Multiple CUDA issues. Seg fault caused by pyrit cowpatty and aircrack.

[InsertCleverNameHere]

When i execute pyrit inside X my screen turns black, also it does not respond to me trying to close X. Have to restart box to restore X. Upon start up I tried pyrit selftest/benchmark which were all great. 12000 k/s. Self test ok. However I get a seg fault with anything I run bascially. I have tried nearly every application that has cuda support with the same result. Aircrack-ng was giving me a different error iirc. However with X dying everytime i try to use CUDA I do not have a screen shot of the error. Hopefully when I fix CUDA with the other applications it will work with aircrack-ng as well.
Nvidia 260
Q9550
BT r1

Thanks in Advance.

[Archangel-Amael]

We are working to get this issue resolved, please stay tuned.

[purehate]

Actually your problem is not backtrack related. the Nvidia 260 drivers are buggy and do not work well with any CUDA tools. If you want them to work you will need to revert back to 256 drivers which is what i base the backtrack builds on.

[InsertCleverNameHere]

I was also missing pycuda but all is working now... I wish I had realised everything I needed was in the repos... thanks for the help.

well it was working now im getting this..

File "/usr/bin/pyrit", line 6, in <module>
pyrit_cli.Pyrit_CLI().initFromArgv()
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 110, in initFromArgv
func(self, **options)
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 148, in new_f
f(*args, **kwds)
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 709, in attack_passthrough
with cpyrit.util.PassthroughIterator(essid, reader) as rstiter:
File "/usr/lib/python2.5/site-packages/cpyrit/util.py", line 175, in __init__
self.cp = cpyrit.CPyrit()
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 450, in __init__
self.cores.append(CUDACore(queue=self, dev_idx=dev_idx))
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 243, in __init__
_cpyrit_cuda.CUDADevice.__init__(self, dev_idx)
SystemError: CUDA_ERROR_UNKNOWN

[jonathan11]

I was also missing pycuda but all is working now... I wish I had realised everything I needed was in the repos... thanks for the help.

well it was working now im getting this..

File "/usr/bin/pyrit", line 6, in <module>
pyrit_cli.Pyrit_CLI().initFromArgv()
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 110, in initFromArgv
func(self, **options)
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 148, in new_f
f(*args, **kwds)
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 709, in attack_passthrough
with cpyrit.util.PassthroughIterator(essid, reader) as rstiter:
File "/usr/lib/python2.5/site-packages/cpyrit/util.py", line 175, in __init__
self.cp = cpyrit.CPyrit()
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 450, in __init__
self.cores.append(CUDACore(queue=self, dev_idx=dev_idx))
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 243, in __init__
_cpyrit_cuda.CUDADevice.__init__(self, dev_idx)
SystemError: CUDA_ERROR_UNKNOWN

I get a similar error to what you get, see;

Traceback (most recent call last):
  File "/usr/bin/pyrit", line 6, in <module>
    pyrit_cli.Pyrit_CLI().initFromArgv()
  File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 116, in initFromArg      v
    func(self, **options)
  File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 164, in new_f
    f(*args, **kwds)
  File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 989, in attack_batc      h
    with cpyrit.util.StorageIterator(storage, essid) as dbiter:
  File "/usr/lib/python2.5/site-packages/cpyrit/util.py", line 119, in __init__
    self.cp = cpyrit.CPyrit() if yieldNewResults else None
  File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 450, in __init_      _
    self.cores.append(CUDACore(queue=self, dev_idx=dev_idx))
  File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 243, in __init_      _
    _cpyrit_cuda.CUDADevice.__init__(self, dev_idx)
SystemError: CUDA_ERROR_UNKNOWN

The last parts is the same at least..
I found a workaround that works for me, and that is to stop X and reinstall the nVidia drivers. The error can occur again after x amount of reboots of the system.
Maybe it works for you?
Do you have any take on what might be wrong purehate?

[cseven]

I think I just broke something also:

Traceback (most recent call last):
File "/usr/bin/pyrit", line 6, in <module>
pyrit_cli.Pyrit_CLI().initFromArgv()
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 116, in initFromArgv
func(self, **options)
File "/usr/lib/python2.5/site-packages/pyrit_cli.py", line 1274, in verify
with cpyrit.cpyrit.CPyrit() as cp:
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 450, in __init__
self.cores.append(CUDACore(queue=self, dev_idx=dev_idx))
File "/usr/lib/python2.5/site-packages/cpyrit/cpyrit.py", line 243, in __init__
_cpyrit_cuda.CUDADevice.__init__(self, dev_idx)
SystemError: CUDA_ERROR_UNKNOWN

This is with Pyrit 0.3.1-dev (svn r281) using a Nvidia GTX 260M 1GB card. I'm not sure where I went wrong. At first I thought it was a garbled database......hmm as I'm typing this I was switching kernels with my NVIDIA driver (256.53), I wonder if that messed something up.

It was driver related in my case. I hope this helps someone out if they come across this error. I installed the beta NVIDIA-Linux-x86-260.19.04 and it seems to work. I'm not sure the driver is all that great since it is in beta but it fixed the problem

[purehate]

Whatever you guys are doing its not the driver, I have used pyrit with all the last nvidia-driver versions this year.

[espreto]

I know that my post not has the solution to the problem, but I found interesting to put the link here.
Had posted on my twitter this morning!

CUDA, Supercomputing for the Masses: Part 1 until Part 20
http://www.drdobbs.com/architecture-...sign/207200659

@Mods:
Can delete the post if they feel necessary.

Regards,