Mark Kettenis <[email protected]> writes:
> Certainly looks like a smoking gun to me.
>
> One thing to look at is amdgpu_acpi_is_s0ix_active() in
> dev/pci/drm/amd/amdgpu/amdgpu_acpi.c. That function should return
> true, but may be returning false because we don't #define CONFIG_AMD_PMC.
> So maybe try what happens if you force that function to return true.
The patch below doesn't seem to have any effect. I still see the same
unfinished sleep the second time around. The last dmesg entry then
is still
> config_suspend: wsdisplay0 0
At that point acpi0's WCHAN is dmafenc:
0 53191 0 1 -22 0 0 0 dmafenc DK ?? 0:23.66 (acpi0)
X is in kqread:
0 84130 1 6 18 0 1012 1952 sigsusp I ?? 0:00.00
/usr/X11R6/bin/xenodm
35 81802 84130 0 2 0 68936 97272 kqread I ?? 0:00.75
/usr/X11R6/bin/X :0 vt05 -auth /etc/X11/xenodm/authdir/authfiles/A:0-InfYGC
(Xorg)
0 48574 81802 3 2 0 2976 1468 sbwait IpU ?? 0:00.00 X:
[priv] (Xorg)
0 81121 1 20 2 0 972 1032 sbwait I ?? 0:00.00
/usr/X11R6/bin/xconsole
35 33692 1 0 2 0 1320 3692 kqread Ip ?? 0:00.80
/usr/X11R6/bin/xconsole
Per deraadt@ suggestion I used `pkill X` to see what happens:
0 48574 81802 0 -22 0 0 0 - Zp ?? 0:00.00 (Xorg)
0 84130 1 0 18 0 1012 1960 sigsusp I ?? 0:00.00
/usr/X11R6/bin/xenodm
35 81802 84130 0 0 0 66868 95332 drmioc D ?? 0:00.76
/usr/X11R6/bin/X :0 vt05 -auth /etc/X11/xenodm/authdir/authfiles/A:0-InfYGC
(Xorg)
No otherwise observable symptoms besides X becoming D. When zzz is not
anywhere in the picture, kill'ing X over ssh leads to its normal restart.
I'll try to run other experiments with not having started X at all.
modified sys/dev/pci/drm/amd/amdgpu/amdgpu_acpi.c
@@ -1514,10 +1514,10 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device
*adev)
{
if (!(adev->flags & AMD_IS_APU) ||
(pm_suspend_target_state != PM_SUSPEND_TO_IDLE))
- return false;
+ return true;
if (adev->asic_type < CHIP_RAVEN)
- return false;
+ return true;
/*
* If ACPI_FADT_LOW_POWER_S0 is not set in the FADT, it is generally
@@ -1529,13 +1529,13 @@ bool amdgpu_acpi_is_s0ix_active(struct amdgpu_device
*adev)
dev_err_once(adev->dev,
"Power consumption will be higher as BIOS has not
been configured for suspend-to-idle.\n"
"To use suspend-to-idle change the sleep mode in
BIOS setup.\n");
- return false;
+ return true;
}
#if !IS_ENABLED(CONFIG_AMD_PMC)
dev_err_once(adev->dev,
"Power consumption will be higher as the kernel has not
been compiled with CONFIG_AMD_PMC.\n");
- return false;
+ return true;
#else
return true;
#endif /* CONFIG_AMD_PMC */
