Skip to contentSkip to conversation

msm8996: sdhci-msm: apq8096-db820c sdhci fails to init - "Timeout waiting for hardware interrupt."

DISCUSSION Paolo Pisati 4 participants 7 messages Oct 10, 2019 – Oct 11, 2019 Arm Qualcomm mailing list

lore
Share & tools
lore
#0 lore raw

Sdhci consistenlty fails to initialize (and thus work) on my apq8096-db820c.

The issue is present since v5.0[*] mainline up to latest v5.4-rc2, using defconfig and:

CONFIG_SCSI_UFS_QCOM=y
CONFIG_PHY_QCOM_QMP=y
CONFIG_PHY_QCOM_UFS=y
CONFIG_ATL1C=y

but can be 100% reproduced with a clean defconfig too.

During boot, when it's time to mount the sdcard, mmc0 spits out a lot of:

...

[   13.683059] mmc0: Timeout waiting for hardware interrupt.
[   13.683095] mmc0: sdhci: ============ SDHCI REGISTER DUMP ===========
[   13.687441] mmc0: sdhci: Sys addr:  0x00000000 | Version:  0x00004902
[   13.693861] mmc0: sdhci: Blk size:  0x00004200 | Blk cnt:  0x00000000
[   13.700285] mmc0: sdhci: Argument:  0x00012444 | Trn mode: 0x00000033
[   13.706707] mmc0: sdhci: Present:   0x01680206 | Host ctl: 0x0000001f
[   13.713131] mmc0: sdhci: Power:     0x00000001 | Blk gap:  0x00000000
[   13.719555] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    0x00000007
[   13.725979] mmc0: sdhci: Timeout:   0x0000000a | Int stat: 0x00000000
[   13.732403] mmc0: sdhci: Int enab:  0x03ff900b | Sig enab: 0x03ff100b
[   13.738824] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
[   13.745249] mmc0: sdhci: Caps:      0x322dc8b2 | Caps_1:   0x00008007
[   13.751673] mmc0: sdhci: Cmd:       0x0000123a | Max curr: 0x00000000
[   13.758097] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  0x5b590000
[   13.764519] mmc0: sdhci: Resp[2]:   0x76b27f80 | Resp[3]:  0x0a404012
[   13.770944] mmc0: sdhci: Host ctl2: 0x00000000
[   13.777365] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0x00000001588be200
[   13.781708] mmc0: sdhci: ============================================
[   13.888927] mmc0: Reset 0x4 never completed.
...
[   14.004327] mmc0: Controller never released inhibit bit(s).

in between several sdhci register dumps.

Has anyone seen that before? Is sdhci-msm support broken upstream or am i missing something config-wise?

Full boot logs here: https://pastebin.ubuntu.com/p/BtRrgnjV7J/

*: nothing earlier then v5.0 boots on this board, so i couldn't test it.

Signature · 2 lines
bye,
p.
#1 lore raw

I'm using sdhci-msm on msm8998, so its not completely broken upstream.

What speed card are you trying to use?

Quoted 48 lines · Paolo Pisati

Sdhci consistenlty fails to initialize (and thus work) on my apq8096-db820c.

The issue is present since v5.0[*] mainline up to latest v5.4-rc2, using defconfig and:

CONFIG_SCSI_UFS_QCOM=y
CONFIG_PHY_QCOM_QMP=y
CONFIG_PHY_QCOM_UFS=y
CONFIG_ATL1C=y

but can be 100% reproduced with a clean defconfig too.

During boot, when it's time to mount the sdcard, mmc0 spits out a lot of:

...

[   13.683059] mmc0: Timeout waiting for hardware interrupt.
[   13.683095] mmc0: sdhci: ============ SDHCI REGISTER DUMP ===========
[   13.687441] mmc0: sdhci: Sys addr:  0x00000000 | Version:  0x00004902
[   13.693861] mmc0: sdhci: Blk size:  0x00004200 | Blk cnt:  0x00000000
[   13.700285] mmc0: sdhci: Argument:  0x00012444 | Trn mode: 0x00000033
[   13.706707] mmc0: sdhci: Present:   0x01680206 | Host ctl: 0x0000001f
[   13.713131] mmc0: sdhci: Power:     0x00000001 | Blk gap:  0x00000000
[   13.719555] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    0x00000007
[   13.725979] mmc0: sdhci: Timeout:   0x0000000a | Int stat: 0x00000000
[   13.732403] mmc0: sdhci: Int enab:  0x03ff900b | Sig enab: 0x03ff100b
[   13.738824] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
[   13.745249] mmc0: sdhci: Caps:      0x322dc8b2 | Caps_1:   0x00008007
[   13.751673] mmc0: sdhci: Cmd:       0x0000123a | Max curr: 0x00000000
[   13.758097] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  0x5b590000
[   13.764519] mmc0: sdhci: Resp[2]:   0x76b27f80 | Resp[3]:  0x0a404012
[   13.770944] mmc0: sdhci: Host ctl2: 0x00000000
[   13.777365] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0x00000001588be200
[   13.781708] mmc0: sdhci: ============================================
[   13.888927] mmc0: Reset 0x4 never completed.
...
[   14.004327] mmc0: Controller never released inhibit bit(s).

in between several sdhci register dumps.

Has anyone seen that before? Is sdhci-msm support broken upstream or am i missing something config-wise?

Full boot logs here: https://pastebin.ubuntu.com/p/BtRrgnjV7J/

*: nothing earlier then v5.0 boots on this board, so i couldn't test it. --
bye,
p.

#3 lore raw
In reply to Paolo Pisati

Hi Paolo,

Quoted 36 lines · Paolo Pisati

Sdhci consistenlty fails to initialize (and thus work) on my apq8096-db820c.

The issue is present since v5.0[*] mainline up to latest v5.4-rc2, using defconfig and:

CONFIG_SCSI_UFS_QCOM=y
CONFIG_PHY_QCOM_QMP=y
CONFIG_PHY_QCOM_UFS=y
CONFIG_ATL1C=y

but can be 100% reproduced with a clean defconfig too.

During boot, when it's time to mount the sdcard, mmc0 spits out a lot of:

...

[   13.683059] mmc0: Timeout waiting for hardware interrupt.
[   13.683095] mmc0: sdhci: ============ SDHCI REGISTER DUMP ===========
[   13.687441] mmc0: sdhci: Sys addr:  0x00000000 | Version:  0x00004902
[   13.693861] mmc0: sdhci: Blk size:  0x00004200 | Blk cnt:  0x00000000
[   13.700285] mmc0: sdhci: Argument:  0x00012444 | Trn mode: 0x00000033
[   13.706707] mmc0: sdhci: Present:   0x01680206 | Host ctl: 0x0000001f
[   13.713131] mmc0: sdhci: Power:     0x00000001 | Blk gap:  0x00000000
[   13.719555] mmc0: sdhci: Wake-up:   0x00000000 | Clock:    0x00000007
[   13.725979] mmc0: sdhci: Timeout:   0x0000000a | Int stat: 0x00000000
[   13.732403] mmc0: sdhci: Int enab:  0x03ff900b | Sig enab: 0x03ff100b
[   13.738824] mmc0: sdhci: ACmd stat: 0x00000000 | Slot int: 0x00000000
[   13.745249] mmc0: sdhci: Caps:      0x322dc8b2 | Caps_1:   0x00008007
[   13.751673] mmc0: sdhci: Cmd:       0x0000123a | Max curr: 0x00000000
[   13.758097] mmc0: sdhci: Resp[0]:   0x00000900 | Resp[1]:  0x5b590000
[   13.764519] mmc0: sdhci: Resp[2]:   0x76b27f80 | Resp[3]:  0x0a404012
[   13.770944] mmc0: sdhci: Host ctl2: 0x00000000
[   13.777365] mmc0: sdhci: ADMA Err:  0x00000000 | ADMA Ptr: 0x00000001588be200
[   13.781708] mmc0: sdhci: ============================================
[   13.888927] mmc0: Reset 0x4 never completed.
...
[   14.004327] mmc0: Controller never released inhibit bit(s).
Quoted 4 lines

in between several sdhci register dumps.

Has anyone seen that before? Is sdhci-msm support broken upstream or am i missing something config-wise?

I encountered that same error working on the Nexus 5 support upstream. Here's the fix:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=03864e57770a9541e7ff3990bacf2d9a2fffcd5d

Brian

#4 lore raw
Quoted 4 lines · Brian Masney

I encountered that same error working on the Nexus 5 support upstream. Here's the fix:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=03864e57770a9541e7ff3990bacf2d9a2fffcd5d

No dice, same exact problem.

But the patch is present downstream[1]:

commit c26727f853308dc4a6645dad09e9565429f8604f

Author: Loic Poulain <[email protected]>
Date:   Wed Dec 12 17:51:48 2018 +0100

arm64: dts: apq8096-db820c: Increase load on l21 for SDCARD

In the same way as for msm8974-hammerhead, l21 load, used for SDCARD VMMC, needs to be increased in order to prevent any voltage drop issues (due to limited current) happening with some SDCARDS or during specific operations (e.g. write).

Fixes 660a9763c6a9 (arm64: dts: qcom: db820c: Add pm8994 regulator node)

so it's probably worth carrying it.

1:
https://git.linaro.org/landing-teams/working/qualcomm/kernel.git/commit/arch/arm64/boot/dts/qcom?h=release/qcomlt-4.14&id=e6415afc1aef2ac9361437ff583ba1be5a932b78

Signature · 2 lines
bye,
p.
#5 lore raw

Hi,

Quoted 22 lines · Paolo Pisati
Quoted 4 lines · Brian Masney

I encountered that same error working on the Nexus 5 support upstream. Here's the fix:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=03864e57770a9541e7ff3990bacf2d9a2fffcd5d

No dice, same exact problem.

But the patch is present downstream[1]:

commit c26727f853308dc4a6645dad09e9565429f8604f

Author: Loic Poulain <[email protected]>
Date:   Wed Dec 12 17:51:48 2018 +0100

arm64: dts: apq8096-db820c: Increase load on l21 for SDCARD

In the same way as for msm8974-hammerhead, l21 load, used for SDCARD VMMC, needs to be increased in order to prevent any voltage drop issues (due to limited current) happening with some SDCARDS or during specific operations (e.g. write).

Quoted 5 lines
Fixes 660a9763c6a9 (arm64: dts: qcom: db820c: Add pm8994 regulator node)

so it's probably worth carrying it.

I've sent it to LKML, but it has never landed (and I've never followed-up).

Regards,
Loic

#6 lore raw
Quoted 15 lines · Loic Poulain
Quoted 15 lines

No dice, same exact problem.

But the patch is present downstream[1]:

commit c26727f853308dc4a6645dad09e9565429f8604f

Author: Loic Poulain <[email protected]>
Date:   Wed Dec 12 17:51:48 2018 +0100

arm64: dts: apq8096-db820c: Increase load on l21 for SDCARD

In the same way as for msm8974-hammerhead, l21 load, used for SDCARD VMMC, needs to be increased in order to prevent any voltage drop issues (due to limited current) happening with some SDCARDS or during specific operations (e.g. write).

Quoted 7 lines
Quoted 5 lines
Fixes 660a9763c6a9 (arm64: dts: qcom: db820c: Add pm8994 regulator node)

so it's probably worth carrying it.

I've sent it to LKML, but it has never landed (and I've never followed-up).

I see - btw, do you have a recent kernel where the mmc works on the db820c?

Signature · 2 lines
bye,
p.