Hi,
We (Fedora) have been getting bugreports that brcmfmac wifi drops out and needs a rmmod + modprobe
to restore functionality starting with kernel 4.19:
https://bugzilla.redhat.com/show_bug.cgi?id=1658178
Disabling powersaving reportedly helps, but leads to an oops (which might be a hint?) :
https://bugzilla.redhat.com/show_bug.cgi?id=1661768
Regards
Hans
On 1/11/2019 4:34 PM, Arend Van Spriel wrote:
> - list (HTML filtering)
>
> Op vr 11 jan. 2019 16:18 schreef Hans de Goede <[email protected]
> <mailto:[email protected]>:
>
> Hi,
>
> We (Fedora) have been getting bugreports that brcmfmac wifi drops
> out and needs a rmmod + modprobe
> to restore functionality starting with kernel 4.19:
>
> https://bugzilla.redhat.com/show_bug.cgi?id=1658178
>
>
> Will look into this. Maybe it is similar to what Russell King reported.
Probably not. The bug report above is with PCIe device. The dmesg
attached to the bug report shows user is doing system suspend/resume
over and over until the following occurs:
[28151.874963] brcmfmac: brcmf_psm_watchdog_notify: PSM's watchdog has
fired!
[29354.351337] brcmfmac: brcmf_psm_watchdog_notify: PSM's watchdog has
fired!
PSM is the hardware block which processes the 802.11 packets received
from firmware and passes them to the phy layer. When this watchdog barks
the firmware will end up in a trap. We do collect a devcoredump upon the
watchdog, but unfortunately that will not tell much about the failure in
the hardware block. On monday I will try to come up with a patch to
obtain more information from the device regarding this watchdog.
Regards,
Arend
On 12/01/2019 13:29, Arend Van Spriel wrote:
> On 1/11/2019 4:34 PM, Arend Van Spriel wrote:
>> - list (HTML filtering)
>>
>> Op vr 11 jan. 2019 16:18 schreef Hans de Goede <[email protected]
>> <mailto:[email protected]>:
>>
>> Hi,
>>
>> We (Fedora) have been getting bugreports that brcmfmac wifi drops
>> out and needs a rmmod + modprobe
>> to restore functionality starting with kernel 4.19:
>>
>> https://bugzilla.redhat.com/show_bug.cgi?id=1658178
>>
>>
>> Will look into this. Maybe it is similar to what Russell King reported.
>
> Probably not. The bug report above is with PCIe device. The dmesg
> attached to the bug report shows user is doing system suspend/resume
> over and over until the following occurs:
>
> [28151.874963] brcmfmac: brcmf_psm_watchdog_notify: PSM's watchdog has
> fired!
> [29354.351337] brcmfmac: brcmf_psm_watchdog_notify: PSM's watchdog has
> fired!
>
> PSM is the hardware block which processes the 802.11 packets received
> from firmware and passes them to the phy layer. When this watchdog barks
> the firmware will end up in a trap. We do collect a devcoredump upon the
> watchdog, but unfortunately that will not tell much about the failure in
> the hardware block. On monday I will try to come up with a patch to
> obtain more information from the device regarding this watchdog.
Arend, did you manage to look any further into this? I've also noticed
problems with brcmfmac dropping out since 4.19, though haven't seen any
notices about PSM: https://bugzilla.kernel.org/show_bug.cgi?id=201853.