A true mainline Linux Kernel for the Pinebook Pro

[spikerguy]

@tsys

is usbc charging working in manjaro? i am receiving this error after compiling your 5.7 branch whien plugging in power, and it is not charging.

[   44.562406] cdn-dp fec00000.dp: [drm:cdn_dp_pd_event_work [rockchipdrm]] Not connected. Disabling cdn

Yes USB-C Charging is working on Manjaro. Are you on Manjaro arm-stable branch ?

[tsys]

is usbc charging working in manjaro? i am receiving this error after compiling your 5.7 branch whien plugging in power, and it is not charging.

I've noticed something strange going on with USB PD, too. It seems there is a bug in the state machine that causes the port to change between power source and sink role at a pretty high rate. I'll investigate this further and see if I can trace down what exactly broke.

[tsys]

@tsys

is usbc charging working in manjaro? i am receiving this error after compiling your 5.7 branch whien plugging in power, and it is not charging.

[   44.562406] cdn-dp fec00000.dp: [drm:cdn_dp_pd_event_work [rockchipdrm]] Not connected. Disabling cdn

Just fixed this in the latest commit on the 5.7 branch. Seems chargers are actually putting the USB-C sink into device role. Thanks for reporting.

[xmixahlx]

fix is tested wokring. thanks, @tsys !!!

i'm always on the mainline stable/rc + hwaccel patches, as y'all are probably aware.

... i did quick compile a clean 5.7 branch prior to reporting the issue, btw.

[as400]

I can confirm - usb-c charging works with your fix @tsys .

[timmes]

s3 also doesn't seem to work with the presence of an NVMe device.

I'm having the exact same issue as other users are describing, with a 256GB WD SN520 NVMe SSD. Appears to go to sleep, then comes straight back.

dmesg output:

Code:

[Jun 9 23:51] vbus_5vout: disabling
[  +0.000020] LDO_REG1: disabling
[  +0.210595] PM: suspend entry (s2idle)
[  +0.004079] Filesystems sync: 0.004 seconds
[  +0.001594] dwmmc_rockchip fe310000.mmc: pre_suspend failed for non-removable host: -38
[  +0.000017] Freezing user space processes ... (elapsed 0.002 seconds) done.
[  +0.002241] OOM killer disabled.
[  +0.000003] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[  +0.001508] printk: Suspending console(s) (use no_console_suspend to debug)
[  +0.185511] busb rockchip_usb2phy_exit
[  +0.355728] busb rockchip_usb2phy_exit
[  +0.000211] busb rockchip_usb2phy_exit
[  +0.040727] busb rockchip_usb2phy_exit
[  +5.487775] rockchip-pcie f8000000.pcie: PCIe link enter L2 timeout!
[  +0.000021] PM: dpm_run_callback(): rockchip_pcie_suspend_noirq+0x0/0x138 returns -110
[  +0.000005] PM: Device f8000000.pcie failed to suspend noirq: error -110
[  +0.030373] PM: noirq suspend of devices failed
[  +0.233519] busb rport->port_id=1
[  +0.000020] busb port power on
[  +0.001395] busb rport->port_id=1
[  +0.000010] busb port power on
[  +0.000043] busb rport->port_id=0
[  +0.000013] busb port power on
[  +0.006113] usb usb5: root hub lost power or was reset
[  +0.000007] usb usb6: root hub lost power or was reset
[  +0.089414] busb rport->port_id=0
[  +0.000020] busb port power on
[  +0.004017] usb usb7: root hub lost power or was reset
[  +0.000008] usb usb8: root hub lost power or was reset
[  +0.170668] cdn-dp fec00000.dp: [drm:cdn_dp_pd_event_work [rockchipdrm]] Not connected. Disabling cdn
[  +0.068467] nvme nvme0: 6/0/0 default/read/poll queues
[  +0.097397] OOM killer enabled.
[  +0.000005] Restarting tasks ... done.
[  +0.005636] PM: suspend exit

I think

Code:

rockchip-pcie f8000000.pcie: PCIe link enter L2 timeout!

Is the critical part here, especially as it waits a couple of seconds before this happens. lspci -vvv reports ASPM L2 is not supported. That could be a crucial difference with the NVMe SSD's that do not have this problem? If anyone who has this working could post their lspci results that would be great. Either way this SSD suspends just fine in a regular laptop.

Enabling ASPM power management works fine by the way (and appears to actually save ~250mW, but it is kind of hard to measure), but it doesn't make it work either way. Haven't tried forcing it off with the kernel parameter yet.

Running a stock 5.6.0-2-MANJARO-ARM kernel.

I'm more than willing to try out some suggestions. The Pinebook Pro is a great laptop especially because it is so easy to charge from existing 5V infrastructure (like USB sockets in public transport), but having to shutdown and restart every time is a major PITA. I'll have to remove the SSD if this cannot be fixed.

[xmixahlx]

hi @tsys and cc: @spikerguy

i have the following build error with 5.8-rc1 branch:

CC drivers/soc/rockchip/rockchip_pm_config.o
drivers/soc/rockchip/rockchip_pm_config.c: In function ‘rockchip_pm_virt_pwroff_prepare’:
drivers/soc/rockchip/rockchip_pm_config.c:102:10: error: implicit declaration of function ‘disable_nonboot_cpus’ [-Werror=implicit-function-declaration]
102 | error = disable_nonboot_cpus();
| ^~~~~~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
make[6]: *** [scripts/Makefile.build:281: drivers/soc/rockchip/rockchip_pm_config.o] Error 1
make[5]: *** [scripts/Makefile.build:497: drivers/soc/rockchip] Error 2
make[4]: *** [scripts/Makefile.build:497: drivers/soc] Error 2
make[4]: *** Waiting for unfinished jobs....
make[3]: *** [Makefile:1764: drivers] Error 2

[tsys]

hi @tsys and cc: @spikerguy

i have the following build error with 5.8-rc1 branch:

Thanks, I had not tested the branch myself yet. Fixed on latest v5.8-rc1

[xmixahlx]

@tsys compiles fine and working (well?) so far. FYI i have numerous debug messages similar to: (using PREEMPT and DEBUG_PREEMPT)

[ 87.534574] BUG: using smp_processor_id() in preemptible [00000000] code: kworker/u12:1/89
[ 87.534588] caller is debug_smp_processor_id+0x20/0x30
[ 87.534595] CPU: 3 PID: 89 Comm: kworker/u12:1 Not tainted 5.8.0-rc1-0.39-1-pinebookpro-hwaccel+ #1
[ 87.534597] Hardware name: Pine64 Pinebook Pro (DT)
[ 87.534605] Workqueue: writeback wb_workfn (flush-259:0)
[ 87.534612] Call trace:
[ 87.534618] dump_backtrace+0x0/0x1e0
[ 87.534622] show_stack+0x18/0x30
[ 87.534631] dump_stack+0xbc/0x114
[ 87.534635] check_preemption_disabled+0xf8/0x110
[ 87.534638] debug_smp_processor_id+0x20/0x30
[ 87.534644] ext4_mb_new_blocks+0x1b0/0x4a0
[ 87.534651] ext4_ext_map_blocks+0x618/0xce0
[ 87.534654] ext4_map_blocks+0xec/0x580
[ 87.534659] ext4_writepages+0x598/0xa80
[ 87.534666] do_writepages+0x54/0x110
[ 87.534673] __writeback_single_inode+0x3c/0x1d0
[ 87.534678] writeback_sb_inodes+0x1f8/0x3e0
[ 87.534682] __writeback_inodes_wb+0x4c/0xf0
[ 87.534686] wb_writeback+0x1dc/0x1f0
[ 87.534690] wb_workfn+0x30c/0x350
[ 87.534694] process_one_work+0x190/0x320
[ 87.534698] worker_thread+0x4c/0x400
[ 87.534704] kthread+0x104/0x140
[ 87.534709] ret_from_fork+0x10/0x24

[schanzen]

s3 also doesn't seem to work with the presence of an NVMe device.

I'm having the exact same issue as other users are describing, with a 256GB WD SN520 NVMe SSD. Appears to go to sleep, then comes straight back.

dmesg output:

Code:

[Jun 9 23:51] vbus_5vout: disabling
[  +0.000020] LDO_REG1: disabling
[  +0.210595] PM: suspend entry (s2idle)
[  +0.004079] Filesystems sync: 0.004 seconds
[  +0.001594] dwmmc_rockchip fe310000.mmc: pre_suspend failed for non-removable host: -38
[  +0.000017] Freezing user space processes ... (elapsed 0.002 seconds) done.
[  +0.002241] OOM killer disabled.
[  +0.000003] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[  +0.001508] printk: Suspending console(s) (use no_console_suspend to debug)
[  +0.185511] busb rockchip_usb2phy_exit
[  +0.355728] busb rockchip_usb2phy_exit
[  +0.000211] busb rockchip_usb2phy_exit
[  +0.040727] busb rockchip_usb2phy_exit
[  +5.487775] rockchip-pcie f8000000.pcie: PCIe link enter L2 timeout!
[  +0.000021] PM: dpm_run_callback(): rockchip_pcie_suspend_noirq+0x0/0x138 returns -110
[  +0.000005] PM: Device f8000000.pcie failed to suspend noirq: error -110
[  +0.030373] PM: noirq suspend of devices failed
[  +0.233519] busb rport->port_id=1
[  +0.000020] busb port power on
[  +0.001395] busb rport->port_id=1
[  +0.000010] busb port power on
[  +0.000043] busb rport->port_id=0
[  +0.000013] busb port power on
[  +0.006113] usb usb5: root hub lost power or was reset
[  +0.000007] usb usb6: root hub lost power or was reset
[  +0.089414] busb rport->port_id=0
[  +0.000020] busb port power on
[  +0.004017] usb usb7: root hub lost power or was reset
[  +0.000008] usb usb8: root hub lost power or was reset
[  +0.170668] cdn-dp fec00000.dp: [drm:cdn_dp_pd_event_work [rockchipdrm]] Not connected. Disabling cdn
[  +0.068467] nvme nvme0: 6/0/0 default/read/poll queues
[  +0.097397] OOM killer enabled.
[  +0.000005] Restarting tasks ... done.
[  +0.005636] PM: suspend exit

I think

Code:

rockchip-pcie f8000000.pcie: PCIe link enter L2 timeout!

Is the critical part here, especially as it waits a couple of seconds before this happens. lspci -vvv reports ASPM L2 is not supported. That could be a crucial difference with the NVMe SSD's that do not have this problem? If anyone who has this working could post their lspci results that would be great. Either way this SSD suspends just fine in a regular laptop.

Enabling ASPM power management works fine by the way (and appears to actually save ~250mW, but it is kind of hard to measure), but it doesn't make it work either way. Haven't tried forcing it off with the kernel parameter yet.

Running a stock 5.6.0-2-MANJARO-ARM kernel.

I'm more than willing to try out some suggestions. The Pinebook Pro is a great laptop especially because it is so easy to charge from existing 5V infrastructure (like USB sockets in public transport), but having to shutdown and restart every time is a major PITA. I'll have to remove the SSD if this cannot be fixed.

I can confirm that I have the same issue with my Intel Intel 660p M.2 1TB. Also no L2 capability reported in lspci.
For reference, this is a know problem [1]. I guess this needs to be resolved in the kernel at some point and and is no configuration issue.

[1] https://forum.manjaro.org/t/pinebook-pro.../145159/13