On 9/5/24 9:27 AM, Kalle Valo wrote:
Marek Vasut <marex@xxxxxxx> writes:
Are you also taking into account hibernation? During hibernation the
device will be powered off. I can't remember the details right now but
wanted to mention this.
I don't think I am. Isn't hibernation actually a full shutdown, so the
hardware does get reinitialized ?
I don't know how it works exactly nor what you exactly mean with
reinitalized.
My understanding is that hibernation is suspend-to-disk, isn't it ?
Yes, that's what I mean.
(that's not something that is even available on my hardware)
Yeah, I'm guessing wilc1000 is used more in embedded enviroments where
hibernation is not really an important, it's used more in laptops. So I
guess it won't be a big problem if wilc1000 doesn't support hibernation
but please try to keep it in mind still.
Doesn't the hardware get completely turned OFF during suspend-to-disk
and then turned ON (and therefore initialized again) on
resume-from-disk?
I'm not sure how it works exactly but my experience is that for
hibernation a driver cannot assume that the firmware is running during
resume.
But at least with ath11k hibernation didn't work when it
left the firmware running during suspend.
Is there a thread on lore or some details of this you could point me to?
See commit 166a490f59ac ("wifi: ath11k: support hibernation") and the
commits before that. I'm sure that there's more info about hibernation
but don't have any pointers.
Ah, thank you.
The firmware reloading support should be addressed by some sort of
future variant of patch Ajay sent as attachment in [1], that should then
also deal with the hibernation part. I asked them to post the patch
properly to the ML.
What about this patch, it fixes suspend/resume for me and with a fix to
the atmel controller I suggested to Alexis in this thread, it should not
have any adverse effects .
[1]
https://lore.kernel.org/all/2bbdc690-aec5-4a11-893e-01270c6d5b84@xxxxxxxxxxxxx/
