Closed Bug 1031553 Opened 10 years ago Closed 10 years ago

b2g crashes during enabling data connection

Categories

(Core :: Graphics: Layers, defect, P1)

32 Branch
ARM
Gonk (Firefox OS)
defect

Tracking

()

RESOLVED FIXED
mozilla33
blocking-b2g 2.0+
Tracking Status
firefox31 --- wontfix
firefox32 --- fixed
firefox33 --- fixed
b2g-v2.0 --- fixed
b2g-v2.1 --- fixed

People

(Reporter: tkundu, Assigned: sotaro)

References

Details

(Keywords: crash, Whiteboard: [caf-crash 240][caf priority: p1][CR 686673][b2g-crash])

Attachments

(3 files, 1 obsolete file)

Attached file stack trace
Test steps:
1. Run music, send sms, call another phone from test device in automation.
2. Enable data connection on the device.
3. Device is rebooting and mini dumps are generated on the phone.

STR will be very difficult to follow but if you can give us right direction then we can enable additional logs which may help debugging this crash
blocking-b2g: --- → 2.0?
Flags: needinfo?(msreckovic)
Whiteboard: [CR 686673]
Whiteboard: [CR 686673] → [caf priority: p1][CR 686673]
Crash observed on: 

Device: msm8610
Gonk Version: AU_LINUX_GECKO_B2G_KK_3.6.01.04.00.000.016
Moz BuildID: 20140623000201
B2G Version: 2.0
Gecko Version: 32.0a2
Gaia:  http://git.mozilla.org/?p=releases/gaia.git;a=commit;h=729f214b887ce8efe7d870145d31acb2c6427817
Gecko: http://git.mozilla.org/?p=releases/gecko.git;a=commit;h=f3721e7a9961f09722975d0d76af31130f8847ef
Component: General → Graphics: Layers
Keywords: crash
Product: Firefox OS → Core
Whiteboard: [caf priority: p1][CR 686673] → [caf priority: p1][CR 686673][b2g-crash]
Version: unspecified → 32 Branch
This also reproduces on v2.0 7x27a builds.
(Seen 90 times in the last week on v2.0 across all devices)
blocking-b2g: 2.0? → 2.0+
Observed on: 

Device: msm8610
Gonk Version: AU_LINUX_GECKO_B2G_KK_3.6.01.04.00.000.020
Moz BuildID: 20140628000201
B2G Version: 2.0
Gecko Version: 32.0a2
Gaia:  http://git.mozilla.org/?p=releases/gaia.git;a=commit;h=91958625774ebe6c425adf322b09f1edf906578d
Gecko: http://git.mozilla.org/?p=releases/gecko.git;a=commit;h=eab4f8ff82758f08f049e14d305e29fcc50cb292
(In reply to Tapas Kumar Kundu from comment #0)
> Created attachment 8447427 [details]
> stack trace

The crash happens withing egl library. We do not have a source code of egl library :-(
Assignee: nobody → sotaro.ikeda.g
Bug 1029856 might affect to this bug.
Flags: needinfo?(msreckovic) → needinfo?(milan)
It seems very difficult to reproduce the problem. one possibility is "gralloc buffer is null". As another possible cause, Bug 1029856 or Bug 1031527 might related to this bug.
(In reply to Sotaro Ikeda [:sotaro] from comment #6)
> (In reply to Tapas Kumar Kundu from comment #0)
> > Created attachment 8447427 [details]
> > stack trace
> 
> The crash happens withing egl library. We do not have a source code of egl
> library :-(

I also agree :). I am wondering whether any operation of gecko can lead to this crash in our code or not. 
I will try with test patch from bug 1029856 and bug 1031527 to see whether that is the root cause of this issue or not.
Attachment #8448239 - Flags: review?(jmuizelaar)
Comment on attachment 8448239 [details] [diff] [review]
patch - Add more validity checks to GrallocTextureSourceOGL

Review of attachment 8448239 [details] [diff] [review]:
-----------------------------------------------------------------

This patch worries me. I'm sort of worried that this patch is covering up a more serious issue and will make it harder to find. Do we have any idea why the GrallocTextureHost is no longer valid in these cases?
(In reply to Jeff Muizelaar [:jrmuizel] from comment #11)
> Comment on attachment 8448239 [details] [diff] [review]
> patch - Add more validity checks to GrallocTextureSourceOGL
> 
> Review of attachment 8448239 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> This patch worries me. I'm sort of worried that this patch is covering up a
> more serious issue and will make it harder to find. Do we have any idea why
> the GrallocTextureHost is no longer valid in these cases?

There are two possibliteis.
- TextureHost does not have gralloc buffer since it's construction.
- TextureClient/TextureParent is destroyed and TextureClient had TextureFlags::DEALLOCATE_CLIENT flag.
Comment on attachment 8448239 [details] [diff] [review]
patch - Add more validity checks to GrallocTextureSourceOGL

Clear review request. I am going to address the concern.
Attachment #8448239 - Flags: review?(jmuizelaar)
Observed on: 

Device: msm8610
Gonk Version: AU_LINUX_GECKO_B2G_KK_3.6.01.04.00.000.017
Moz BuildID: 20140624000201
B2G Version: 2.0
Gecko Version: 32.0a2
Gaia:  http://git.mozilla.org/?p=releases/gaia.git;a=commit;h=9d2f7bd16a8dc0c74c97c5a40d2f0731f3dfff4b
Gecko: http://git.mozilla.org/?p=releases/gecko.git;a=commit;h=32c226e5a7adbb95e9c4ee003dc9e64699da03e1
Observed on: 

Device: msm8610
Gonk Version: AU_LINUX_GECKO_B2G_KK_3.6.01.04.00.000.021
Moz BuildID: 20140630000201
B2G Version: 2.0
Gecko Version: 32.0a2
Gaia:  http://git.mozilla.org/?p=releases/gaia.git;a=commit;h=c0c8ad187c0466285f2580531e09f8322996f561
Gecko: http://git.mozilla.org/?p=releases/gecko.git;a=commit;h=ef0781044f85ee7f364236c074ab0b7644c38ff8
Status: NEW → ASSIGNED
Change the patch based on the comment. Did the following changes.
- TextureClient create TextureHost only when the TextureClient allocated buffer.
- Create GrallocTextureSourceOGL only when GrallocTextureHostOGL has a gralloc buffer
- GrallocTextureHostOGL forgets GrallocTextureSourceOGL when GrallocTextureHostOGL forgets the gralloc buffer.
- Add pointer checks to GrallocTextureHostOGL
Attachment #8448239 - Attachment is obsolete: true
Attachment #8449671 - Flags: review?(nical.bugzilla)
Attachment #8449671 - Flags: review?(jmuizelaar)
Flags: needinfo?(milan)
Attachment #8449671 - Flags: review?(nical.bugzilla) → review+
Comment on attachment 8449671 [details] [diff] [review]
patch - Create GrallocTextureSourceOGL only when GrallocTextureHostOGL has a gralloc buffer

Cancel jeff's review. nical's review+ is enough.
Attachment #8449671 - Flags: review?(jmuizelaar)
https://hg.mozilla.org/mozilla-central/rev/b7f8089105b2
Status: ASSIGNED → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla33
Whiteboard: [caf priority: p1][CR 686673][b2g-crash] → [caf-crash 240][caf priority: p1][CR 686673][b2g-crash]
New test case needs to be added. There is no existing test case.
QA Whiteboard: [QAnalyst-Triage?]
Flags: needinfo?(ktucker)
Flags: in-moztrap?(rmead)
This is done with Automation. No test case added to moztrap.
QA Whiteboard: [QAnalyst-Triage?] → [QAnalyst-Triage+]
Flags: needinfo?(ktucker)
Flags: in-moztrap?(rmead)
Flags: in-moztrap-
Unable to verify this issue because it is an automation crash.
QA Whiteboard: [QAnalyst-Triage+] → [QAnalyst-verify-][QAnalyst-Triage?]
Flags: needinfo?(ktucker)
QA Whiteboard: [QAnalyst-verify-][QAnalyst-Triage?] → [QAnalyst-verify-][QAnalyst-Triage+]
Flags: needinfo?(ktucker)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: