Hello maintainers,I've recently been doing some suspend/resume testing related to an s2idle bug in amdgpu (https://gitlab.freedesktop.org/drm/amd/-/issues/1879) and I've come across some problems with C-states.
After resuming from s2idle, Pkg(HW) will no longer enter any state C2 or deeper. Suspending to s3 would not trigger this issue on 5.16.2, but after trying out 5.16.4 I'm facing a similar problem where Pkg(HW) will rarely if ever go deeper than C3. Now, the reason I am contacting you is because I was playing around today and found that unloading i2c-i801 seems to fix the issue.
# CPU: Intel(R) Core(TM) i7-8665U CPU @ 1.90GHz# GPU: Intel Corporation WhiskeyLake-U GT2 [UHD Graphics 620], Advanced Micro Devices, Inc. [AMD/ATI] Lexa XT [Radeon PRO WX 3100] # 00:15.0 Serial bus controller: Intel Corporation Cannon Point-LP Serial IO I2C Controller #0 (rev 30) # 00:15.1 Serial bus controller: Intel Corporation Cannon Point-LP Serial IO I2C Controller #1 (rev 300:19.0 Serial bus controller: Intel Corporation Cannon Point-LP Serial IO I2C Host Controller (rev 30)
dmesg output when loading the module after an unload: [ 1002.961091] i801_smbus 0000:00:1f.4: SPD Write Disable is set [ 1002.961171] i801_smbus 0000:00:1f.4: SMBus using PCI interrupt[ 1002.961321] iTCO_wdt iTCO_wdt: unable to reset NO_REBOOT flag, device disabled by hardware/BIOS [ 1002.975399] i801_smbus 0000:00:1f.4: Accelerometer lis3lv02d is present on SMBus but its address is unknown, skipping registration
[ 1002.975406] i2c i2c-10: 2/2 memory slots populated (from DMI)[ 1002.976713] ee1004 10-0050: 512 byte EE1004-compliant SPD EEPROM, read-only
[ 1002.976808] i2c i2c-10: Successfully instantiated SPD at 0x50This is my first LKML bug report, so I could use some guidance in how to provide more information.
Regards, Kim Nilsson PS. Sorry about the initial HTML version!
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature
