Skip to contentSkip to conversation

drm/i915: Fix crash when failing to parse MIPI VBT

PATCH Applied [email protected] 3 participants 5 messages Jul 24, 2014 – Jul 25, 2014

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: Fix crash when failing to parse MIPI VBT

From: Rafael Barbalho <[email protected]>

This particular nasty presented itself while trying to register the intelfb device (intel_fbdev.c). During the process of registering the device the driver will disable the crtc via i9xx_crtc_disable. These will also disable the panel using the generic mipi panel functions in dsi_mod_vbt_generic.c. The stale MIPI generic data sequence pointers would cause a crash within those functions. However, all of this is happening while console_lock is held from do_register_framebuffer inside fbcon.c. Which means that you got kernel log and just the device appearing to reboot/hang for no apparent reason.

The fault started from the FB_EVENT_FB_REGISTERED event using the fb_notifier_call_chain call in fbcon.c.

---
 drivers/gpu/drm/i915/intel_bios.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_bios.c +1 -1
index 608ed30..a669550 100644
@@ -878,7 +878,7 @@ err:
 
 	/* error during parsing so set all pointers to null
 	 * because of partial parsing */
-	memset(dev_priv->vbt.dsi.sequence, 0, MIPI_SEQ_MAX);
+	memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence));
 }
 
 static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
Signature · 1 line
2.0.2
Reviewed-by on v1 1/1
#1 lore raw
Quoted 26 lines · On 7/24/2014 7:46 PM, [email protected] wrote:

From: Rafael Barbalho <[email protected]>

This particular nasty presented itself while trying to register the intelfb device (intel_fbdev.c). During the process of registering the device the driver will disable the crtc via i9xx_crtc_disable. These will also disable the panel using the generic mipi panel functions in dsi_mod_vbt_generic.c. The stale MIPI generic data sequence pointers would cause a crash within those functions. However, all of this is happening while console_lock is held from do_register_framebuffer inside fbcon.c. Which means that you got kernel log and just the device appearing to reboot/hang for no apparent reason.

The fault started from the FB_EVENT_FB_REGISTERED event using the fb_notifier_call_chain call in fbcon.c.

---
  drivers/gpu/drm/i915/intel_bios.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_bios.c +0 -0
index 608ed30..a669550 100644

drivers/gpu/drm/i915/intel_bios.c err: · line 881

Quoted 5 lines
drivers/gpu/drm/i915/intel_bios.c +1 -1
@@ -878,7 +878,7 @@ err:
 
  	/* error during parsing so set all pointers to null
  	 * because of partial parsing */
-	memset(dev_priv->vbt.dsi.sequence, 0, MIPI_SEQ_MAX);
+	memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence));

Ouch !! This mistake hurts. This is manifesting now because the VBT probably you tested had sequences not supported by the driver. I am influencing a TLV based VBT structure design for the sequence which when done will ensure proper parsing for all that is known and unknown pointers will remain NULL. But for now

Regards
Shobhit

#2 lore raw
In reply to [email protected]
Quoted 11 lines · On Thu, Jul 24, 2014 at 03:16:12PM +0100, [email protected] wrote:

From: Rafael Barbalho <[email protected]>

This particular nasty presented itself while trying to register the intelfb device (intel_fbdev.c). During the process of registering the device the driver will disable the crtc via i9xx_crtc_disable. These will also disable the panel using the generic mipi panel functions in dsi_mod_vbt_generic.c. The stale MIPI generic data sequence pointers would cause a crash within those functions. However, all of this is happening while console_lock is held from do_register_framebuffer inside fbcon.c. Which means that you got kernel log and just the device appearing to reboot/hang for no apparent reason.

CONFIG_I915_FBDEV=n for when the console_lock gets in the way.

Quoted 5 lines

The fault started from the FB_EVENT_FB_REGISTERED event using the fb_notifier_call_chain call in fbcon.c.

Queued for -next, thanks for the patch.
-Daniel

Quoted 24 lines
---
 drivers/gpu/drm/i915/intel_bios.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_bios.c +1 -1
index 608ed30..a669550 100644
@@ -878,7 +878,7 @@ err:
 
 	/* error during parsing so set all pointers to null
 	 * because of partial parsing */
-	memset(dev_priv->vbt.dsi.sequence, 0, MIPI_SEQ_MAX);
+	memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence));
 }
 
 static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
Signature · 6 lines
2.0.2

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#3 lore raw
Quoted 15 lines · Daniel Vetter
Quoted 11 lines · On Thu, Jul 24, 2014 at 03:16:12PM +0100, [email protected] wrote:

From: Rafael Barbalho <[email protected]>

This particular nasty presented itself while trying to register the intelfb device (intel_fbdev.c). During the process of registering the device the driver will disable the crtc via i9xx_crtc_disable. These will also disable the panel using the generic mipi panel functions in dsi_mod_vbt_generic.c. The stale MIPI generic data sequence pointers would cause a crash within those functions. However, all of this is happening while console_lock is held from do_register_framebuffer inside fbcon.c. Which means that you got kernel log and just the device appearing to reboot/hang for no apparent reason.

CONFIG_I915_FBDEV=n for when the console_lock gets in the way.

Quoted 7 lines
Quoted 5 lines

The fault started from the FB_EVENT_FB_REGISTERED event using the fb_notifier_call_chain call in fbcon.c.

Queued for -next, thanks for the patch.

Actually this is for fixes since 3.16 has dsi support. Also for regressions please cite the commit that introduced the offending behaviour. I've added that.
-Daniel

Quoted 30 lines

-Daniel

Quoted 24 lines
---
 drivers/gpu/drm/i915/intel_bios.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/i915/intel_bios.c +1 -1
index 608ed30..a669550 100644
@@ -878,7 +878,7 @@ err:
 
 	/* error during parsing so set all pointers to null
 	 * because of partial parsing */
-	memset(dev_priv->vbt.dsi.sequence, 0, MIPI_SEQ_MAX);
+	memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence));
 }
 
 static void parse_ddi_port(struct drm_i915_private *dev_priv, enum port port,
Signature · 6 lines
2.0.2

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
#4 lore raw
Quoted 19 lines · Daniel Vetter
Quoted 18 lines · Daniel Vetter
Quoted 11 lines · On Thu, Jul 24, 2014 at 03:16:12PM +0100, [email protected] wrote:

From: Rafael Barbalho <[email protected]>

This particular nasty presented itself while trying to register the intelfb device (intel_fbdev.c). During the process of registering the device the driver will disable the crtc via i9xx_crtc_disable. These will also disable the panel using the generic mipi panel functions in dsi_mod_vbt_generic.c. The stale MIPI generic data sequence pointers would cause a crash within those functions. However, all of this is happening while console_lock is held from do_register_framebuffer inside fbcon.c. Which means that you got kernel log and just the device appearing to reboot/hang for no apparent reason.

CONFIG_I915_FBDEV=n for when the console_lock gets in the way.

Quoted 3 lines

The fault started from the FB_EVENT_FB_REGISTERED event using the fb_notifier_call_chain call in fbcon.c.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

Queued for -next, thanks for the patch.

Actually this is for fixes since 3.16 has dsi support. Also for regressions please cite the commit that introduced the offending behaviour. I've added that.

Also this reminds me that there is still a WARN dump in 3.16 which will be fixed by - [v2] drm/i915: Add correct hw/sw config check for DSI encoder

Assuming this can go in -fixes if it okay, this is waiting for review

Regards
Shobhit