[Bug 216516] s2ram freezes screen (Ryzen-5650U incl. Radeon GPU)

bugzilla-daemon at kernel.org bugzilla-daemon at kernel.org
Fri Sep 23 15:52:31 UTC 2022


https://bugzilla.kernel.org/show_bug.cgi?id=216516

--- Comment #6 from Mario Limonciello (AMD) (mario.limonciello at amd.com) ---
> Tested s2idle with the openSUSE-15.4 stock kernel v5.14 and it consumed
> MASSIVELY more power than S3.

TONS and TONS of fixes have landed for s2idle since then.  5.15.y is the first
kernel that you should be confident to use it.

Please re-test with s2idle using 6.0-rc6.  If that doesn't work let's fix it.

> Is "echo s2idle > /sys/power/mem_sleep" equivalent to enabling s2idle via
> ACPI?
So I wouldn't need to reboot to change ACPI for testing s2idle.

No, on AMD systems there are Pre-OS changes that occur when you use modern
standby/s2idle.  If you use s2idle when the firmware is configured for S3 you
will never get into the proper power state.


> Actually S3 is very well hidden. I don't know a way to enable it directly in
> the BIOS. But it's in the ACPI tables.

As you're patching the firmware this is not a regression.  It takes more to
properly enable S3 in Pre-OS code than changing a variable evaluated by the
Linux kernel's ACPI interpreter..

Thorsten - this is not a regression, patching the BIOS like this is completely
off the beating path.  We can't possibly support S3 in Linux if the OEM doesn't
offer it in their BIOS.

kolAflash - Let's focus on fixing s2idle for your system.  With 6.0-rc6 and an
"unpatched" BIOS let's see if you have any remaining problems with it.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.


More information about the dri-devel mailing list