Comment 39 for bug 1112652

@Francois

Thanks for pointing out the commits in comment #24. There a few bugs very similar to this, which all involve e1000e. It seemw this regression was introduced in an audio commit specific to the device: 8086:1c20

This bug was fixed in v3.9-rc3. However, some report it is fixed in v3.9-rc2. It would be helpful to confirm with exact release fixes this. Once that is known, we can perform a reverse bisect to identify the commit that fixes this.

Before starting the reverse bisect, I will built a test kernel with the commits you mention in comment #24. I'll post a link to that test kernel shortly.

Bugs that may be duplicates:
bug 1165248
bug 1169244
bug 1131616
bug 1142961