+++ This bug was initially created as a clone of Bug #1274132 +++ adding a couple of more dll versions to the blacklist based on data from 47.0b8: igd10iumd32.dll igd11dxva32.dll@0x2549b6 -> 10.18.15.4248 igd10iumd32.dll@0x2aade6 -> 10.18.10.3958 TppValidateCleanupGroupMember -> 10.18.10.3496 igd10iumd32.dll@0x3e2cfe -> 9.18.10.3006 igd10umd32.dll @0x0 | igd10umd32.dll@0x17874f -> 9.17.10.3040 atidxx32.dll atidxx32.dll@0x74527 -> 8.17.10.531 RtlpFreeHeap | RtlFreeHeap | atiumdva.dll@0x1328a8 -> 8.17.10.453 nvwgf2um.dll RtlpLowFragHeapAllocFromContext | RtlpAllocateHeapInternal | RtlAllocateHeap | nvwgf2um.dll@0x909fb2 -> 10.18.13.6510 nvwgf2um.dll@0xbfce -> 9.18.13.4195

I find this highly suspicious all those dlls crashing. How well are they doing with other program? Are they also blacklisted in chrome? Could it just be due to the way we are using d3d11 in our code?

Comment on attachment 8756584 [details] [diff] [review] bug1275739.patch Review of attachment 8756584 [details] [diff] [review]: ----------------------------------------------------------------- Thank you for finding these.

(In reply to Jean-Yves Avenard [:jya] from comment #2) > I find this highly suspicious all those dlls crashing. How well are they > doing with other program? > > Are they also blacklisted in chrome? > > Could it just be due to the way we are using d3d11 in our code? I agree that it could very well be our use that's somehow incorrect, but Chris and I thought it'd still be worth trying to reduce crash reports quickly before the 47 release. Having blacklisted DLLs in a pref allows us & others to easily re-enable these drivers for local debugging.

Comment on attachment 8756584 [details] [diff] [review] bug1275739.patch Approval Request Comment [Feature/regressing bug #]: bug 1248496 [User impact if declined]: firefox crashes on media playback for users of certain graphics drivers [Describe test coverage new/current, TreeHerder]: none [Risks and why]: low; this makes use of the purpose-built blacklisting pref from bug 1273691 which landed in 47.0b8 and worked well there. the patch is just covering a few more crashy dll versions that surfaced in 47.0b8 crash data [String/UUID change made/needed]: none

This is something we *do* want fixed in Fx47 as this is related to release blocking bug 1270686.

Comment on attachment 8756584 [details] [diff] [review] bug1275739.patch This fix is meant to help with a potentially release blocking startup crash. Aurora48+, Beta47+