drm/i915: prevent division by zero when asking for chipset power
Subsystems (2)
PATCH 1/1 drm/i915: prevent division by zero when asking for chipset power
This prevents an in-kernel division by zero which happens when we are asking for i915_chipset_val too quickly, or within a race condition between the power monitoring thread and userspace accesses via debugfs.
The issue can be reproduced easily via the following command:
while ``; do cat /sys/kernel/debug/dri/0/i915_emon_status; done
This is particularly dangerous because it can be triggered by a non-privileged user by just reading the debugfs entry.
This issue was also found independently by Konstantin Belousov <[email protected]>, who proposed a similar patch.
--- drivers/gpu/drm/i915/i915_dma.c | 10 ++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + 2 files changed, 11 insertions(+), 0 deletions(-)
drivers/gpu/drm/i915/i915_dma.c +10 -0
@@ -1453,6 +1453,14 @@ unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) diff1 = now - dev_priv->last_time1; +/* Prevent division-by-zero if we are asking too fast. +* Also, we don't get interesting results in we are polling +* faster than once in 10ms, so just return the saved value +* in such cases. +*/ +if (diff1 <=10 ) +return dev_priv->chipset_power; + count1 =I915_READ (DMIEC ); count2 =I915_READ (DDREC ); count3 =I915_READ (CSIEC );
@@ -1483,6 +1491,8 @@ unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) dev_priv->last_count1 = total_count; dev_priv->last_time1 = now; + dev_priv->chipset_power = ret; +return ret; }
drivers/gpu/drm/i915/i915_drv.h +1 -0
@@ -707,6 +707,7 @@ typedef struct drm_i915_private {u64 last_count1;unsigned long last_time1; +unsigned long chipset_power;u64 last_count2;struct timespec last_time2;unsigned long gfx_power;
Signature · 1 line
1.7.7.1
Quoted 10 lines · Eugeni Dodonov
This prevents an in-kernel division by zero which happens when we are asking for
i915_chipset_valtoo quickly, or within a race condition between the power monitoring thread and userspace accesses via debugfs.The issue can be reproduced easily via the following command:
while ``; do cat/sys/kernel/debug/dri/0/i915_emon_status; doneThis is particularly dangerous because it can be triggered by a non-privileged user by just reading the debugfs entry.
Quoted 8 lines
This issue was also found independently by Konstantin Belousov <[email protected]>, who proposed a similar patch.
Reported-by Konstantin Belousov <[email protected]>Acked-by Jesse Barnes <[email protected]>Acked-by Keith Packard <[email protected]>Cc <[email protected]>Signed-off-by Eugeni Dodonov <[email protected]>
Just spotted a typo...
Quoted 10 lines
--- drivers/gpu/drm/i915/i915_dma.c | 10 ++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + 2 files changed, 11 insertions(+), 0 deletions(-)drivers/gpu/drm/i915/i915_dma.c +0
-0
drivers/gpu/drm/i915/i915_dma.c unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) · line 1457
Quoted 5 lines
drivers/gpu/drm/i915/i915_dma.c +2
-0@@ -1453,6 +1453,14 @@ unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) diff1 = now - dev_priv->last_time1; +/* Prevent division-by-zero if we are asking too fast. +* Also, we don't get interesting results in we are polling
s/in/if/
-Chris
Signature · 1 line
Chris Wilson, Intel Open Source Technology Centre
PATCH 1/1 drm/i915: prevent division by zero when asking for chipset power
This prevents an in-kernel division by zero which happens when we are asking for i915_chipset_val too quickly, or within a race condition between the power monitoring thread and userspace accesses via debugfs.
The issue can be reproduced easily via the following command:
while ``; do cat /sys/kernel/debug/dri/0/i915_emon_status; done
This is particularly dangerous because it can be triggered by a non-privileged user by just reading the debugfs entry.
This issue was also found independently by Konstantin Belousov <[email protected]>, who proposed a similar patch.
--- drivers/gpu/drm/i915/i915_dma.c | 10 ++++++++++ drivers/gpu/drm/i915/i915_drv.h | 1 + 2 files changed, 11 insertions(+), 0 deletions(-)
drivers/gpu/drm/i915/i915_dma.c +10 -0
@@ -1453,6 +1453,14 @@ unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) diff1 = now - dev_priv->last_time1; +/* Prevent division-by-zero if we are asking too fast. +* Also, we don't get interesting results if we are polling +* faster than once in 10ms, so just return the saved value +* in such cases. +*/ +if (diff1 <=10 ) +return dev_priv->chipset_power; + count1 =I915_READ (DMIEC ); count2 =I915_READ (DDREC ); count3 =I915_READ (CSIEC );
@@ -1483,6 +1491,8 @@ unsigned long i915_chipset_val(struct drm_i915_private *dev_priv) dev_priv->last_count1 = total_count; dev_priv->last_time1 = now; + dev_priv->chipset_power = ret; +return ret; }
drivers/gpu/drm/i915/i915_drv.h +1 -0
@@ -707,6 +707,7 @@ typedef struct drm_i915_private {u64 last_count1;unsigned long last_time1; +unsigned long chipset_power;u64 last_count2;struct timespec last_time2;unsigned long gfx_power;
Signature · 1 line
1.7.7.1
Quoted 13 lines · On Thu, 10 Nov 2011 10:51:26 -0200 Eugeni Dodonov <[email protected]>…
This prevents an in-kernel division by zero which happens when we are asking for
i915_chipset_valtoo quickly, or within a race condition between the power monitoring thread and userspace accesses via debugfs.The issue can be reproduced easily via the following command:
while ``; do cat/sys/kernel/debug/dri/0/i915_emon_status; doneThis is particularly dangerous because it can be triggered by a non-privileged user by just reading the debugfs entry.
This issue was also found independently by Konstantin Belousov <[email protected]>, who proposed a similar patch.
Quoted 6 lines
Reported-by Konstantin Belousov <[email protected]>Acked-by Jesse Barnes <[email protected]>Acked-by Keith Packard <[email protected]>Cc <[email protected]>Signed-off-by Eugeni Dodonov <[email protected]>---
Yep, looks fine, thanks for pushing this Eugeni.
Signature · 1 line
Jesse Barnes, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx