That sounds like it would work, but I'm not familiar with how get-mesa.sh works. My modified scripts will simply clone the git master with '--depth 1' as a temporary build directory and your manual clone where you are bisecting in is left untouched. If the build succeeds it will make a package of the latest git master or what commit you specified.

My workflow would be like this.

1. Using the attached script above I would build and test the current master with: ./mesa.SlackBuild
2. Assuming the issue is still present: git clone git://anongit.freedesktop.org/mesa/mesa
3. cd mesa
4. git bisect start
5. git bisect good foo
6. git bisect bad bar
7. It gives a commit to test now and I build that commit and make a package with: COMMIT=3e27b37 ./mesa.SlackBuild (Using the script I attached above)
8. Use upgradepkg on the new package in /tmp/.
9. Test it and then in the original clone do 'git bisect good', 'git bisect bad' or 'git bisect skip' depending on if it works, fails, or can not be tested due to unrelated issues.
10. Repeat until git tells you what the first bad commit was.
11. Share the bisect result and the specific issue upstream on the mesa issue tracker or in irc, #dri-devel @ freenode is a good general graphics channel.

Again, assuming you are not using nouveau and do try my scripts, you should modify the configure arguments accordingly. You can look in Pat's script or './configure --help' for reference.

I have bisected the Mesa code base and opened a bug with our friends at Mesa

https://bugs.freedesktop.org/show_bug.cgi?id=106835

For reference