Bug#934781: firmware-iwlwifi: iwl4965: Microcode SW error detected

[Celejar <celejar@gmail.com>]

On Fri, 13 Dec 2019 15:29:23 +1030
Andrew Bettison <andrew@iverin.com.au> wrote:

> On Sun, 08 Sep 2019 23:43:04 -0400 Celejar <celejar@gmail.com> wrote:
> 
>  > Package: firmware-iwlwifi
>  > Version: 20190717-2
>  > Followup-For: Bug #934781
>  >
>  > I did some further digging - in my case, at least, the problem seems to
>  > be triggered by some relatively recent kernel change. I combed through
>  > the system journal for the last few months, and the problem first starts
>  > appearing in the logs a few days after I began running kernel 5.2.x
>  > (from 5.2.7 through 5.2.9). Previously, while running 4.19.x, the
>  > problem never appears.
>  >
>  > I haven't done a bisection, but it seems pretty clear at this point that
>  > there's a microcode bug that has begun to be triggered by something in
>  > newer kernels.

> As an experiment, I downloaded the tarball from 
> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tag/?h=20190815 
> (as recommended in 
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=939262 ), and manually 
> copied its intel/* and iwlwifi-* files into /lib/firmware (overwriting 
> the ones installed by the firmware-iwlwifi package).
> 
> After rebooting kernel 5.3.0-2 (amd64) I still get "iwlwifi: Microcode 
> SW error" messages accompanied by "ieee80211 phy0: Hardware restart was 
> requested" during periods of high Wi-Fi throughput, but they do not 
> cause the Wi-Fi to hang.  So the newer firmware appears to rectify the 
> worst part of the problem (Wi-Fi freeze) but does not fully resolve the 
> problem.

On my system, the problem last appeared on Sep. 17 (with a Debian
kernel version 5.2.9-2). Since then, I have not seen any "Microcode SW"
errors. I've been running 4.19.72 (self-built), 5.2.17-1, 5.3.7-1,
5.3.9 (-1, -2, -3), and 5.3.15-1 (Debian).

[I examined the boot logs with:

~# journalctl -S 2019-06-01 | grep "Linux ver\|Microcode SW"
]

Celejar