[-- Attachment #1: Type: text/plain, Size: 1413 bytes --] New issue by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806 Description: On my Thinkpad T460s (i5-6300U), even after running `powertop --auto-tune`, the package cannot go below pc3 state. This seems specific to Void, as on Arch Linux it can go below that. ``` $ xuname Void 5.10.5_1 x86_64-musl GenuineIntel uptodate hold rrnDDFFF ``` ``` $ sudo powertop --auto-tune Password: modprobe cpufreq_stats failedLoaded 147 prior measurements RAPL device for cpu 0 RAPL Using PowerCap Sysfs : Domain Mask f RAPL device for cpu 0 RAPL Using PowerCap Sysfs : Domain Mask f Devfreq not enabled glob returned GLOB_NOMATCH To show power estimates do 192 measurement(s) connected to battery only Leaving PowerTOP ``` Incidentally, I've been trying to use Arch Linux kernel in Void, with the final goal of pinning this powertop issue down to some kernel config difference between Arch and Void, but while I can successfully build it (after changing some part about signing modules, vary GCC version and add `zstd` as `hostmakedepends`), when I try to boot it fails at decrypting the partition, with ``` encrypted_key: failed to alloc_cypher (-2) modprobe: ERROR: could not insert 'dm_crypt': Unknown symbol in module, or unknown parameter (see dmesg) device-mapper: reload ioctl on failed: Invalid argument ``` Does anyone know what I'm missing? Thanks.
[-- Attachment #1: Type: text/plain, Size: 220 bytes --] New comment by abenson on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758183134 Comment: > kernel config difference Why not just compare the kernel configurations?
[-- Attachment #1: Type: text/plain, Size: 281 bytes --] New comment by ericonr on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758184832 Comment: You're probably getting conflicts between the `/usr/lib/modules` folders. How did you determine you weren't entering states below PC3 ?
[-- Attachment #1: Type: text/plain, Size: 971 bytes --] New comment by ericonr on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758204397 Comment: Fow what it's worth, I have a diff in https://gist.github.com/ericonr/a49d4fa9980aab0a5862a2c50bec6c3a The only difference that I think could apply is ` SATA_MOBILE_LPM_POLICY 3 -> 0`, which means ``` config SATA_MOBILE_LPM_POLICY int "Default SATA Link Power Management policy for mobile chipsets" range 0 4 default 0 depends on SATA_AHCI help Select the Default SATA Link Power Management (LPM) policy to use for mobile / laptop variants of chipsets / "South Bridges". The value set has the following meanings: 0 => Keep firmware settings 1 => Maximum performance 2 => Medium power 3 => Medium power with Device Initiated PM enabled 4 => Minimum power Note "Minimum power" is known to cause issues, including disk corruption, with some disks and should not be used. ```
[-- Attachment #1: Type: text/plain, Size: 299 bytes --] New comment by abenson on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758271011 Comment: I'm currently building a kernel with this change to test on my X270, which has the same CPU as the T460s mentioned earlier and seems to be in the same boat.
[-- Attachment #1: Type: text/plain, Size: 261 bytes --] New comment by abenson on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758290974 Comment: C2 (pc2) 6.8% C3 (pc3) 0.1% C6 (pc6) 58.2% C7 (pc7) 55.7% That seems to have been it!
[-- Attachment #1: Type: text/plain, Size: 547 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-758508543 Comment: Wow, many thanks @ericonr for your quick and to the point reply and to @abenson for testing. I looked at the diff before, but was unable to pinpoint the relevant lines. I was testing the pc states with powertop. I will also build with the suggested change later and report, just to confirm. As far as building Arch kernel, that may not be necessary anymore, but I may come back to it later just for fun.
[-- Attachment #1: Type: text/plain, Size: 527 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-759411125 Comment: Even with the proposed modification ``` $ cat config-5.10.6_2 | grep SATA_MOBIL CONFIG_SATA_MOBILE_LPM_POLICY=3 $ xuname Void 5.10.6_2 x86_64-musl GenuineIntel uptodate hold rrnDDFFF ``` and after running `powertop --auto-tune` and `powertop --calibrate`, I still cannot get below pc3 state. I tested from tty, with many services stopped. Could it be tlp is interfering with this?
[-- Attachment #1: Type: text/plain, Size: 528 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-759411125 Comment: Even with the proposed modification ``` $ cat config-5.10.6_2 | grep SATA_MOBIL CONFIG_SATA_MOBILE_LPM_POLICY=3 $ xuname Void 5.10.6_2 x86_64-musl GenuineIntel uptodate hold rrnDDFFF ``` and after running `powertop --auto-tune` and `powertop --calibrate`, I still cannot go below pc3 state. I tested from tty, with many services stopped. Could it be `tlp` is interfering with this?
[-- Attachment #1: Type: text/plain, Size: 187 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-759436797 Comment: @abenson may I ask if you're also using `musl`?
[-- Attachment #1: Type: text/plain, Size: 196 bytes --] New comment by abenson on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-759778565 Comment: I'm neither using musl nor tlp on the tested systems.
[-- Attachment #1: Type: text/plain, Size: 262 bytes --] New comment by abenson on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-759778565 Comment: I'm neither using musl nor tlp on the tested systems. - X270 (i5-6300U) - Dell Precision 7720 (some Xeon something)
[-- Attachment #1: Type: text/plain, Size: 226 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-760121240 Comment: Thanks. I tried disabling `tlp`, still not good. Will try soon with a `glibc` install.
[-- Attachment #1: Type: text/plain, Size: 214 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-761584345 Comment: I can confirm the suggested fix works with `glibc` install. Thanks again.
[-- Attachment #1: Type: text/plain, Size: 1471 bytes --] Closed issue by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806 Description: On my Thinkpad T460s (i5-6300U), even after running `powertop --auto-tune`, the package cannot go below pc3 state. This seems specific to Void, as on Arch Linux it can go below that. ``` $ xuname Void 5.10.5_1 x86_64-musl GenuineIntel uptodate hold rrnDDFFF ``` ``` $ sudo powertop --auto-tune Password: modprobe cpufreq_stats failedLoaded 147 prior measurements RAPL device for cpu 0 RAPL Using PowerCap Sysfs : Domain Mask f RAPL device for cpu 0 RAPL Using PowerCap Sysfs : Domain Mask f Devfreq not enabled glob returned GLOB_NOMATCH To show power estimates do 192 measurement(s) connected to battery only Leaving PowerTOP ``` Incidentally, I've been trying to use Arch Linux kernel in Void, with the final goal of pinning this powertop issue down to some kernel config difference between Arch and Void, but while I can successfully build it (after changing some part about signing modules, vary GCC version and add `zstd` as `hostmakedepends`), when I try to boot it fails at decrypting the partition, with ``` encrypted_key: failed to alloc_cypher (-2) modprobe: ERROR: could not insert 'dm_crypt': Unknown symbol in module, or unknown parameter (see dmesg) device-mapper: table: 254:0: crypt: unknown target type device-mapper: reload ioctl on failed: Invalid argument ``` Does anyone know what I'm missing? Thanks.
[-- Attachment #1: Type: text/plain, Size: 205 bytes --] New comment by travankor on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-761718809 Comment: Would be interesting to find out what is different for musl.
[-- Attachment #1: Type: text/plain, Size: 259 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-761758686 Comment: I still have musl on my machine (at least until grub 2.06 comes out), so I can run some tests. Do you have any ideas?
[-- Attachment #1: Type: text/plain, Size: 267 bytes --] New comment by dkwo on void-packages repository https://github.com/void-linux/void-packages/issues/27806#issuecomment-764971473 Comment: I stand corrected: with the new kernel config, also in musl it can reach pc9 power state. I'm not sure why I missed it before.