Skip to contentSkip to conversation

PROBLEM: mmc_select_hs400es failed, error -110

DISCUSSION Milan P. Stanić 5 participants 14 messages Mar 18, 2020 – Aug 13, 2021 1 message not in the archive

Subsystems (2)
lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

#1 lore raw

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Signature · 1 line
TIA
Quoted 66 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Thank in advance

--
Kind regards

[...]

#2 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi Milan

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

Quoted 22 lines · Milan P. Stanić

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 56 lines
Quoted 53 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Thank in advance

--
Kind regards

[...]

#3 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi Shawn,

Quoted 25 lines · Shawn Lin

Hi Milan

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

Quoted 19 lines · Milan P. Stanić

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 10 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

Quoted 6 lines
Quoted 3 lines
Quoted 3 lines

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Sorry if I was not clear in previous mail. suspend-to-ram and resume works fine if the machine doesn't mount internal emmc card to use it in way.

If it boots from external mmc card or ssd disk attached over usb-c interface and use root filesystem (or any other filesystem, /home for example) on them (external mmc or disk on usb-c) it can go numerous times to suspend-resume cycles without any problem.

But if any filesystem from emmc mounted then in can suspend and even resume, but when resumed after some random time it start to make
problems: programs segfault, cant read or write files and similar issue.

As I wrote in previous mail I can apply patches, build kernel and (maybe) tweak some parameters in kernel source (if I know which one should help me) to help subsystem developers to fix this problem.

Signature · 1 line
thanks
Quoted 57 lines
Quoted 55 lines
Quoted 52 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Thank in advance

--
Kind regards

[...]

#4 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

In reply to Shawn Lin

Hi Shawn,

Quoted 4 lines · Shawn Lin

Hi Milan

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

Robin.

Quoted 90 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 48 lines
Quoted 48 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr:  0x00000008 | Version:  0x00001002
mmc1: sdhci: Blk size:  0x00007200 | Blk cnt:  0x00000008
mmc1: sdhci: Argument:  0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present:   0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power:     0x0000000b | Blk gap:  0x00000080
mmc1: sdhci: Wake-up:   0x00000000 | Clock:    0x0000bc07
mmc1: sdhci: Timeout:   0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab:  0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps:      0x44edc880 | Caps_1:   0x801020f7
mmc1: sdhci: Cmd:       0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]:   0x00000000 | Resp[1]:  0x373300bd
mmc1: sdhci: Resp[2]:   0x35303030 | Resp[3]:  0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and  output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

Signature · 1 line
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#5 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr:  0x00000008 | Version:  0x00001002
mmc1: sdhci: Blk size:  0x00007200 | Blk cnt:  0x00000008
mmc1: sdhci: Argument:  0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present:   0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power:     0x0000000b | Blk gap:  0x00000080
mmc1: sdhci: Wake-up:   0x00000000 | Clock:    0x0000bc07
mmc1: sdhci: Timeout:   0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab:  0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps:      0x44edc880 | Caps_1:   0x801020f7
mmc1: sdhci: Cmd:       0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]:   0x00000000 | Resp[1]:  0x373300bd
mmc1: sdhci: Resp[2]:   0x35303030 | Resp[3]:  0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and  output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

Signature · 1 line
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#6 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi,

Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Quoted 119 lines · Milan P. Stanić

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr:  0x00000008 | Version:  0x00001002
mmc1: sdhci: Blk size:  0x00007200 | Blk cnt:  0x00000008
mmc1: sdhci: Argument:  0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present:   0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power:     0x0000000b | Blk gap:  0x00000080
mmc1: sdhci: Wake-up:   0x00000000 | Clock:    0x0000bc07
mmc1: sdhci: Timeout:   0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab:  0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps:      0x44edc880 | Caps_1:   0x801020f7
mmc1: sdhci: Cmd:       0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]:   0x00000000 | Resp[1]:  0x373300bd
mmc1: sdhci: Resp[2]:   0x35303030 | Resp[3]:  0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and  output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

Signature · 1 line
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#7 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Quoted 7 lines · Milan P. Stanić

Hi,

Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 3 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#8 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

Signature · 1 line
Kind regards
Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#9 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi all,

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

Signature · 1 line
Regards
Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

Signature · 1 line
Kind regards
Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#10 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Quoted 11 lines · Milan P. Stanić

Hi all,

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

That delta is more than 4 years of development. I am not sure investigating the delta would make sense, however I leave that to Rock-chips people to decide - or whomever that are willing to help to debug this.

Kind regards
Uffe

Quoted 174 lines

--
Regards

Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

--
Kind regards

Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#11 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Quoted 10 lines · Ulf Hansson
Quoted 8 lines · Milan P. Stanić

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

Quoted 4 lines

That delta is more than 4 years of development. I am not sure investigating the delta would make sense, however I leave that to Rock-chips people to decide - or whomever that are willing to help to debug this.

I think it is not needed to look at full delta from this working chromeOS kernel and current mainline but maybe to 5.0 or 5.1 releases.

But I'm not sure and because I don't know what to look I can only hope that rockchip developers will find what is wrong.

Anyway, I use this rk3399 gru-kevin chromebook with external mmc card and it works fine if I don't 'touch' emmc. But is slower than when emmc is used for root and home filesystem, significantly, and would be nice if someone can find fix or just point me where and what to look.

Signature · 1 line
Kind regards
Quoted 177 lines

Kind regards
Uffe

Quoted 174 lines

--
Regards

Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

--
Kind regards

Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#12 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Hi,

Quoted 19 lines · Milan P. Stanić
Quoted 10 lines · Ulf Hansson
Quoted 8 lines · Milan P. Stanić

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

Quoted 4 lines

That delta is more than 4 years of development. I am not sure investigating the delta would make sense, however I leave that to Rock-chips people to decide - or whomever that are willing to help to debug this.

I think it is not needed to look at full delta from this working chromeOS kernel and current mainline but maybe to 5.0 or 5.1 releases.

Quoted 7 lines

But I'm not sure and because I don't know what to look I can only hope that rockchip developers will find what is wrong.

Anyway, I use this rk3399 gru-kevin chromebook with external mmc card and it works fine if I don't 'touch' emmc. But is slower than when emmc is used for root and home filesystem, significantly, and would be nice if someone can find fix or just point me where and what to look.

More than a one year passed but this bug still persist on gru-kevin chromebook. I'm testing this with every kernel release in hope it will be fixed.
Now I'm tested it on 5.14.0-rc5

Signature · 1 line
Kind regards
Quoted 180 lines
Signature · 1 line
Kind regards
Quoted 177 lines

Kind regards
Uffe

Quoted 174 lines

--
Regards

Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

--
Kind regards

Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#13 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Quoted 31 lines · Milan P. Stanić

Hi,

Quoted 19 lines · Milan P. Stanić
Quoted 10 lines · Ulf Hansson
Quoted 8 lines · Milan P. Stanić

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

Quoted 4 lines

That delta is more than 4 years of development. I am not sure investigating the delta would make sense, however I leave that to Rock-chips people to decide - or whomever that are willing to help to debug this.

I think it is not needed to look at full delta from this working chromeOS kernel and current mainline but maybe to 5.0 or 5.1 releases.

Quoted 7 lines

But I'm not sure and because I don't know what to look I can only hope that rockchip developers will find what is wrong.

Anyway, I use this rk3399 gru-kevin chromebook with external mmc card and it works fine if I don't 'touch' emmc. But is slower than when emmc is used for root and home filesystem, significantly, and would be nice if someone can find fix or just point me where and what to look.

Quoted 4 lines

More than a one year passed but this bug still persist on gru-kevin chromebook. I'm testing this with every kernel release in hope it will be fixed.
Now I'm tested it on 5.14.0-rc5

Does this happen if you limit the emmc from going into high speed modes? (disable hs-200 and hs-400 modes)
I'm curious if it's an issue resuming the emmc at all, or if it's an issue jumping back into high speed mode.
Depending on if the emmc gets power cycled and how they handle power management, this can lead to a mode mismatch between the controller and the emmc.
Also does it eventually sort itself out? For example if rootfs isn't on the emmc and resume completes, does the emmc eventually become unwedged?

I haven't started playing with power management on rk3399 yet, but at least this will help isolate down where the issue is.

Very Respectfully,
Peter Geis

Quoted 189 lines

--
Kind regards

Quoted 180 lines

--
Kind regards

Quoted 177 lines

Kind regards
Uffe

Quoted 174 lines

--
Regards

Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

--
Kind regards

Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry if do not write or express correctly. And sorry if I posted bug report to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz, actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by echo mem > /sys/power/state from ACPI power and LID button handlers.

This only happens when I boot and use internal emmc card and never when boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state) this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which use to build kernel, output of the awk -f scripts/ver_linux as file ver_linux.txt and output of dmesg as file mmc-err.txt (from which I deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply patches and rebuild kernel or whatever you ask me (of course if my understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

#14 lore raw

Re: PROBLEM: mmc_select_hs400es failed, error -110【请注意,邮件由[email protected]代发】

Quoted 29 lines · Peter Geis
Quoted 28 lines · Milan P. Stanić
Quoted 26 lines · Milan P. Stanić
Quoted 10 lines · Ulf Hansson
Quoted 8 lines · Milan P. Stanić

I built chromeOS kernel 4.4.174 and after three days with few suspend-to-ram and resume mmc driver works fine, i.e. no problem occured.

Could someone who know programming mmc drivers look at differences between this chromeOS kernel (4.4.174) and some of mainline kernels to find cause of the problem and maybe to create fix or patch to send for testing.

Quoted 4 lines

That delta is more than 4 years of development. I am not sure investigating the delta would make sense, however I leave that to Rock-chips people to decide - or whomever that are willing to help to debug this.

I think it is not needed to look at full delta from this working chromeOS kernel and current mainline but maybe to 5.0 or 5.1 releases.

But I'm not sure and because I don't know what to look I can only hope that rockchip developers will find what is wrong.

Anyway, I use this rk3399 gru-kevin chromebook with external mmc card and it works fine if I don't 'touch' emmc. But is slower than when emmc is used for root and home filesystem, significantly, and would be nice if someone can find fix or just point me where and what to look.

Quoted 4 lines
Quoted 4 lines

More than a one year passed but this bug still persist on gru-kevin chromebook. I'm testing this with every kernel release in hope it will be fixed.
Now I'm tested it on 5.14.0-rc5

Let me describe what happens (sorry for bad English, I'm self taught).

This never happens with external mmc.

This happens only with internal emmc and only after resume from suspend to ram if the internal emmc mounted or the machine is booted with emmc only.

This happens on random time after resume, sometimes after 5-10 minutes, sometimes hours and sometimes even after day or two.

Never happens if the machine isn't suspended to ram.

Quoted 2 lines

Does this happen if you limit the emmc from going into high speed modes? (disable hs-200 and hs-400 modes)

Could you tel me how can I disable these modes (I'm not kernel hacker).

Quoted 2 lines

I'm curious if it's an issue resuming the emmc at all, or if it's an issue jumping back into high speed mode.

Here is excerpt from dmesg output:
----------------------------------
usb usb5: root hub lost power or was reset
usb usb6: root hub lost power or was reset
usb usb7: root hub lost power or was reset
usb usb8: root hub lost power or was reset
usb 1-1: reset high-speed USB device number 2 using ehci-platform
OOM killer enabled.
Restarting tasks ... done.
PM: suspend exit
mmc0: mmc_select_hs400es failed, error -110
mmc0: error -110 doing runtime resume
mmc_erase: group start error -110, status 0x0
mmc0: cache flush error -110
mmc0: mmc_select_hs400es failed, error -110
mmc0: tried to HW reset card, got error -110
blk_update_request: I/O error, dev mmcblk0, sector 18022496 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18022512 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18036192 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18383152 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18522184 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18522216 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18758024 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc_erase: group start error -110, status 0x0
blk_update_request: I/O error, dev mmcblk0, sector 18758080 op 0x3:(DISCARD) flags 0x0 phys_seg 1 prio class 0
mmc0: Got data interrupt 0x00000002 even though no data operation was in progress.
mmc0: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc0: sdhci: Sys addr: 0x00000020 | Version: 0x00001002
mmc0: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000020
mmc0: sdhci: Argument: 0x00000000 | Trn mode: 0x00000033
mmc0: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc0: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc0: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc0: sdhci: Timeout: 0x0000000b | Int stat: 0x00000000
mmc0: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc0: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc0: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc0: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc0: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc0: sdhci: Host ctl2: 0x00000000
mmc0: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x02d9d200
mmc0: sdhci: ============================================
mmc0: Got data interrupt 0x00000002 even though no data operation was in progress.
----------------------------------

Quoted 7 lines

Depending on if the emmc gets power cycled and how they handle power management, this can lead to a mode mismatch between the controller and the emmc.

Also does it eventually sort itself out? For example if rootfs isn't on the emmc and resume completes, does the emmc eventually become unwedged?

No, if the rootds is on emmc after some time when this happens trying to run any command they segfaults.

Quoted 2 lines

I haven't started playing with power management on rk3399 yet, but at least this will help isolate down where the issue is.

Thank you

Signature · 1 line
Kind regards
Quoted 192 lines

Very Respectfully,
Peter Geis

Quoted 189 lines

--
Kind regards

Quoted 180 lines

--
Kind regards

Quoted 177 lines

Kind regards
Uffe

Quoted 174 lines

--
Regards

Quoted 169 lines · Milan P. Stanić

Hi,

Quoted 18 lines · Ulf Hansson
Quoted 5 lines · Milan P. Stanić

Hi,
Anyone looked at this problem?

Or, is there better url or mailing list where I should send this bug report?

This is the mmc-list and since the rock-chips list was added by Shawn
- you have directed your question correctly.

The problem is, this isn't a list where everyone can get free support. People may be working on other platforms, for example.

The best option for you, is probably to ping Shawn or to reach out to some other Rock-chips people, as it seems like these guys needs to have a closer look.

Quoted 1 line
Quoted 7 lines
Quoted 2 lines

Or, could someone tell me what I could try to fix it, which file and parameters to change?

Well, if you can narrow down the problem that is always helpful. So for example, is there any upstream kernel that works - or is it only working through the chrome-os tree (which likely contains vendor specific changes that are not upstream).

Problem is in that the in last 5-6 months chrome-os kernels even don't boot, i.e. I only get blank display. I patched it for few releases locally but for about 3-4 months even my patch doesn't work. Maybe I could find chrome-os kernel about year or two old and try, but even then I don't know where to look for this problem.

Other option is to try with mainline kernels from 4.20.x and up one by one major release to look on which version this problem started. This will take time because problem occurs irregularly, sometimes in a few hours but sometimes it works fine for a few days until it appears.

Thank you for pointing me whom to ask and what to try.

--
Kind regards

Quoted 124 lines

Kind regards
Uff

Quoted 121 lines · Milan P. Stanić
Quoted 119 lines

Hi,

Quoted 5 lines · Robin Murphy

Hi Shawn,

Quoted 2 lines · Shawn Lin

Hi Milan

Quoted 7 lines
Quoted 2 lines

[+linux-rockchip to see if someone has a Samsung chromebook one plus and could confirm if it works]

FWIW I've also tried suspend on my NanoPC-T4 and seen that the eMMC (also HS400-ES) fails to come back properly on resume (thus resume never completes due to the missing root filesystem). IIRC it might even have been reproducible with suspend-to-idle, but I'd have to double-check that.

I forgot to tell that emmc worked without problem on this machine with ChromeOS kernel 4.4.xx downloaded from
https://chromium.googlesource.com/chromiumos/third_party/kernel/+/chromeos-4.4
and with patches from
https://github.com/archlinuxarm/PKGBUILDs/tree/master/core/linux-gru
for about two years (iirc).
Problem started when I switched to mainline kernels, somewhere around 5.1.xx

Quoted 94 lines

Robin.

Quoted 92 lines · Milan P. Stanić
Quoted 22 lines

Hello,

Sorry to annoy again, but could you tell me if I sent this bug report to right mail address or I should send it somewhere else.

Also, did I sent bug report correctly or I did some mistakes which caused it to be ignored.

Quoted 13 lines · Milan P. Stanić

Hello,

I'm not native English speaker and I'm self taught in English so sorry
if do not write or express correctly. And sorry if I posted bug report
to wrong address.

I'm running linux 5.6.0-rc3 without any patches on Samsung chromebook
one plus, Arm64 rockchip rk3399 based model name:
Machine model: Google Kevin

I build kernels from upstream git.kernel.org for this machine for some
time (iirc, from 5.2.1 and up) but I'm getting error messages in kernel
after machine resumes from suspend-to-ram.

It sounds to me suspend-to-ram never works for this machine, at least since 5.2.1. Am I right?

Quoted 50 lines
Quoted 50 lines

excerpt from dmesg output:
-----------------------------------------------------------------------
Restarting tasks ... done.
PM: suspend exit
mmc_host mmc0: Bus speed (slot 0) = 400000Hz (slot req 400000Hz,
actual 400000HZ div = 0)
mmc1: mmc_select_hs400es failed, error -110
mmc1: error -110 doing runtime resume
mmc1: Got data interrupt 0x00000002 even though no data
operation was in progress.
mmc1: sdhci: ============ SDHCI REGISTER DUMP ===========
mmc1: sdhci: Sys addr: 0x00000008 | Version: 0x00001002
mmc1: sdhci: Blk size: 0x00007200 | Blk cnt: 0x00000008
mmc1: sdhci: Argument: 0x00000000 | Trn mode: 0x00000023
mmc1: sdhci: Present: 0x1fff0001 | Host ctl: 0x00000035
mmc1: sdhci: Power: 0x0000000b | Blk gap: 0x00000080
mmc1: sdhci: Wake-up: 0x00000000 | Clock: 0x0000bc07
mmc1: sdhci: Timeout: 0x0000000d | Int stat: 0x00000000
mmc1: sdhci: Int enab: 0x03ff000b | Sig enab: 0x03ff000b
mmc1: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
mmc1: sdhci: Caps: 0x44edc880 | Caps_1: 0x801020f7
mmc1: sdhci: Cmd: 0x00000c1a | Max curr: 0x00000000
mmc1: sdhci: Resp[0]: 0x00000000 | Resp[1]: 0x373300bd
mmc1: sdhci: Resp[2]: 0x35303030 | Resp[3]: 0x00000000
mmc1: sdhci: Host ctl2: 0x00000000
mmc1: sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0xed41e200
-----------------------------------------------------------------------

I invoke suspend-to-ram by `echo mem > /sys/power/state` from ACPI power
and LID button handlers.

This only happens when I boot and use internal emmc card and never when
boot and use external mmc card.

If suspend-to-ram is not invoked (machine is always in normal state)
this problem never happen (or I missed it somehow).

I'm attaching kernel .config (file config-5.6.0-rc3-1-gru.conf) which
use to build kernel, output of the `awk -f scripts/ver_linux` as file
ver_linux.txt and output of dmesg as file mmc-err.txt (from which I
deleted wifi connection logs).

Sorry if I did something bad or wrong with this bug report, I don't have
much experience with bug reporting, especially for kernel.

I'm ready to send you more data, and investigate this more, apply
patches and rebuild kernel or whatever you ask me (of course if my
understanding and knowledge is enough for this job).

Quoted 8 lines
Quoted 5 lines

Thank in advance

--
Kind regards

[...]

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip

_______________________________________________
Linux-rockchip mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-rockchip