compiz crashes with segfault after recent updates

Just a screen shot I captured one time.

Status changed to 'Confirmed' because the bug affects multiple users.

See also #1283671.

This bug has been reported on the Ubuntu ISO testing tracker.

A list of all reports related to this bug can be found here:
http://iso.qa.ubuntu.com/qatracker/reports/bugs/1282342

I had the trusty daily iso from 2014.02.15. It installs and runs fine.
I'm saying that this issue was introduced on approximately 2014.02.16 or 2014.02.17
The bug 1283671 referred to above says 2014.02.21, but I think it was a few days prior.

These are the commands I used to create and install new VM's today:

1.) Using daily iso from 2014.02.15 (works fine)

sudo virt-install -n desk_tt2 -r 8192 --disk path=/media/newhd/desk_tt2.img,bus=virtio,size=50 -c trusty-desktop-amd64-20140215.iso --network bridge=br0,model=virtio --video=vmvga --graphics vnc,listen=0.0.0.0 --noautoconsole -v --vcpus=4

2.) Using daily iso from 2014.02.24 (does not work due to issues of this bug report)

sudo virt-install -n desk_tt -r 8192 --disk path=/media/newhd/desk_tt.img,bus=virtio,size=50 -c trusty-desktop-amd64-20140224.iso --network bridge=br0,model=virtio --video=vmvga --graphics vnc,listen=0.0.0.0 --noautoconsole -v --vcpus=4

The bug 1283671 is for a 32 bit installation. My 32 bit trusty VM has this issue also.

I have also tried: One CPU instead of four; Less memory, down to 2 gigabytes; The cirrus video driver instead of vmvga.

Without a stack trace from the coredump, we cannot be sure what is causing this crash. I have a feeling this is the same crash that is afflicting many users (myself included), but as I said, without the stack trace, I cannot be sure.

Until you can provide a valid stack trace of your failure, I'm going to mark this Incomplete. If you do produce one, you can update this bug, and I will look over it.

This is the entire file, and I assume contains what you want. Otherwise just let me know, I'll to anything to help with this.

Hi Doug,

Thanks for attaching the full crash file. I took a look over it, but unfortunately, the stack trace has no good information in it, so I still can't determine what the cause of the crash is.

If you get another crash, perhaps we can get a little better information from the core dump.

Well thanks very much for looking. I actually have several of those files from about 4 different VM's.
For whatever reason those files do not get generated for every occurrence of the issue.
Since posting the full file earlier, I have been learning about how to use apport-retrace (and gdb) to extract the stack trace for myself, so maybe I can eventually come back with something useful. So far it isn't working though.

My system seems to have hung again while "installing extra package compiz-core to get Executable path", or similar the terminal window has actually disappeared now. However, I did get this just now and via a SSH session:

...
Core was generated by `compiz'.
Program terminated with signal SIGSEGV, Segmentation fault.
#0 0x00007f75e06e9c24 in __memmove_sse2 (dest=0x26a8d7e0, src=<optimized out>, len=973293064) at ../string/memmove.c:99
99 ../string/memmove.c: No such file or directory.
(gdb)

I'll also add an attachment in a minute, but I have to delete the cor dump becuase the file is over 50 megabytes..

This time including the entire file, while still a large file it is much smaller than the last one with the core dump included.

Hi Doug,

Thanks for digging into this. There is enough info in your last attachment to determine this is a duplicate of bug #1284536 and I'm going to mark it as such. I saw some chatter on IRC that this could be an issue in Mesa.