Card reader problem with kernel 2.6.38.6-1

seqizz · 2011-05-11 09:34:43

Hi, just upgraded the kernel to 2.6.38.6-1 , system didn't detect the sd card. Getting error messages in dmesg like this:

[   98.133386] mmc0: Timeout waiting for hardware interrupt.
[   98.133401] sdhci: =========== REGISTER DUMP (mmc0)===========
[   98.133415] sdhci: Sys addr: 0x46dade80 | Version:  0x0000b001
[   98.133431] sdhci: Blk size: 0x00007040 | Blk cnt:  0x00000001
[   98.133438] sdhci: Argument: 0x00000200 | Trn mode: 0x00000013
[   98.133446] sdhci: Present:  0x01ff0001 | Host ctl: 0x00000013
[   98.133453] sdhci: Power:    0x0000000f | Blk gap:  0x00000000
[   98.133461] sdhci: Wake-up:  0x00000000 | Clock:    0x00000100
[   98.133468] sdhci: Timeout:  0x0000000a | Int stat: 0x00000000
[   98.133476] sdhci: Int enab: 0x02ff00cb | Sig enab: 0x02ff00cb
[   98.133483] sdhci: AC12 err: 0x00000000 | Slot int: 0x00000000
[   98.133491] sdhci: Caps:     0x014832b2 | Caps_1:   0x00000000
[   98.133498] sdhci: Cmd:      0x0000101a | Max curr: 0x00000000
[   98.133506] sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x7a61e800
[   98.133509] sdhci: ===========================================
[  108.160078] mmc0: Timeout waiting for hardware interrupt.
[  108.160093] sdhci: =========== REGISTER DUMP (mmc0)===========
[  108.160106] sdhci: Sys addr: 0x46dade80 | Version:  0x0000b001
[  108.160118] sdhci: Blk size: 0x00007040 | Blk cnt:  0x00000001
[  108.160133] sdhci: Argument: 0x00000200 | Trn mode: 0x00000013
[  108.160140] sdhci: Present:  0x01ff0001 | Host ctl: 0x00000013
[  108.160148] sdhci: Power:    0x0000000f | Blk gap:  0x00000000
[  108.160155] sdhci: Wake-up:  0x00000000 | Clock:    0x00000100
[  108.160163] sdhci: Timeout:  0x0000000a | Int stat: 0x00000000
[  108.160171] sdhci: Int enab: 0x02ff00cb | Sig enab: 0x02ff00cb
[  108.160178] sdhci: AC12 err: 0x00000000 | Slot int: 0x00000000
[  108.160186] sdhci: Caps:     0x014832b2 | Caps_1:   0x00000000
[  108.160193] sdhci: Cmd:      0x0000101a | Max curr: 0x00000000
[  108.160201] sdhci: ADMA Err: 0x00000000 | ADMA Ptr: 0x7a61e800
[  108.160204] sdhci: ===========================================

And if I remove the card:

[   86.233372] mmc0: Card removed during transfer!
[   86.233386] mmc0: Resetting controller.
[   86.233460] mmcblk0: unable to set block size to 512: -123
[   86.233569] mmcblk: probe of mmc0:0002 failed with error -22
[   86.433495] mmc0: card 0002 removed

Hardware? It's a sony vaio netbook:

02:00.0 System peripheral: JMicron Technology Corp. SD/MMC Host Controller (rev 80)

Probably will downgrade, ideas?

jlacroix · 2011-05-12 21:51:39

It's a bit worse for me. On my laptop, inserting a card into my card reader crashes my ENTIRE laptop. I mean no tty's or anything. If I boot up with an SD card inserted, the boot process will completely fail. I reverted back to 2.6.38.5 (the previous kernel) and the card reader works fine again.

Sara · 2011-05-13 05:52:26

jlacroix wrote:

It's a bit worse for me. On my laptop, inserting a card into my card reader crashes my ENTIRE laptop. I mean no tty's or anything. If I boot up with an SD card inserted, the boot process will completely fail. I reverted back to 2.6.38.5 (the previous kernel) and the card reader works fine again.

I've had both happen to me: I get a TTY and then I can't login properly, or I don't get a TTY and my tablet PC crashes entirely. I was about to downgrade via the Live CD (not having known the SD card was the problem, not having checked the forums), but then I accidentally booted into Arch, with my USB stick in (in addition to the SD card I usually have in), and the boot worked normally. This is pretty annoying...

Here's the info on my SD card:
48:03.2 SD Host controller: O2 Micro, Inc. Integrated MMC/SD Controller (rev 02)

Last edited by Sara (2011-05-13 05:56:36)

GogglesGuy · 2011-05-13 13:19:42

https://bugs.archlinux.org/task/23778 ?

ugaciaka · 2011-05-13 20:58:37

I same problem here ...

dif · 2011-05-14 08:25:23

Same thing here on HP dv5. Four times out of five the boot process fails. In most cases it is during loading the nvidia driver.
The boot process is successful every time my SDHC 4GB "Transcend" card is out of the reader.
The reader is manufactured by JMicron Technology Corp.

System: kernel26 2.6.38.6-2

@GogglesGuy
I don't think it is the same. Our problem is that we cannot start up the system at all, or system freezes dead the moment a card is put into the reader.

Last edited by dif (2011-05-14 08:32:24)

ugaciaka · 2011-05-14 09:05:24

PS

i have asus laptop

08:03.1 SD Host controller: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 22)

bangkok_manouel · 2011-05-14 09:10:12

there were 4 commits related to mmc between .38.5 and .38.6
http://git.kernel.org/?p=linux/kernel/g … 04758487e3
http://git.kernel.org/?p=linux/kernel/g … 1f9db3af33
http://git.kernel.org/?p=linux/kernel/g … 8c70a67038
http://git.kernel.org/?p=linux/kernel/g … 6d36350543
you should try to revert them one by one. once you are sure about the guilty commit, then you may want to email the author, he'll be happy to fix it with you.

Last edited by bangkok_manouel (2011-05-14 11:09:22)

kgas · 2011-05-14 17:29:17

in hp g6 10600ee laptop the card reader is not at all detected even after manually loading and unloading the module. this thread also deals with it.
lsusb / dmesg shows no changes before and after the sd card insertion.

Archieman · 2011-05-15 04:50:48

bangkok_manouel wrote:

there were 4 commits related to mmc between .38.5 and .38.6
http://git.kernel.org/?p=linux/kernel/g … 04758487e3
http://git.kernel.org/?p=linux/kernel/g … 1f9db3af33
http://git.kernel.org/?p=linux/kernel/g … 8c70a67038
http://git.kernel.org/?p=linux/kernel/g … 6d36350543
you should try to revert them one by one. once you are sure about the guilty commit, then you may want to email the author, he'll be happy to fix it with you.

If only I knew how to do what you are saying I would do it. How do I revert commits?

bangkok_manouel · 2011-05-15 09:52:02

Archieman wrote:

bangkok_manouel wrote:
there were 4 commits related to mmc between .38.5 and .38.6
http://git.kernel.org/?p=linux/kernel/g … 04758487e3
http://git.kernel.org/?p=linux/kernel/g … 1f9db3af33
http://git.kernel.org/?p=linux/kernel/g … 8c70a67038
http://git.kernel.org/?p=linux/kernel/g … 6d36350543
you should try to revert them one by one. once you are sure about the guilty commit, then you may want to email the author, he'll be happy to fix it with you.
If only I knew how to do what you are saying I would do it. How do I revert commits?

make sure you have what you need

# pacman -S base-devel abs

update the ABS tree

# abs

copy the stuff we need, download the needed patches.

$ cd
$ cp -r /var/abs/core/kernel26 .
$ cd kernel26
$ wget https://patchwork.kernel.org/patch/758692/raw/ -O 15.patch
$ wget https://patchwork.kernel.org/patch/758682/raw/ -O 16.patch
$ wget https://patchwork.kernel.org/patch/758672/raw/ -O 17.patch
$ wget https://patchwork.kernel.org/patch/758652/raw/ -O 18.patch

edit the PKGBUILD to add the patches to the source:

source=(ftp://ftp.kernel.org/pub/linux/kernel/v2.6/linux-$_basekernel.tar.bz2
        ftp://ftp.archlinux.org/other/kernel26/${_patchname}.bz2
        # the main kernel config files
        config config.x86_64
        15.patch 16.patch 17.patch 18.patch
        # standard config files for mkinitcpio ramdisk
        kernel26.preset)

add the reverting process, it should look like this:

  cd ${srcdir}/linux-$_basekernel
  # Add -ARCH patches
  # See http://projects.archlinux.org/linux-2.6-ARCH.git/
  patch -Np1 -i ${srcdir}/${_patchname}
  patch -Np1 -R -i ${srcdir}/15.patch
  patch -Np1 -R -i ${srcdir}/16.patch
  patch -Np1 -R -i ${srcdir}/17.patch
  patch -Np1 -R -i ${srcdir}/18.patch
  if [ "$CARCH" = "x86_64" ]; then

to revert them one by one, just comment out the patch command lines. to revert only 15.patch for example:

  cd ${srcdir}/linux-$_basekernel
  # Add -ARCH patches
  # See http://projects.archlinux.org/linux-2.6-ARCH.git/
  patch -Np1 -i ${srcdir}/${_patchname}
  patch -Np1 -R -i ${srcdir}/15.patch
#  patch -Np1 -R -i ${srcdir}/16.patch
#  patch -Np1 -R -i ${srcdir}/17.patch
#  patch -Np1 -R -i ${srcdir}/18.patch
  if [ "$CARCH" = "x86_64" ]; then

update md5sums:

$ makepkg -g >> PKGBUILD

then build your kernel:

$ makepkg

and install it with pacman, reboot, sacrifice a goat and test it.
once you have found the guilty commit, you can email the author and the people who signed-off and reported/tested or post that stuff on LKML.

zersaa · 2011-05-15 13:11:46

Have the exact same problem with a ThinkPad T510,
0d:00.0 SD Host controller: Ricoh Co Ltd MMC/SD Host Controller (rev 01)

jlacroix · 2011-05-15 13:12:50

Those seem like a lot of steps, I think I'll just wait for 2.6.39 which should be out very soon.

bangkok_manouel · 2011-05-15 15:26:40

jlacroix wrote:

Those seem like a lot of steps, I think I'll just wait for 2.6.39 which should be out very soon.

for a patch to go to -stable, it has to go upstream first (i.e. linus' branch)

jlacroix · 2011-05-15 15:29:46

bangkok_manouel wrote:

jlacroix wrote:
Those seem like a lot of steps, I think I'll just wait for 2.6.39 which should be out very soon.
for a patch to go to -stable, it has to go upstream first (i.e. linus' branch)

Right, but 2.6.39 will be final in a couple of days more than likely, which isn't enough time to get it patched before final release. I don't think it's worth the effort now until we find out whether or not its still an issue in 2.6.39. If it is, then we should report it. There's always a good chance it could already be taken care of.

bangkok_manouel · 2011-05-15 15:35:11

you should try .39-rc7-git9 then
edit: all those commits are at the top of history in linus' tree...
http://git.kernel.org/?p=linux/kernel/g … 14;hb=HEAD
http://git.kernel.org/?p=linux/kernel/g … ac;hb=HEAD
http://git.kernel.org/?p=linux/kernel/g … 0a;hb=HEAD
http://git.kernel.org/?p=linux/kernel/g … d6;hb=HEAD

Last edited by bangkok_manouel (2011-05-15 15:43:52)

laloch · 2011-05-15 23:20:33

Same problem here. The guilty commit is 3fe962c04...
I'll report the issue upstream.

nae64 · 2011-05-16 12:36:19

Can you post your working test package please, to bridge the time gap to 2.6.39?

laloch · 2011-05-16 13:43:06

nae64 wrote:

Can you post your working test package please, to bridge the time gap to 2.6.39?

Here it is...
x86_64: http://dl.dropbox.com/u/25169171/kernel … pkg.tar.xz
CFLAGS="-march=x86-64 -mtune=generic -O2 -pipe"
i686: http://dl.dropbox.com/u/25169171/kernel … pkg.tar.xz
CFLAGS="-m32 -march=i686 -mtune=generic -O2 -pipe"

Edit:
If you don't have a Dropbox account, please consider registering using this link. We'll both get 250MB of extra space for using the link. Thanks.

Edit2:
i686 package link

Last edited by laloch (2011-05-17 17:04:27)

bangkok_manouel · 2011-05-16 17:41:16

https://lkml.org/lkml/2011/5/16/330

bangkok_manouel · 2011-05-17 03:00:04

http://git.kernel.org/?p=linux/kernel/g … c8844fb98b

tacticalbread · 2011-05-17 03:28:08

I get the same problem, although I can use the SD card for a little bit of time, before my system crashes completely.

nae64 · 2011-05-17 11:08:48

@laloch: I already have a Dropbox account, sorry. Also I am using i686, so I will have to compile it myself, but thank you very much for posting it.

@tacticalbread: I think this behaviour is affected by the device you are using. I get the same behaviour when I am using a microSD card (Kingston) in an adapter card, but not for my standard Toshiba one. Either way, it doesn't work.

laloch · 2011-05-17 12:50:11

nae64 wrote:

@laloch: I already have a Dropbox account, sorry. Also I am using i686, so I will have to compile it myself, but thank you very much for posting it.

Sorry for not being of any help. Perhaps I should learn to make cross-compiled packages at last

nae64 wrote:

@tacticalbread: I think this behaviour is affected by the device you are using. I get the same behaviour when I am using a microSD card (Kingston) in an adapter card, but not for my standard Toshiba one. Either way, it doesn't work.

I can confirm this - two SD cards, one almost always triggers the divide error in the sdhci module, the other doesn't. Neither one works though.

We should post as many logs and backtraces as we can before 2.6.39 hits the repo. Archlinux is the only distro brave enough to push kernels this fresh and thus the only distro suffering from this problem, so the Archlinux users should collect as much info as possible.

tacticalbread · 2011-05-17 16:50:01

Here's my laptop's SD slot:

01:09.1 SD Host controller: Ricoh Co Ltd R5C822 SD/SDIO/MMC/MS/MSPro Host Adapter (rev 22)
01:09.2 System peripheral: Ricoh Co Ltd R5C592 Memory Stick Bus Host Adapter (rev 12)
01:09.3 System peripheral: Ricoh Co Ltd xD-Picture Card Controller (rev 12)

I've got an old desktop with an SD slot that I installed Arch onto, I'll try it with that today as well.

edit:
Desktop's card slot:

Bus 002 Device 002: ID 058f:9360 Alcor Micro Corp. 8-in-1 Media Card Reader

This computer won't boot if an SD card is in the slot. Once it's booted, I can mount the SD card, but it only mounts read only, even if I mount -o rw as root.

Last edited by tacticalbread (2011-05-17 18:06:46)

Arch Linux

#1 2011-05-11 09:34:43

Card reader problem with kernel 2.6.38.6-1

#2 2011-05-12 21:51:39

Re: Card reader problem with kernel 2.6.38.6-1

#3 2011-05-13 05:52:26

Re: Card reader problem with kernel 2.6.38.6-1

#4 2011-05-13 13:19:42

Re: Card reader problem with kernel 2.6.38.6-1

#5 2011-05-13 20:58:37

Re: Card reader problem with kernel 2.6.38.6-1

#6 2011-05-14 08:25:23

Re: Card reader problem with kernel 2.6.38.6-1

#7 2011-05-14 09:05:24

Re: Card reader problem with kernel 2.6.38.6-1

#8 2011-05-14 09:10:12

Re: Card reader problem with kernel 2.6.38.6-1

#9 2011-05-14 17:29:17

Re: Card reader problem with kernel 2.6.38.6-1

#10 2011-05-15 04:50:48

Re: Card reader problem with kernel 2.6.38.6-1

#11 2011-05-15 09:52:02

Re: Card reader problem with kernel 2.6.38.6-1

#12 2011-05-15 13:11:46

Re: Card reader problem with kernel 2.6.38.6-1

#13 2011-05-15 13:12:50

Re: Card reader problem with kernel 2.6.38.6-1

#14 2011-05-15 15:26:40

Re: Card reader problem with kernel 2.6.38.6-1

#15 2011-05-15 15:29:46

Re: Card reader problem with kernel 2.6.38.6-1

#16 2011-05-15 15:35:11

Re: Card reader problem with kernel 2.6.38.6-1

#17 2011-05-15 23:20:33

Re: Card reader problem with kernel 2.6.38.6-1

#18 2011-05-16 12:36:19

Re: Card reader problem with kernel 2.6.38.6-1

#19 2011-05-16 13:43:06

Re: Card reader problem with kernel 2.6.38.6-1

#20 2011-05-16 17:41:16

Re: Card reader problem with kernel 2.6.38.6-1

#21 2011-05-17 03:00:04

Re: Card reader problem with kernel 2.6.38.6-1

#22 2011-05-17 03:28:08

Re: Card reader problem with kernel 2.6.38.6-1

#23 2011-05-17 11:08:48

Re: Card reader problem with kernel 2.6.38.6-1

#24 2011-05-17 12:50:11

Re: Card reader problem with kernel 2.6.38.6-1

#25 2011-05-17 16:50:01

Re: Card reader problem with kernel 2.6.38.6-1

Board footer