You mentioned this bug did not happen with the 3.0 kernel. If that is the case, we can perform a kernel bisect to identify the commit that introduced this regression. That would require testing of 10 to 12 kernels that I build. Would you be able to assist in testing?
You mentioned this bug did not happen with the 3.0 kernel. If that is the case, we can perform a kernel bisect to identify the commit that introduced this regression. That would require testing of 10 to 12 kernels that I build. Would you be able to assist in testing?