diff mbox series

drm: Mark up racy check of drm_gem_object.handle_count

Message ID 20200309120151.7675-1-chris@chris-wilson.co.uk (mailing list archive)
State New, archived
Headers show
Series drm: Mark up racy check of drm_gem_object.handle_count | expand

Commit Message

Chris Wilson March 9, 2020, 12:01 p.m. UTC
[ 1715.899800] BUG: KCSAN: data-race in drm_gem_handle_create_tail / drm_gem_object_handle_put_unlocked
[ 1715.899838]
[ 1715.899861] write to 0xffff8881830f3604 of 4 bytes by task 7834 on cpu 1:
[ 1715.899896]  drm_gem_handle_create_tail+0x62/0x250
[ 1715.899927]  drm_gem_open_ioctl+0xc1/0x160
[ 1715.899956]  drm_ioctl_kernel+0xe4/0x120
[ 1715.899981]  drm_ioctl+0x297/0x4c7
[ 1715.900003]  ksys_ioctl+0x89/0xb0
[ 1715.900027]  __x64_sys_ioctl+0x42/0x60
[ 1715.900052]  do_syscall_64+0x6e/0x2c0
[ 1715.900079]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[ 1715.900100]
[ 1715.900119] read to 0xffff8881830f3604 of 4 bytes by task 8137 on cpu 0:
[ 1715.900149]  drm_gem_object_handle_put_unlocked+0x31/0x130
[ 1715.900180]  drm_gem_object_release_handle+0x93/0xe0
[ 1715.900208]  drm_gem_handle_delete+0x7b/0xe0
[ 1715.900235]  drm_gem_close_ioctl+0x61/0x80
[ 1715.900264]  drm_ioctl_kernel+0xe4/0x120
[ 1715.900291]  drm_ioctl+0x297/0x4c7
[ 1715.900316]  ksys_ioctl+0x89/0xb0
[ 1715.900340]  __x64_sys_ioctl+0x42/0x60
[ 1715.900363]  do_syscall_64+0x6e/0x2c0
[ 1715.900388]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/drm_gem.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Daniel Vetter March 16, 2020, 10:03 a.m. UTC | #1
On Mon, Mar 09, 2020 at 12:01:51PM +0000, Chris Wilson wrote:
> [ 1715.899800] BUG: KCSAN: data-race in drm_gem_handle_create_tail / drm_gem_object_handle_put_unlocked
> [ 1715.899838]
> [ 1715.899861] write to 0xffff8881830f3604 of 4 bytes by task 7834 on cpu 1:
> [ 1715.899896]  drm_gem_handle_create_tail+0x62/0x250
> [ 1715.899927]  drm_gem_open_ioctl+0xc1/0x160
> [ 1715.899956]  drm_ioctl_kernel+0xe4/0x120
> [ 1715.899981]  drm_ioctl+0x297/0x4c7
> [ 1715.900003]  ksys_ioctl+0x89/0xb0
> [ 1715.900027]  __x64_sys_ioctl+0x42/0x60
> [ 1715.900052]  do_syscall_64+0x6e/0x2c0
> [ 1715.900079]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> [ 1715.900100]
> [ 1715.900119] read to 0xffff8881830f3604 of 4 bytes by task 8137 on cpu 0:
> [ 1715.900149]  drm_gem_object_handle_put_unlocked+0x31/0x130
> [ 1715.900180]  drm_gem_object_release_handle+0x93/0xe0
> [ 1715.900208]  drm_gem_handle_delete+0x7b/0xe0
> [ 1715.900235]  drm_gem_close_ioctl+0x61/0x80
> [ 1715.900264]  drm_ioctl_kernel+0xe4/0x120
> [ 1715.900291]  drm_ioctl+0x297/0x4c7
> [ 1715.900316]  ksys_ioctl+0x89/0xb0
> [ 1715.900340]  __x64_sys_ioctl+0x42/0x60
> [ 1715.900363]  do_syscall_64+0x6e/0x2c0
> [ 1715.900388]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

I'm impressed. The dream that we all get replaced by some scripts might be
real :-)

> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>

> ---
>  drivers/gpu/drm/drm_gem.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
> index a9e4a610445a..37627d06fb06 100644
> --- a/drivers/gpu/drm/drm_gem.c
> +++ b/drivers/gpu/drm/drm_gem.c
> @@ -218,7 +218,7 @@ drm_gem_object_handle_put_unlocked(struct drm_gem_object *obj)
>  	struct drm_device *dev = obj->dev;
>  	bool final = false;
>  
> -	if (WARN_ON(obj->handle_count == 0))
> +	if (WARN_ON(READ_ONCE(obj->handle_count) == 0))
>  		return;
>  
>  	/*
> -- 
> 2.20.1
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
diff mbox series

Patch

diff --git a/drivers/gpu/drm/drm_gem.c b/drivers/gpu/drm/drm_gem.c
index a9e4a610445a..37627d06fb06 100644
--- a/drivers/gpu/drm/drm_gem.c
+++ b/drivers/gpu/drm/drm_gem.c
@@ -218,7 +218,7 @@  drm_gem_object_handle_put_unlocked(struct drm_gem_object *obj)
 	struct drm_device *dev = obj->dev;
 	bool final = false;
 
-	if (WARN_ON(obj->handle_count == 0))
+	if (WARN_ON(READ_ONCE(obj->handle_count) == 0))
 		return;
 
 	/*