Flash on Rawhide (with nvidia's binary driver)

[Will Woods]

On Fri, 2009-04-17 at 22:00 -0400, Brian Vuyk wrote:
> I've been having some minor issues with the proprietary Flash on Rawhide 
> (32-bit).
> 
> It seems that whenever I make a flash file fullscreen, Firefox crashes. 
> When running it through GDB, I get the following printed out just prior 
> to the backtrace starting:

> 
> I would like to get a sense of how many people are experiencing the same 
> crash. From conversation with others in #fedora-qa, it seems to be 
> pretty widespread. It can be easily tested by visiting any video site 
> (youtube, megavideo etc.), opening a video, and clicking the 'enlarge' 
> or 'fullscreen' button.
> 
> I should note that I used this same version of Flash (10.0.22.87) with 
> no fulllscreen issues on Ubuntu prior to beginning to test Rawhide.
> 
> There is an existing bug in bugzilla about it, to which I've added my 
> information. However, it doesn't seem to be drawing a lot of attention, 
> and I think this is a fairly important bug to be quashed, since so many 
> people use the proprietary flash.
> 
> Here is the bugzilla link: 
> https://bugzilla.redhat.com/show_bug.cgi?id=477110

We talked about this on IRC earlier, but for the record - please note
the files in the backtrace for this crash:
  /lib/libc.so.6[0xafea41]                                                                                                                                         
  /usr/lib/nvidia/libGL.so.1[0x1d3f8e5]   

And the backtrace itself:
  #5  0x01d3f8e5 in ?? () from /usr/lib/nvidia/libGL.so.1
  [...]
  #25 0x9c93b9f6 in _init () from /usr/lib/nvidia/libGLcore.so.1

The proprietary, closed-source nvidia driver is almost undoubtedly the
cause of your problem. If anyone who *isn't* using closed-source
graphics drivers can reproduce this problem, we might have some hope of
debugging that. Otherwise all you can really do is ask nvidia to fix
their code.

> *** glibc detected *** /usr/lib/firefox-3.1b3/firefox: free(): invalid 
> pointer: 0xa4958b00 ***
> 
> Is this possibly related to the recent malloc changes in glibc?

Can't really tell - but even if it was, only nvidia can fix it. So talk
to them.

-w