didn't happen, it should return IRQ_HANDLED. I think
that if none of the enabled interrupts happened and got handled, we
should return IRQ_NONE -- that way the kernel IRQ code recoginizes
a spurious interrupt and masks it off pretty quickly...
Fixes: 7729c7a232a9 ("mmc: tmio: Provide separate interrupt handlers")
Signed-off-by: Sergei Shtylyov <sergei.shtylyov(a)cogentembedded.com>
---
The patch is against the 'linux-4.9.y' branch of the -stable repo.
drivers/mmc/host/tmio_mmc_pio.c | 11 +++++++----
1 file changed, 7 insertions(+), 4 deletions(-)
Index: linux-stable/drivers/mmc/host/tmio_mmc_pio.c
===================================================================
--- linux-stable.orig/drivers/mmc/host/tmio_mmc_pio.c
+++ linux-stable/drivers/mmc/host/tmio_mmc_pio.c
@@ -675,7 +675,7 @@ static bool __tmio_mmc_sdcard_irq(struct
return false;
}
-static void tmio_mmc_sdio_irq(int irq, void *devid)
+static bool tmio_mmc_sdio_irq(int irq, void *devid)
{
struct tmio_mmc_host *host = devid;
struct mmc_host *mmc = host->mmc;
@@ -684,7 +684,7 @@ static void tmio_mmc_sdio_irq(int irq, v
unsigned int sdio_status;
if (!(pdata->flags & TMIO_MMC_SDIO_IRQ))
- return;
+ return false;
status = sd_ctrl_read16(host, CTL_SDIO_STATUS);
ireg = status & TMIO_SDIO_MASK_ALL & ~host->sdcard_irq_mask;
@@ -697,6 +697,8 @@ static void tmio_mmc_sdio_irq(int irq, v
if (mmc->caps & MMC_CAP_SDIO_IRQ && ireg & TMIO_SDIO_STAT_IOIRQ)
mmc_signal_sdio_irq(mmc);
+
+ return ireg;
}
irqreturn_t tmio_mmc_irq(int irq, void *devid)
@@ -718,9 +720,10 @@ irqreturn_t tmio_mmc_irq(int irq, void *
if (__tmio_mmc_sdcard_irq(host, ireg, status))
return IRQ_HANDLED;
- tmio_mmc_sdio_irq(irq, devid);
+ if (tmio_mmc_sdio_irq(irq, devid))
+ return IRQ_HANDLED;
- return IRQ_HANDLED;
+ return IRQ_NONE;
}
EXPORT_SYMBOL(tmio_mmc_irq);
commit 51232df5e4b268936beccde5248f312a316800be upstream.
When the managed cache is enabled, the last reference count
of a workgroup must be used for its workstation.
Otherwise, it could lead to incorrect (un)freezes in
the reclaim path, and it would be harmful.
A typical race as follows:
Thread 1 (In the reclaim path) Thread 2
workgroup_freeze(grp, 1) refcnt = 1
...
workgroup_unfreeze(grp, 1) refcnt = 1
workgroup_get(grp) refcnt = 2 (x)
workgroup_put(grp) refcnt = 1 (x)
...unexpected behaviors
* grp is detached but still used, which violates cache-managed
freeze constraint.
Reviewed-by: Chao Yu <yuchao0(a)huawei.com>
Signed-off-by: Gao Xiang <gaoxiang25(a)huawei.com>
---
change log v3:
- avoid using xa_tag apis since Matthew refuses the first XArray patch
in the previous version again;
RESEND in order to ping due to lack of response from Matthew and Greg
for half a month, patch 2/2 is needed to be backported to linux-4.19.
change log v2:
- fix tagged pointer apis to make it work properly pointed out by Matthew;
Thanks,
Gao Xiang
drivers/staging/erofs/internal.h | 1 +
drivers/staging/erofs/utils.c | 139 ++++++++++++++++++++++++++++-----------
2 files changed, 101 insertions(+), 39 deletions(-)
diff --git a/drivers/staging/erofs/internal.h b/drivers/staging/erofs/internal.h
index c70f0c5237ea..58d8cbc3f921 100644
--- a/drivers/staging/erofs/internal.h
+++ b/drivers/staging/erofs/internal.h
@@ -260,6 +260,7 @@ static inline bool erofs_workgroup_get(struct erofs_workgroup *grp, int *ocnt)
}
#define __erofs_workgroup_get(grp) atomic_inc(&(grp)->refcount)
+#define __erofs_workgroup_put(grp) atomic_dec(&(grp)->refcount)
extern int erofs_workgroup_put(struct erofs_workgroup *grp);
diff --git a/drivers/staging/erofs/utils.c b/drivers/staging/erofs/utils.c
index dd2ac9dbc4b4..2d96820da62e 100644
--- a/drivers/staging/erofs/utils.c
+++ b/drivers/staging/erofs/utils.c
@@ -87,12 +87,21 @@ int erofs_register_workgroup(struct super_block *sb,
grp = (void *)((unsigned long)grp |
1UL << RADIX_TREE_EXCEPTIONAL_SHIFT);
- err = radix_tree_insert(&sbi->workstn_tree,
- grp->index, grp);
+ /*
+ * Bump up reference count before making this workgroup
+ * visible to other users in order to avoid potential UAF
+ * without serialized by erofs_workstn_lock.
+ */
+ __erofs_workgroup_get(grp);
- if (!err) {
- __erofs_workgroup_get(grp);
- }
+ err = radix_tree_insert(&sbi->workstn_tree,
+ grp->index, grp);
+ if (unlikely(err))
+ /*
+ * it's safe to decrease since the workgroup isn't visible
+ * and refcount >= 2 (cannot be freezed).
+ */
+ __erofs_workgroup_put(grp);
erofs_workstn_unlock(sbi);
radix_tree_preload_end();
@@ -101,19 +110,99 @@ int erofs_register_workgroup(struct super_block *sb,
extern void erofs_workgroup_free_rcu(struct erofs_workgroup *grp);
+static void __erofs_workgroup_free(struct erofs_workgroup *grp)
+{
+ atomic_long_dec(&erofs_global_shrink_cnt);
+ erofs_workgroup_free_rcu(grp);
+}
+
int erofs_workgroup_put(struct erofs_workgroup *grp)
{
int count = atomic_dec_return(&grp->refcount);
if (count == 1)
atomic_long_inc(&erofs_global_shrink_cnt);
- else if (!count) {
- atomic_long_dec(&erofs_global_shrink_cnt);
- erofs_workgroup_free_rcu(grp);
- }
+ else if (!count)
+ __erofs_workgroup_free(grp);
return count;
}
+#ifdef EROFS_FS_HAS_MANAGED_CACHE
+/* for cache-managed case, customized reclaim paths exist */
+static void erofs_workgroup_unfreeze_final(struct erofs_workgroup *grp)
+{
+ erofs_workgroup_unfreeze(grp, 0);
+ __erofs_workgroup_free(grp);
+}
+
+bool erofs_try_to_release_workgroup(struct erofs_sb_info *sbi,
+ struct erofs_workgroup *grp,
+ bool cleanup)
+{
+ void *entry;
+
+ /*
+ * for managed cache enabled, the refcount of workgroups
+ * themselves could be < 0 (freezed). So there is no guarantee
+ * that all refcount > 0 if managed cache is enabled.
+ */
+ if (!erofs_workgroup_try_to_freeze(grp, 1))
+ return false;
+
+ /*
+ * note that all cached pages should be unlinked
+ * before delete it from the radix tree.
+ * Otherwise some cached pages of an orphan old workgroup
+ * could be still linked after the new one is available.
+ */
+ if (erofs_try_to_free_all_cached_pages(sbi, grp)) {
+ erofs_workgroup_unfreeze(grp, 1);
+ return false;
+ }
+
+ /*
+ * it is impossible to fail after the workgroup is freezed,
+ * however in order to avoid some race conditions, add a
+ * DBG_BUGON to observe this in advance.
+ */
+ entry = radix_tree_delete(&sbi->workstn_tree, grp->index);
+ DBG_BUGON((void *)((unsigned long)entry &
+ ~RADIX_TREE_EXCEPTIONAL_ENTRY) != grp);
+
+ /*
+ * if managed cache is enable, the last refcount
+ * should indicate the related workstation.
+ */
+ erofs_workgroup_unfreeze_final(grp);
+ return true;
+}
+
+#else
+/* for nocache case, no customized reclaim path at all */
+bool erofs_try_to_release_workgroup(struct erofs_sb_info *sbi,
+ struct erofs_workgroup *grp,
+ bool cleanup)
+{
+ int cnt = atomic_read(&grp->refcount);
+ void *entry;
+
+ DBG_BUGON(cnt <= 0);
+ DBG_BUGON(cleanup && cnt != 1);
+
+ if (cnt > 1)
+ return false;
+
+ entry = radix_tree_delete(&sbi->workstn_tree, grp->index);
+ DBG_BUGON((void *)((unsigned long)entry &
+ ~RADIX_TREE_EXCEPTIONAL_ENTRY) != grp);
+
+ /* (rarely) could be grabbed again when freeing */
+ erofs_workgroup_put(grp);
+ return true;
+}
+
+#endif
+
unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
unsigned long nr_shrink,
bool cleanup)
@@ -130,44 +219,16 @@ unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
batch, first_index, PAGEVEC_SIZE);
for (i = 0; i < found; ++i) {
- int cnt;
struct erofs_workgroup *grp = (void *)
((unsigned long)batch[i] &
~RADIX_TREE_EXCEPTIONAL_ENTRY);
first_index = grp->index + 1;
- cnt = atomic_read(&grp->refcount);
- BUG_ON(cnt <= 0);
-
- if (cleanup)
- BUG_ON(cnt != 1);
-
-#ifndef EROFS_FS_HAS_MANAGED_CACHE
- else if (cnt > 1)
-#else
- if (!erofs_workgroup_try_to_freeze(grp, 1))
-#endif
+ /* try to shrink each valid workgroup */
+ if (!erofs_try_to_release_workgroup(sbi, grp, cleanup))
continue;
- if (radix_tree_delete(&sbi->workstn_tree,
- grp->index) != grp) {
-#ifdef EROFS_FS_HAS_MANAGED_CACHE
-skip:
- erofs_workgroup_unfreeze(grp, 1);
-#endif
- continue;
- }
-
-#ifdef EROFS_FS_HAS_MANAGED_CACHE
- if (erofs_try_to_free_all_cached_pages(sbi, grp))
- goto skip;
-
- erofs_workgroup_unfreeze(grp, 1);
-#endif
- /* (rarely) grabbed again when freeing */
- erofs_workgroup_put(grp);
-
++freed;
if (unlikely(!--nr_shrink))
break;
--
2.14.5
The patch below does not apply to the 4.19-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 227ad6d957898a88b1746e30234ece64d305f066 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris(a)chris-wilson.co.uk>
Date: Sun, 30 Dec 2018 12:28:42 +0000
Subject: [PATCH] drm: Reorder set_property_atomic to avoid returning with an
active ww_ctx
Delay the drm_modeset_acquire_init() until after we check for an
allocation failure so that we can return immediately upon error without
having to unwind.
WARNING: lock held when returning to user space!
4.20.0+ #174 Not tainted
------------------------------------------------
syz-executor556/8153 is leaving the kernel with locks still held!
1 lock held by syz-executor556/8153:
#0: 000000005100c85c (crtc_ww_class_acquire){+.+.}, at:
set_property_atomic+0xb3/0x330 drivers/gpu/drm/drm_mode_object.c:462
Reported-by: syzbot+6ea337c427f5083ebdf2(a)syzkaller.appspotmail.com
Fixes: 144a7999d633 ("drm: Handle properties in the core for atomic drivers")
Signed-off-by: Chris Wilson <chris(a)chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Sean Paul <sean(a)poorly.run>
Cc: David Airlie <airlied(a)linux.ie>
Cc: <stable(a)vger.kernel.org> # v4.14+
Reviewed-by: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181230122842.21917-1-chris@…
diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c
index bb1dd46496cd..a9005c1c2384 100644
--- a/drivers/gpu/drm/drm_mode_object.c
+++ b/drivers/gpu/drm/drm_mode_object.c
@@ -459,12 +459,13 @@ static int set_property_atomic(struct drm_mode_object *obj,
struct drm_modeset_acquire_ctx ctx;
int ret;
- drm_modeset_acquire_init(&ctx, 0);
-
state = drm_atomic_state_alloc(dev);
if (!state)
return -ENOMEM;
+
+ drm_modeset_acquire_init(&ctx, 0);
state->acquire_ctx = &ctx;
+
retry:
if (prop == state->dev->mode_config.dpms_property) {
if (obj->type != DRM_MODE_OBJECT_CONNECTOR) {
The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 227ad6d957898a88b1746e30234ece64d305f066 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris(a)chris-wilson.co.uk>
Date: Sun, 30 Dec 2018 12:28:42 +0000
Subject: [PATCH] drm: Reorder set_property_atomic to avoid returning with an
active ww_ctx
Delay the drm_modeset_acquire_init() until after we check for an
allocation failure so that we can return immediately upon error without
having to unwind.
WARNING: lock held when returning to user space!
4.20.0+ #174 Not tainted
------------------------------------------------
syz-executor556/8153 is leaving the kernel with locks still held!
1 lock held by syz-executor556/8153:
#0: 000000005100c85c (crtc_ww_class_acquire){+.+.}, at:
set_property_atomic+0xb3/0x330 drivers/gpu/drm/drm_mode_object.c:462
Reported-by: syzbot+6ea337c427f5083ebdf2(a)syzkaller.appspotmail.com
Fixes: 144a7999d633 ("drm: Handle properties in the core for atomic drivers")
Signed-off-by: Chris Wilson <chris(a)chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Sean Paul <sean(a)poorly.run>
Cc: David Airlie <airlied(a)linux.ie>
Cc: <stable(a)vger.kernel.org> # v4.14+
Reviewed-by: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181230122842.21917-1-chris@…
diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c
index bb1dd46496cd..a9005c1c2384 100644
--- a/drivers/gpu/drm/drm_mode_object.c
+++ b/drivers/gpu/drm/drm_mode_object.c
@@ -459,12 +459,13 @@ static int set_property_atomic(struct drm_mode_object *obj,
struct drm_modeset_acquire_ctx ctx;
int ret;
- drm_modeset_acquire_init(&ctx, 0);
-
state = drm_atomic_state_alloc(dev);
if (!state)
return -ENOMEM;
+
+ drm_modeset_acquire_init(&ctx, 0);
state->acquire_ctx = &ctx;
+
retry:
if (prop == state->dev->mode_config.dpms_property) {
if (obj->type != DRM_MODE_OBJECT_CONNECTOR) {
The patch below does not apply to the 5.0-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 227ad6d957898a88b1746e30234ece64d305f066 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris(a)chris-wilson.co.uk>
Date: Sun, 30 Dec 2018 12:28:42 +0000
Subject: [PATCH] drm: Reorder set_property_atomic to avoid returning with an
active ww_ctx
Delay the drm_modeset_acquire_init() until after we check for an
allocation failure so that we can return immediately upon error without
having to unwind.
WARNING: lock held when returning to user space!
4.20.0+ #174 Not tainted
------------------------------------------------
syz-executor556/8153 is leaving the kernel with locks still held!
1 lock held by syz-executor556/8153:
#0: 000000005100c85c (crtc_ww_class_acquire){+.+.}, at:
set_property_atomic+0xb3/0x330 drivers/gpu/drm/drm_mode_object.c:462
Reported-by: syzbot+6ea337c427f5083ebdf2(a)syzkaller.appspotmail.com
Fixes: 144a7999d633 ("drm: Handle properties in the core for atomic drivers")
Signed-off-by: Chris Wilson <chris(a)chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Sean Paul <sean(a)poorly.run>
Cc: David Airlie <airlied(a)linux.ie>
Cc: <stable(a)vger.kernel.org> # v4.14+
Reviewed-by: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181230122842.21917-1-chris@…
diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c
index bb1dd46496cd..a9005c1c2384 100644
--- a/drivers/gpu/drm/drm_mode_object.c
+++ b/drivers/gpu/drm/drm_mode_object.c
@@ -459,12 +459,13 @@ static int set_property_atomic(struct drm_mode_object *obj,
struct drm_modeset_acquire_ctx ctx;
int ret;
- drm_modeset_acquire_init(&ctx, 0);
-
state = drm_atomic_state_alloc(dev);
if (!state)
return -ENOMEM;
+
+ drm_modeset_acquire_init(&ctx, 0);
state->acquire_ctx = &ctx;
+
retry:
if (prop == state->dev->mode_config.dpms_property) {
if (obj->type != DRM_MODE_OBJECT_CONNECTOR) {
The patch below does not apply to the 4.20-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From 227ad6d957898a88b1746e30234ece64d305f066 Mon Sep 17 00:00:00 2001
From: Chris Wilson <chris(a)chris-wilson.co.uk>
Date: Sun, 30 Dec 2018 12:28:42 +0000
Subject: [PATCH] drm: Reorder set_property_atomic to avoid returning with an
active ww_ctx
Delay the drm_modeset_acquire_init() until after we check for an
allocation failure so that we can return immediately upon error without
having to unwind.
WARNING: lock held when returning to user space!
4.20.0+ #174 Not tainted
------------------------------------------------
syz-executor556/8153 is leaving the kernel with locks still held!
1 lock held by syz-executor556/8153:
#0: 000000005100c85c (crtc_ww_class_acquire){+.+.}, at:
set_property_atomic+0xb3/0x330 drivers/gpu/drm/drm_mode_object.c:462
Reported-by: syzbot+6ea337c427f5083ebdf2(a)syzkaller.appspotmail.com
Fixes: 144a7999d633 ("drm: Handle properties in the core for atomic drivers")
Signed-off-by: Chris Wilson <chris(a)chris-wilson.co.uk>
Cc: Daniel Vetter <daniel.vetter(a)ffwll.ch>
Cc: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Cc: Sean Paul <sean(a)poorly.run>
Cc: David Airlie <airlied(a)linux.ie>
Cc: <stable(a)vger.kernel.org> # v4.14+
Reviewed-by: Maarten Lankhorst <maarten.lankhorst(a)linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20181230122842.21917-1-chris@…
diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c
index bb1dd46496cd..a9005c1c2384 100644
--- a/drivers/gpu/drm/drm_mode_object.c
+++ b/drivers/gpu/drm/drm_mode_object.c
@@ -459,12 +459,13 @@ static int set_property_atomic(struct drm_mode_object *obj,
struct drm_modeset_acquire_ctx ctx;
int ret;
- drm_modeset_acquire_init(&ctx, 0);
-
state = drm_atomic_state_alloc(dev);
if (!state)
return -ENOMEM;
+
+ drm_modeset_acquire_init(&ctx, 0);
state->acquire_ctx = &ctx;
+
retry:
if (prop == state->dev->mode_config.dpms_property) {
if (obj->type != DRM_MODE_OBJECT_CONNECTOR) {
The patch below was submitted to be applied to the 5.0-stable tree.
I fail to see how this patch meets the stable kernel rules as found at
Documentation/process/stable-kernel-rules.rst.
I could be totally wrong, and if so, please respond to
<stable(a)vger.kernel.org> and let me know why this patch should be
applied. Otherwise, it is now dropped from my patch queues, never to be
seen again.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
>From e552f0851070fe4975d610a99910be4e9bf5d7bd Mon Sep 17 00:00:00 2001
From: Ben Dooks <ben.dooks(a)codethink.co.uk>
Date: Fri, 1 Mar 2019 12:00:46 +0000
Subject: [PATCH] drm: add __user attribute to ptr_to_compat()
The ptr_to_compat() call takes a "void __user *", so cast
the compat drm calls that use it to avoid the following
warnings from sparse:
drivers/gpu/drm/drm_ioc32.c:188:39: warning: incorrect type in argument 1 (different address spaces)
drivers/gpu/drm/drm_ioc32.c:188:39: expected void [noderef] <asn:1>*uptr
drivers/gpu/drm/drm_ioc32.c:188:39: got void *[addressable] [assigned] handle
drivers/gpu/drm/drm_ioc32.c:529:41: warning: incorrect type in argument 1 (different address spaces)
drivers/gpu/drm/drm_ioc32.c:529:41: expected void [noderef] <asn:1>*uptr
drivers/gpu/drm/drm_ioc32.c:529:41: got void *[addressable] [assigned] handle
Cc: stable(a)vger.kernel.org
Signed-off-by: Ben Dooks <ben.dooks(a)codethink.co.uk>
Signed-off-by: Sean Paul <seanpaul(a)chromium.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190301120046.26961-1-ben.do…
diff --git a/drivers/gpu/drm/drm_ioc32.c b/drivers/gpu/drm/drm_ioc32.c
index 67b1fca39aa6..0e3043e08c69 100644
--- a/drivers/gpu/drm/drm_ioc32.c
+++ b/drivers/gpu/drm/drm_ioc32.c
@@ -185,7 +185,7 @@ static int compat_drm_getmap(struct file *file, unsigned int cmd,
m32.size = map.size;
m32.type = map.type;
m32.flags = map.flags;
- m32.handle = ptr_to_compat(map.handle);
+ m32.handle = ptr_to_compat((void __user *)map.handle);
m32.mtrr = map.mtrr;
if (copy_to_user(argp, &m32, sizeof(m32)))
return -EFAULT;
@@ -216,7 +216,7 @@ static int compat_drm_addmap(struct file *file, unsigned int cmd,
m32.offset = map.offset;
m32.mtrr = map.mtrr;
- m32.handle = ptr_to_compat(map.handle);
+ m32.handle = ptr_to_compat((void __user *)map.handle);
if (map.handle != compat_ptr(m32.handle))
pr_err_ratelimited("compat_drm_addmap truncated handle %p for type %d offset %x\n",
map.handle, m32.type, m32.offset);
@@ -526,7 +526,7 @@ static int compat_drm_getsareactx(struct file *file, unsigned int cmd,
if (err)
return err;
- req32.handle = ptr_to_compat(req.handle);
+ req32.handle = ptr_to_compat((void __user *)req.handle);
if (copy_to_user(argp, &req32, sizeof(req32)))
return -EFAULT;
This is a note to let you know that I've just added the patch titled
staging: olpc_dcon_xo_1: add missing 'const' qualifier
to my staging git tree which can be found at
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git
in the staging-linus branch.
The patch will show up in the next release of the linux-next tree
(usually sometime within the next 24 hours during the week.)
The patch will hopefully also be merged in Linus's tree for the
next -rc kernel release.
If you have any questions about this process, please let me know.
>From ae0a6d2017f733781dcc938a471ccc2d05f9bee6 Mon Sep 17 00:00:00 2001
From: Arnd Bergmann <arnd(a)arndb.de>
Date: Mon, 4 Mar 2019 20:42:33 +0100
Subject: staging: olpc_dcon_xo_1: add missing 'const' qualifier
gcc noticed a mismatch between the type qualifiers after a recent
cleanup:
drivers/staging/olpc_dcon/olpc_dcon_xo_1.c: In function 'dcon_init_xo_1':
drivers/staging/olpc_dcon/olpc_dcon_xo_1.c:48:26: error: initialization discards 'const' qualifier from pointer target type [-Werror=discarded-qualifiers]
Add the 'const' keyword that should have been there all along.
Fixes: 2159fb372929 ("staging: olpc_dcon: olpc_dcon_xo_1.c: Switch to the gpio descriptor interface")
Signed-off-by: Arnd Bergmann <arnd(a)arndb.de>
Cc: stable <stable(a)vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
drivers/staging/olpc_dcon/olpc_dcon_xo_1.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/staging/olpc_dcon/olpc_dcon_xo_1.c b/drivers/staging/olpc_dcon/olpc_dcon_xo_1.c
index 80b8d4153414..a54286498a47 100644
--- a/drivers/staging/olpc_dcon/olpc_dcon_xo_1.c
+++ b/drivers/staging/olpc_dcon/olpc_dcon_xo_1.c
@@ -45,7 +45,7 @@ static int dcon_init_xo_1(struct dcon_priv *dcon)
{
unsigned char lob;
int ret, i;
- struct dcon_gpio *pin = &gpios_asis[0];
+ const struct dcon_gpio *pin = &gpios_asis[0];
for (i = 0; i < ARRAY_SIZE(gpios_asis); i++) {
gpios[i] = devm_gpiod_get(&dcon->client->dev, pin[i].name,
--
2.21.0
This is a note to let you know that I've just added the patch titled
staging: speakup_soft: Fix alternate speech with other synths
to my staging git tree which can be found at
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git
in the staging-linus branch.
The patch will show up in the next release of the linux-next tree
(usually sometime within the next 24 hours during the week.)
The patch will hopefully also be merged in Linus's tree for the
next -rc kernel release.
If you have any questions about this process, please let me know.
>From 45ac7b31bc6c4af885cc5b5d6c534c15bcbe7643 Mon Sep 17 00:00:00 2001
From: Samuel Thibault <samuel.thibault(a)ens-lyon.org>
Date: Thu, 7 Mar 2019 23:06:57 +0100
Subject: staging: speakup_soft: Fix alternate speech with other synths
When switching from speakup_soft to another synth, speakup_soft would
keep calling synth_buffer_getc() from softsynthx_read.
Let's thus make synth.c export the knowledge of the current synth, so
that speakup_soft can determine whether it should be running.
speakup_soft also needs to set itself alive, otherwise the switch would
let it remain silent.
Signed-off-by: Samuel Thibault <samuel.thibault(a)ens-lyon.org>
Cc: stable <stable(a)vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
drivers/staging/speakup/speakup_soft.c | 16 +++++++++++-----
drivers/staging/speakup/spk_priv.h | 1 +
drivers/staging/speakup/synth.c | 6 ++++++
3 files changed, 18 insertions(+), 5 deletions(-)
diff --git a/drivers/staging/speakup/speakup_soft.c b/drivers/staging/speakup/speakup_soft.c
index edff6ce85655..9d85a3a1af4c 100644
--- a/drivers/staging/speakup/speakup_soft.c
+++ b/drivers/staging/speakup/speakup_soft.c
@@ -210,12 +210,15 @@ static ssize_t softsynthx_read(struct file *fp, char __user *buf, size_t count,
return -EINVAL;
spin_lock_irqsave(&speakup_info.spinlock, flags);
+ synth_soft.alive = 1;
while (1) {
prepare_to_wait(&speakup_event, &wait, TASK_INTERRUPTIBLE);
- if (!unicode)
- synth_buffer_skip_nonlatin1();
- if (!synth_buffer_empty() || speakup_info.flushing)
- break;
+ if (synth_current() == &synth_soft) {
+ if (!unicode)
+ synth_buffer_skip_nonlatin1();
+ if (!synth_buffer_empty() || speakup_info.flushing)
+ break;
+ }
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
if (fp->f_flags & O_NONBLOCK) {
finish_wait(&speakup_event, &wait);
@@ -235,6 +238,8 @@ static ssize_t softsynthx_read(struct file *fp, char __user *buf, size_t count,
/* Keep 3 bytes available for a 16bit UTF-8-encoded character */
while (chars_sent <= count - bytes_per_ch) {
+ if (synth_current() != &synth_soft)
+ break;
if (speakup_info.flushing) {
speakup_info.flushing = 0;
ch = '\x18';
@@ -331,7 +336,8 @@ static __poll_t softsynth_poll(struct file *fp, struct poll_table_struct *wait)
poll_wait(fp, &speakup_event, wait);
spin_lock_irqsave(&speakup_info.spinlock, flags);
- if (!synth_buffer_empty() || speakup_info.flushing)
+ if (synth_current() == &synth_soft &&
+ (!synth_buffer_empty() || speakup_info.flushing))
ret = EPOLLIN | EPOLLRDNORM;
spin_unlock_irqrestore(&speakup_info.spinlock, flags);
return ret;
diff --git a/drivers/staging/speakup/spk_priv.h b/drivers/staging/speakup/spk_priv.h
index c8e688878fc7..ac6a74883af4 100644
--- a/drivers/staging/speakup/spk_priv.h
+++ b/drivers/staging/speakup/spk_priv.h
@@ -74,6 +74,7 @@ int synth_request_region(unsigned long start, unsigned long n);
int synth_release_region(unsigned long start, unsigned long n);
int synth_add(struct spk_synth *in_synth);
void synth_remove(struct spk_synth *in_synth);
+struct spk_synth *synth_current(void);
extern struct speakup_info_t speakup_info;
diff --git a/drivers/staging/speakup/synth.c b/drivers/staging/speakup/synth.c
index 25f259ee4ffc..3568bfb89912 100644
--- a/drivers/staging/speakup/synth.c
+++ b/drivers/staging/speakup/synth.c
@@ -481,4 +481,10 @@ void synth_remove(struct spk_synth *in_synth)
}
EXPORT_SYMBOL_GPL(synth_remove);
+struct spk_synth *synth_current(void)
+{
+ return synth;
+}
+EXPORT_SYMBOL_GPL(synth_current);
+
short spk_punc_masks[] = { 0, SOME, MOST, PUNC, PUNC | B_SYM };
--
2.21.0
This is a note to let you know that I've just added the patch titled
staging: comedi: ni_mio_common: Fix divide-by-zero for DIO cmdtest
to my staging git tree which can be found at
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git
in the staging-linus branch.
The patch will show up in the next release of the linux-next tree
(usually sometime within the next 24 hours during the week.)
The patch will hopefully also be merged in Linus's tree for the
next -rc kernel release.
If you have any questions about this process, please let me know.
>From bafd9c64056cd034a1174dcadb65cd3b294ff8f6 Mon Sep 17 00:00:00 2001
From: Ian Abbott <abbotti(a)mev.co.uk>
Date: Mon, 4 Mar 2019 14:33:54 +0000
Subject: staging: comedi: ni_mio_common: Fix divide-by-zero for DIO cmdtest
`ni_cdio_cmdtest()` validates Comedi asynchronous commands for the DIO
subdevice (subdevice 2) of supported National Instruments M-series
cards. It is called when handling the `COMEDI_CMD` and `COMEDI_CMDTEST`
ioctls for this subdevice. There are two causes for a possible
divide-by-zero error when validating that the `stop_arg` member of the
passed-in command is not too large.
The first cause for the divide-by-zero is that calls to
`comedi_bytes_per_scan()` are only valid once the command has been
copied to `s->async->cmd`, but that copy is only done for the
`COMEDI_CMD` ioctl. For the `COMEDI_CMDTEST` ioctl, it will use
whatever was left there by the previous `COMEDI_CMD` ioctl, if any.
(This is very likely, as it is usual for the application to use
`COMEDI_CMDTEST` before `COMEDI_CMD`.) If there has been no previous,
valid `COMEDI_CMD` for this subdevice, then `comedi_bytes_per_scan()`
will return 0, so the subsequent division in `ni_cdio_cmdtest()` of
`s->async->prealloc_bufsz / comedi_bytes_per_scan(s)` will be a
divide-by-zero error. To fix this error, call a new function
`comedi_bytes_per_scan_cmd(s, cmd)`, based on the existing
`comedi_bytes_per_scan(s)` but using a specified `struct comedi_cmd` for
its calculations. (Also refactor `comedi_bytes_per_scan()` to call the
new function.)
Once the first cause for the divide-by-zero has been fixed, the second
cause is that `comedi_bytes_per_scan_cmd()` can legitimately return 0 if
the `scan_end_arg` member of the `struct comedi_cmd` being tested is 0.
Fix it by only performing the division (and validating that `stop_arg`
is no more than the maximum value) if `comedi_bytes_per_scan_cmd()`
returns a non-zero value.
The problem was reported on the COMEDI mailing list here:
https://groups.google.com/forum/#!topic/comedi_list/4t9WlHzMhKM
Reported-by: Ivan Vasilyev <grabesstimme(a)gmail.com>
Tested-by: Ivan Vasilyev <grabesstimme(a)gmail.com>
Fixes: f164cbf98fa8 ("staging: comedi: ni_mio_common: add finite regeneration to dio output")
Cc: <stable(a)vger.kernel.org> # 4.6+
Cc: Spencer E. Olson <olsonse(a)umich.edu>
Signed-off-by: Ian Abbott <abbotti(a)mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
drivers/staging/comedi/comedidev.h | 2 ++
drivers/staging/comedi/drivers.c | 33 ++++++++++++++++---
.../staging/comedi/drivers/ni_mio_common.c | 10 ++++--
3 files changed, 38 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/comedi/comedidev.h b/drivers/staging/comedi/comedidev.h
index a7d569cfca5d..0dff1ac057cd 100644
--- a/drivers/staging/comedi/comedidev.h
+++ b/drivers/staging/comedi/comedidev.h
@@ -1001,6 +1001,8 @@ int comedi_dio_insn_config(struct comedi_device *dev,
unsigned int mask);
unsigned int comedi_dio_update_state(struct comedi_subdevice *s,
unsigned int *data);
+unsigned int comedi_bytes_per_scan_cmd(struct comedi_subdevice *s,
+ struct comedi_cmd *cmd);
unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s);
unsigned int comedi_nscans_left(struct comedi_subdevice *s,
unsigned int nscans);
diff --git a/drivers/staging/comedi/drivers.c b/drivers/staging/comedi/drivers.c
index eefa62f42c0f..5a32b8fc000e 100644
--- a/drivers/staging/comedi/drivers.c
+++ b/drivers/staging/comedi/drivers.c
@@ -394,11 +394,13 @@ unsigned int comedi_dio_update_state(struct comedi_subdevice *s,
EXPORT_SYMBOL_GPL(comedi_dio_update_state);
/**
- * comedi_bytes_per_scan() - Get length of asynchronous command "scan" in bytes
+ * comedi_bytes_per_scan_cmd() - Get length of asynchronous command "scan" in
+ * bytes
* @s: COMEDI subdevice.
+ * @cmd: COMEDI command.
*
* Determines the overall scan length according to the subdevice type and the
- * number of channels in the scan.
+ * number of channels in the scan for the specified command.
*
* For digital input, output or input/output subdevices, samples for
* multiple channels are assumed to be packed into one or more unsigned
@@ -408,9 +410,9 @@ EXPORT_SYMBOL_GPL(comedi_dio_update_state);
*
* Returns the overall scan length in bytes.
*/
-unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
+unsigned int comedi_bytes_per_scan_cmd(struct comedi_subdevice *s,
+ struct comedi_cmd *cmd)
{
- struct comedi_cmd *cmd = &s->async->cmd;
unsigned int num_samples;
unsigned int bits_per_sample;
@@ -427,6 +429,29 @@ unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
}
return comedi_samples_to_bytes(s, num_samples);
}
+EXPORT_SYMBOL_GPL(comedi_bytes_per_scan_cmd);
+
+/**
+ * comedi_bytes_per_scan() - Get length of asynchronous command "scan" in bytes
+ * @s: COMEDI subdevice.
+ *
+ * Determines the overall scan length according to the subdevice type and the
+ * number of channels in the scan for the current command.
+ *
+ * For digital input, output or input/output subdevices, samples for
+ * multiple channels are assumed to be packed into one or more unsigned
+ * short or unsigned int values according to the subdevice's %SDF_LSAMPL
+ * flag. For other types of subdevice, samples are assumed to occupy a
+ * whole unsigned short or unsigned int according to the %SDF_LSAMPL flag.
+ *
+ * Returns the overall scan length in bytes.
+ */
+unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
+{
+ struct comedi_cmd *cmd = &s->async->cmd;
+
+ return comedi_bytes_per_scan_cmd(s, cmd);
+}
EXPORT_SYMBOL_GPL(comedi_bytes_per_scan);
static unsigned int __comedi_nscans_left(struct comedi_subdevice *s,
diff --git a/drivers/staging/comedi/drivers/ni_mio_common.c b/drivers/staging/comedi/drivers/ni_mio_common.c
index 5edf59ac6706..b04dad8c7092 100644
--- a/drivers/staging/comedi/drivers/ni_mio_common.c
+++ b/drivers/staging/comedi/drivers/ni_mio_common.c
@@ -3545,6 +3545,7 @@ static int ni_cdio_cmdtest(struct comedi_device *dev,
struct comedi_subdevice *s, struct comedi_cmd *cmd)
{
struct ni_private *devpriv = dev->private;
+ unsigned int bytes_per_scan;
int err = 0;
/* Step 1 : check if triggers are trivially valid */
@@ -3579,9 +3580,12 @@ static int ni_cdio_cmdtest(struct comedi_device *dev,
err |= comedi_check_trigger_arg_is(&cmd->convert_arg, 0);
err |= comedi_check_trigger_arg_is(&cmd->scan_end_arg,
cmd->chanlist_len);
- err |= comedi_check_trigger_arg_max(&cmd->stop_arg,
- s->async->prealloc_bufsz /
- comedi_bytes_per_scan(s));
+ bytes_per_scan = comedi_bytes_per_scan_cmd(s, cmd);
+ if (bytes_per_scan) {
+ err |= comedi_check_trigger_arg_max(&cmd->stop_arg,
+ s->async->prealloc_bufsz /
+ bytes_per_scan);
+ }
if (err)
return 3;
--
2.21.0
This is a note to let you know that I've just added the patch titled
staging: erofs: fix to handle error path of erofs_vmap()
to my staging git tree which can be found at
git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging.git
in the staging-linus branch.
The patch will show up in the next release of the linux-next tree
(usually sometime within the next 24 hours during the week.)
The patch will hopefully also be merged in Linus's tree for the
next -rc kernel release.
If you have any questions about this process, please let me know.
>From 8bce6dcede65139a087ff240127e3f3c01363eed Mon Sep 17 00:00:00 2001
From: Chao Yu <yuchao0(a)huawei.com>
Date: Mon, 11 Mar 2019 23:10:10 +0800
Subject: staging: erofs: fix to handle error path of erofs_vmap()
erofs_vmap() wrapped vmap() and vm_map_ram() to return virtual
continuous memory, but both of them can failed due to a lot of
reason, previously, erofs_vmap()'s callers didn't handle them,
which can potentially cause NULL pointer access, fix it.
Fixes: 3883a79abd02 ("staging: erofs: introduce VLE decompression support")
Fixes: 0d40d6e399c1 ("staging: erofs: add a generic z_erofs VLE decompressor")
Cc: <stable(a)vger.kernel.org> # 4.19+
Signed-off-by: Gao Xiang <gaoxiang25(a)huawei.com>
Signed-off-by: Chao Yu <yuchao0(a)huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org>
---
drivers/staging/erofs/unzip_vle.c | 4 ++++
drivers/staging/erofs/unzip_vle_lz4.c | 7 +++++--
2 files changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/staging/erofs/unzip_vle.c b/drivers/staging/erofs/unzip_vle.c
index 8715bc50e09c..c7b3b21123c1 100644
--- a/drivers/staging/erofs/unzip_vle.c
+++ b/drivers/staging/erofs/unzip_vle.c
@@ -1029,6 +1029,10 @@ static int z_erofs_vle_unzip(struct super_block *sb,
skip_allocpage:
vout = erofs_vmap(pages, nr_pages);
+ if (!vout) {
+ err = -ENOMEM;
+ goto out;
+ }
err = z_erofs_vle_unzip_vmap(compressed_pages,
clusterpages, vout, llen, work->pageofs, overlapped);
diff --git a/drivers/staging/erofs/unzip_vle_lz4.c b/drivers/staging/erofs/unzip_vle_lz4.c
index 48b263a2731a..0daac9b984a8 100644
--- a/drivers/staging/erofs/unzip_vle_lz4.c
+++ b/drivers/staging/erofs/unzip_vle_lz4.c
@@ -136,10 +136,13 @@ int z_erofs_vle_unzip_fast_percpu(struct page **compressed_pages,
nr_pages = DIV_ROUND_UP(outlen + pageofs, PAGE_SIZE);
- if (clusterpages == 1)
+ if (clusterpages == 1) {
vin = kmap_atomic(compressed_pages[0]);
- else
+ } else {
vin = erofs_vmap(compressed_pages, clusterpages);
+ if (!vin)
+ return -ENOMEM;
+ }
preempt_disable();
vout = erofs_pcpubuf[smp_processor_id()].data;
--
2.21.0
`ni_cdio_cmdtest()` validates Comedi asynchronous commands for the DIO
subdevice (subdevice 2) of supported National Instruments M-series
cards. It is called when handling the `COMEDI_CMD` and `COMEDI_CMDTEST`
ioctls for this subdevice. There are two causes for a possible
divide-by-zero error when validating that the `stop_arg` member of the
passed-in command is not too large.
The first cause for the divide-by-zero is that calls to
`comedi_bytes_per_scan()` are only valid once the command has been
copied to `s->async->cmd`, but that copy is only done for the
`COMEDI_CMD` ioctl. For the `COMEDI_CMDTEST` ioctl, it will use
whatever was left there by the previous `COMEDI_CMD` ioctl, if any.
(This is very likely, as it is usual for the application to use
`COMEDI_CMDTEST` before `COMEDI_CMD`.) If there has been no previous,
valid `COMEDI_CMD` for this subdevice, then `comedi_bytes_per_scan()`
will return 0, so the subsequent division in `ni_cdio_cmdtest()` of
`s->async->prealloc_bufsz / comedi_bytes_per_scan(s)` will be a
divide-by-zero error. To fix this error, call a new function
`comedi_bytes_per_scan_cmd(s, cmd)`, based on the existing
`comedi_bytes_per_scan(s)` but using a specified `struct comedi_cmd` for
its calculations. (Also refactor `comedi_bytes_per_scan()` to call the
new function.)
Once the first cause for the divide-by-zero has been fixed, the second
cause is that `comedi_bytes_per_scan_cmd()` can legitimately return 0 if
the `scan_end_arg` member of the `struct comedi_cmd` being tested is 0.
Fix it by only performing the division (and validating that `stop_arg`
is no more than the maximum value) if `comedi_bytes_per_scan_cmd()`
returns a non-zero value.
The problem was reported on the COMEDI mailing list here:
https://groups.google.com/forum/#!topic/comedi_list/4t9WlHzMhKM
Fixes: f164cbf98fa8 ("staging: comedi: ni_mio_common: add finite regeneration to dio output")
Cc: <stable(a)vger.kernel.org> # 4.6+
Cc: Spencer E. Olson <olsonse(a)umich.edu>
Signed-off-by: Ian Abbott <abbotti(a)mev.co.uk>
---
drivers/staging/comedi/comedidev.h | 2 ++
drivers/staging/comedi/drivers.c | 33 ++++++++++++++++---
.../staging/comedi/drivers/ni_mio_common.c | 10 ++++--
3 files changed, 38 insertions(+), 7 deletions(-)
diff --git a/drivers/staging/comedi/comedidev.h b/drivers/staging/comedi/comedidev.h
index a7d569cfca5d..0dff1ac057cd 100644
--- a/drivers/staging/comedi/comedidev.h
+++ b/drivers/staging/comedi/comedidev.h
@@ -1001,6 +1001,8 @@ int comedi_dio_insn_config(struct comedi_device *dev,
unsigned int mask);
unsigned int comedi_dio_update_state(struct comedi_subdevice *s,
unsigned int *data);
+unsigned int comedi_bytes_per_scan_cmd(struct comedi_subdevice *s,
+ struct comedi_cmd *cmd);
unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s);
unsigned int comedi_nscans_left(struct comedi_subdevice *s,
unsigned int nscans);
diff --git a/drivers/staging/comedi/drivers.c b/drivers/staging/comedi/drivers.c
index eefa62f42c0f..5a32b8fc000e 100644
--- a/drivers/staging/comedi/drivers.c
+++ b/drivers/staging/comedi/drivers.c
@@ -394,11 +394,13 @@ unsigned int comedi_dio_update_state(struct comedi_subdevice *s,
EXPORT_SYMBOL_GPL(comedi_dio_update_state);
/**
- * comedi_bytes_per_scan() - Get length of asynchronous command "scan" in bytes
+ * comedi_bytes_per_scan_cmd() - Get length of asynchronous command "scan" in
+ * bytes
* @s: COMEDI subdevice.
+ * @cmd: COMEDI command.
*
* Determines the overall scan length according to the subdevice type and the
- * number of channels in the scan.
+ * number of channels in the scan for the specified command.
*
* For digital input, output or input/output subdevices, samples for
* multiple channels are assumed to be packed into one or more unsigned
@@ -408,9 +410,9 @@ EXPORT_SYMBOL_GPL(comedi_dio_update_state);
*
* Returns the overall scan length in bytes.
*/
-unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
+unsigned int comedi_bytes_per_scan_cmd(struct comedi_subdevice *s,
+ struct comedi_cmd *cmd)
{
- struct comedi_cmd *cmd = &s->async->cmd;
unsigned int num_samples;
unsigned int bits_per_sample;
@@ -427,6 +429,29 @@ unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
}
return comedi_samples_to_bytes(s, num_samples);
}
+EXPORT_SYMBOL_GPL(comedi_bytes_per_scan_cmd);
+
+/**
+ * comedi_bytes_per_scan() - Get length of asynchronous command "scan" in bytes
+ * @s: COMEDI subdevice.
+ *
+ * Determines the overall scan length according to the subdevice type and the
+ * number of channels in the scan for the current command.
+ *
+ * For digital input, output or input/output subdevices, samples for
+ * multiple channels are assumed to be packed into one or more unsigned
+ * short or unsigned int values according to the subdevice's %SDF_LSAMPL
+ * flag. For other types of subdevice, samples are assumed to occupy a
+ * whole unsigned short or unsigned int according to the %SDF_LSAMPL flag.
+ *
+ * Returns the overall scan length in bytes.
+ */
+unsigned int comedi_bytes_per_scan(struct comedi_subdevice *s)
+{
+ struct comedi_cmd *cmd = &s->async->cmd;
+
+ return comedi_bytes_per_scan_cmd(s, cmd);
+}
EXPORT_SYMBOL_GPL(comedi_bytes_per_scan);
static unsigned int __comedi_nscans_left(struct comedi_subdevice *s,
diff --git a/drivers/staging/comedi/drivers/ni_mio_common.c b/drivers/staging/comedi/drivers/ni_mio_common.c
index 5edf59ac6706..b04dad8c7092 100644
--- a/drivers/staging/comedi/drivers/ni_mio_common.c
+++ b/drivers/staging/comedi/drivers/ni_mio_common.c
@@ -3545,6 +3545,7 @@ static int ni_cdio_cmdtest(struct comedi_device *dev,
struct comedi_subdevice *s, struct comedi_cmd *cmd)
{
struct ni_private *devpriv = dev->private;
+ unsigned int bytes_per_scan;
int err = 0;
/* Step 1 : check if triggers are trivially valid */
@@ -3579,9 +3580,12 @@ static int ni_cdio_cmdtest(struct comedi_device *dev,
err |= comedi_check_trigger_arg_is(&cmd->convert_arg, 0);
err |= comedi_check_trigger_arg_is(&cmd->scan_end_arg,
cmd->chanlist_len);
- err |= comedi_check_trigger_arg_max(&cmd->stop_arg,
- s->async->prealloc_bufsz /
- comedi_bytes_per_scan(s));
+ bytes_per_scan = comedi_bytes_per_scan_cmd(s, cmd);
+ if (bytes_per_scan) {
+ err |= comedi_check_trigger_arg_max(&cmd->stop_arg,
+ s->async->prealloc_bufsz /
+ bytes_per_scan);
+ }
if (err)
return 3;
--
2.20.1
From: Matthew Wilcox <willy(a)infradead.org>
commit 3159f943aafdbacb2f94c38fdaadabf2bbde2a14 upstream.
Introduce xarray value entries and tagged pointers to replace radix
tree exceptional entries. This is a slight change in encoding to allow
the use of an extra bit (we can now store BITS_PER_LONG - 1 bits in a
value entry). It is also a change in emphasis; exceptional entries are
intimidating and different. As the comment explains, you can choose
to store values or pointers in the xarray and they are both first-class
citizens.
Signed-off-by: Matthew Wilcox <willy(a)infradead.org>
Reviewed-by: Josef Bacik <jbacik(a)fb.com>
[ take the minimum subset of tagged pointer support only. ]
Cc: Matthew Wilcox <willy(a)infradead.org>
Signed-off-by: Gao Xiang <gaoxiang25(a)huawei.com>
---
RESEND in order to ping due to lack of response from Matthew and Greg
for half a month, patch 2/2 is needed to be backported to linux-4.19.
change log v2:
- fix tagged pointer apis to make it work properly pointed out by Matthew;
Thanks,
Gao Xiang
drivers/staging/erofs/utils.c | 18 +++++----------
include/linux/xarray.h | 52 +++++++++++++++++++++++++++++++++++++++++++
2 files changed, 58 insertions(+), 12 deletions(-)
diff --git a/drivers/staging/erofs/utils.c b/drivers/staging/erofs/utils.c
index dd2ac9dbc4b4..c466e8c8ea90 100644
--- a/drivers/staging/erofs/utils.c
+++ b/drivers/staging/erofs/utils.c
@@ -32,7 +32,6 @@ static atomic_long_t erofs_global_shrink_cnt;
#ifdef CONFIG_EROFS_FS_ZIP
-/* radix_tree and the future XArray both don't use tagptr_t yet */
struct erofs_workgroup *erofs_find_workgroup(
struct super_block *sb, pgoff_t index, bool *tag)
{
@@ -44,9 +43,8 @@ struct erofs_workgroup *erofs_find_workgroup(
rcu_read_lock();
grp = radix_tree_lookup(&sbi->workstn_tree, index);
if (grp != NULL) {
- *tag = radix_tree_exceptional_entry(grp);
- grp = (void *)((unsigned long)grp &
- ~RADIX_TREE_EXCEPTIONAL_ENTRY);
+ *tag = xa_pointer_tag(grp);
+ grp = xa_untag_pointer(grp);
if (erofs_workgroup_get(grp, &oldcount)) {
/* prefer to relax rcu read side */
@@ -83,9 +81,7 @@ int erofs_register_workgroup(struct super_block *sb,
sbi = EROFS_SB(sb);
erofs_workstn_lock(sbi);
- if (tag)
- grp = (void *)((unsigned long)grp |
- 1UL << RADIX_TREE_EXCEPTIONAL_SHIFT);
+ grp = xa_tag_pointer(grp, tag);
err = radix_tree_insert(&sbi->workstn_tree,
grp->index, grp);
@@ -131,9 +127,7 @@ unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
for (i = 0; i < found; ++i) {
int cnt;
- struct erofs_workgroup *grp = (void *)
- ((unsigned long)batch[i] &
- ~RADIX_TREE_EXCEPTIONAL_ENTRY);
+ struct erofs_workgroup *grp = xa_untag_pointer(batch[i]);
first_index = grp->index + 1;
@@ -150,8 +144,8 @@ unsigned long erofs_shrink_workstation(struct erofs_sb_info *sbi,
#endif
continue;
- if (radix_tree_delete(&sbi->workstn_tree,
- grp->index) != grp) {
+ if (xa_untag_pointer(radix_tree_delete(&sbi->workstn_tree,
+ grp->index)) != grp) {
#ifdef EROFS_FS_HAS_MANAGED_CACHE
skip:
erofs_workgroup_unfreeze(grp, 1);
diff --git a/include/linux/xarray.h b/include/linux/xarray.h
index 2dfc8006fe64..c89d90515939 100644
--- a/include/linux/xarray.h
+++ b/include/linux/xarray.h
@@ -9,6 +9,58 @@
#include <linux/spinlock.h>
+/**
+ * xa_tag_pointer() - Create an XArray entry for a tagged pointer.
+ * @p: Plain pointer.
+ * @tag: Tag value (0 or 1).
+ *
+ * If the user of the XArray prefers, they can tag their pointers instead
+ * of storing value entries. Two tags are available (0 and 1).
+ * These are distinct from the xa_mark_t as they are not replicated up
+ * through the array and cannot be searched for.
+ *
+ * Context: Any context.
+ * Return: An XArray entry.
+ */
+static inline void *xa_tag_pointer(void *p, unsigned long tag)
+{
+ if (__builtin_constant_p(tag))
+ BUILD_BUG_ON(tag > 1);
+ else
+ BUG_ON(tag > 1);
+ return (void *)((unsigned long)p | (tag << 1));
+}
+
+/**
+ * xa_untag_pointer() - Turn an XArray entry into a plain pointer.
+ * @entry: XArray entry.
+ *
+ * If you have stored a tagged pointer in the XArray, call this function
+ * to get the untagged version of the pointer.
+ *
+ * Context: Any context.
+ * Return: A pointer.
+ */
+static inline void *xa_untag_pointer(void *entry)
+{
+ return (void *)((unsigned long)entry & ~3UL);
+}
+
+/**
+ * xa_pointer_tag() - Get the tag stored in an XArray entry.
+ * @entry: XArray entry.
+ *
+ * If you have stored a tagged pointer in the XArray, call this function
+ * to get the tag of that pointer.
+ *
+ * Context: Any context.
+ * Return: A tag.
+ */
+static inline unsigned int xa_pointer_tag(void *entry)
+{
+ return ((unsigned long)entry & 3UL) >> 1;
+}
+
#define xa_trylock(xa) spin_trylock(&(xa)->xa_lock)
#define xa_lock(xa) spin_lock(&(xa)->xa_lock)
#define xa_unlock(xa) spin_unlock(&(xa)->xa_lock)
--
2.14.5
Current ALSA firewire-motu driver uses the value of 'model' field
of unit directory in configuration ROM for modalias for MOTU
FireWire models. However, as long as I checked, Pre8 and
828mk3(Hybrid) have the same value for the field (=0x100800).
unit | version | model
--------------- | --------- | ----------
828mkII | 0x000003 | 0x101800
Traveler | 0x000009 | 0x107800
Pre8 | 0x00000f | 0x100800 <-
828mk3(FW) | 0x000015 | 0x106800
AudioExpress | 0x000033 | 0x104800
828mk3(Hybrid) | 0x000035 | 0x100800 <-
When updating firmware for MOTU 8pre FireWire from v1.0.0 to v1.0.3,
I got change of the value from 0x100800 to 0x103800. On the other
hand, the value of 'version' field is fixed to 0x00000f. As a quick
glance, the higher 12 bits of the value of 'version' field represent
firmware version, while the lower 12 bits is unknown.
By induction, the value of 'version' field represents actual model.
This commit changes modalias to match the value of 'version' field,
instead of 'model' field. For degug, long name of added sound card
includes hexadecimal value of 'model' field.
Cc: <stable(a)vger.kernel.org> # v4.19+
Fixes: 6c5e1ac0e144 ("ALSA: firewire-motu: add support for Motu Traveler")
Signed-off-by: Takashi Sakamoto <o-takashi(a)sakamocchi.jp>
---
sound/firewire/motu/motu.c | 20 ++++++++++----------
1 file changed, 10 insertions(+), 10 deletions(-)
diff --git a/sound/firewire/motu/motu.c b/sound/firewire/motu/motu.c
index 220e61926ea4..513291ba0ab0 100644
--- a/sound/firewire/motu/motu.c
+++ b/sound/firewire/motu/motu.c
@@ -36,7 +36,7 @@ static void name_card(struct snd_motu *motu)
fw_csr_iterator_init(&it, motu->unit->directory);
while (fw_csr_iterator_next(&it, &key, &val)) {
switch (key) {
- case CSR_VERSION:
+ case CSR_MODEL:
version = val;
break;
}
@@ -46,7 +46,7 @@ static void name_card(struct snd_motu *motu)
strcpy(motu->card->shortname, motu->spec->name);
strcpy(motu->card->mixername, motu->spec->name);
snprintf(motu->card->longname, sizeof(motu->card->longname),
- "MOTU %s (version:%d), GUID %08x%08x at %s, S%d",
+ "MOTU %s (version:%06x), GUID %08x%08x at %s, S%d",
motu->spec->name, version,
fw_dev->config_rom[3], fw_dev->config_rom[4],
dev_name(&motu->unit->device), 100 << fw_dev->max_speed);
@@ -237,20 +237,20 @@ static const struct snd_motu_spec motu_audio_express = {
#define SND_MOTU_DEV_ENTRY(model, data) \
{ \
.match_flags = IEEE1394_MATCH_VENDOR_ID | \
- IEEE1394_MATCH_MODEL_ID | \
- IEEE1394_MATCH_SPECIFIER_ID, \
+ IEEE1394_MATCH_SPECIFIER_ID | \
+ IEEE1394_MATCH_VERSION, \
.vendor_id = OUI_MOTU, \
- .model_id = model, \
.specifier_id = OUI_MOTU, \
+ .version = model, \
.driver_data = (kernel_ulong_t)data, \
}
static const struct ieee1394_device_id motu_id_table[] = {
- SND_MOTU_DEV_ENTRY(0x101800, &motu_828mk2),
- SND_MOTU_DEV_ENTRY(0x107800, &snd_motu_spec_traveler),
- SND_MOTU_DEV_ENTRY(0x106800, &motu_828mk3), /* FireWire only. */
- SND_MOTU_DEV_ENTRY(0x100800, &motu_828mk3), /* Hybrid. */
- SND_MOTU_DEV_ENTRY(0x104800, &motu_audio_express),
+ SND_MOTU_DEV_ENTRY(0x000003, &motu_828mk2),
+ SND_MOTU_DEV_ENTRY(0x000009, &snd_motu_spec_traveler),
+ SND_MOTU_DEV_ENTRY(0x000015, &motu_828mk3), /* FireWire only. */
+ SND_MOTU_DEV_ENTRY(0x000035, &motu_828mk3), /* Hybrid. */
+ SND_MOTU_DEV_ENTRY(0x000033, &motu_audio_express),
{ }
};
MODULE_DEVICE_TABLE(ieee1394, motu_id_table);
--
2.19.1