I have a laptop where there are two unusable AHCI SATA controllers on PCI: 00:08.2/04:00.0 SATA controller: Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode] (rev 81) 00:08.2/04:00.1 SATA controller: Advanced Micro Devices, Inc. [AMD] FCH SATA Controller [AHCI mode] (rev 81) Since they are unused, I would like them to go to runtime suspend (allowing the PCI bridge they are behind to suspend also). I have the following udev rule: ACTION=="add", SUBSYSTEM=="pci", TEST=="power/control", ATTR{power/control}="auto" Unfortunately, they do not autosuspend: /sys/devices/pci0000:00/0000:00:08.2/0000:04:00.0 $ cat power/runtime_status active There is a 'ata1' subnode for which runtime PM is not enabled: /sys/devices/pci0000:00/0000:00:08.2/0000:04:00.0 $ cat ata1/power/control on ata1/uevent is empty, so there doesn't seem to be ab obvious way to modify power/control for this node via udev rules. Why is runtime PM disabled for this node, and what's the proper way to enable it?
re: /sys/devices/pci0000:00/0000:00:08.2/0000:04:00.0 $ cat ata1/power/control on can we set it to auto manually? if yes, can we enable the SATA controller runtime PM then? please attach the dmesg output after boot.
Created attachment 295643 [details] complete dmesg
(In reply to Zhang Rui from comment #1) > re: /sys/devices/pci0000:00/0000:00:08.2/0000:04:00.0 $ cat > ata1/power/control > on > > can we set it to auto manually? if yes, can we enable the SATA controller > runtime PM then? Yes, writing 'auto' in that file works, and after that SATA controller is able to enter runtime autosuspend. > please attach the dmesg output after boot. Attached.
You edited the bug title to mention the CPU model, but as far as I can tell this is not specific to AMD. On Intel-based systems I have access to, it's the same: ata1/power/control is 'on' after boot. Isn't it the same for all AHCI controllers? You left the bug in NEEDINFO state, is there something else I need to provide?
(In reply to Alexander Monakov from comment #4) > You edited the bug title to mention the CPU model, but as far as I can tell > this is not specific to AMD. On Intel-based systems I have access to, it's > the same: ata1/power/control is 'on' after boot. Isn't it the same for all > AHCI controllers? Usually, I always prefer to add the laptop/cpu model as part of the bug report title. It sometimes remind me what the problem is about. > > You left the bug in NEEDINFO state, is there something else I need to > provide? No.