It is a new crash signature that first appeared in 4.0b12pre/20110216. It is #8 top crasher in this build. Signature nvd3dum.dll@0x38ba73 UUID a2ee7ef7-2c75-49af-8749-f94f12110217 Time 2011-02-17 02:33:00.741836 Uptime 480 Last Crash 494 seconds (8.2 minutes) before submission Install Age 1482 seconds (24.7 minutes) since version was first installed. Product Firefox Version 4.0b12pre Build ID 20110216030352 Branch 2.0 OS Windows NT OS Version 6.1.7600 CPU x86 CPU Info GenuineIntel family 6 model 26 stepping 4 Crash Reason EXCEPTION_ACCESS_VIOLATION_READ Crash Address 0x4 App Notes AdapterVendorID: 10de, AdapterDeviceID: 1200, AdapterDriverVersion: 8.17.12.7028 Frame Module Signature [Expand] Source 0 nvd3dum.dll nvd3dum.dll@0x38ba73 1 nvd3dum.dll nvd3dum.dll@0x3c9a74 2 d3d9.dll CBatchFilterI::ProcessBatch 3 d3d9.dll CBatchFilterI::FlushBatchWorkerThread 4 d3d9.dll CBatchFilterI::LHBatchDestroyResource 5 d3d9.dll DestroyResource 6 d3d9.dll DdDestroySurfaceLH 7 d3d9.dll CMipMap::~CMipMap 8 d3d9.dll CMipMap::`vector deleting destructor' 9 d3d9.dll CBaseObject::~CBaseObject The regression range is: http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=2c646d10b9c7&tochange=b4aa47ca42c1 More reports at: https://crash-stats.mozilla.com/report/list?product=Firefox&range_value=4&range_unit=weeks&signature=nvd3dum.dll%400x38ba73

Driver needs blacklisting?

Probably not, after all. Like bug 629265, this is probably actually a bug on our side: bug 634366.

If it is was a driver version issue, it would have also happened in 4.0b11 at the driver release date.

These are all crashes on Youtube. Plugin-related?

Strange. On the surface bug 628872 seems suspicious, but looks innocuous. There's a known-broken push in that range, bug 631058, which was backed out today, but should only affect <video> and <audio>. It might result in corrupted driver state though. Bug 631575 is also a suspect. I don't see how it would cause problems like this, but bug 622585 touched d3d10 code. Unless someone is able to repro in the meantime, I think we should see if these magically go away with the backout of 631058, and if not, back out 631575 and see if they go away then. :/

Calling this a betaN+ hardblocker for now, but hopefully the backout of bug 631058 fixed it

all the crashes are within a 4 hour window so far and all are on build 20110216030352 maybe possible dups in this batch.

This is also a really small number of crashes to be a blocker.

I doubt that bug 631058 would cause this, it doesn't touch gfx at all.

(In reply to comment #9) > I doubt that bug 631058 would cause this, it doesn't touch gfx at all. I doubt so too. The only reason 631058 might be suspect is that it might cause media code to corrupt userspace driver state from use-after-free-type bugs. Just wild speculation though.

There have been no crashes since Feb 17, 2011 04:35 PST. I'd suggest that this not hardblock

(In reply to comment #11) > There have been no crashes since Feb 17, 2011 04:35 PST. I'd suggest that this > not hardblock Crash processing was busted. The cluster is back up, but not sure if it is all caught up...

The last 4 or so crashes have the 3dvisionlive@nvidia.com extension installed on the same processor...probably the same person

these drivers have not been released outside of nvidia (or its partners) so i expect the reporter could be with nvidia.

Still only 18 crashes total and on a unreleased driver further suggests that this doesn't hardblock.

There have no crashes for the last four weeks. I close it as WFM.