drm/fb-helper: Fix a locking bug in an error path
Subsystems (2)
PATCH 1/1 drm/fb-helper: Fix a locking bug in an error path
The name of the function __drm_fb_helper_initial_config_and_unlock() and also the comment above that function make it clear that all code paths in this function should unlock fb_helper->lock before returning. Add a mutex_unlock() call in the only code path where it is missing. This has been detected by the Clang thread-safety analyzer.
--- drivers/gpu/drm/drm_fb_helper.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/gpu/drm/drm_fb_helper.c +3 -1
@@ -1641,8 +1641,10 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)drm_client_modeset_probe (&fb_helper->client, width, height); info =drm_fb_helper_alloc_info (fb_helper); -if (IS_ERR (info)) +if (IS_ERR (info)) { +mutex_unlock (&fb_helper->lock);return PTR_ERR (info); + } ret =drm_fb_helper_single_fb_probe (fb_helper);if (ret <0 ) {
Hi
Am 03.04.26 um 22:53 schrieb Bart Van Assche:
Quoted 6 lines
The name of the function
__drm_fb_helper_initial_config_and_unlock()and also the comment above that function make it clear that all code paths in this function should unlockfb_helper->lockbefore returning. Add amutex_unlock()call in the only code path where it is missing. This has been detected by the Clang thread-safety analyzer.
Quoted 6 lines
Cc Thomas Zimmermann <[email protected]>Cc Christian König <[email protected]> # radeonCc Dmitry Baryshkov <[email protected]> # msmCc Javier Martinez Canillas <[email protected]>Fixes 63c971af4036 ("drm/fb-helper: Allocate and release fb_info in single place")Signed-off-by Bart Van Assche <[email protected]>
Thanks a lot.
Best regards
Thomas
Quoted 20 lines
--- drivers/gpu/drm/drm_fb_helper.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)drivers/gpu/drm/drm_fb_helper.c +3
-1@@ -1641,8 +1641,10 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)drm_client_modeset_probe (&fb_helper->client, width, height); info =drm_fb_helper_alloc_info (fb_helper); -if (IS_ERR (info)) +if (IS_ERR (info)) { +mutex_unlock (&fb_helper->lock);return PTR_ERR (info); + } ret =drm_fb_helper_single_fb_probe (fb_helper);if (ret <0 ) {
Signature · 6 lines
-- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
Thomas Zimmermann
maintainer(cc: dri-devel)
Am 07.04.26 um 09:44 schrieb Thomas Zimmermann:
Quoted 43 lines
Hi
Am 03.04.26 um 22:53 schrieb Bart Van Assche:
Quoted 6 lines
The name of the function
__drm_fb_helper_initial_config_and_unlock()and also the comment above that function make it clear that all code paths in this function should unlockfb_helper->lockbefore returning. Add amutex_unlock()call in the only code path where it is missing. This has been detected by the Clang thread-safety analyzer.Quoted 6 lines
Cc Thomas Zimmermann <[email protected]>Cc Christian König <[email protected]> # radeonCc Dmitry Baryshkov <[email protected]> # msmCc Javier Martinez Canillas <[email protected]>Fixes 63c971af4036 ("drm/fb-helper: Allocate and release fb_info in single place")Signed-off-by Bart Van Assche <[email protected]>Reviewed-by Thomas Zimmermann <[email protected]>Thanks a lot.
Best regards
ThomasQuoted 19 lines
---
drivers/gpu/drm/drm_fb_helper.c| 4 +++-1 file changed, 3 insertions(+), 1 deletion(-)drivers/gpu/drm/drm_fb_helper.c +3
-1@@ -1641,8 +1641,10 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper) drm_client_modeset_probe(&fb_helper->client, width, height); info = drm_fb_helper_alloc_info(fb_helper); - if (IS_ERR(info)) + if (IS_ERR(info)) { + mutex_unlock(&fb_helper->lock); return PTR_ERR(info); + } ret = drm_fb_helper_single_fb_probe(fb_helper); if (ret < 0) {
Signature · 6 lines
-- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
Quoted 1 line · Thomas Zimmermann
(cc: dri-devel)
Hi Thomas,
Is dri-devel the right mailing list to Cc for this patch? If so, how about adding an L: line to the drivers/gpu/drm/ entry in the MAINTAINERS file? That entry currently doesn't have an L: line as one can see below:
DRM DRIVERS AND MISC GPU PATCHES
M: Maarten Lankhorst <[email protected]>
M: Maxime Ripard <[email protected]>
M: Thomas Zimmermann <[email protected]>
S: Maintained
W: https://drm.pages.freedesktop.org/maintainer-tools/drm-misc.html
T: git https://gitlab.freedesktop.org/drm/misc/kernel.git
F: Documentation/devicetree/bindings/display/
F: Documentation/devicetree/bindings/gpu/
F: Documentation/gpu/
F: drivers/gpu/drm/
F: drivers/gpu/vga/
F: include/drm/drm
F: include/linux/vga*
F: include/uapi/drm/
X: drivers/gpu/drm/amd/
X: drivers/gpu/drm/armada/
X: drivers/gpu/drm/etnaviv/
X: drivers/gpu/drm/exynos/
X: drivers/gpu/drm/i915/
X: drivers/gpu/drm/kmb/
X: drivers/gpu/drm/mediatek/
X: drivers/gpu/drm/msm/
X: drivers/gpu/drm/nova/
X: drivers/gpu/drm/radeon/
X: drivers/gpu/drm/tegra/
X: drivers/gpu/drm/tyr/
X: drivers/gpu/drm/xe/
Thanks,
Bart.
Thomas Zimmermann
maintainerHi
Am 07.04.26 um 17:17 schrieb Bart Van Assche:
Quoted 7 lines · Thomas Zimmermann
Quoted 1 line
(cc: dri-devel)
Hi Thomas,
Is dri-devel the right mailing list to Cc for this patch? If so, how about adding an L: line to the
drivers/gpu/drm/entry in theMAINTAINERSfile? That entry currently doesn't have an L: line as one can see below:
It is covered by the general "DRM DRIVERS" entry and tools seem to get it right.
But yeah, that L: entry seems missing here. And it's so obviously missing that I wonder if it's intentional.
Best regards
Thomas
Quoted 33 lines
DRM DRIVERS AND MISC GPU PATCHES
M: Maarten Lankhorst <[email protected]> M: Maxime Ripard <[email protected]>
M: Thomas Zimmermann <[email protected]>
S: Maintained
W: https://drm.pages.freedesktop.org/maintainer-tools/drm-misc.html
T: git https://gitlab.freedesktop.org/drm/misc/kernel.git
F:Documentation/devicetree/bindings/display/
F:Documentation/devicetree/bindings/gpu/
F:Documentation/gpu/
F:drivers/gpu/drm/
F:drivers/gpu/vga/
F:include/drm/drm
F:include/linux/vga*
F:include/uapi/drm/
X:drivers/gpu/drm/amd/
X:drivers/gpu/drm/armada/
X:drivers/gpu/drm/etnaviv/
X:drivers/gpu/drm/exynos/
X:drivers/gpu/drm/i915/
X:drivers/gpu/drm/kmb/
X:drivers/gpu/drm/mediatek/
X:drivers/gpu/drm/msm/
X:drivers/gpu/drm/nova/
X:drivers/gpu/drm/radeon/
X:drivers/gpu/drm/tegra/
X:drivers/gpu/drm/tyr/
X:drivers/gpu/drm/xe/Thanks,
Bart.
Signature · 6 lines
-- Thomas Zimmermann Graphics Driver Developer SUSE Software Solutions Germany GmbH Frankenstr. 146, 90461 Nürnberg, Germany, www.suse.com GF: Jochen Jaser, Andrew McDonald, Werner Knoblich, (HRB 36809, AG Nürnberg)
Quoted 6 lines · Bart Van Assche
The name of the function
__drm_fb_helper_initial_config_and_unlock()and also the comment above that function make it clear that all code paths in this function should unlockfb_helper->lockbefore returning. Add amutex_unlock()call in the only code path where it is missing. This has been detected by the Clang thread-safety analyzer.
Quoted 7 lines
Cc Thomas Zimmermann <[email protected]>Cc Christian König <[email protected]> # radeonCc Dmitry Baryshkov <[email protected]> # msmCc Javier Martinez Canillas <[email protected]>Fixes 63c971af4036 ("drm/fb-helper: Allocate and release fb_info in single place")Signed-off-by Bart Van Assche <[email protected]>---
Indeed, thanks for fixing this.
Signature · 5 lines
Best regards, Javier Martinez Canillas Core Platforms Red Hat