1021265 - DisplayLink (dlumd32.dll) Startup crash in @0x0 | CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER, unsigned int, unsigned int)

Reporter

Description

•

11 years ago

This bug was filed from the Socorro interface and is report bp-7553c1b0-b28e-45a2-837d-494972140530. ============================================================= This signature started appearing around 5/21, and all of these are reported on Windows 7. Most area also duplicates from a handful of installations. There are no comments so far. 0 @0x0 1 d3d11.dll CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER,unsigned int,unsigned int) 2 igd10umd32.dll igd10umd32.dll@0x20dd0a 3 igd10umd32.dll igd10umd32.dll@0x201e20 4 dlumd32.dll dlumd32.dll@0xfd3d 5 d3d11.dll CResource<ID3D11Resource>::Map<1,5>(CContext *,CResource<ID3D11Resource> *,unsigned int,D3D11_MAP,unsigned int,D3D11_MAPPED_SUBRESOURCE *) 6 d3d11.dll CContext::ID3D11DeviceContext1_Map_<4>(ID3D11DeviceContext1 *,ID3D11Resource *,unsigned int,D3D11_MAP,unsigned int,D3D11_MAPPED_SUBRESOURCE *) 7 xul.dll mozilla::layers::CompositorD3D11::UpdateConstantBuffers() gfx/layers/d3d11/CompositorD3D11.cpp 8 xul.dll mozilla::layers::CompositorD3D11::DrawQuad(mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const &,mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const &,mozilla::layers::EffectChain const &,float,mozilla::gfx::Matrix4x4 const &) gfx/layers/d3d11/CompositorD3D11.cpp 9 xul.dll mozilla::layers::ContentHostBase::Composite(mozilla::layers::EffectChain &,float,mozilla::gfx::Matrix4x4 const &,mozilla::gfx::Filter const &,mozilla::gfx::RectTyped<mozilla::gfx::UnknownUnits> const &,nsIntRegion const *,mozilla::layers::TiledLayerProperties *) gfx/layers/composite/ContentHost.cpp 10 xul.dll mozilla::layers::ThebesLayerComposite::RenderLayer(nsIntRect const &) gfx/layers/composite/ThebesLayerComposite.cpp 11 xul.dll mozilla::layers::ContainerRender<mozilla::layers::ContainerLayerComposite>(mozilla::layers::ContainerLayerComposite *,mozilla::layers::LayerManagerComposite *,nsIntRect const &) gfx/layers/composite/ContainerLayerComposite.cpp 12 xul.dll mozilla::layers::ContainerLayerComposite::RenderLayer(nsIntRect const &) gfx/layers/composite/ContainerLayerComposite.cpp 13 xul.dll mozilla::layers::LayerManagerComposite::Render() gfx/layers/composite/LayerManagerComposite.cpp 14 xul.dll mozilla::layers::LayerManagerComposite::EndTransaction(void (*)(mozilla::layers::ThebesLayer *,gfxContext *,nsIntRegion const &,mozilla::layers::DrawRegionClip,nsIntRegion const &,void *),void *,mozilla::layers::LayerManager::EndTransactionFlags) gfx/layers/composite/LayerManagerComposite.cpp 15 xul.dll mozilla::layers::LayerManagerComposite::EndEmptyTransaction(mozilla::layers::LayerManager::EndTransactionFlags) gfx/layers/composite/LayerManagerComposite.cpp 16 xul.dll mozilla::layers::CompositorParent::CompositeToTarget(mozilla::gfx::DrawTarget *) gfx/layers/ipc/CompositorParent.cpp 17 xul.dll mozilla::layers::CompositorParent::Composite() gfx/layers/ipc/CompositorParent.cpp 18 xul.dll MessageLoop::RunTask(Task *) ipc/chromium/src/base/message_loop.cc 19 xul.dll MessageLoop::DeferOrRunPendingTask(MessageLoop::PendingTask const &) ipc/chromium/src/base/message_loop.cc 20 xul.dll MessageLoop::DoWork() ipc/chromium/src/base/message_loop.cc 21 xul.dll base::MessagePumpDefault::Run(base::MessagePump::Delegate *) ipc/chromium/src/base/message_pump_default.cc 22 xul.dll MessageLoop::RunHandler() ipc/chromium/src/base/message_loop.cc 23 xul.dll MessageLoop::Run() ipc/chromium/src/base/message_loop.cc 24 xul.dll base::Thread::ThreadMain() ipc/chromium/src/base/thread.cc 25 xul.dll `anonymous namespace'::ThreadFunc(void *) ipc/chromium/src/base/platform_thread_win.cc 26 kernel32.dll BaseThreadInitThunk 27 ntdll.dll __RtlUserThreadStart 28 ntdll.dll _RtlUserThreadStart

(Away)

Comment 1

•

11 years ago

Given the timing, likely OMTC. All the reports are Intel adapter 0x0116. As Juan notes, these crashes are coming from a small number of installations hitting this many times. I feel bad for the user who sent us 17 crash reports! This could be a pretty serious problem if there is a subset of users who basically can't start their browser now.

Summary: crash in @0x0 | CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER, unsigned int, unsigned int) → Startup crash in @0x0 | CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER, unsigned int, unsigned int)

Masayuki Nakano [:masayuki] (he/him)(JST, +0900)

Comment 2

•

10 years ago

I know a victim of this bug. He said that this crash reproduces at every start up of Fx 33. However, in safe mode, he succeeded to launch Firefox. So, people meeting this crash cannot update Firefox in normal mode even if this is fixed soon.

Tracy Walker [:tracy]

Comment 3

•

10 years ago

[Tracking Requested - why for this release]: Doesn't seem like much, but 627 crashes on 408 installs since release. A startup top crasher that accounts for just over 1% of all crashes on Fx 33. bug 1026893 related?

status-firefox33: --- → affected

tracking-firefox33: --- → ?

Keywords: topcrash-win

Sylvestre Ledru [:Sylvestre]

Updated

•

10 years ago

tracking-firefox33: ? → +

Sylvestre Ledru [:Sylvestre]

Comment 4

•

10 years ago

Bas, Nical, if you have an idea!

Flags: needinfo?(nical.bugzilla)

Flags: needinfo?(bas)

Nicolas Silva [:nical]

Comment 5

•

10 years ago

9.17.10.3347 is from 2013, it's not going to be in the blacklist that we are updating. This looks like Bug 1044975 which got backed out from beta.

Flags: needinfo?(nical.bugzilla)

Bas Schouten (:bas.schouten)

Comment 6

•

10 years ago

(In reply to Nicolas Silva [:nical] from comment #5) > 9.17.10.3347 is from 2013, it's not going to be in the blacklist that we are > updating. > > This looks like Bug 1044975 which got backed out from beta. We could try including in the chemspill, I'm pretty sure it didn't cause the problems we had on RC.

Flags: needinfo?(bas)

Sylvestre Ledru [:Sylvestre]

Comment 7

•

10 years ago

(In reply to Bas Schouten (:bas.schouten) from comment #6) > > We could try including in the chemspill, I'm pretty sure it didn't cause the > problems we had on RC. Do we have a way to confirm this statement? Not that I don't trust you, we need a confirmation.

Tracy Walker [:tracy]

Comment 8

•

10 years ago

This was in the Fx33b99(RC1) build. But at 10% of the volume that it is occurring on release.

Tracy Walker [:tracy]

Comment 9

•

10 years ago

The driver correlations are same as https://bugzilla.mozilla.org/show_bug.cgi?id=1026893#c7 Graphics Adapter Report: The number of crashes for a particular signature by graphics vendor and chipset. Vendor Adapter Report Count Percentage 0x8086 0x0116 225 35.885 % 0x8086 0x0102 200 31.898 % 0x8086 0x0126 158 25.199 % 0x8086 0x0106 33 5.263 % 0x8086 0x0112 9 1.435 % 0x8086 0x010a 2 0.319 %

Sylvestre Ledru [:Sylvestre]

Comment 10

•

10 years ago

This bug has been identified as a top crash of the 33 release. Bas, Nical, could you help here?

Flags: needinfo?(nical.bugzilla)

Flags: needinfo?(bas)

Nicolas Silva [:nical]

Comment 11

•

10 years ago

Did we end up relanding the fix for bug 1044975? I can't think of anything else I can do about this crash (as long as I don't have hardware that reproduces the bug).

Flags: needinfo?(nical.bugzilla)

Sylvestre Ledru [:Sylvestre]

Comment 12

•

10 years ago

Nical, yes but for 33.0.1, not in 33.0 So, we will see in this case what happens once we release 33.0.1

Nicolas Silva [:nical]

Updated

•

10 years ago

Blocks: 1088034

Sylvestre Ledru [:Sylvestre]

Comment 13

•

10 years ago

Tracy, do you still see it in 33.0.1/33.0.2? Thanks

Flags: needinfo?(twalker)

Sylvestre Ledru [:Sylvestre]

Updated

•

10 years ago

Flags: needinfo?(bas)

Tracy Walker [:tracy]

Comment 14

•

10 years ago

Yes, it is still present. numbers for the past 7 days. Product Version Number Of Crashes Firefox 33.0 414 Firefox 33.0.2 56 Firefox 33.0.1 54

Flags: needinfo?(twalker)

Tracy Walker [:tracy]

Comment 15

•

10 years ago

but note this and bug 1026893 (#47) have been sliding down the top crash list. this one to #73

Sylvestre Ledru [:Sylvestre]

Comment 16

•

10 years ago

OK. I am going to wontfix it. Do you know if it impacts 34 too?

status-firefox32: affected → wontfix

status-firefox33: affected → wontfix

status-firefox34: --- → ?

status-firefox35: --- → ?

Keywords: topcrash-win

Tracy Walker [:tracy]

Comment 17

•

10 years ago

yes, though in lower numbers, decreasing with each beta. Product Version Percentage Number Of Crashes Firefox 33.0 70.53 % 414 Firefox 33.0.2 9.54 % 56 Firefox 33.0.1 9.20 % 54 Firefox 34.0b2 4.94 % 29 Firefox 34.0b3 2.90 % 17 Firefox 33.0b3 0.68 % 4 Firefox 35.0a2 0.68 % 4 Firefox 33.0b1 0.51 % 3 Firefox 34.0b4 0.34 % 2 Firefox 34.0b1 0.34 % 2 Firefox 33.0b99 0.17 % 1 Firefox 33.0b9 0.17 % 1

status-firefox34: ? → affected

status-firefox35: ? → affected

Staszek Kempa

Comment 18

•

10 years ago

For the numbers "decreasing with each beta" please also consider the "real life" scenario - where with upgrade to 33.0 browser stopped to work at all - people simply switch to other. Some of those did try next stable release, some of them also did try beta... after 2 or 3 failed attempts - user simply give up.. and will wait for product which at least starts. In this case the fact that numbers go down do not describe "better quality of the code" , but unfortunately only the dying patient of users.

Rolf Niggemann

Comment 19

•

10 years ago

I have this bug on my computer running windows 7 since yesterday. The bug is present in 33.0.2 and even the latest nightly which I installed yesterday. Basically all actual Firefox versions only run in save mode on this machine. It is really annoying and makes Firefox practically useless. In my opinion, the bug started after I upgraded Adobe Reader to the newest Adobe Reader XI. When trying to upgrade plugins, there is "Adobe Acrobat NPAPI Plug-in" that indicates status "vulnerable 11.0.9.29" but does not upgrade to newer version. It only reopens the same "plugin check&upgrade" tab again and again when I click on the "update now" button next to it. Not shure if that has to say anything, but maybe it is something to look at?

Milan Sreckovic [:milan] (needinfo for best results)

Updated

•

10 years ago

Flags: needinfo?(jmuizelaar)

Jeff Muizelaar [:jrmuizel]

Comment 20

•

10 years ago

(In reply to Rolf Niggemann from comment #19) > I have this bug on my computer running windows 7 since yesterday. > The bug is present in 33.0.2 and even the latest nightly which I installed > yesterday. > Basically all actual Firefox versions only run in save mode on this machine. > It is really annoying and makes Firefox practically useless. > > In my opinion, the bug started after I upgraded Adobe Reader to the newest > Adobe Reader XI. > When trying to upgrade plugins, there is "Adobe Acrobat NPAPI Plug-in" that > indicates status "vulnerable 11.0.9.29" but does not upgrade to newer > version. It only reopens the same "plugin check&upgrade" tab again and again > when I click on the "update now" button next to it. > Not shure if that has to say anything, but maybe it is something to look at? Just to confirm some things. The crash happens at startup and it happens everytime? Can you post a link to the crash from about:crashes?

Flags: needinfo?(Rolf.Niggemann)

Rolf Niggemann

Comment 21

•

10 years ago

Flags: needinfo?(Rolf.Niggemann)

Jeff Muizelaar [:jrmuizel]

Comment 22

•

10 years ago

(In reply to Rolf Niggemann from comment #21) > > > Can you post a link to the crash from about:crashes? > Here are some from 33.0.2 and 36.0a1 as well. Can you try updating your driver to see if it fixes the problem?

Flags: needinfo?(jmuizelaar) → needinfo?(Rolf.Niggemann)

Rolf Niggemann

Comment 23

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #22) > Can you try updating your driver to see if it fixes the problem? Windows tells me I have the newest driver when I try to do an automatic update. Intel(R) HD Graphics 3000 Intel Corporation Date: 26.09.2012 Version: 9.17.10.2867 Can I give you more informations?

Flags: needinfo?(Rolf.Niggemann)

Staszek Kempa

Comment 24

•

10 years ago

Rolf Niggemann

Comment 25

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #22) > Can you try updating your driver to see if it fixes the problem? After my reply in comment #23, I found an driver update from Intel which I was able to install. The actual driver is now: Intel(R) HD Graphics 3000 Intel Corporation Date: 20.03.2014 Version: 9.17.10.3517 Firefox still crashes after the driver update immediately at startup: Submitted Crash Reports Report ID Date Submitted bp-8574e8c6-3509-4cc6-9203-2648f2141031 31.10.2014 23:03

Flags: needinfo?(jmuizelaar)

(Away)

Comment 26

•

10 years ago

I would like to echo Staszek's statement from comment 18: decreasing volume does not always mean the crash went away; sometimes the users went away. We should be extra-sensitive to perma-statup-crashes that will make people stop using Firefox. Should we reconsider 33 tracking for this? Maybe it's still wontfix (it might be too late already for the lost users) but I just wanted to make sure the decision has all the data.

Flags: needinfo?(sledru)

Robert Kaiser

Comment 27

•

10 years ago

[Tracking Requested - why for this release]: We should at least track 34 if we can't get this into some 33.x release. I'm pretty sure that this means lost users as it's clearly a startup crash and comments tell that people cannot get Firefox to run any more after the update to 33.

tracking-firefox34: --- → ?

Sylvestre Ledru [:Sylvestre]

Comment 28

•

10 years ago

David, I agree with you and I share your concern but we don't have a patch yet...

status-firefox33: wontfix → affected

tracking-firefox34: ? → +

Flags: needinfo?(sledru)

Nicolas Silva [:nical]

Comment 29

•

10 years ago

Note that the crashes recorded by Rolf happen on the main thread, and interestingly there seems to be a mix of BasicLayers and D2D, which is odd (and sounds like trouble). The other crashes on this bug happen in the compositor thread.

Nicolas Silva [:nical]

Comment 30

•

10 years ago

(In reply to Nicolas Silva [:nical] from comment #29) > Note that the crashes recorded by Rolf happen on the main thread, and > interestingly there seems to be a mix of BasicLayers and D2D, which is odd > (and sounds like trouble). > > The other crashes on this bug happen in the compositor thread. FTR, this is the code that decides to create a BasicLayermanager to paint an inactive layer (while the rest of is done normally on a ClientLayer. http://hg.mozilla.org/mozilla-central/annotate/e0b505a37b1c/layout/base/FrameLayerBuilder.cpp#l2595 So the use of BasicLayerManager actually looks somewhat legit, although I don't know for sure how/why that works.

Bas Schouten (:bas.schouten)

Comment 31

•

10 years ago

Can someone go buy/find a machine (surely we have one lying around in an office?) with device 0x0116 and install the relevant driver?

Jeff Muizelaar [:jrmuizel]

Comment 32

•

10 years ago

(In reply to Bas Schouten (:bas.schouten) from comment #31) > Can someone go buy/find a machine (surely we have one lying around in an > office?) with device 0x0116 and install the relevant driver? I have one and will do so.

Jeff Muizelaar [:jrmuizel]

Comment 33

•

10 years ago

I have a machine with a 0x116 and 9.17.10.3517 driver and don't get crashes on startup. So there must be something else going on here.

Flags: needinfo?(jmuizelaar)

Jeff Muizelaar [:jrmuizel]

Comment 34

•

10 years ago

My guess is that this is related to dlumd32.dll which is a displaylink dll

Jeff Muizelaar [:jrmuizel]

Comment 35

•

10 years ago

All of the crashes that I looked at had this dll in the call stack. I don't know why the correlation report is not showing this.

Arthur K. (he/him)

Comment 36

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #34) > My guess is that this is related to dlumd32.dll which is a displaylink dll Yes and some of these Displaylink devices aren't related to the IGP but are third party add-on USB-based devices so we're possibly looking at the wrong thing. It's possible that if this user updated to v7.7 M0 of the displaylink software it might update his dlumd32.dll.

Arthur K. (he/him)

Comment 37

•

10 years ago

(In reply to Rolf Niggemann from comment #25) > (In reply to Jeff Muizelaar [:jrmuizel] from comment #22) > > > Can you try updating your driver to see if it fixes the problem? > > After my reply in comment #23, I found an driver update from Intel which I > was able to install. > The actual driver is now: > Intel(R) HD Graphics 3000 > Intel Corporation > Date: 20.03.2014 > Version: 9.17.10.3517 > > Firefox still crashes after the driver update immediately at startup: > > Submitted Crash Reports > Report ID Date Submitted > bp-8574e8c6-3509-4cc6-9203-2648f2141031 31.10.2014 23:03 Rolf, Do you think you could update your Displaylink software to v7.7 M0 from http://www.displaylink.com/support/sla.php?fileid=131 and see if it stops crashing?

Jeff Muizelaar [:jrmuizel]

Comment 38

•

10 years ago

(In reply to Arthur K. from comment #37) > (In reply to Rolf Niggemann from comment #25) > > (In reply to Jeff Muizelaar [:jrmuizel] from comment #22) > > > > > Can you try updating your driver to see if it fixes the problem? > > > > After my reply in comment #23, I found an driver update from Intel which I > > was able to install. > > The actual driver is now: > > Intel(R) HD Graphics 3000 > > Intel Corporation > > Date: 20.03.2014 > > Version: 9.17.10.3517 > > > > Firefox still crashes after the driver update immediately at startup: > > > > Submitted Crash Reports > > Report ID Date Submitted > > bp-8574e8c6-3509-4cc6-9203-2648f2141031 31.10.2014 23:03 > > Rolf, > > Do you think you could update your Displaylink software to v7.7 M0 from > http://www.displaylink.com/support/sla.php?fileid=131 and see if it stops > crashing? If you could report the current version that you have installed that would be valuable too. Look in Add & Remove programs for the version number.

Jeff Muizelaar [:jrmuizel]

Updated

•

10 years ago

Summary: Startup crash in @0x0 | CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER, unsigned int, unsigned int) → DisplayLink (dlumd32.dll) Startup crash in @0x0 | CContext::UMQueryHS_ConstBuf_(D3D10DDI_HRTCORELAYER, unsigned int, unsigned int)

Jeff Muizelaar [:jrmuizel]

Comment 39

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #38) If could tell us the type of displaylink device you have that would be helpful too.

(Away)

Comment 40

•

10 years ago

Something's wrong with our DLL correlation tools. KaiRo has pinged rhelmer for help. In the meantime I checked ~20 reports by hand. I've seen versions 8.5.5.27797 to 8.6.1.36484 of dlumd32.dll. The latter has timestamp of Febuary 2012, which gives me hope that this may be fixed in a newer version.

Arthur K. (he/him)

Comment 41

•

10 years ago

(In reply to David Major [:dmajor] (UTC+13) from comment #40) > Something's wrong with our DLL correlation tools. KaiRo has pinged rhelmer > for help. In the meantime I checked ~20 reports by hand. I've seen versions > 8.5.5.27797 to 8.6.1.36484 of dlumd32.dll. The latter has timestamp of > Febuary 2012, which gives me hope that this may be fixed in a newer version. I can confirm that the updated Displaylink 7.7 M0 software installs version 9.18.7.7 of dlumd32.dll and it's dated 9/26/2014. Hopefully it's as simple as that.

Arthur K. (he/him)

Comment 42

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #39) > (In reply to Jeff Muizelaar [:jrmuizel] from comment #38) > > If could tell us the type of displaylink device you have that would be > helpful too. Just from playing around with this Displaylink software (don't have the hardware to test), it seems to hook into the display driver in a similar fashion as the Jaws screen reading software does.

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 43

•

10 years ago

Attached patch mozilla-release patch making us regard d3d11 as broken with displaylink, and fall back to basic layers (obsolete) — Details — Splinter Review

Attachment #8516239 - Flags: review?(jmuizelaar)

(Away)

Comment 44

•

10 years ago

I suspect that the fallback can be limited to: (dlumd32 <= 8.6.1.36484 && igd10umd32 >= 9.17.10.2843) -- but it should be confirmed with rhelmer's correlations when they're ready.

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 45

•

10 years ago

Attached patch mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers — Details — Splinter Review

Attachment #8516239 - Attachment is obsolete: true

Attachment #8516239 - Flags: review?(jmuizelaar)

Attachment #8516271 - Flags: review?(jmuizelaar)

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 46

•

10 years ago

Is 64-bit Firefox also affected? Is the DLL version there named something like "dlumd64.dll" ?

Rolf Niggemann

Comment 47

•

10 years ago

(In reply to Arthur K. from comment #37) > Rolf, > > Do you think you could update your Displaylink software to v7.7 M0 from > http://www.displaylink.com/support/sla.php?fileid=131 and see if it stops > crashing? I am going to try the update now. (In reply to Jeff Muizelaar [:jrmuizel] from comment #38) > If you could report the current version that you have installed that would > be valuable too. > > Look in Add & Remove programs for the version number. My current dlumb32.dll is: (is that what you are looking for?) Fileversion: 8.5.5.27797 Productname: DisplayLink Core Software v5.5.27797.0 Productversion: 5.5.27797.0 Change Date: 21.09.2010 15:03 Language: English (Great Britain) (In reply to Jeff Muizelaar [:jrmuizel] from comment #39) > If could tell us the type of displaylink device you have that would be > helpful too. My machine is a HP EliteBook 2760p. The display is an internal Intel(R) HD Graphics 3000 Hardware IDs: PCI\VEN_8086&DEV_0116&SUBSYS_162A103C&REV_09 PCI\VEN_8086&DEV_0116&SUBSYS_162A103C PCI\VEN_8086&DEV_0116&CC_030000 PCI\VEN_8086&DEV_0116&CC_0300 I hope I got the right information. Please let me know if you look for more or different stuff. I will go and try the DisplayLink update now and report back here what happened.

Jeff Muizelaar [:jrmuizel]

Comment 48

•

10 years ago

(In reply to Rolf Niggemann from comment #47) > > (In reply to Jeff Muizelaar [:jrmuizel] from comment #39) > > If could tell us the type of displaylink device you have that would be > > helpful too. > > My machine is a HP EliteBook 2760p. > The display is an internal Intel(R) HD Graphics 3000 > Hardware IDs: > PCI\VEN_8086&DEV_0116&SUBSYS_162A103C&REV_09 > PCI\VEN_8086&DEV_0116&SUBSYS_162A103C > PCI\VEN_8086&DEV_0116&CC_030000 > PCI\VEN_8086&DEV_0116&CC_0300 > > I hope I got the right information. Please let me know if you look for more > or different stuff. Do you have a USB to VGA or USB to DVI adapter? I'm looking for the model of that device.

Jeff Muizelaar [:jrmuizel]

Comment 49

•

10 years ago

Comment on attachment 8516271 [details] [diff] [review] mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers Review of attachment 8516271 [details] [diff] [review]: ----------------------------------------------------------------- Let's try this on trunk to start.

Attachment #8516271 - Flags: review?(jmuizelaar) → review+

(Away)

Comment 50

•

10 years ago

(In reply to Benoit Jacob [:bjacob] from comment #46) > Is 64-bit Firefox also affected? Is the DLL version there named something > like "dlumd64.dll" ? There are no 64-bit crashes on this signature. In the old correlation data (from before it broke) I don't see any other signatures with a dlumd64.dll.

Rolf Niggemann

Comment 51

•

10 years ago

(In reply to Rolf Niggemann from comment #47) > (In reply to Arthur K. from comment #37) > > Rolf, > > > > Do you think you could update your Displaylink software to v7.7 M0 from > > http://www.displaylink.com/support/sla.php?fileid=131 and see if it stops > > crashing? > > I will go and try the DisplayLink update now and report back here what > happened. The update did run fine and Nightly starts now without crashing. I'll download and install the actual stable now and see what's happening with it. (In reply to Jeff Muizelaar [:jrmuizel] from comment #48) > Do you have a USB to VGA or USB to DVI adapter? I'm looking for the model of > that device. Let me see if I can find out something on that. The machine is a notebook, there is nothing externally attached, so i assume there might be something internally on the motherboard. There are some USB Controllers that I can see, but none has some display attached as far as I can see it: #### While I was writing this comment, at this point, Nightly crashed but restarted. https://crash-stats.mozilla.com/report/index/bp-07c98fd2-b7c0-4e34-92f4-33ef82141103 #### * Generic USB Hub: USB\VID_8087&PID_0024&REV_0000 - Validity Sensor (VFS471) - HP hs2340 HSPA+ Mobile Broadband Module * Generic USB Hub: USB\VID_8087&PID_0024&REV_0000 - USB Composite Device - USB Composite Device - Broadcom 2070 Bluetooth * HP hs2340 HSPA+ Mobile Broadband Module USB Device: USB\VID_03F0&PID_3A1D&REV_0000 * Intel(R) 6 Series/C200 Series Chipset Family USB Enhanced Host Controller - 1C2D: PCI\VEN_8086&DEV_1C2D&SUBSYS_162A103C&REV_04 * Intel(R) 6 Series/C200 Series Chipset Family USB Enhanced Host Controller - 1C26: PCI\VEN_8086&DEV_1C26&SUBSYS_162A103C&REV_04 * USB-Root-Hub: USB\ROOT_HUB20&VID8086&PID1C2D&REV0004 - Generic USB Hub (6 Port) * USB-Root-Hub: USB\ROOT_HUB20&VID8086&PID1C26&REV0004 - Generic USB Hub (8 Port) * USB Composite Device: USB\VID_046D&PID_C52B&REV_1201 * USB Composite Device: USB\VID_04F2&PID_B242&REV_9990 I am sorry, I am not sure if I could provide what you were looking for. Please let me know if you need other or more informations.

Jeff Muizelaar [:jrmuizel]

Comment 52

•

10 years ago

Comment on attachment 8516271 [details] [diff] [review] mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers Review of attachment 8516271 [details] [diff] [review]: ----------------------------------------------------------------- ::: gfx/thebes/gfxWindowsPlatform.cpp @@ +1525,5 @@ > + } > + if (displayLinkModuleVersion <= GFX_DRIVER_VERSION(8,6,1,36484)) { > + return false; > + } > + } dmajor suggested doing an additional check for igd10umd32 because the problem only happens on intel

Rolf Niggemann

Comment 53

•

10 years ago

(In reply to Rolf Niggemann from comment #51) > I'll download and install the actual stable now and see what's happening > with it. Done that now, downloaded and installed 33.0.2 and its starting fine, no startup crash. Thanks to all of you for looking into the issue and fixing it!!! Great!!! Is there anything else I can help with?

(Away)

Comment 54

•

10 years ago

> The machine is a notebook, there is nothing externally attached, so i assume > there might be something internally on the motherboard. > There are some USB Controllers that I can see, but none has some display > attached as far as I can see it: Interesting. Looking at [1] I would guess that either the external VGA is driven by a DisplayLink chip, or they pre-install DisplayLink in case you buy the docking station with DisplayPort. Maybe in your use with nothing connected, you might not have needed the DisplayLink software at all. Although now that the driver is updated, might as well leave it installed. [1] http://www.tabletpcreview.com/tabletreview/hp-elitebook-2760p-review/

Rolf Niggemann

Comment 55

•

10 years ago

(In reply to David Major [:dmajor] (UTC+13) from comment #54) > > The machine is a notebook, there is nothing externally attached, so i assume > Interesting. Looking at [1] I would guess that either the external VGA is > driven by a DisplayLink chip, or they pre-install DisplayLink in case you > buy the docking station with DisplayPort. Maybe in your use with nothing > connected, you might not have needed the DisplayLink software at all. > Although now that the driver is updated, might as well leave it installed. > [1] http://www.tabletpcreview.com/tabletreview/hp-elitebook-2760p-review/ It runs now better than before, at least the driver update did stop Firefox crashing. I definately leave the update as it is now. Do not touch a running system and so... ;-) Again thank you to all of you!

Arthur K. (he/him)

Comment 56

•

10 years ago

(In reply to David Major [:dmajor] (UTC+13) from comment #54) > > The machine is a notebook, there is nothing externally attached, so i assume > > there might be something internally on the motherboard. > > There are some USB Controllers that I can see, but none has some display > > attached as far as I can see it: > Interesting. Looking at [1] I would guess that either the external VGA is > driven by a DisplayLink chip, or they pre-install DisplayLink in case you > buy the docking station with DisplayPort. Maybe in your use with nothing > connected, you might not have needed the DisplayLink software at all. > Although now that the driver is updated, might as well leave it installed. > [1] http://www.tabletpcreview.com/tabletreview/hp-elitebook-2760p-review/ My hunch is that Rolf's PC must have had this Displaylink device at some point as the software controlling the external Displaylink device is present. If there is no Displaylink hardware any longer, he can likely uninstall the Displaylink software altogether and not hit this issue some time down the line when/if things change.

Arthur K. (he/him)

Comment 57

•

10 years ago

(In reply to Rolf Niggemann from comment #55) > (In reply to David Major [:dmajor] (UTC+13) from comment #54) > > > The machine is a notebook, there is nothing externally attached, so i assume > > Interesting. Looking at [1] I would guess that either the external VGA is > > driven by a DisplayLink chip, or they pre-install DisplayLink in case you > > buy the docking station with DisplayPort. Maybe in your use with nothing > > connected, you might not have needed the DisplayLink software at all. > > Although now that the driver is updated, might as well leave it installed. > > [1] http://www.tabletpcreview.com/tabletreview/hp-elitebook-2760p-review/ > > It runs now better than before, at least the driver update did stop Firefox > crashing. I definately leave the update as it is now. Do not touch a running > system and so... ;-) > Again thank you to all of you! Rolf, if you don't have this external Displaylink hardware any longer, you can likely remove the Displaylink software altogether.

(Away)

Comment 58

•

10 years ago

Here are the correlations thanks to rhelmer: https://gist.github.com/rhelmer/a309212497969998aedb The numbers confirm that the crashes are (dlumd32 <= 8.6.1.36484 && igd10umd32 >= 9.17.10.2843).

Loic

Comment 59

•

10 years ago

A user on the French community board is hitting the same crashes. Do we have a known workaround?

(Away)

Comment 60

•

10 years ago

> Do we have a known workaround? Update the DisplayLink driver: http://www.displaylink.com/support/sla.php?fileid=131

Rolf Niggemann

Comment 61

•

10 years ago

(In reply to Arthur K. from comment #57) > Rolf, if you don't have this external Displaylink hardware any longer, you > can likely remove the Displaylink software altogether. I was interested in finding out which device I could have, that requires the DisplayLink driver and thought, it would pop up in System Manager, when I uninstall the driver: --> Just nothing happened after the uninstall. Firefox still works fine as does the rest of the system. If I ever find out which device it was that caused the trouble, I'll let you know.

Arthur K. (he/him)

Comment 62

•

10 years ago

(In reply to Rolf Niggemann from comment #61) > (In reply to Arthur K. from comment #57) > > Rolf, if you don't have this external Displaylink hardware any longer, you > > can likely remove the Displaylink software altogether. > > I was interested in finding out which device I could have, that requires the > DisplayLink driver and thought, it would pop up in System Manager, when I > uninstall the driver: --> Just nothing happened after the uninstall. > Firefox still works fine as does the rest of the system. > If I ever find out which device it was that caused the trouble, I'll let you > know. Yes, this is what I suspected. The Displaylink device would have been an external USB-to-VGA type of device and the Displaylink software hooks into the onboard VGA and, I think, creates a secondary output through the Displaylink device. Similar to dual monitor display or some other similar multiplexing. Displaylink seems to be totally different from Displayport. Thanks much for working with us to gather information and test things. I think the removal of the software was safe to do.

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 63

•

10 years ago

OK, updated the patch to only block on Intel. Didn't add the Intel version check per conversation with Jeff. https://hg.mozilla.org/integration/mozilla-inbound/rev/3d0169c5795f

Arthur K. (he/him)

Comment 64

•

10 years ago

(In reply to Benoit Jacob [:bjacob] from comment #63) > OK, updated the patch to only block on Intel. Didn't add the Intel version > check per conversation with Jeff. > https://hg.mozilla.org/integration/mozilla-inbound/rev/3d0169c5795f I can confirm on an Optiplex 960 I have with igd10umd32.dll version 8.15.10.2869 that I have no issues.

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 65

•

10 years ago

bustage fix, https://hg.mozilla.org/integration/mozilla-inbound/rev/0b298d91fbdd

Benoit Jacob [:bjacob] (mostly away)

Assignee

Comment 66

•

10 years ago

Adding AppNotes about that, https://hg.mozilla.org/integration/mozilla-inbound/rev/1b07e2892e9d

Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)

Comment 67

•

10 years ago

https://hg.mozilla.org/mozilla-central/rev/3d0169c5795f https://hg.mozilla.org/mozilla-central/rev/0b298d91fbdd

Assignee: nobody → bjacob

Status: NEW → RESOLVED

Closed: 10 years ago

Resolution: --- → FIXED

Target Milestone: --- → mozilla36

Sylvestre Ledru [:Sylvestre]

Comment 68

•

10 years ago

Benoit, can we have the uplift requests? Thanks

Flags: needinfo?(bjacob)

Nicolas Silva [:nical]

Comment 69

•

10 years ago

Comment on attachment 8516271 [details] [diff] [review] mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers Approval Request Comment [Feature/regressing bug #]: [User impact if declined]: Startup top crasher [Describe test coverage new/current, TBPL]: none [Risks and why]: Very low, at worst we blacklist more people than we should and they loose D3D11 compositing. [String/UUID change made/needed]: none

Attachment #8516271 - Flags: approval-mozilla-release?

Attachment #8516271 - Flags: approval-mozilla-beta?

Attachment #8516271 - Flags: approval-mozilla-aurora?

Sylvestre Ledru [:Sylvestre]

Comment 70

•

10 years ago

Comment on attachment 8516271 [details] [diff] [review] mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers Also in GECKO331_2014103013_RELBRANCH Thanks

Flags: needinfo?(bjacob)

Attachment #8516271 - Flags: approval-mozilla-release?

Attachment #8516271 - Flags: approval-mozilla-release+

Attachment #8516271 - Flags: approval-mozilla-beta?

Attachment #8516271 - Flags: approval-mozilla-beta+

Attachment #8516271 - Flags: approval-mozilla-aurora?

Attachment #8516271 - Flags: approval-mozilla-aurora+

Ryan VanderMeulen [:RyanVM]

Comment 71

•

10 years ago

https://hg.mozilla.org/releases/mozilla-aurora/rev/ebe40cdbec21 https://hg.mozilla.org/releases/mozilla-beta/rev/57c47cb49c03 https://hg.mozilla.org/releases/mozilla-release/rev/d4a8d739eb5e https://hg.mozilla.org/releases/mozilla-release/rev/7496596b3075

status-firefox33: affected → fixed

status-firefox34: affected → fixed

status-firefox35: affected → fixed

status-firefox36: --- → fixed

Ryan VanderMeulen [:RyanVM]

Comment 72

•

10 years ago

https://hg.mozilla.org/mozilla-central/rev/1b07e2892e9d

Ryan VanderMeulen [:RyanVM]

Comment 73

•

10 years ago

Also landed on GECKO330_2014101104_RELBRANCH for 33.0.3: https://hg.mozilla.org/releases/mozilla-release/rev/63daea50bacd

Robert Kaiser

Comment 74

•

10 years ago

There are still crashes with that signature in 33.0.3, which should have the blocklist: https://crash-stats.mozilla.com/report/list?signature=%400x0+%7C+CContext%3A%3AUMQueryHS_ConstBuf_%28D3D10DDI_HRTCORELAYER%2C+unsigned+int%2C+unsigned+int%29&product=Firefox&query_type=contains&range_unit=weeks&process_type=browser&version=Firefox%3A33.0.3&hang_type=crash&date=2014-11-10+23%3A59%3A59&range_value=2#reports

Robert Kaiser

Comment 75

•

10 years ago

That said, if the fix needs some of the checkins from bug 1064864, bug 1076910, bug 1064107, bug 1021265, bug 1093863, or bug 1089183 to be effective, those landed on a different relbranch than 33.0.3 was built from. They are in 33.1, though.

Robert Kaiser

Comment 76

•

10 years ago

This is definitely still happening in early 33.1 data.

Robert Kaiser

Comment 77

•

10 years ago

See https://crash-stats.mozilla.com/report/list?signature=%400x0+|+CContext%3A%3AUMQueryHS_ConstBuf_%28D3D10DDI_HRTCORELAYER%2C+unsigned+int%2C+unsigned+int%29&build_id=20141106120505&product=Firefox&version=Firefox%3A33.1#tab-reports for reports of this in the 33.1 release.

Sylvestre Ledru [:Sylvestre]

Comment 78

•

10 years ago

I guess this still impacts other branches...

Status: RESOLVED → REOPENED

status-firefox33: fixed → affected

status-firefox34: fixed → affected

status-firefox35: fixed → affected

status-firefox36: fixed → affected

Resolution: FIXED → ---

Jeff Muizelaar [:jrmuizel]

Comment 79

•

10 years ago

Is it happening less?

Jeff Muizelaar [:jrmuizel]

Comment 80

•

10 years ago

I don't understand why are blacklisting code would not catch this problem.

Arthur K. (he/him)

Comment 81

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #80) > I don't understand why are blacklisting code would not catch this problem. Are we blocking dlumd32.dll that is <= 8.6.1.36484 or *only* 8.6.1.36484? I am seeing 8.5.5.29055, 8.5.6.31854, and also dlumd10.dll/dlumd11.dll in the reports.

Robert Kaiser

Comment 82

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #79) > Is it happening less? We don't have enough data yet to tell. Also, as this is a startup crash, it's hard to even get that data, as people will leave us (or that release) when they can't launch Firefox and then we won't see any data from them any more.

Jeff Muizelaar [:jrmuizel]

Comment 83

•

10 years ago

Should be less than or equal: http://hg.mozilla.org/releases/mozilla-release/file/a4d51da32fe7/gfx/thebes/gfxWindowsPlatform.cpp#l1526

Jeff Muizelaar [:jrmuizel]

Comment 84

•

10 years ago

https://hg.mozilla.org/releases/mozilla-beta/rev/252c3ab238d0

Jeff Muizelaar [:jrmuizel]

Comment 85

•

10 years ago

Attached patch Fix driver version parsing — Details — Splinter Review

Approval Request Comment [Feature/regressing bug #]: OMTC [User impact if declined]: Crash on startup [Describe test coverage new/current, TBPL]: None [Risks and why]: Pretty minimal the current code is broken.

Attachment #8520170 - Flags: review+

Attachment #8520170 - Flags: approval-mozilla-release?

Attachment #8520170 - Flags: approval-mozilla-beta+

Attachment #8520170 - Flags: approval-mozilla-aurora?

Jeff Muizelaar [:jrmuizel]

Updated

•

10 years ago

Flags: needinfo?(lmandel)

Lawrence Mandel [:lmandel] (use needinfo)

Comment 86

•

10 years ago

Comment on attachment 8520170 [details] [diff] [review] Fix driver version parsing I approved for Beta in person. Adding Aurora approval. Sylvestre - We should review whether it is worth spinning another 33 release for this trivial fix.

Flags: needinfo?(lmandel) → needinfo?(sledru)

Attachment #8520170 - Flags: approval-mozilla-aurora? → approval-mozilla-aurora+

Sylvestre Ledru [:Sylvestre]

Updated

•

10 years ago

Flags: needinfo?(sledru)

Attachment #8520170 - Flags: approval-mozilla-release? → approval-mozilla-release+

Ryan VanderMeulen [:RyanVM]

Comment 87

•

10 years ago

https://hg.mozilla.org/integration/mozilla-inbound/rev/96f0df721845 https://hg.mozilla.org/releases/mozilla-aurora/rev/2068eecf1f2a https://hg.mozilla.org/releases/mozilla-release/rev/12ec9a7c648a (GECKO331_2014103013_RELBRANCH - 33.1.x) https://hg.mozilla.org/releases/mozilla-release/rev/f2bdbc945da2 (default - 33.x)

status-firefox33: affected → fixed

status-firefox34: affected → fixed

status-firefox35: affected → fixed

status-firefox36: affected → fixed

Carsten Book [:Tomcat]

Comment 88

•

10 years ago

https://hg.mozilla.org/mozilla-central/rev/96f0df721845

Status: REOPENED → RESOLVED

Closed: 10 years ago → 10 years ago

Resolution: --- → FIXED

[SV Manager] Florin Mezei, QA (:FlorinMezei)

Updated

•

10 years ago

Flags: qe-verify+

Robert Kaiser

Comment 89

•

10 years ago

I can confirm that we so far do not see this crash in 33.1.1 \o/

Michael Buckley

Comment 90

•

10 years ago

Don't know if this would be the right place to mention it, but it seems with the update to 33.1.1 Firefox will now hang on start up if my extension installed. Could this bug be related? There is some info here https://codefisher.org/forum/thread/1628/#p6298

(Away)

Comment 91

•

10 years ago

Michael: Please file a new bug for that issue. Given the mention of "friends that have dissimilar PCs" I doubt it's connected to DisplayLink. FWIW I couldn't reproduce by installing Toolbar Buttons on 33.1.1, nor by installing it on 33.1 and letting it update.

Nathan Kidd

Comment 92

•

10 years ago

I have another class of hardware with what appears to be the same underlying problem as this bug, but FF 33.1.1 doesn't resolve it. This is a Dell Precision M6700 laptop with an Intel HD 4000/NVIDIA Quadro K3000M Optimus configuration running Windows 7 (64-bit), SP1 Symptom is FF will crash, usually when loading a page, then the restart will be stuck in a "FF Crashed, restart Y/N? Y, immediately crash" loop. Eventually it offers Safe Mode, which always works. The problem started in FF 33. I downgraded to 32 for a while (which never had this issues), saw this bug listed as fixed in 33.1.1 and was tempted enough to try upgrading again. (C.f. :dmajor 's prescient note that "decreasing volume does not always mean the crash went away, sometimes the users went away") The bottom stack frames are: nvwgf2um.dll nvumdshim.dll d3d11.dll CShaderResourceView::CLS::FinalConstruct(CContext*, CShaderResourceView::TConstructorArgs const*) xul.dll DoesD3D11DeviceWork(ID3D11Device*) about:crashes says: bp-16a036f3-0937-44d2-9e54-e96562141119 11/18/2014 7:39 PM bp-be3fb669-df7f-439c-8555-b298a2141118 11/18/2014 5:18 PM bp-8fcfccde-b76f-4ac5-947d-4883e2141118 11/18/2014 5:18 PM bp-ed0d148f-183e-4629-93a2-c707a2141118 11/18/2014 5:18 PM bp-ef807d15-9c01-4f2a-ba3e-6cb8e2141118 11/18/2014 5:18 PM bp-0bfc240e-db1f-408a-8d6a-675022141118 11/18/2014 5:18 PM bp-5570d629-2edb-4fa8-99aa-dac782141113 11/13/2014 11:00 AM bp-8db28c5f-63c3-48f5-9327-f7b5b2141113 11/13/2014 11:00 AM bp-f296999a-d0ff-481f-bdf1-880bd2141113 11/13/2014 11:00 AM bp-a92d775e-fd88-4b00-a428-097b92141113 11/13/2014 11:00 AM bp-2e344137-87e8-4d28-aebf-ab0212141113 11/12/2014 8:36 PM bp-e4cfcdda-2da4-4553-a936-4ba762141113 11/12/2014 8:20 PM bp-d3388f88-b1cc-4c09-bd06-d604e2141113 11/12/2014 8:20 PM bp-5480161f-f2e6-4e4d-8653-a35012141106 11/5/2014 10:03 PM bp-5467f837-01f8-4ece-b9cc-34f902141106 11/5/2014 9:38 PM bp-9fabbad0-f587-4c2f-a1a2-cf6832141104 11/3/2014 9:43 PM bp-1638c60d-6ff5-436d-8914-efba02141104 11/3/2014 7:46 PM bp-32751736-2813-4e5a-abc1-6a7862141104 11/3/2014 7:19 PM bp-86a37d9c-95cf-4408-a130-d77e62141103 11/3/2014 4:12 PM bp-bb07bdab-81f7-40d1-990b-09e1e2141102 11/1/2014 10:03 PM bp-1fb79a57-799c-4b4e-bfce-e9c002141031 10/31/2014 3:59 PM bp-0e5a008b-99be-49dd-a05b-671dd2141031 10/31/2014 2:59 PM bp-edf5824d-5dcf-42f2-bf23-755d72141031 10/31/2014 2:59 PM bp-cc4b594f-1798-42c2-831c-5d2432141031 10/31/2014 2:59 PM bp-99d07d9b-a6ee-455c-b199-590b82141031 10/31/2014 2:59 PM bp-c7d652e4-a3a5-441d-ad93-904542141031 10/31/2014 2:59 PM bp-f8128ef5-061a-45d6-8ab2-3052d2141031 10/31/2014 2:59 PM bp-327af5b7-5bfc-48c5-bc32-6d6092141031 10/31/2014 2:59 PM bp-d7835741-08b7-4a8c-96c5-95aa22141031 10/31/2014 2:28 PM 9c54d0a6-9aed-426a-a482-32a22d3927fd 9/17/2014 9:17 AM bp-9b694596-3215-471f-b167-918042140717 7/17/2014 2:03 PM bp-84bac606-28a8-4464-a664-810d72140711 7/11/2014 7:44 PM bp-a525c86b-e9f1-494f-a3c7-b42952140510 5/10/2014 5:19 PM bp-2fe6838c-d6a0-4999-9bfa-bb81b2140509 5/9/2014 6:19 PM bp-8e2e69dc-6c93-4372-8437-556752140227 2/27/2014 10:12 AM bp-a5c30c0e-95c6-49b8-99fa-d488d2140219 2/18/2014 7:15 PM bp-d63da1fe-b188-4f5d-8732-79f602140122 1/22/2014 4:15 PM ce21b38f-7143-4bb0-a407-69751bb168e1 1/21/2014 8:58 PM 12d76d22-12d8-41d7-b82a-1194b51dcbf9 12/28/2013 4:19 PM This is the part where you'll hate me: When you say "have you tried updating your video drivers" I'm going to say back, "I'd really like to, but have video-performance-sensitive benchmarking data on this box that will be invalidated if I update the driver right now." :( I'm running NVIDIA 311.10, Intel 9.17.10.2932 drivers, which are what the Dell support site recommended for this box as recently as last month. If there is other information or tests that are useful please let me know.

Jeff Muizelaar [:jrmuizel]

Comment 93

•

10 years ago

(In reply to Nathan Kidd from comment #92) > I have another class of hardware with what appears to be the same underlying > problem as this bug, but FF 33.1.1 doesn't resolve it. > > This is a Dell Precision M6700 laptop with an Intel HD 4000/NVIDIA Quadro > K3000M Optimus configuration running Windows 7 (64-bit), SP1 > > Symptom is FF will crash, usually when loading a page, then the restart will > be stuck in a "FF Crashed, restart Y/N? Y, immediately crash" loop. > Eventually it offers Safe Mode, which always works. This crash doesn't seem to have anything to do with the DisplayLink crash that we were seeing. Please file a new bug.

Cornel Ionce [:noni] [Hubs QA]

Comment 94

•

10 years ago

For the last 7 days the rates look as it follows: - 2 crashes in 36.0a1 and 2 in 35.0a2 - 3 crashes in 33.1.1 (1408 crashes in 33.1 in the same time period)

Nathan Kidd

Comment 95

•

10 years ago

(In reply to Jeff Muizelaar [:jrmuizel] from comment #93) > (In reply to Nathan Kidd from comment #92) > > I have another class of hardware with what appears to be the same underlying > > problem as this bug, but FF 33.1.1 doesn't resolve it. > > > > This is a Dell Precision M6700 laptop with an Intel HD 4000/NVIDIA Quadro > > K3000M Optimus configuration running Windows 7 (64-bit), SP1 > > > > Symptom is FF will crash, usually when loading a page, then the restart will > > be stuck in a "FF Crashed, restart Y/N? Y, immediately crash" loop. > > Eventually it offers Safe Mode, which always works. > > This crash doesn't seem to have anything to do with the DisplayLink crash > that we were seeing. Please file a new bug. In that case, I have a hard time in good conscience to file the bug without updating/testing with the latest NVIDIA driver first. Hopefully I'll be able to eventually, but not soon. Thanks.

[SV Manager] Florin Mezei, QA (:FlorinMezei)

Comment 96

•

10 years ago

(In reply to Cornel Ionce [QA] from comment #94) > For the last 7 days the rates look as it follows: > - 2 crashes in 36.0a1 and 2 in 35.0a2 > - 3 crashes in 33.1.1 (1408 crashes in 33.1 in the same time period) I did another quick check today and there is still a very low level of occurrence reported (+ no crashes in 34 beta 8 or later). I think it's safe to mark this as verified.

Status: RESOLVED → VERIFIED

status-firefox33: fixed → verified

status-firefox34: fixed → verified

status-firefox35: fixed → verified

status-firefox36: fixed → verified

John Ramirez

Comment 97

•

10 years ago

I just installed the lastest version offered on the site v.33 today 11-26-14. It updated fine - closed and restarted. Now after I closed it, it will not open at all. It hangs and just keeps waiting and waiting with "not responding". I already attempted the fix 33.1, but that seems to have not done anything.

John Ramirez

Comment 98

•

10 years ago

Addendum When I tried to close the program - top right X, it proceeded to want to entirely shut down the machine, awaiting for me to confirm shutting down. The patch 33.1.1.exe, which is what I've tried failed

Tarık Özgün Güner

Comment 99

•

10 years ago

I encountured this startup crash a month before and give you time to fix it. After the last update firefox starts as expected. Great work.

John Ramirez

Comment 100

•

10 years ago

I have removed the latest version and attempted to reinstall using previous versions to no avail. The desktop icon now shows a blue/gold shield which prompts me to OK when launching, then it either does not launch or launches and the homepage or anything else loads.

Milan Sreckovic [:milan] (needinfo for best results)

Comment 101

•

10 years ago

John, do you see this with 34? Can you run in the safe mode?

John Ramirez

Comment 102

•

10 years ago

The problem continued with 34.0.5 even with clean reinstall. Then I tricked it to work by opening with shift key down prompting me to restore or fix? I chose fix, closed FF in task manager and launched it again. This got it to run, but once I closed the browser I would have to do the above all over again. Now suddenly, the browser seems to be working like normal again. Thanks

Tyler Downer [He/Him]

Comment 103

•

10 years ago

We have a user reporting this bug in Firefox 37.0.2, which is a bit concerning.

Arthur K. (he/him)

Comment 104

•

10 years ago

(In reply to Tyler Downer [:Tyler] from comment #103) > We have a user reporting this bug in Firefox 37.0.2, which is a bit > concerning. Is there a crash report somewhere? DisplayLink has updated their software to 7.7 M4 (Feb 2015): http://www.displaylink.com/support/sla.php?fileid=144

Tyler Downer [He/Him]

Comment 105

•

10 years ago

Bah, sorry, thought I put that in the comment. https://crash-stats.mozilla.com/report/index/c3a26c8c-9b87-4309-8dd5-af2f02150427

Arthur K. (he/him)

Comment 106

•

10 years ago

(In reply to Tyler Downer [:Tyler] from comment #105) > Bah, sorry, thought I put that in the comment. > > https://crash-stats.mozilla.com/report/index/c3a26c8c-9b87-4309-8dd5- > af2f02150427 Thanks Tyler. I am seeing reference to the crashing DLL version (8.5.5.30661) in bug 828911 comment 12 but on NVidia hardware. I imagine it's still just due to an old version of this software and not specific to VGA adapter.

(Away)

Comment 107

•

10 years ago

We look for "dlumd32.dll" in order to activate our workaround. The machine in bp-c3a26c8c-9b87-4309-8dd5-af2f02150427 has files called "dlumd10.dll" and "dlumd11.dll" instead, so the fix doesn't kick in. Looking back over the old correlation files, I'm going to estimate that 2-3% of the original crashes are in this situation (since the dlumd32 correlations were 97-98% in early November). Tyler, could you see if the user could try updating the DisplayLink driver? Presumably, getting onto a newer driver version should stop the crashes, regardless of what the files are called. Jeff, should we add the other dlumd's?

Flags: needinfo?(jmuizelaar)

Milan Sreckovic [:milan] (needinfo for best results)

Comment 108

•

10 years ago

We should be OK to just add those two DLLs, and if we find them, and have them be of a "bad" version, deal with them. We should be careful (looking at the comment 52) about going Intel only, or assuming that those DLLs always exist on Intel machines.

Jeff Muizelaar [:jrmuizel]

Comment 109

•

10 years ago

(In reply to David Major [:dmajor] from comment #107) > We look for "dlumd32.dll" in order to activate our workaround. The machine > in bp-c3a26c8c-9b87-4309-8dd5-af2f02150427 has files called "dlumd10.dll" > and "dlumd11.dll" instead, so the fix doesn't kick in. > > Looking back over the old correlation files, I'm going to estimate that 2-3% > of the original crashes are in this situation (since the dlumd32 > correlations were 97-98% in early November). > > Tyler, could you see if the user could try updating the DisplayLink driver? > Presumably, getting onto a newer driver version should stop the crashes, > regardless of what the files are called. > > Jeff, should we add the other dlumd's? Yes, that sounds reasonable.

Flags: needinfo?(jmuizelaar)

Milan Sreckovic [:milan] (needinfo for best results)

Comment 110

•

10 years ago

I'll clone this bug so that we can update the list of DLLs we look at.

Milan Sreckovic [:milan] (needinfo for best results)

Updated

•

10 years ago

Blocks: 1160295

Milan Sreckovic [:milan] (needinfo for best results)

Updated

•

10 years ago

Blocks: 1195844

mozilla-release patch making us regard d3d11 as broken with displaylink, and fall back to basic layers 10 years ago Benoit Jacob [:bjacob] (mostly away) 3.31 KB, patch		Details \| Diff \| Splinter Review
mozilla-release patch making us regard d3d11 as broken with displaylink on versions <= 8.6.1.36484, and fall back to basic layers 10 years ago Benoit Jacob [:bjacob] (mostly away) 4.34 KB, patch	jrmuizel : review+ Sylvestre : approval-mozilla-aurora+ Sylvestre : approval-mozilla-beta+ Sylvestre : approval-mozilla-release+	Details \| Diff \| Splinter Review
Fix driver version parsing 10 years ago Jeff Muizelaar [:jrmuizel] 1.18 KB, patch	jrmuizel : review+ lmandel : approval-mozilla-aurora+ jrmuizel : approval-mozilla-beta+ Sylvestre : approval-mozilla-release+	Details \| Diff \| Splinter Review