Comment 33 for bug 1071209

Revision history for this message
Chris J Arges (arges) wrote : Re: memtest86 test #7 false positives (random number sequence error)

Linked a branch with a fix that works for me in raring.

There seems to be a couple approaches for fixing this. The Suse bug has the small change to clobber the additional register and that seems to work. While the RH bug uses the C code in test.c instead of the optimized assembler to generate the proper binaries. I took the Suse approach for fixing this bug.

I noticed that there are uncommitted changes in the bzr development branch. So adding quilt didn't work so well.
To build I had to do an additional dpkg-source --commit to add these changes to a patch before building.
What's the proper way to do this so the package is in proper shape for patching?

You can test the raring version here:
http://people.canonical.com/~arges/lp1071209/