Closed Bug 1255732 Opened 4 years ago Closed 3 years ago

Firefox crashes on Vimeo on AMD hardware

Categories

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

x86
Windows 8.1
defect

Tracking

()

RESOLVED FIXED
mozilla49
Tracking Status
firefox47 --- fixed
firefox48 --- fixed
firefox49 --- fixed

People

(Reporter: simona.marcu, Assigned: gerald)

References

Details

(Keywords: crash, Whiteboard: gfx-noted)

Crash Data

Attachments

(1 file)

[Note]:
- Initially reproduced this crash normally browsing while testing on a Nightly try build on a Child account 
https://archive.mozilla.org/pub/firefox/try-builds/dkeeler@mozilla.com-68763a0d343f41ad9370f145abacf55068334806/

[Affected versions]:
Firefox 44.0.2
Firefox 47.0a1
Firefox 48.0a1

[Affected platforms]:
Windows 8.1

[Steps to reproduce]:
1. Launch Firefox
2. Navigate YouTube and start playing any video (https://www.youtube.com/)
3. Navigate to Yahoo movies and start playing any trailer (https://www.yahoo.com/movies/)
4. Navigate to Vimeo start playing any video and click repeatedly on the Share button (https://vimeo.com/)


[Expected result]:
The tab where the Vimeo video is opened should not crash (on the latest Nightly).
Firefox should not crash (Firefox 44.0.2).

[Actual result]:

The Vimeo tab is crashing with the signatures:

- The 47.0a1 try build provided above crashes with the signature:
https://crash-stats.mozilla.com/report/index/8c47030e-e45f-4954-8128-de6042160310

- The regular nightly 48.0a1 downloaded from (https://archive.mozilla.org/pub/firefox/nightly/latest-mozilla-central/) crashes with the signature:
https://crash-stats.mozilla.com/report/index/294063a9-eb80-4f09-a12a-8b5332160310

- Firefox 44.0.2 crashes with the signature:
https://crash-stats.mozilla.com/report/index/590a21ee-a002-4255-a950-033682160310


[Regression range]: 
- According with the Socorro reports the crash is present since Firefox 18 [@ atidxx32.dll@0x742c7] and Firefox 25 [@ atiumdva.dll@0x14137 ]. If a more in depth regression is needed, please let me know.

[Additional notes]:
- the crash is reproducible also on an administrator account.
- the crash is more easy to reproduce on a child account on the latest Nightly.

Graphics
Adapter Description	AMD Radeon HD 5450
Adapter Description (GPU #2)	Intel(R) HD Graphics 4600
Adapter Drivers	aticfx32 aticfx32 aticfx32 atiumdag atidxx32 atiumdva
Adapter Drivers (GPU #2)	igdumdim32 igd10iumd32 igd10iumd32
Adapter RAM	1024
Adapter RAM (GPU #2)	Unknown
Asynchronous Pan/Zoom	wheel input enabled; touch input enabled
Device ID	0x68f9
Device ID (GPU #2)	0x0412
Direct2D Enabled	true
DirectWrite Enabled	true (6.3.9600.18123)
Driver Date	7-4-2014
Driver Date (GPU #2)	8-4-2015
Driver Version	13.251.9001.1001
Driver Version (GPU #2)	10.18.14.4264
GPU #2 Active	false
GPU Accelerated Windows	1/1 Direct3D 11 (OMTC)
Subsys ID	254a1458
Subsys ID (GPU #2)	0000000c
Supports Hardware H264 Decoding	Yes
Vendor ID	0x1002
Vendor ID (GPU #2)	0x8086
WebGL Renderer	Google Inc. -- ANGLE (AMD Radeon HD 5450 Direct3D11 vs_5_0 ps_5_0)
windowLayerManagerRemote	true
AzureCanvasBackend	direct2d 1.1
AzureContentBackend	direct2d 1.1
AzureFallbackCanvasBackend	cairo
AzureSkiaAccelerated	0
Summary: Firefox crashes on Vimeo → Firefox crashes on Vimeo on AMD hardware
I'll try to reproduce this when I get back to Toronto
Flags: needinfo?(jmuizelaar)
I wasn't able to reproduce this on a AMD Radeon HD 5670 with the 19.12.0.0 driver. Paul the crash is at atiumdva.dll@0x14137 can you give us an idea of what the driver thinks is going wrong here?
Flags: needinfo?(jmuizelaar) → needinfo?(paul.blinzer)
The issue apparently occurs on a Hybrid Graphics/PowerXpress system and the crash occurs in the Video Acceleration driver, the location is roughly in the DecodeFrame() function though I can't fully match up the offset right now. 
Need to set up a repro here on a debugger.
Flags: needinfo?(paul.blinzer)
Does this repro on any more recent hardware?
(In reply to Paul Blinzer from comment #4)
> Does this repro on any more recent hardware?

Here is a list of the top devices reported, sorted by codename. 

Codename [Device] (ID)                                          Crashes Percent
======================                                          ======= =======
Barts LE [Radeon HD 6790] (0x673e)                              1       0.51%
Barts XT [Radeon HD 6870] (0x6738)                              1       0.51%
BeaverCreek [Radeon HD 6520G] (0x9647)                          2       1.03%
BeaverCreek [Radeon HD 6530D] (0x964a)                          3       1.54%
Broadway XT [Mobility Radeon HD 5870] (0x68a0)                  1       0.51%
Caicos [Radeon HD 6450/7450/8450 / R5 230 OEM] (0x6779)         2       1.03%
Caicos PRO [Radeon HD 7450] (0x677b)                            1       0.51%
Caicos XT [Radeon HD 7470/8470 / R5 235/310 OEM] (0x6778)       13      6.67%
Caicos XTX [Radeon HD 8490 / R5 235X OEM] (0x6771)              1       0.51%
Cape Verde LE [Radeon HD 7730/8730] (0x6837)                    2       1.03%
Cape Verde PRO [Radeon HD 7750/8740 / R7 250E] (0x683f)         2       1.03%
Cape Verde XT [Radeon HD 7770/8760 / R7 250X] (0x683d)          3       1.54%
Cedar [Radeon HD 5000/6000/7350/8350 Series] (0x68f9)           21      10.77%
Curacao PRO [Radeon R9 270/370] (0x6811)                        1       0.51%
Curacao XT [Radeon R9 270X] (0x6810)                            2       1.03%
Cypress PRO [Radeon HD 5850] (0x6899)                           1       0.51%
Hawaii XT [Radeon R9 290X] (0x67b0)                             3       1.54%
Juniper PRO [Radeon HD 6750] (0x68bf)                           1       0.51%
Juniper XT [Radeon HD 6770] (0x68ba)                            2       1.03%
Kabini [Radeon HD 8210] (0x9834)                                3       1.54%
Kabini [Radeon HD 8240 / R3 Series] (0x9838)                    3       1.54%
Kabini [Radeon HD 8330] (0x9832)                                1       0.51%
Kabini [Radeon HD 8400 / R3 Series] (0x9830)                    1       0.51%
Madison [Mobility Radeon HD 5570/6550A] (0x68c7)                2       1.03%
Madison [Mobility Radeon HD 5650/5750 / 6530M/6550M] (0x68c1)   2       1.03%
Oland [Radeon HD 8570 / R7 240/340 OEM] (0x6611)                11      5.64%
Oland PRO [Radeon R7 240] (0x6613)                              5       2.56%
Oland XT [Radeon HD 8670 / R5 340X / R7 250/350X] (0x6610)      3       1.54%
Park [Mobility Radeon HD 5430] (0x68e1)                         2       1.03%
Park [Mobility Radeon HD 5430/5450/5470] (0x68e0)               3       1.54%
Pitcairn PRO [Radeon HD 7850] (0x6819)                          2       1.03%
Pitcairn XT [Radeon HD 7870 GHz Edition] (0x6818)               3       1.54%
Redwood XT [Radeon HD 5670/5690/5730] (0x68d8)                  4       2.05%
Richland [Radeon HD 8450G] (0x9995)                             1       0.51%
Richland [Radeon HD 8470D] (0x9996)                             1       0.51%
Richland [Radeon HD 8510G] (0x9999)                             2       1.03%
Richland [Radeon HD 8550G] (0x990d)                             2       1.03%
Richland [Radeon HD 8570D] (0x990e)                             4       2.05%
Richland [Radeon HD 8610G] (0x990f)                             2       1.03%
Richland [Radeon HD 8650G] (0x990b)                             4       2.05%
AMD Bug is filed and investigated
the signatures in this bug might be related to the enabling of d3d11 dxva as well (the signatures are just the most popular ones - there will be a number of other ones for amd gpus still).
Crash Signature: [@ atiumdva.dll@0x14137 ] [@ atidxx32.dll@0x742c7 ] → [@ atiumdva.dll@0x14137 ] [@ atidxx32.dll@0x742c7 ] [@ atidxx32.dll@0x72e57 ] [@ atidxx32.dll@0x708a7 ] [@ NDXGI::CDevice::SignalSynchronizationObjectCB ] [@ NDXGI::CDevice::WaitForSynchronizationObjectCB | atiuxpag.dll@0xc65a ] [@ NDXGI::CDevice::W…
Flags: needinfo?(gsquelart)
Crash Signature: NDXGI::CDevice::WaitForSynchronizationObjectCB | atiuxpag.dll@0xb28a ] → NDXGI::CDevice::WaitForSynchronizationObjectCB | atiuxpag.dll@0xb28a ] [@ atiumdva.dll@0xfca76]
Crash Signature: NDXGI::CDevice::WaitForSynchronizationObjectCB | atiuxpag.dll@0xb28a ] [@ atiumdva.dll@0xfca76] → NDXGI::CDevice::WaitForSynchronizationObjectCB | atiuxpag.dll@0xb28a ] [@ atiumdva.dll@0xfca76] [@ RtlpFreeHeap | RtlFreeHeap | atiumdva.dll@0x1c8ac8]
See Also: → 1269204
(In reply to philipp from comment #7)
> the signatures in this bug might be related to the enabling of d3d11 dxva as
> well (the signatures are just the most popular ones - there will be a number
> of other ones for amd gpus still).

I'm working on bug 1273691 first, which will make it easier to blacklist more DLLs as needed.

Keeping reminder to self, to add these suggested DLLs later on.
Flags: needinfo?(gsquelart)
Flags: needinfo?(gsquelart)
Depends on: 1273504
Flags: needinfo?(gsquelart)
Comment on attachment 8754620 [details]
MozReview Request: Bug 1255732 - Disable D3D11 for some atidxx32/64 DLLs - r?cpearce

https://reviewboard.mozilla.org/r/54102/#review50816
Attachment #8754620 - Flags: review?(cpearce) → review+
Assignee: nobody → gsquelart
Priority: -- → P1
Comment on attachment 8754620 [details]
MozReview Request: Bug 1255732 - Disable D3D11 for some atidxx32/64 DLLs - r?cpearce

Approval Request Comment
[Feature/regressing bug #]: D3D11 crashes in 47+
[User impact if declined]: many crashes per week in beta 47.
[Describe test coverage new/current, TreeHerder]: No new try as it's just a pref change adding more DLLs.
[Risks and why]: Only a pref change that adds some DLLs; Tiny risk in disabling too many DLLs, but we're falling back to proven D3D9, so it's not so bad.
[String/UUID change made/needed]: None.

Requires uplift of bug 1257028, bug 1273691, bug 1274115, bug 1274127, bug 1274132, bug 1273504 first.
Attachment #8754620 - Flags: approval-mozilla-beta?
Attachment #8754620 - Flags: approval-mozilla-aurora?
Comment on attachment 8754620 [details]
MozReview Request: Bug 1255732 - Disable D3D11 for some atidxx32/64 DLLs - r?cpearce

Let's uplift this to Aurora48 to stabilize it before uplifting to Beta47.
Attachment #8754620 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+
Comment on attachment 8754620 [details]
MozReview Request: Bug 1255732 - Disable D3D11 for some atidxx32/64 DLLs - r?cpearce

Disable D3D11 for some device drivers, this code stabilized on Aurora over the weekend, Beta47+
Attachment #8754620 - Flags: approval-mozilla-beta? → approval-mozilla-beta+
https://hg.mozilla.org/mozilla-central/rev/928fa0c9a879
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla49
i bought a rx 480 recently and since i installed it on my pc firefox crashes every hour. 
here three example links to the last crashes:
https://crash-stats.mozilla.com/report/index/09e74f17-b67c-41af-a017-27ad42160831
https://crash-stats.mozilla.com/report/index/c6a3a601-4630-4c87-af16-a8aae2160831
https://crash-stats.mozilla.com/report/index/655b2856-a8c3-4a54-b24a-f4d5e2160831

all mention this bug so i thought you maybe can reopen it to fix my problem as well.
im using the most recent amd driver 16.8.2 on windows 10. (i dont have any issues on ubuntu 16.04)
(In reply to Martin from comment #18)
> i bought a rx 480 recently and since i installed it on my pc firefox crashes
> every hour. 
> here three example links to the last crashes:
> https://crash-stats.mozilla.com/report/index/09e74f17-b67c-41af-a017-
> 27ad42160831
> https://crash-stats.mozilla.com/report/index/c6a3a601-4630-4c87-af16-
> a8aae2160831
> https://crash-stats.mozilla.com/report/index/655b2856-a8c3-4a54-b24a-
> f4d5e2160831
> 
> all mention this bug so i thought you maybe can reopen it to fix my problem
> as well.
> im using the most recent amd driver 16.8.2 on windows 10. (i dont have any
> issues on ubuntu 16.04)

Is it related to video playing? If it isn't, can you open a new bug and comment again referencing it?
it certainly happens on youtube. but i think at least one or two of the last 15 crashes were not connected so i guess i will do what you recommend.
(In reply to Martin from comment #20)
> it certainly happens on youtube. but i think at least one or two of the last
> 15 crashes were not connected so i guess i will do what you recommend.

If you have clear steps to reproduce the crash, please detail them when you open the new bug.
Also, attach to your bug report the data from about:support (in particular the Graphics section).
okay did that here is the bug for your reference as well: https://bugzilla.mozilla.org/show_bug.cgi?id=1299520
You need to log in before you can comment on or make changes to this bug.