Skip to content
Snippets Groups Projects
New look: Off

pine64-pinephone: cpu-idle driver is being loaded, but still not used.

    • Closed Imported Issue created by Administrator @root

    Describe your issue

    What's the expected behaviour?

    After installing cpupower, issue the command: "cpupower idle-info", and it should spit out something like:

    CPUidle driver: psci_idle
    CPUidle governor: menu
    analyzing CPU 0:

    Number of idle states: 3
    Available idle states: WFI cpu-sleep cluster-sleep
    WFI:
    Flags/Description: ARM WFI
    Latency: 1
    Usage: 42393
    Duration: 72730631
    cpu-sleep:
    Flags/Description: cpu-sleep
    Latency: 2300
    Usage: 4411
    Duration: 25070974
    cluster-sleep:
    Flags/Description: cluster-sleep
    Latency: 2350
    Usage: 2506
    Duration: 16476184

    #### What's the current behaviour? cpupower idle-info returns 0 use values for the cpu-sleep and cluster-sleep states.

    CPUidle driver: psci_idle
    CPUidle governor: menu
    analyzing CPU 0:

    Number of idle states: 3
    Available idle states: WFI cpu-sleep cluster-sleep
    WFI:
    Flags/Description: ARM WFI
    Latency: 1
    Usage: 166755
    Duration: 164166753
    cpu-sleep:
    Flags/Description: cpu-sleep
    Latency: 2300
    Usage: 0
    Duration: 0
    cluster-sleep:
    Flags/Description: cluster-sleep
    Latency: 2350
    Usage: 0
    Duration: 0

    How to reproduce your issue?

    Just install the latest egde build and install cpupower.
    Issue a cpupower idle-info command and it outputs the current stats for the cpu-idle driver.

    What device are you using?

    pine64-pinephone

    On what postmarketOS version did you encounter the issue?

    • edge (master branch)

    • v22.06 (supported until 2023-01-18)

    • v22.12

    • I confirm that the issue still is present after running sudo apk upgrade -a

    On what environment did you encounter the issue?

    Environments

    • Phosh
    • Plasma Mobile
    • Sxmo (Wayland/Sway) Please post the output of sxmo_version.sh
    • Other: Please fill out

    How did you get postmarketOS image?

    What's the build date of the image? (in yyyy-mm-dd format)

    Additional information

    4 of 11 checklist items completed
    To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information

    Child items ...

    • Activity

    • Administrator
      Administrator @root · Imported
      Author Owner

      @BobbyTheBuilder, @megous, @emulti: Copied in because of your participation in the previous bug report #1901 (closed)
      And also @Arnavion, because you have been working on trying to debug this in Pine64 PinePhone matrix channel together with Megi and Norayr.
      We can keep track of the issue and attempts over here to get some good record keeping.
      What I already tried so far was building a kernel with disabled configuration CONFIG_ARM_ALLWINNER_SUN50I_CPUFREQ_NVMEM and CONFIG_ARM_SCPI_CPUFREQ, which seem to be unset in Megi's personal pp-6.1 kernel configuration. I was thinking that maybe one of these modules, which relate to cpu-idle as well interfered.
      This does not seem to be the case. So i'll try to investigate the kernel configuration differences some more.

      By Jan Jasper de Kroon on 2023-02-11T18:26:20

    • Administrator
      Administrator @root · Imported
      Author Owner

      sysfs dump from when I was talking to megi yesterday:

      Standard properties:

      /sys/devices/system/cpu/cpu0/cpuidle/state0/name:1:WFI
/sys/devices/system/cpu/cpu0/cpuidle/state0/desc:1:ARM WFI
/sys/devices/system/cpu/cpu0/cpuidle/state0/default_status:1:enabled
/sys/devices/system/cpu/cpu0/cpuidle/state0/disable:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/latency:1:1
/sys/devices/system/cpu/cpu0/cpuidle/state0/power:1:4294967295
/sys/devices/system/cpu/cpu0/cpuidle/state0/residency:1:1

/sys/devices/system/cpu/cpu0/cpuidle/state1/name:1:cpu-sleep
/sys/devices/system/cpu/cpu0/cpuidle/state1/desc:1:cpu-sleep
/sys/devices/system/cpu/cpu0/cpuidle/state1/default_status:1:enabled
/sys/devices/system/cpu/cpu0/cpuidle/state1/disable:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state1/latency:1:2300
/sys/devices/system/cpu/cpu0/cpuidle/state1/power:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state1/residency:1:25000

/sys/devices/system/cpu/cpu0/cpuidle/state2/name:1:cluster-sleep
/sys/devices/system/cpu/cpu0/cpuidle/state2/desc:1:cluster-sleep
/sys/devices/system/cpu/cpu0/cpuidle/state2/default_status:1:enabled
/sys/devices/system/cpu/cpu0/cpuidle/state2/disable:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/latency:1:2350
/sys/devices/system/cpu/cpu0/cpuidle/state2/power:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/residency:1:50000

      megi's phone:

      /sys/devices/system/cpu/cpu0/cpuidle/state0/above:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/below:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/rejected:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/time:1:3295122667
/sys/devices/system/cpu/cpu0/cpuidle/state0/usage:1:1210799

/sys/devices/system/cpu/cpu0/cpuidle/state1/above:1:259725
/sys/devices/system/cpu/cpu0/cpuidle/state1/below:1:8189
/sys/devices/system/cpu/cpu0/cpuidle/state1/rejected:1:25088
/sys/devices/system/cpu/cpu0/cpuidle/state1/time:1:1580989453
/sys/devices/system/cpu/cpu0/cpuidle/state1/usage:1:279594

/sys/devices/system/cpu/cpu0/cpuidle/state2/above:1:196388
/sys/devices/system/cpu/cpu0/cpuidle/state2/below:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/rejected:1:27914
/sys/devices/system/cpu/cpu0/cpuidle/state2/time:1:152177889
/sys/devices/system/cpu/cpu0/cpuidle/state2/usage:1:208275

      My phone:

      /sys/devices/system/cpu/cpu0/cpuidle/state0/above:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/below:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/rejected:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state0/time:1:587745283
/sys/devices/system/cpu/cpu0/cpuidle/state0/usage:1:537655

/sys/devices/system/cpu/cpu0/cpuidle/state1/above:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state1/below:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state1/rejected:1:34993088
/sys/devices/system/cpu/cpu0/cpuidle/state1/time:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state1/usage:1:0

/sys/devices/system/cpu/cpu0/cpuidle/state2/above:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/below:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/rejected:1:10053422
/sys/devices/system/cpu/cpu0/cpuidle/state2/time:1:0
/sys/devices/system/cpu/cpu0/cpuidle/state2/usage:1:0

      About "rejected", the cpuidle driver docs say:

      Generally, an interrupt received when trying to enter an idle state causes the idle state entry request to be rejected, in which case the CPUIdle driver may return an error code to indicate that this was the case. The usage and rejected files report the number of times the given idle state was entered successfully or rejected, respectively.

      By Arnav Singh on 2023-02-11T19:09:42

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      Also, based on talking to him yesterday which was basically the same as discussed in #1901 (closed) , and also while I was investigating a way to use a different device tree because of #1945 , I wonder what would happen if we made u-boot use the dtb from the kernel and let the TF-A modifications remain. AFAICT @BobbyTheBuilder had a way of doing that in #1901 (comment 1238125161) . The problem with that attempt was that u-boot's own DTS is not as good as the kernel's, which is why I'm proposing lifting the DTSes from the kernel and transplanting them into the u-boot build first. That way the net result would be u-boot uses up-to-date DTB from kernel -> T-FA applies modifications -> kernel uses this TF-A modified DTB.

      If nothing else that would at least bring it closer to what megi's p-boot does?

      By Arnav Singh on 2023-02-11T19:30:40

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      That should be done in a separate fork then is my guess.
      uBoot makes use of the upstream mainline DTS (for compatibility reasons)
      A lot of additions which are in Megi's branch can't simply be copied over, because this would probably break the mainline support.

      By Jan Jasper de Kroon on 2023-02-11T19:55:44

    • Administrator
      Administrator @root · Imported
      Author Owner

      We sort of already have that fork - PP is supposed to use u-boot-pinephone and that already does a bunch of PP-specific patches to u-boot (two patches plus a sed in the APKBUILD). I actually did experiment with adding the 1.2b DTS from the kernel to it yesterday and got it to compile the 1.2b DTB, though I missed something because it didn't actually switch to the new DT after boot.

      By Arnav Singh on 2023-02-11T20:14:05

    • Administrator
      Administrator @root · Imported
      Author Owner

      I'm not able to replicate BobbyTheBuilder's method of booting with the TF-A modified DTB.

      Taking the current uboot-script.cmd, I removed all the stuff for RAM speed and pcsi to keep everything simple:

      -echo Detecting psci idle state
-fdt addr ${fdtcontroladdr}
-fdt get name pscifdt /cpus/idle-states /
-if test $? -eq 0; then
-	echo PSCI idle state enabled;
-	setenv iscpscienabled 1;
-else
-	echo PSCI idle state disabled;
-fi
-
 echo Loading DTB
 load mmc ${mmc_bootdev}:1 ${fdt_addr_r} ${fdtfile}
 fdt addr ${fdt_addr_r}
-fdt resize 2048
+fdt resize
-
-if printenv ram_freq; then
-	echo Adding FTD RAM clock
-	fdt mknode / memory
-	fdt set /memory ram_freq ${ram_freq}
-fi
-
-if test ${iscpscienabled} -eq 1; then
-	echo Applying PSCI DTBO;
-	load mmc ${mmc_bootdev}:1 ${fdtoverlay_addr_r} sun50i-a64-psci.dtbo
-	fdt apply ${fdtoverlay_addr_r}
-fi

      At this point I still had a working boot.

      Then I replaced:

      -load mmc ${mmc_bootdev}:1 ${fdt_addr_r} ${fdtfile}
-fdt addr ${fdt_addr_r}
+cp.b ${fdtcontroladdr} ${fdt_addr_r} f000
+fdt addr ${fdt_addr_r} f000

      ... per BobbyTheBuilder's script. With this, booting shows the pmOS splash screen and then "initramfs-extra not found". Of course /boot/initramfs-extra exists and is fine. Reverting back to load mmc makes boot work again.

      Edit: Even this simple script that just passes ${fdtcontroladdr} directly to the kernel has the same issue:

      echo Loading Initramfs
load mmc ${mmc_bootdev}:1 ${ramdisk_addr_r} initramfs
setenv ramdisk_size ${filesize}

echo Loading Kernel
load mmc ${mmc_bootdev}:1 ${kernel_addr_r} vmlinuz

gpio set 115


echo Booting kernel
gpio set 116
gpio clear 98
booti ${kernel_addr_r} ${ramdisk_addr_r}:${ramdisk_size} ${fdtcontroladdr}

      Commands like fdt addr ${fdtcontroladdr} succeed, indicating there is a valid FDT at that address. But I don't have a serial so I can't print its properties to inspect what's causing init.sh to not find the pmOS_boot partition.

      I suspect some other changes to the u-boot config (pinephone_defconfig) are needed that BobbyTheBuilder didn't mention. I already tried appended CONFIG_OF_CONTROL=y since the documentation indicates it's necessary, but that didn't help.

      By Arnav Singh on 2023-02-12T07:34:46

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      But I don't have a serial so I can't print its properties to inspect what's causing init.sh to not find the pmOS_boot partition.

      I finally got a serial and figured out the problem. Boot device is being mixed up between mmcblk1 and mmcblk2, so pmos initrd was trying to mount boot from mmcblk2 but actual device is mmcblk1 when booted using original DTB. I assume BobbyTheBuilder didn't have this problem because, based on the steps in his post, they were booting from SD card.

      So setting bootdev to 1 instead of 2 fixes that problem and I'm now able to boot with original DTB.

      By Arnav Singh on 2023-02-28T00:37:55

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      You have a serial cable?
      I still need to obtain one myself, it seems like it's very handy to debug early boot troubles.
      So we need to alter the boot script to automatically load from the correct location.
      Force setting it isn't the way to go forward as some people have pmOS on eMMC and some are running it in SD.

      It suggests something in this conditional test isn't working properly:
      if test ${mmc_bootdev} -eq 0 ; then
      echo "Booting from SD";
      setenv bootdev 0;
      else
      echo "Booting from eMMC";
      setenv bootdev 2;
      fi;

      
 *By Jan Jasper de Kroon on 2023-02-28T09:55:53*
      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      So we need to alter the boot script to automatically load from the correct location. Force setting it isn't the way to go forward as some people have pmOS on eMMC and some are running it in SD.

      No, I'm not proposing changing anything right now. I was just documenting the problem I had for the sake of anyone else attempting it.

      First I need to get it working with proper FDT instead of u-boot's default neutered one, and confirm that the original issue of CPU idle states is fixed by this effort. Only then do we need to think about changes to the u-boot script, if any.

      By Arnav Singh on 2023-02-28T10:01:24

    • Administrator
      Administrator @root · Imported
      Author Owner

      One last wrinkle I had was that my changes to u-boot seemed to not be getting picked up even though I was installing it with update-u-boot. I was previously using tow-boot but I assumed that update-u-boot would overwrite it. I realized that assumption was wrong - update-u-boot writes to /dev/mmcblk# but tow-boot wrote itself to /dev/mmcblk#boot0 which takes precedence. I had to boot the tow-boot installer image via SD card and "Erase the eMMC contents" aka wipe /dev/mmcblk#boot0, then my phone reverted to booting via u-boot and my changes were picked up.

      So currently I can confirm that, when booted this way:

      1. cpupower idle-info mentions all three states WFI, cpu-sleep and cluster-sleep. Both cpu-sleep and cluster-sleep now have >0 Duration.

        $ cpupower idle-info
CPUidle driver: psci_idle
CPUidle governor: menu
analyzing CPU 2:

Number of idle states: 3
Available idle states: WFI cpu-sleep cluster-sleep
WFI:
Flags/Description: ARM WFI
Latency: 1
Usage: 94773
Duration: 141866485
cpu-sleep:
Flags/Description: cpu-sleep
Latency: 2300
Usage: 4688
Duration: 100018982
cluster-sleep:
Flags/Description: cluster-sleep
Latency: 2350
Usage: 2535
Duration: 70985966

      2. /sys/devices/system/cpu/cpu*/cpuidle/state{1,2}/{time,usage} are now >0.

        $ grep -Hn '' /sys/devices/system/cpu/cpu*/cpuidle/state{1,2}/{time,usage}

/sys/devices/system/cpu/cpu0/cpuidle/state1/time:1:191465233
/sys/devices/system/cpu/cpu1/cpuidle/state1/time:1:220504841
/sys/devices/system/cpu/cpu2/cpuidle/state1/time:1:213502165
/sys/devices/system/cpu/cpu3/cpuidle/state1/time:1:203952133
/sys/devices/system/cpu/cpu0/cpuidle/state1/usage:1:38551
/sys/devices/system/cpu/cpu1/cpuidle/state1/usage:1:9294
/sys/devices/system/cpu/cpu2/cpuidle/state1/usage:1:10278
/sys/devices/system/cpu/cpu3/cpuidle/state1/usage:1:7404
/sys/devices/system/cpu/cpu0/cpuidle/state2/time:1:44232873
/sys/devices/system/cpu/cpu1/cpuidle/state2/time:1:154982845
/sys/devices/system/cpu/cpu2/cpuidle/state2/time:1:124901154
/sys/devices/system/cpu/cpu3/cpuidle/state2/time:1:168485522
/sys/devices/system/cpu/cpu0/cpuidle/state2/usage:1:12825
/sys/devices/system/cpu/cpu1/cpuidle/state2/usage:1:5236
/sys/devices/system/cpu/cpu2/cpuidle/state2/usage:1:5085
/sys/devices/system/cpu/cpu3/cpuidle/state2/usage:1:4963

      3. BobbyTheBuilder had mentioned "low screen brightness", "USB-C port not functional" and "laggy UI" issues in his attempt in the other issue. These issues are not present, so that's good. However, the modem is not detected.

        $ lsusb

Bus 002 Device 001: ID 1d6b:0002 Linux 6.1.9 ehci_hcd EHCI Host Controller
Bus 004 Device 001: ID 1d6b:0001 Linux 6.1.9 ohci_hcd Generic Platform OHCI controller
Bus 001 Device 001: ID 1d6b:0002 Linux 6.1.9 ehci_hcd EHCI Host Controller
Bus 003 Device 001: ID 1d6b:0001 Linux 6.1.9 ohci_hcd Generic Platform OHCI controller

        This has happened despite booting with the DTB from the kernel. I did note that there are some comments in the DTS like "power is already turned on by the bootloader", and in p-boot's source I see some modifications it makes to the FDT, so those might need to be replicated in boot.scr or in the built-in DTB. I'll check with megi.

      If anyone else wants to try it, I've pushed my changes to https://gitlab.com/Arnavion/pmaports/-/compare/master...uboot

      To summarize, the current state of that branch is:

      1. Good: It is demonstrably possible to have u-boot boot with a full FDT based on the kernel DTS, let TF-A modify it to add sleep states etc, and then have the kernel use that FDT as-is. CPU sleep states work in the booted kernel.

      2. Good: All other devices (IIO, charging through USB-C port, connecting devices to USB-C port) seem to work, except...

      3. Needs investigation: Modem doesn't work, for unknown reason. This is a blocker.

      4. Needs investigation: Some invisible things related to power and safety might be broken, based on comments in the DTS and on p-boot code.

      5. Bad: The branch uses a hacky patch to import the DTS from kernel tree into u-boot.

        Possible solution: It might work to have u-boot-pinephone have a makedepends on linux-postmarketos-allwinner, so that it can do EXT_DTB=/kernel/buildroot/.../pinephone.dtb to import the DTB from kernel build output. That way we don't have to worry about patching u-boot. However...

      6. Bad: The branch only uses the pinephone-1.2.dtb for the sake of simplicitly, so it will break for 1.1 PinePhones.

        The EXT_DTB solution in the previous point will probably not work for embedding both DTBs, based on my reading of u-boot source.

      7. Kinda bad: This change to u-boot only takes effect when the phone is booted using u-boot. The changes to the boot.scr will be incompatible with booting from tow-boot - the mmc device is numbered different so the init.sh won't find the /boot partition, and even if it does get found it'll use the FDT built from upstream u-boot's DTS and thus be missing most functionality.

        Possible solution: The boot.scr can probably test the FDT to determine if the one it got from bootloader is good enough to use, or if it should ignore it and load mmc the full one from disk like it does today. So tow-boot users will still have a working device, just without the sleep states etc.

      8. Kinda bad: All this effort will need to be revisited if / when the PinePhone switches from u-boot to tow-boot + UEFI (!3068 (closed))

      By Arnav Singh on 2023-03-01T07:24:23

      Edited by Administrator
    • Administrator mentioned in issue #1945 · Imported

      mentioned in issue #1945

      By Arnav Singh on 2023-03-02T16:40:53

    • Administrator
      Administrator @root · Imported
      Author Owner
      1. Needs investigation: Modem doesn't work, for unknown reason. This is a blocker.

      This is fixed. I diffed /proc/devicetree between original and modified u-boot and realized I'd made a typo copying the patched DTS over.

      1. Needs investigation: Some invisible things related to power and safety might be broken, based on comments in the DTS and on p-boot code.

      I went through p-boot source and matched up the modifications it makes to the fdt, and they all seem to be no-ops. So this is also fine.

      As an aside, diffing the devicetrees also gave me the ability to check for differences in how the cpus/idle-states are defined between original and modified trees. And the answer is that there's no difference, so I'm not sure why the modified tree lets the CPU go into sleep states but the original does not.

      I've updated my branch. At this point my phone is fully functional and can go into lower sleep states.

      I'll send an MR tomorrow with my branch, and then we can bikeshed about how hacky it is and what better ways there are of achieving this.

      By Arnav Singh on 2023-03-04T09:43:34

    • Administrator
      Administrator @root · Imported
      Author Owner

      I found a nice (IMO) approach of using dtc to create the DTS from kernel package's DTB instead of needing to patch it in. I've opened !3926 with that.

      That approach has problems with the mount matrices of accelerometer and magnetometer, so I opened !3931 with an alternative approach.

      By Arnav Singh on 2023-03-06T21:55:36

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      I do not replicate the mentioned behavior, here's the boot log of one of my PP on Edge with all packages update as of today

      Detecting psci idle state
Working FDT set to f9f39ee0
PSCI idle state enabled
Loading DTB
53815 bytes read in 5 ms (10.3 MiB/s)
Working FDT set to 4fa00000
ram_freq=624
Adding FTD RAM clock
Applying PSCI DTBO
1130 bytes read in 2 ms (551.8 KiB/s)
Loading Initramfs
      pine64-pinephone:~# cpupower idle-info
CPUidle driver: psci_idle
CPUidle governor: menu
analyzing CPU 2:

Number of idle states: 3
Available idle states: WFI cpu-sleep cluster-sleep
WFI:
Flags/Description: ARM WFI
Latency: 1
Usage: 21648
Duration: 18468425
cpu-sleep:
Flags/Description: cpu-sleep
Latency: 2300
Usage: 377
Duration: 5770600
cluster-sleep:
Flags/Description: cluster-sleep
Latency: 2350
Usage: 190
Duration: 4775805

      Looking from your initial post, the PSCI idles states are properly initialized but they are not "used".

      Most likely there's CPU activity preventing the CPU to fall asleep ?

      Do you have a full boot log from the serial cable to share ?

      By Bobby The Builder on 2023-03-12T17:27:56

    • Administrator
      Administrator @root · Imported
      Author Owner

      I do not replicate the mentioned behavior, here's the boot log of one of my PP on Edge with all packages update as of today

      Are you booting from eMMC or from SD card?

      Most likely there's CPU activity preventing the CPU to fall asleep ?

      No. If that was the case this MR would not have made a difference.

      Do you have a full boot log from the serial cable to share ?

      I can share it in a few hours.

      By Arnav Singh on 2023-03-12T18:50:12

    • Administrator
      Administrator @root · Imported
      Author Owner

      I tested both configuration:

      Here's the microSD [mmcblk0] 
      pine64-pinephone:~# cat /proc/cmdline 
init=/init.sh rw console=tty0 console=ttyS0,115200 earlycon=uart,mmio32,0x01c28000 panic=10 consoleblank=0 loglevel=1 PMOS_NO_OUTPUT_REDIRECT PMOS_FORCE_PARTITION_RESIZE pmos_boot=/dev/mmcblk0p1 pmos_root=/dev/mmcblk0p2
pine64-pinephone:~# cpupower idle-info
CPUidle driver: psci_idle
CPUidle governor: menu
analyzing CPU 0:

Number of idle states: 3
Available idle states: WFI cpu-sleep cluster-sleep
WFI:
Flags/Description: ARM WFI
Latency: 1
Usage: 96436
Duration: 98745342
cpu-sleep:
Flags/Description: cpu-sleep
Latency: 2300
Usage: 10332
Duration: 73434536
cluster-sleep:
Flags/Description: cluster-sleep
Latency: 2350
Usage: 2972
Duration: 20236153
      Here's the eMMC [mmcblk2] 
      pine64-pinephone:~# cat /proc/cmdline 
init=/init.sh rw console=tty0 console=ttyS0,115200 earlycon=uart,mmio32,0x01c28000 panic=10 consoleblank=0 loglevel=1 PMOS_NO_OUTPUT_REDIRECT PMOS_FORCE_PARTITION_RESIZE pmos_boot=/dev/mmcblk2p1 pmos_root=/dev/mmcblk2p2
pine64-pinephone:~# cpupower idle-info
CPUidle driver: psci_idle
CPUidle governor: menu
analyzing CPU 0:

Number of idle states: 3
Available idle states: WFI cpu-sleep cluster-sleep
WFI:
Flags/Description: ARM WFI
Latency: 1
Usage: 56119
Duration: 46869340
cpu-sleep:
Flags/Description: cpu-sleep
Latency: 2300
Usage: 2888
Duration: 14925020
cluster-sleep:
Flags/Description: cluster-sleep
Latency: 2350
Usage: 798
Duration: 2571878

      I can share it in a few hours.

      Thanks :ok_hand:

      By Bobby The Builder on 2023-03-12T19:47:42

    • Administrator
      Administrator @root · Imported
      Author Owner

      Do you have a full boot log from the serial cable to share ?

      Default packages:

      linux-postmarketos-allwinner-6.1.9_git20230201-r1
u-boot-pinephone-2023.01-r0
device-pine64-pinephone-0.46-r0
device-pine64-pinephone-nonfree-firmware-0.46-r0
device-pine64-pinephone-upower-0.46-r0
device-pine64-pinephone-gnome-0.46-r0
device-pine64-pinephone-camera-0.46-r0
      Boot log 
      U-Boot SPL 2023.01 (Jan 14 2023 - 18:27:51 +0000)
DRAM: 3072 MiB
Trying to boot from MMC2
NOTICE:  BL31: v2.8(release):
NOTICE:  BL31: Built : 18:16:06, Jan 14 2023
NOTICE:  BL31: Detected Allwinner A64/H64/R18 SoC (1689)
NOTICE:  BL31: Found U-Boot DTB at 0x208fb48, model: Pine64 PinePhone (1.2)


U-Boot 2023.01 (Jan 14 2023 - 18:27:51 +0000) Allwinner Technology

CPU:   Allwinner A64 (SUN50I)
Model: Pine64 PinePhone (1.2)
DRAM:  3 GiB @ 552 MHz
Core:  71 devices, 18 uclasses, devicetree: separate
WDT:   Not starting watchdog@1c20ca0
MMC:   mmc@1c0f000: 0, mmc@1c10000: 2, mmc@1c11000: 1
Loading Environment from FAT... Unable to use mmc 1:1...
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
In:    serial
Out:   serial
Err:   serial
Net:   No ethernet found.
Hit any key to stop autoboot:  0 
switch to partitions #0, OK
mmc1(part 0) is current device
Scanning mmc 1:1...
Found U-Boot script /boot.scr
1627 bytes read in 2 ms (793.9 KiB/s)
## Executing script at 4fc00000
gpio: pin 98 (gpio 98) value is 1
gpio: pin 114 (gpio 114) value is 1
Booting from eMMC
arch=arm
baudrate=115200
board=sunxi
board_name=sunxi
boot_a_script=load ${devtype} ${devnum}:${distro_bootpart} ${scriptaddr} ${prefix}${script}; source ${scriptaddr}
boot_efi_binary=load ${devtype} ${devnum}:${distro_bootpart} ${kernel_addr_r} efi/boot/bootaa64.efi; if fdt addr -q ${fdt_addr_r}; then bootefi ${kernel_addr_r} ${fdt_addr_r};else bootefi ${kernel_addr_r} ${fdtcontroladdr};fi
boot_efi_bootmgr=if fdt addr -q ${fdt_addr_r}; then bootefi bootmgr ${fdt_addr_r};else bootefi bootmgr;fi
boot_extlinux=sysboot ${devtype} ${devnum}:${distro_bootpart} any ${scriptaddr} ${prefix}${boot_syslinux_conf}
boot_prefixes=/ /boot/
boot_script_dhcp=boot.scr.uimg
boot_scripts=boot.scr.uimg boot.scr
boot_syslinux_conf=extlinux/extlinux.conf
boot_targets=fel mmc_auto pxe dhcp 
bootargs=init=/init.sh rw console=tty0 console=ttyS0,115200 earlycon=uart,mmio32,0x01c28000 panic=10 consoleblank=0 loglevel=1 PMOS_NO_OUTPUT_REDIRECT PMOS_FORCE_PARTITION_RESIZE pmos_boot=/dev/mmcblk2p1 pmos_root=/dev/mmcblk2p2
bootcmd=run distro_bootcmd
bootcmd_dhcp=devtype=dhcp; if dhcp ${scriptaddr} ${boot_script_dhcp}; then source ${scriptaddr}; fi;setenv efi_fdtfile ${fdtfile}; setenv efi_old_vci ${bootp_vci};setenv efi_old_arch ${bootp_arch};setenv bootp_vci PXEClient:Arch:00011:UNDI:003000;setenv bootp_arch 0xb;if d;
bootcmd_fel=if test -n ${fel_booted} && test -n ${fel_scriptaddr}; then echo '(FEL boot)'; source ${fel_scriptaddr}; fi
bootcmd_mmc0=devnum=0; run mmc_boot
bootcmd_mmc1=devnum=1; run mmc_boot
bootcmd_mmc_auto=if test ${mmc_bootdev} -eq 1; then run bootcmd_mmc1; run bootcmd_mmc0; elif test ${mmc_bootdev} -eq 0; then run bootcmd_mmc0; run bootcmd_mmc1; fi
bootcmd_pxe=dhcp; if pxe get; then pxe boot; fi
bootdelay=0
bootdev=2
bootfstype=ext4
bootm_size=0xa000000
console=ttyS0,115200
cpu=armv8
devplist=1
dfu_alt_info_ram=kernel ram 0x40080000 0x1000000;fdt ram 0x4FA00000 0x100000;ramdisk ram 0x4FF00000 0x4000000
distro_bootcmd=for target in ${boot_targets}; do run bootcmd_${target}; done
efi_dtb_prefixes=/ /dtb/ /dtb/current/
ethaddr=02:ba:0c:1b:a1:48
fdt_addr_r=0x4FA00000
fdtcontroladdr=f9f39ee0
fdtfile=allwinner/sun50i-a64-pinephone-1.2.dtb
fdtoverlay_addr_r=0x4FE00000
fileaddr=4fc00000
filesize=65b
kernel_addr_r=0x40080000
kernel_comp_addr_r=0x44000000
kernel_comp_size=0xb000000
load_efi_dtb=load ${devtype} ${devnum}:${distro_bootpart} ${fdt_addr_r} ${prefix}${efi_fdtfile}
loadaddr=0x42000000
mmc_boot=if mmc dev ${devnum}; then devtype=mmc; run scan_dev_for_boot_part; fi
mmc_bootdev=1
partitions=name=loader1,start=8k,size=32k,uuid=${uuid_gpt_loader1};name=loader2,size=984k,uuid=${uuid_gpt_loader2};name=esp,size=128M,bootable,uuid=${uuid_gpt_esp};name=system,size=-,uuid=${uuid_gpt_system};
pxefile_addr_r=0x4FD00000
ram_freq=552
ramdisk_addr_r=0x4FF00000
scan_dev_for_boot=echo Scanning ${devtype} ${devnum}:${distro_bootpart}...; for prefix in ${boot_prefixes}; do run scan_dev_for_extlinux; run scan_dev_for_scripts; done;run scan_dev_for_efi;
scan_dev_for_boot_part=part list ${devtype} ${devnum} -bootable devplist; env exists devplist || setenv devplist 1; for distro_bootpart in ${devplist}; do if fstype ${devtype} ${devnum}:${distro_bootpart} bootfstype; then run scan_dev_for_boot; fi; done; setenv devplist
scan_dev_for_efi=setenv efi_fdtfile ${fdtfile}; for prefix in ${efi_dtb_prefixes}; do if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${efi_fdtfile}; then run load_efi_dtb; fi;done;run boot_efi_bootmgr;if test -e ${devtype} ${devnum}:${distro_bootpart} efi/boote
scan_dev_for_extlinux=if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${boot_syslinux_conf}; then echo Found ${prefix}${boot_syslinux_conf}; run boot_extlinux; echo EXTLINUX FAILED: continuing...; fi
scan_dev_for_scripts=for script in ${boot_scripts}; do if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${script}; then echo Found U-Boot script ${prefix}${script}; run boot_a_script; echo SCRIPT FAILED: continuing...; fi; done
scriptaddr=0x4FC00000
serial#=92c0a9ba0c1ba148
soc=sunxi
stderr=serial,vidconsole
stdin=serial
stdout=serial,vidconsole
uuid_gpt_esp=c12a7328-f81f-11d2-ba4b-00a0c93ec93b
uuid_gpt_system=b921b045-1df0-41c3-af44-4c6f280d3fae

Environment size: 4678/65532 bytes
Detecting psci idle state
Working FDT set to f9f39ee0
PSCI idle state enabled
Loading DTB
53815 bytes read in 4 ms (12.8 MiB/s)
Working FDT set to 4fa00000
ram_freq=552
Adding FTD RAM clock
Applying PSCI DTBO
1130 bytes read in 2 ms (551.8 KiB/s)
Loading Initramfs
2348624 bytes read in 58 ms (38.6 MiB/s)
Loading Kernel
19021832 bytes read in 426 ms (42.6 MiB/s)
gpio: pin 115 (gpio 115) value is 1
Loading user script
Failed to load 'user.scr'
No user script found
Booting kernel
gpio: pin 116 (gpio 116) value is 1
gpio: pin 98 (gpio 98) value is 0
Moving Image from 0x40080000 to 0x40200000, end=414a0000
## Flattened Device Tree blob at 4fa00000
   Booting using the fdt blob at 0x4fa00000
Working FDT set to 4fa00000
   Loading Ramdisk to 49dc2000, end 49fff650 ... OK
   Loading Device Tree to 0000000049daf000, end 0000000049dc1fff ... OK
Working FDT set to 49daf000
Warning: HDMI PHY init timeout!

Starting kernel ...

[    0.000000] Booting Linux on physical CPU 0x0000000000 [0x410fd034]
[    0.000000] Linux version 6.1.9 (pmos@build) (aarch64-alpine-linux-musl-gcc (Alpine 12.2.1_git20220924-r9) 12.2.1 20220924, GNU ld (GNU Binutils) 2.40) #2-postmarketos-allwinner SMP PREEMPT Sun Feb 19 20:40:18 UTC 20
[    0.000000] Machine model: Pine64 PinePhone (1.2)
[    0.000000] earlycon: uart0 at MMIO32 0x0000000001c28000 (options '')
[    0.000000] printk: bootconsole [uart0] enabled
### postmarketOS initramfs ###
Configuring kernel firmware image search path
NOTE: Waiting 10 seconds for the framebuffer /dev/fb0.
If your device does not have a framebuffer, disable this with:
no_framebuffer=true in <https://postmarketos.org/deviceinfo>
Setting framebuffer mode to: U:720x1440p-0
720x1440 @ 68x136mm, dpi=268, logo_size_px=450.000000
Setup usb network
  /sys/class/android_usb does not exist, skipping android_usb
  Setting up an USB gadget through configfs
Starting unudhcpd
  Using interface usb0
  Starting the DHCP daemon
Trying to start server with parameters: Server IP addr: 172.16.42.1:67, client IP addr: 172.16.42.2, interface: usb0
Trying to bind to interface: usb0
Server started!
Mount boot partition (/dev/mmcblk2p1) to /boot (read-only)
Detected ext filesystem
Extract /boot/initramfs-extra
cpio: lib/ld-musl-aarch64.so.1 not created: newer or same age file exists
cpio: lib/libc.musl-aarch64.so.1 not created: newer or same age file exists
cpio: lib/libdevmapper.so.1.02 not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.alias not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.alias.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin.modinfo not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.dep not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.dep.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.order not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.symbols not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.symbols.bin not created: newer or same age file exists
34156 blocks
Received DHCP DISCOVER from client: b6:a1:3d:d4:49:94
Received DHCP REQUEST from client: b6:a1:3d:d4:49:94
unable to add device to libinput context: No such file or directory
720x1440 @ 68x136mm, dpi=268, logo_size_px=450.000000
Mount root partition (/dev/mapper/root) to /sysroot (read-only)
Detected ext4 filesystem
Mount boot partition (/dev/mmcblk2p1) to /sysroot/boot (read-write)
Detected ext filesystem
umount: can't unmount /dev: Invalid argument

   OpenRC 0.46 is starting up Linux 6.1.9 (aarch64)

/lib/rc/sh/init.sh: line 15: can't create /dev/null: Read-only file system
 * md5sum is missing, which suggests /usr is not mounted
 * If you have separate /usr, it must be mounted by initramfs
 * If not, you should check coreutils is installed correctly
 * Mounting /proc ... [ ok ]
 * Mounting /run ... * /run/openrc: creating directory
 * /run/lock: creating directory
 * /run/lock: correcting owner
/lib/rc/sh/init.sh: line 98: can't create /dev/null: Read-only file system
/lib/rc/sh/gendepends.sh: line 53: can't create /dev/null: Read-only file system
zram swap: activating with size: 1497 MB
 * Caching service dependencies ... [ ok ]
 * Mounting devtmpfs on /dev ... [ ok ]
 * Mounting /dev/mqueue ... [ ok ]
 * Mounting /dev/pts ... [ ok ]
 * Mounting /dev/shm ... [ ok ]
 * Mounting /sys ... [ ok ]
 * Mounting security filesystem ... [ ok ]
 * Mounting debug filesystem ... [ ok ]
 * Mounting config filesystem ... [ ok ]
 * Mounting fuse control filesystem ... [ ok ]
 * Starting udev ... [ ok ]
 * Generating a rule to create a /dev/root symlink ... [ ok ]
 * Populating /dev with existing devices through uevents ... [ ok ]
 * Waiting for uevents to be processed ... [ ok ]
 * Loading modules ... [ ok ]
 * Setting system clock using the hardware clock [UTC] ... [ ok ]
 * Checking local filesystems  ... [ ok ]
 * Remounting root filesystem read/write ... [ ok ]
 * Remounting filesystems ... [ ok ]
 * Mounting local filesystems ... [ ok ]
 * Configuring kernel parameters ...sysctl: error: 'kernel.unprivileged_bpf_disabled' is an unknown key
 [ ok ]
 * Creating user login records ... [ ok ]
 * Setting hostname ... [ ok ]
 * Starting busybox syslog ... [ ok ]
 * /run/dbus: creating directory
 * /run/dbus: correcting owner
 * Starting System Message Bus ... [ ok ]
 * Loading nftables rules and starting firewall ... [ ok ]
 * Starting haveged ... [ ok ]
 * /var/run/wpa_supplicant: creating directory
 * Starting WPA Supplicant ... [ ok ]
 * Starting networkmanager ... [ ok ]
 * Starting chronyd ... [ ok ]
 * Starting modemmanager ... [ ok ]
 * Starting eg25-manager ... [ ok ]
 * Bringing up network interface lo ...RTNETLINK answers: File exists
 [ ok ]
 * Starting gpsd ... [ ok ]
 * Starting HKDM ... [ ok ]
 * Starting iio-sensor-proxy ... [ ok ]
 * Starting kill-pbsplash ... [ ok ]
 * Starting postmarketos-tweakd ...Scanning /usr/share/postmarketos-tweaks
  Loading /usr/share/postmarketos-tweaks/unlocker.yml
  Loading /usr/share/postmarketos-tweaks/pinephone.yml
  Loading /usr/share/postmarketos-tweaks/fonts.yml
  Loading /usr/share/postmarketos-tweaks/phosh.yml
  Loading /usr/share/postmarketos-tweaks/about.yml
  Loading /usr/share/postmarketos-tweaks/appearance.yml
  Loading /usr/share/postmarketos-tweaks/sound.yml
  Loading /usr/share/postmarketos-tweaks/power.yml
Scanning /etc/postmarketos-tweaks
/sys/class/power_supply/axp20x-battery/voltage_max_design = 4200000
 [ ok ]
 * Restoring rfkill configuration ... [ ok ]
 * Starting sensord ... [ ok ]
 * Starting sshd ... [ ok ]
 * Activating swap file ...Configured swap file size is 0, skipping creation.
 [ ok ]
 * Starting System login manager ... [ ok ]
 * Starting tinydm ... [ ok ]
 * Starting wake daemon ... [ ok ]
 * Starting local ... [ ok ]

Welcome to postmarketOS
Kernel 6.1.9 on an aarch64 (/dev/ttyS0)
shigusa login:

      This MR's packages:

      linux-postmarketos-allwinner-6.1.9_git20230201-r3
u-boot-pinephone-2023.01-r1
device-pine64-pinephone-0.46-r1
device-pine64-pinephone-nonfree-firmware-0.46-r1
device-pine64-pinephone-upower-0.46-r1
device-pine64-pinephone-gnome-0.46-r1
device-pine64-pinephone-camera-0.46-r1
      Boot log 
      U-Boot SPL 2023.01 (Mar 12 2023 - 09:48:01 +0000)
DRAM: 3072 MiB
Trying to boot from MMC2
NOTICE:  BL31: v2.8(release):
NOTICE:  BL31: Built : 18:16:06, Jan 14 2023
NOTICE:  BL31: Detected Allwinner A64/H64/R18 SoC (1689)
NOTICE:  BL31: Found U-Boot DTB at 0x208fb48, model: Pine64 PinePhone (1.2)


U-Boot 2023.01 (Mar 12 2023 - 09:48:01 +0000) Allwinner Technology

CPU:   Allwinner A64 (SUN50I)
Model: Pine64 PinePhone (1.2)
DRAM:  3 GiB @ 552 MHz
Core:  73 devices, 18 uclasses, devicetree: separate
WDT:   Not starting watchdog@1c20ca0
MMC:   mmc@1c0f000: 0, mmc@1c10000: 1, mmc@1c11000: 2
Loading Environment from FAT... Unable to use mmc 2:1...
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
Warning: HDMI PHY init timeout!
In:    serial
Out:   serial
Err:   serial
Net:   No ethernet found.
Hit any key to stop autoboot:  0 
switch to partitions #0, OK
mmc2(part 0) is current device
Scanning mmc 2:1...
Found U-Boot script /boot.scr
1232 bytes read in 1 ms (1.2 MiB/s)
## Executing script at 4fc00000
gpio: pin 98 (gpio 98) value is 1
gpio: pin 114 (gpio 114) value is 1
Booting from eMMC
arch=arm
baudrate=115200
board=sunxi
board_name=sunxi
boot_a_script=load ${devtype} ${devnum}:${distro_bootpart} ${scriptaddr} ${prefix}${script}; source ${scriptaddr}
boot_efi_binary=load ${devtype} ${devnum}:${distro_bootpart} ${kernel_addr_r} efi/boot/bootaa64.efi; if fdt addr -q ${fdt_addr_r}; then bootefi ${kernel_addr_r} ${fdt_addr_r};else bootefi ${kernel_addr_r} ${fdtcontroladdr};fi
boot_efi_bootmgr=if fdt addr -q ${fdt_addr_r}; then bootefi bootmgr ${fdt_addr_r};else bootefi bootmgr;fi
boot_extlinux=sysboot ${devtype} ${devnum}:${distro_bootpart} any ${scriptaddr} ${prefix}${boot_syslinux_conf}
boot_prefixes=/ /boot/
boot_script_dhcp=boot.scr.uimg
boot_scripts=boot.scr.uimg boot.scr
boot_syslinux_conf=extlinux/extlinux.conf
boot_targets=fel mmc_auto pxe dhcp 
bootargs=init=/init.sh rw console=tty0 console=ttyS0,115200 earlycon=uart,mmio32,0x01c28000 panic=10 consoleblank=0 loglevel=1 PMOS_NO_OUTPUT_REDIRECT PMOS_FORCE_PARTITION_RESIZE pmos_boot=/dev/mmcblk2p1 pmos_root=/dev/mmcblk2p2
bootcmd=run distro_bootcmd
bootcmd_dhcp=devtype=dhcp; if dhcp ${scriptaddr} ${boot_script_dhcp}; then source ${scriptaddr}; fi;setenv efi_fdtfile ${fdtfile}; setenv efi_old_vci ${bootp_vci};setenv efi_old_arch ${bootp_arch};setenv bootp_vci PXEClient:Arch:00011:UNDI:003000;setenv bootp_arch 0xb;if d;
bootcmd_fel=if test -n ${fel_booted} && test -n ${fel_scriptaddr}; then echo '(FEL boot)'; source ${fel_scriptaddr}; fi
bootcmd_mmc0=devnum=0; run mmc_boot
bootcmd_mmc2=devnum=2; run mmc_boot
bootcmd_mmc_auto=if test ${mmc_bootdev} -eq 2; then run bootcmd_mmc2; run bootcmd_mmc0; elif test ${mmc_bootdev} -eq 0; then run bootcmd_mmc0; run bootcmd_mmc2; fi
bootcmd_pxe=dhcp; if pxe get; then pxe boot; fi
bootdelay=0
bootdev=2
bootfstype=ext4
bootm_size=0xa000000
console=ttyS0,115200
cpu=armv8
devplist=1
dfu_alt_info_ram=kernel ram 0x40080000 0x1000000;fdt ram 0x4FA00000 0x100000;ramdisk ram 0x4FF00000 0x4000000
distro_bootcmd=for target in ${boot_targets}; do run bootcmd_${target}; done
efi_dtb_prefixes=/ /dtb/ /dtb/current/
ethaddr=02:ba:0c:1b:a1:48
fdt_addr_r=0x4FA00000
fdtcontroladdr=f9f37440
fdtfile=allwinner/sun50i-a64-pinephone-1.2.dtb
fdtoverlay_addr_r=0x4FE00000
fileaddr=4fc00000
filesize=4d0
kernel_addr_r=0x40080000
kernel_comp_addr_r=0x44000000
kernel_comp_size=0xb000000
load_efi_dtb=load ${devtype} ${devnum}:${distro_bootpart} ${fdt_addr_r} ${prefix}${efi_fdtfile}
loadaddr=0x42000000
mmc_boot=if mmc dev ${devnum}; then devtype=mmc; run scan_dev_for_boot_part; fi
mmc_bootdev=2
partitions=name=loader1,start=8k,size=32k,uuid=${uuid_gpt_loader1};name=loader2,size=984k,uuid=${uuid_gpt_loader2};name=esp,size=128M,bootable,uuid=${uuid_gpt_esp};name=system,size=-,uuid=${uuid_gpt_system};
pxefile_addr_r=0x4FD00000
ram_freq=552
ramdisk_addr_r=0x4FF00000
scan_dev_for_boot=echo Scanning ${devtype} ${devnum}:${distro_bootpart}...; for prefix in ${boot_prefixes}; do run scan_dev_for_extlinux; run scan_dev_for_scripts; done;run scan_dev_for_efi;
scan_dev_for_boot_part=part list ${devtype} ${devnum} -bootable devplist; env exists devplist || setenv devplist 1; for distro_bootpart in ${devplist}; do if fstype ${devtype} ${devnum}:${distro_bootpart} bootfstype; then run scan_dev_for_boot; fi; done; setenv devplist
scan_dev_for_efi=setenv efi_fdtfile ${fdtfile}; for prefix in ${efi_dtb_prefixes}; do if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${efi_fdtfile}; then run load_efi_dtb; fi;done;run boot_efi_bootmgr;if test -e ${devtype} ${devnum}:${distro_bootpart} efi/boote
scan_dev_for_extlinux=if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${boot_syslinux_conf}; then echo Found ${prefix}${boot_syslinux_conf}; run boot_extlinux; echo EXTLINUX FAILED: continuing...; fi
scan_dev_for_scripts=for script in ${boot_scripts}; do if test -e ${devtype} ${devnum}:${distro_bootpart} ${prefix}${script}; then echo Found U-Boot script ${prefix}${script}; run boot_a_script; echo SCRIPT FAILED: continuing...; fi; done
scriptaddr=0x4FC00000
serial#=92c0a9ba0c1ba148
soc=sunxi
stderr=serial,vidconsole
stdin=serial
stdout=serial,vidconsole
uuid_gpt_esp=c12a7328-f81f-11d2-ba4b-00a0c93ec93b
uuid_gpt_system=b921b045-1df0-41c3-af44-4c6f280d3fae

Environment size: 4678/65532 bytes
Loading DTB
Working FDT set to f9f37440
Working FDT set to 4fa00000
ram_freq=552
Adding FTD RAM clock
Loading Initramfs
2348624 bytes read in 57 ms (39.3 MiB/s)
Loading Kernel
19021832 bytes read in 426 ms (42.6 MiB/s)
gpio: pin 115 (gpio 115) value is 1
Loading user script
Failed to load 'user.scr'
No user script found
Booting kernel
gpio: pin 116 (gpio 116) value is 1
gpio: pin 98 (gpio 98) value is 0
Moving Image from 0x40080000 to 0x40200000, end=414a0000
## Flattened Device Tree blob at 4fa00000
   Booting using the fdt blob at 0x4fa00000
Working FDT set to 4fa00000
   Loading Ramdisk to 49dc2000, end 49fff650 ... OK
   Loading Device Tree to 0000000049daf000, end 0000000049dc1fff ... OK
Working FDT set to 49daf000
Warning: HDMI PHY init timeout!

Starting kernel ...

[    0.000000] Booting Linux on physical CPU 0x0000000000 [0x410fd034]
[    0.000000] Linux version 6.1.9 (pmos@futaba) (aarch64-alpine-linux-musl-gcc (Alpine 12.2.1_git20220924-r9) 12.2.1 20220924, GNU ld (GNU Binutils) 2.40) #5-postmarketos-allwinner SMP PREEMPT Sun Mar 12 09:47:34 UTC 20
[    0.000000] Machine model: Pine64 PinePhone (1.2)
[    0.000000] earlycon: uart0 at MMIO32 0x0000000001c28000 (options '')
[    0.000000] printk: bootconsole [uart0] enabled
### postmarketOS initramfs ###
Configuring kernel firmware image search path
NOTE: Waiting 10 seconds for the framebuffer /dev/fb0.
If your device does not have a framebuffer, disable this with:
no_framebuffer=true in <https://postmarketos.org/deviceinfo>
Setting framebuffer mode to: U:720x1440p-0
720x1440 @ 68x136mm, dpi=268, logo_size_px=450.000000
Setup usb network
  /sys/class/android_usb does not exist, skipping android_usb
  Setting up an USB gadget through configfs
Starting unudhcpd
  Using interface usb0
  Starting the DHCP daemon
Trying to start server with parameters: Server IP addr: 172.16.42.1:67, client IP addr: 172.16.42.2, interface: usb0
Trying to bind to interface: usb0
Server started!
Mount boot partition (/dev/mmcblk2p1) to /boot (read-only)
Detected ext filesystem
Extract /boot/initramfs-extra
cpio: lib/ld-musl-aarch64.so.1 not created: newer or same age file exists
cpio: lib/libc.musl-aarch64.so.1 not created: newer or same age file exists
cpio: lib/libdevmapper.so.1.02 not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.alias not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.alias.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.builtin.modinfo not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.dep not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.dep.bin not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.order not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.symbols not created: newer or same age file exists
cpio: lib/modules/6.1.9/modules.symbols.bin not created: newer or same age file exists
34156 blocks
Received DHCP DISCOVER from client: a6:75:46:de:6c:df
Received DHCP REQUEST from client: a6:75:46:de:6c:df
unable to add device to libinput context: No such file or directory
720x1440 @ 68x136mm, dpi=268, logo_size_px=450.000000
Mount root partition (/dev/mapper/root) to /sysroot (read-only)
Detected ext4 filesystem
Mount boot partition (/dev/mmcblk2p1) to /sysroot/boot (read-write)
Detected ext filesystem
umount: can't unmount /dev: Invalid argument

   OpenRC 0.46 is starting up Linux 6.1.9 (aarch64)

/lib/rc/sh/init.sh: line 15: can't create /dev/null: Read-only file system
 * md5sum is missing, which suggests /usr is not mounted
 * If you have separate /usr, it must be mounted by initramfs
 * If not, you should check coreutils is installed correctly
 * Mounting /proc ... [ ok ]
 * Mounting /run ... * /run/openrc: creating directory
 * /run/lock: creating directory
 * /run/lock: correcting owner
/lib/rc/sh/init.sh: line 98: can't create /dev/null: Read-only file system
/lib/rc/sh/gendepends.sh: line 53: can't create /dev/null: Read-only file system
zram swap: activating with size: 1497 MB
 * Caching service dependencies ... [ ok ]
 * Mounting devtmpfs on /dev ... [ ok ]
 * Mounting /dev/mqueue ... [ ok ]
 * Mounting /dev/pts ... [ ok ]
 * Mounting /dev/shm ... [ ok ]
 * Mounting /sys ... [ ok ]
 * Mounting security filesystem ... [ ok ]
 * Mounting debug filesystem ... [ ok ]
 * Mounting config filesystem ... [ ok ]
 * Mounting fuse control filesystem ... [ ok ]
 * Starting udev ... [ ok ]
 * Generating a rule to create a /dev/root symlink ... [ ok ]
 * Populating /dev with existing devices through uevents ... [ ok ]
 * Waiting for uevents to be processed ... [ ok ]
 * Loading modules ... [ ok ]
 * Setting system clock using the hardware clock [UTC] ... [ ok ]
 * Checking local filesystems  ... [ ok ]
 * Remounting root filesystem read/write ... [ ok ]
 * Remounting filesystems ... [ ok ]
 * Mounting local filesystems ... [ ok ]
 * Configuring kernel parameters ...sysctl: error: 'kernel.unprivileged_bpf_disabled' is an unknown key
 [ ok ]
 * Creating user login records ... [ ok ]
 * Setting hostname ... [ ok ]
 * Starting busybox syslog ... [ ok ]
 * /run/dbus: creating directory
 * /run/dbus: correcting owner
 * Starting System Message Bus ... [ ok ]
 * Loading nftables rules and starting firewall ... [ ok ]
 * Starting haveged ... [ ok ]
 * /var/run/wpa_supplicant: creating directory
 * Starting WPA Supplicant ... [ ok ]
 * Starting networkmanager ... [ ok ]
 * Starting chronyd ... [ ok ]
 * Starting modemmanager ... [ ok ]
 * Starting eg25-manager ... [ ok ]
 * Bringing up network interface lo ...RTNETLINK answers: File exists
 [ ok ]
 * Starting gpsd ... [ ok ]
 * Starting HKDM ... [ ok ]
 * Starting iio-sensor-proxy ... [ ok ]
 * Starting kill-pbsplash ... [ ok ]
 * Starting postmarketos-tweakd ...Scanning /usr/share/postmarketos-tweaks
  Loading /usr/share/postmarketos-tweaks/unlocker.yml
  Loading /usr/share/postmarketos-tweaks/pinephone.yml
  Loading /usr/share/postmarketos-tweaks/fonts.yml
  Loading /usr/share/postmarketos-tweaks/phosh.yml
  Loading /usr/share/postmarketos-tweaks/about.yml
  Loading /usr/share/postmarketos-tweaks/appearance.yml
  Loading /usr/share/postmarketos-tweaks/sound.yml
  Loading /usr/share/postmarketos-tweaks/power.yml
Scanning /etc/postmarketos-tweaks
/sys/class/power_supply/axp20x-battery/voltage_max_design = 4200000
 [ ok ]
 * Restoring rfkill configuration ... [ ok ]
 * Starting sensord ... [ ok ]
 * Starting sshd ... [ ok ]
 * Activating swap file ...Configured swap file size is 0, skipping creation.
 [ ok ]
 * Starting System login manager ... [ ok ]
 * Starting tinydm ... [ ok ]
 * Starting wake daemon ... [ ok ]
 * Starting local ... [ ok ]

Welcome to postmarketOS
Kernel 6.1.9 on an aarch64 (/dev/ttyS0)
shigusa login:

      I use 552 for the RAM speed (I assume it must've been the default at some point). Default 528 also has the same effect. I never tried 624 like you have.

      Also, as soon as ssh and wlan0 are up, which happens while phosh and gnome-session are still spinning all cores at 100% as part of their startup, cpupower idle-info over ssh already reports >0 usage for the lower states. So difference in user activity cannot be the reason anyway.

      By Arnav Singh on 2023-03-12T20:17:48

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      Thanks Arnav for all the feedback. There's a lot of info in this thread and in the other MRs to digest but we'll get to the bottom of this.

      Can I ask you run one more tedious test:

      1. Make sure the SPI if formatted and holds nothing
      2. Flash the latest edge image available to the eMMC
      3. Shutdown the phone
      4. Remove the battery [Edit*] & mSD card
      5. Press the power button for a good 15/20 seconds
      6. Plug back the battery & the serial cable
      7. Record the serial data stream
      8. Power on the phone
      9. Save a copy of Serial log (up to login prompt is fine)
      10. Let the phone be until it goes to sleep
      11. Wake it up and capture the output of cpupower -c 0,1,2,3 idle-info and cpupower -c 0,1,2,3 frequency-info
      12. Change the governor to conservative: cpupower frequency-set -g conservative
      13. Let the phone be until it goes to sleep
      14. Wake it up and capture the output of cpupower -c 0,1,2,3 idle-info and cpupower -c 0,1,2,3 frequency-info

      At the end the following info should be gathered:

      1. Serial cable log (boot)
      2. cpupower capture
      3. dmesg output at the end of step # 14

      By Bobby The Builder on 2023-03-12T21:50:10

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      I'm probably not going to get to that any time soon. Maybe @PsychoGame can help.

      By Arnav Singh on 2023-03-12T22:11:35

    • Administrator
      Administrator @root · Imported
      Author Owner

      I did decompile DTB from U-Boot and from the kernel to compare them.

      For the record, here they are: kernel.dts uboot.dts

      Here's what's most likely to have an impact on your device (missing definitions in U-Boot DTB):

      sram@40000 node

      syscon

      Symbols

      scpi node

      Scpi

      Symbols2

      I did not find from where those changes do come from in the kernel sources yet...

      Can you check with Megi if it makes sense to him ? I still don't understand why I do not replicate the behavior on my devices (pmOS CE, KDE CE, Mobian CE && PP OG)...

      By Bobby The Builder on 2023-03-13T19:16:47

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      @BobbyTheBuilder,

      Are these changes in both u-Boot DTS's, I mean when booting from SD compared to eMMC.
      Strange thing is that the CPU-Idle gets initiated when booting from Sd.
      When you flash the exact same image to eMMC, CPU-Idle is broken, while it was working for exact the same image on SD.

      By Jan Jasper de Kroon on 2023-03-14T06:44:39

    • Administrator
      Administrator @root · Imported
      Author Owner

      Hello Jan,

      Are these changes in both u-Boot DTS's, I mean when booting from SD compared to eMMC.

      The diff is U-Boot DTS against Kernel DTS, they are different mainly because U-Boot does not need to have the whole hardware initialized.

      I used the same pmOS iso which was flashed to a SD card. I then booted on the SD card and verified that cpu-idle worked. It did. Then I flashed the same pmOS iso on the eMMC, shut down the phone, removed the SD card and let the phone boot. CPU idle did work when booted from the eMMC. Note: I never touched my SPI, it's empty.

      Strange thing is that the CPU-Idle gets initiated when booting from Sd.

      Actually, looking at the logs from Arnav's phone cpu-idle always get initialized whether booted from eMMC or mSD. The issue is that the CPU never enters the idle state when booted from the eMMC.

      When you flash the exact same image to eMMC, CPU-Idle is broken, while it was working for exact the same image on SD.

      Not on any of my devices, and I tested this on 4 different PP.

      Also we did not encounter this behavior when many people tested #1901 (closed) !3806

      I've went through the whole discussion and looked at Arnav's MR and what I understand is, in his situation, if U-Boot DTS is replaced by the one from the kernel then the CPU can enter idle states when booting from the eMMC.

      So, from this information I did decompile both DTB (from U-Boot and kernel) to compare them.

      The above post shows which nodes are the most likely to have an impact on the cpu-idle based on the diff.

      I'm just trying to narrow down the possibilities to help identify from where the issue comes from.

      Do you reliably reproduce the behavior on you PP ? if so can you give this a try first: #1946 (comment 1310647139)

      Maybe we can follow-up in IRC as well, I'm connected so ping me whenever you're available.

      By Bobby The Builder on 2023-03-14T15:50:15

    • Administrator
      Administrator @root · Imported
      Author Owner

      Strange thing is that the CPU-Idle gets initiated when booting from Sd.

      Actually, looking at the logs from Arnav's phone cpu-idle always get initialized whether booted from eMMC or mSD. The issue is that the CPU never enters the idle state when booted from the eMMC.

      To be clear, there are three known people who have the issue that, when booted from eMMC with default pmos packages, the idle states are recognized (due to the hard-coded overlay merged by boot.scr, so this is unsurprising) but never entered (cpupower idle-info and sysfs report 0 duration, which is this issue). Those are the person who initially reported this in #pinephone (norayr on Matrix, mentioned in the first comment in this thread), me and PsychoGame. luigi311 on Discord also reported the same thing except that they use Mobian, and I can't find the source of Mobian's u-boot package to see what it does.

      There are two people who have the experience that, when booted from SD card with a default pmos image, the idle states are recognized and entered. Those are me and PsychoGame. That is what PsychoGame is talking about in that excerpt. See https://gitlab.postmarketos.org/postmarketos/pmaports/-/merge_requests/3931#note_1305793721 and the next three comments.

      And lastly, there is one person for whom !3931 's approach of using kernel DTS in u-boot fixes the problem: me, and there is one person for whom that MR makes the phone not boot: PsychoGame (I need boot logs to know why).

      Not on any of my devices, and I tested this on 4 different PP.

      To be precise, I have a 1.2b. Is that what you referred to as "PP OG" ? I've only heard "OG" used to refer to all models of the PinePhone to distinguish it from the PinePhone Pro.

      By Arnav Singh on 2023-03-14T16:28:56

    • Administrator
      Administrator @root · Imported
      Author Owner

      Hello Arnav,

      The PP OG is the original one which was sold without a Linux OS flavor branding, so late in the chain production.

      To be precise, I have a 1.2b.

      Interesting... Looking at your boot log it's not loading the sun50i-a64-pinephone-1.2b.dtb either from eMMC or mSD

      This is from your boot log:

      fdtfile=allwinner/sun50i-a64-pinephone-1.2.dtb

      And here's where the ftd file is loaded: https://gitlab.postmarketos.org/postmarketos/pmaports/-/blob/master/device/main/device-pine64-pinephone/uboot-script.cmd#L27

      I'm still curious to know if removing the battery to drain capacitors makes any difference...

      Anyways, your MR indicates that a change in the DTB is needed to resolve the situation on your device, at least, and I'm curious to narrow down which one it is. Then we'll be able to dig further on the problem. Once it's fully understood we'll be able to decide what's the best way to handle it.

      Please invite any other users to participate in this thread.

      By Bobby The Builder on 2023-03-14T16:59:10

      Edited by Administrator
    • Administrator
      Administrator @root · Imported
      Author Owner

      Interesting... Looking at your boot log it's not loading the sun50i-a64-pinephone-1.2b.dtb either from eMMC or mSD

      #1945

      1.2b.dtb is a new thing in megi's 6.1 kernel. Neither u-boot nor device-pine64-pinephone know about it.

      By Arnav Singh on 2023-03-14T17:01:09

    • Administrator
      Administrator @root · Imported
      Author Owner

      1.2b.dtb is a new thing in megi's 6.1 kernel. Neither u-boot nor device-pine64-pinephone know about it.

      Ahhh nice, thanks for the info!

      Okay so let's focus on identifying the difference between U-Boot & kernel DTS that makes cpu idle work on your device.

      By Bobby The Builder on 2023-03-14T17:12:12

    • Administrator
      Administrator @root · Imported
      Author Owner

      It might be easier to investigate why the current pmaports situation (upstream u-boot DTS in u-boot + FDT discarded and replaced with full DTS by boot.scr) works when booting from SD card but not when booting from eMMC. ("works" == "idle states are entered", as mentioned in #1946 (comment 1312627115) )

      By Arnav Singh on 2023-03-14T17:41:12

      • Administrator
        Administrator @root · Imported
        Author Owner

        I'm experiencing the same issue. I have OG PinePhone v1.2b, Tow-Boot and pmOS on eMMC.

        By Matteo on 2023-04-01T23:27:33

        Edited by Ghost User
      • Administrator
        Administrator @root · Imported
        Author Owner

        Can you try !3931 ?

        By Arnav Singh on 2023-04-01T23:27:33

      • Administrator
        Administrator @root · Imported
        Author Owner

        How would I go to try it out, can you point me to a wiki or something? I'm using stable v22.12. So in this case I guess I should first install edge and then try your changes?

        By Matteo on 2023-04-16T23:07:41

      • Administrator
        Administrator @root · Imported
        Author Owner

        Yes.

        1. Install Edge to eMMC.

        2. Confirm that you still have the issue. If you don't have the issue stop here.

        3. Uninstall Tow-Boot. Reboot back into pmOS.

        4. Confirm that you still have the issue. If you don't have the issue stop here.

        5. Run apk list | grep device-pine64-pinephone and make a note of all the packages in the output.

        6. Run mrtest add 3931 and choose the u-boot-pinephone and device-pine64-pinephone packages, plus any other packages that showed up in the output of step (5).

        7. Run update-u-boot.

        8. Reboot.

        9. Confirm that you no longer have the issue.

        If you make your phone unbootable at any point, you can recover using one of two methods:

        Method 1: Using mass storage mode plus another computer.

        1. Install Tow-Boot.

        2. Reboot and put the phone in mass storage mode.

        3. Mount the root partition (cryptsetup open /dev/... phone-root; mount /dev/mapper/phone-root /mnt

        4. chroot into the root partition.

        5. Run mrtest zap to uninstall the mrtest dummy package, and apk upgrade -Ua to restore the u-boot-pinephone and device-pine64-pinephone packages back to pmaports versions.

        6. Exit the chroot, umount /mnt; cryptsetup close phone-root, and reboot the phone.

        7. Phone should boot fine now.

        8. Run update-u-boot. This will restore the pmaports u-boot.

        9. At this point you can uninstall Tow-Boot if you want.

        Method 2: Using an SD card.

        1. Make an SD card with pmOS Edge.

        2. Boot from the SD card.

        3. Do steps (3), (4) and (5) as above.

        4. Exit the chroot, umount /mnt; cryptsetup close phone-root.

        5. Run update-u-boot, but make sure that update-u-boot flashes to the eMMC, not the SD card.

        6. Shut down the phone, remove the SD card and reboot the phone. It should boot fine now.

        7. At this point you can uninstall Tow-Boot if you want.

        By Arnav Singh on 2023-04-16T23:54:25

        Edited by Administrator
      • Administrator
        Administrator @root · Imported
        Author Owner

        How do I uninstall Tow-Boot without wiping the entire internal eMMC?

        By Matteo on 2023-04-17T19:00:57

      • Administrator
        Administrator @root · Imported
        Author Owner

        Uninstalling Tow-Boot doesn't wipe the entire eMMC. The option in the uninstaller that you're referring to wipes just the partition that Tow-Boot installed itself to (mmcblk#boot0).

        By Arnav Singh on 2023-04-17T19:18:31

      • Administrator
        Administrator @root · Imported
        Author Owner

        I don't have any Tow-Boot uninstaller, I looked everywhere in PostmarketOS and Tow-Boot docs and nothing. Tow-Boot installer has just the options of installing or erasing the eMMC content which I want to avoid to don't install pmOS edge again. Can I just remove mmcblk#boot0 partition manually? Also what's the other mmcblk#boot1 ro partition that I see in the internal eMMC?

        By Matteo on 2023-04-17T19:29:41

        Edited by Administrator
      • Administrator
        Administrator @root · Imported
        Author Owner

        Tow-Boot installer has just the options of installing or erasing the eMMC content which I want to avoid to don't install pmOS edge again.

        That is the option I mentioned in my comment.

        By Arnav Singh on 2023-04-17T19:36:43

      • Administrator
        Administrator @root · Imported
        Author Owner

        Thanks Arnav, I haven't read it with attention before. The erase eMMC option in Tow-Boot installed refers only to the Boot portion. I stopped at step 3. since I got attached output from cpupower idle-info command. idle_info.txt Please let me know if you want me to test something else. Happy to help.

        By Matteo on 2023-04-17T20:05:26

      • Administrator
        Administrator @root · Imported
        Author Owner

        So just to be clear, that cpupower idle-info was run on a new Edge installation to eMMC, right?

        If so, make sure it continues to work after a reboot. And if it still works, you can gradually add whatever packages / customizations you originally had and rebooting after each one, and if any of those make the idle states stop reporting, that will tell us what breaks it.

        By Arnav Singh on 2023-04-17T20:18:10

      • Administrator
        Administrator @root · Imported
        Author Owner

        Yes exactly, new Edge installation to eMMC (no SD card inserted), no additional packages installed, only cpupower and erased Tow-Boot. I just restarted and re-tested and cpupower idle-info still reports idle states (increased usage). I'll continue testing adding packages and let you know how it goes.

        By Matteo on 2023-04-17T21:34:52

      • Administrator
        Administrator @root · Imported
        Author Owner

        One more thing - do you have FDE or not?

        By Arnav Singh on 2023-04-17T21:38:05

      • Administrator
        Administrator @root · Imported
        Author Owner

        Yes FDE and Phosh as de. If when removing Tow-Boot I got the cpu idle to work, isn't that one the culprit?

        By Matteo on 2023-04-17T21:41:13

      • Administrator
        Administrator @root · Imported
        Author Owner

        No, removing Tow-Boot didn't fix the issue for me, otherwise I wouldn't have needed to go so far as making a custom u-boot in the first place.

        By Arnav Singh on 2023-04-17T21:43:30

      • Administrator
        Administrator @root · Imported
        Author Owner

        I found out that when calling cpupower idle-info, I get a random CPU usage. So I need to use the option --cpu to select a CPU like: cpupower --cpu 0 idle-info. Is it normal that if I keep calling cpupower --cpu 0 idle-info from the same terminal session, repeatedly, I get Usage and Duration values for all 3 states (WFI, cpu-sleep and cluster-sleep) increased?

        By Matteo on 2023-04-17T22:35:39

      • Administrator
        Administrator @root · Imported
        Author Owner

        Is it normal that if I keep calling cpupower --cpu 0 idle-info from the same terminal session, repeatedly, I get Usage and Duration values for all 3 states (WFI, cpu-sleep and cluster-sleep) increased?

        Yes.

        By Arnav Singh on 2023-04-17T22:45:02

        Edited by Ghost User
      • Administrator
        Administrator @root · Imported
        Author Owner

        I installed more packages but nothing changed. So I installed Tow-Boot again and the issue came back, then I erased it and the issue disappeared. Also I noticed that when inserting a newly created pmOS v22.12 SD card, it will boot without the need to press volume down when having Tow-Boot installed. Is this because pmOS still creates images with uboot so Tow-Boot is ignored in that case? My personal conclusion is that this issue occurs for OG PP v1.2b, pmOS Edge and v22.12 stable on eMMC with Tow-Boot.

        By Matteo on 2023-04-18T23:46:27

      • Please register or sign in to reply
    • Administrator added device-pine64-pinephone release-edge release-v22.12 labels · Imported

      added device-pine64-pinephone release-edge release-v22.12 labels

      • Administrator
        Administrator @root · Imported
        Author Owner

        Added edge and v22.12 labels. Pretty sure this affects v22.12 as well, given that we backported what was supposed to fix broken cpuidle from edge to stable in SP 1.

        By Oliver Smith on 2023-04-16T23:08:21

        Edited by Ghost User
      • Administrator
        Administrator @root · Imported
        Author Owner

        Yep, I'm using v22.12 and I have this issue.

        By Matteo on 2023-04-16T23:08:21

      • Administrator
        Administrator @root · Imported
        Author Owner

        I'm on v22.12 (plasma-mobile on emmc, Pinephone 1.2a) and I can't reproduce this issue (ie cpupower idle-info reports usage values for sleep states).

        cpupower idle-info
CPUidle driver: psci_idle
CPUidle governor: menu
analyzing CPU 0:

Number of idle states: 3
Available idle states: WFI cpu-sleep cluster-sleep
WFI:
Flags/Description: ARM WFI
Latency: 1
Usage: 388119
Duration: 370604041
cpu-sleep:
Flags/Description: cpu-sleep
Latency: 2300
Usage: 39218
Duration: 364696000
cluster-sleep:
Flags/Description: cluster-sleep
Latency: 2350
Usage: 17008
Duration: 160999341

        I never installed tow-boot on the system or did anything u-boot specific, just the occasional apk upgrade and update-u-boot (though it is a self-built image, from v21.06 iirc)

        I do have a power related issue, but its probably caused by something different considering my cpupower output -> #2014

        By Zapeth on 2023-04-26T15:37:02

        Edited by Administrator
      • Please register or sign in to reply
    • Administrator mentioned in issue #1901 (closed) · Imported

      mentioned in issue #1901 (closed)

      By Oliver Smith on 2023-04-02T17:07:20

      • Administrator
        Administrator @root · Imported
        Author Owner

        While working on a fix for another issue, I went back from using the packages from !3931 to default packages of linux-postmarketos-allwinner, u-boot-pinephone and device-pine64-pinephone, and installed that default u-boot via update-u-boot. For whatever reason this issue does not repro for me any more, ie the idle states are now reporting >0 usage with default packages from pmaports. To be precise, this is with:

        • device-pine64-pinephone-0.46-r0

        • linux-postmarketos-allwinner-6.3.0_git20230426-r0

        • u-boot-pinephone-2023.01-r0

        From the time ~6 weeks ago when I did have the issue with default packages, the change since then is only the kernel package, since boot.scr and u-boot have not changed since then.

        By Arnav Singh on 2023-05-01T00:47:54

        Edited by Ghost User
      • Administrator
        Administrator @root · Imported
        Author Owner

        Have you tried to install tow-boot? I have the issue as soon as I install it and goes away when I remove it.

        By Matteo on 2023-05-01T00:48:36

        Edited by Administrator
      • Administrator
        Administrator @root · Imported
        Author Owner

        #1946 (comment 1356101198)

        By Arnav Singh on 2023-05-01T01:10:04

      • Please register or sign in to reply
      • Administrator
        Administrator @root · Imported
        Author Owner

        The issue for me seems to be resolved
        It seemed that some remnants of Tow-Boot werd stille installed.
        After erasing the Tow-Boot from eMMC again using the installer image, the cpu-idle works again for me in pmOS on eMMC.
        To start a small discussion as well, I think we should actively start to discontinue support for Tow-Boot in PP and also PPP.
        The package doesn't seem to be actively maintained anymore and laggs behind on upstream uBoot (which incorporates important fixes for both platforms in the meantime)

        By Jan Jasper de Kroon on 2023-06-23T10:44:37

        Edited by Ghost User
      • Administrator
        Administrator @root · Imported
        Author Owner

        I think we should actively start to discontinue support for Tow-Boot in PP and also PPP.

        No, not going to happen. Tow-Boot has maintenance problems yes we agree there, but we're currently actively talking with upstream on resolving that. Tow-Boot solves some actual problems and makes booting a whole lot easier and at least I personally have no intention to revert that.

        I rather have us spend time on fixing cpu-idle with Tow-Boot than dropping it.

        By Bart Ribbers on 2023-06-23T12:33:26

        Edited by Ghost User
      • Please register or sign in to reply
    • Administrator
      Administrator @root · Imported
      Author Owner

      With the 2022.07-006 release that was just cut a few hours ago, the two lower idle states are still broken. Actually there was some weirdness because on the first boot after install the modem and USB-power were also broken, but on reboot they magically fixed themselves, though wifi remained broken. In any case idle counters for the two lower states were still 0.

      By Arnav Singh on 2023-07-06T08:18:14

    • Administrator mentioned in merge request !3068 (closed) · Imported

      mentioned in merge request !3068 (closed)

      By Arnav Singh on 2023-07-07T23:51:33

    • Administrator added release-v23.06 label · Imported

      added release-v23.06 label

    • Administrator removed release-v22.12 label · Imported

      removed release-v22.12 label

    • Administrator added status::confirmed label · Imported

      added status::confirmed label

    • Administrator added device-category::main label · Imported

      added device-category::main label

    • Administrator
      Administrator @root · Imported
      Author Owner

      I'm closing this now, given that it isn't a problem anymore when using the stock kernel / recommended installation methods.

      I've updated the wiki, moved the warning from "installation" section (from the time where this was a problem for all installs) to troubleshooting (now that it is only a problem when tow-boot was installed apparently): https://wiki.postmarketos.org/wiki/PINE64_PinePhone_(pine64-pinephone)#CPU_idle_does_not_work

      Regarding whether to switch the PP to Tow-Boot, or to switch the PPP back from Tow-Boot, I'd say that is something the device maintainers and other people that actually test and do the work on these devices should decide.

      Thanks everybody for the updates to this issue! :tada:

      By Oliver Smith on 2023-12-13T16:59:42

    • Administrator closed · Imported

      closed

    • Administrator mentioned in issue #2635 · Imported

      mentioned in issue #2635

      By Arnav Singh on 2024-03-05T02:22:36

    • Administrator
      Administrator @root · Imported
      Author Owner

      https://github.com/Tow-Boot/Tow-Boot/releases/tag/release-2023.07-007 has the fix.

      By Arnav Singh on 2024-04-19T03:33:31

    Please register or sign in to reply