In https://bugzilla.kernel.org/show_bug.cgi?id=70031, there is a report of a
regression in 3.14-rc1 for a 14e4:4331 device. Kernel 3.13 is OK. Can anyone
with a 4331 confirm that PHY transmission errors are a lot more prevalent with
either mainline 3.14-rc1 or the current wireless-testing than with 3.13?
I do not have that particular device, but I do not see any changes for my
14e4:4315; however, the 4331 has a different PHY than does the 4315.
A review of the commits between 3.13 and 3.14-rc1 shows nothing for bcma that
should cause this effect. For b43, the only one that might cause this is commit
64e5acb entitled "b43: fix the wrong assignment of status.freq in b43_rx()".
That was clearly a bug and it fixed thing for my system, but we must consider if
that bug was also masking some other problem that is now revealed.
I have asked the bug reporter at b.k.o if it is possible to bisect the problem.
Larry