Bug 14006 (AtanasKamburov)

Created attachment 22766 [details]
lspci

Created attachment 22767 [details]
dmesg

oh, and btw, the results are the same if I try to suspend/hibernate with stopped X.
I would be happy to provide more info if needed, or try your suggestions.

Created attachment 22768 [details]
kernel-2.6.30.5  .config

Hi, Atanas
    Do you mean that you can't execute the command of "echo disk > /sys/power/sate" under console mode?
    Please double check it again and confirm whether you have the permission to access it.
    thanks.

I can execute the command "echo disk >
/sys/power/sate" as root, and hibernate begins, but the problem is that after writing some stuff to the disk from the ram (probably) the laptop just hangs.

From the dmesg log in comment #2 it seems that there exist the following warning message:
   >hda: hda1 hda2 hda3
   >hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
   >hda: task_no_data_intr: error=0x04 { DriveStatusError }
   >hda: possibly failed opcode: 0xea
   >hda: wcache flush failed!

   Do you have an opportunity to try another disk and see whether the issue still exists?
   Thanks.

Hi Yakui,

thanks for your attention to my hibernation problems.
You are right, I get this error (task_no_data_intr) with newer kernels but not with the kernel that hibernation works with (2.6.27). Unfortunately, I cannot try another disk. If the problem were because of the disk, then hibernation shouldn't work with 2.6.27, I think...

Btw., I just realized that comment #6 is actually wrong (sorry, I had played around with a lot of kernels and options so I was confused). The result from the command "echo disk > /sys/power/sate" with kernel 2.6.31 is that I get the message "bash: echo: write error: Input/output error" and the laptop hangs (I cannot execute any command or open any window, but I can move the mouse pointer) - I think the disk stops spinning then.  The same command ("echo disk > /sys/power/sate") works with kernel 2.6.27.

Thanks, and please let me know if there is any more info that you need.

Since the issue appears to be reproducible 100% of the time on your system, would it be possible to carry out a bisection of commits between 2.6.27 and the first failing kernel?

thanks for the suggestion.  I would be happy to do that.  I understand that what you mean is to take a succession of kernels from 2.6.27.1 and compile and test every version after that.  Since my OS is gentoo linux, I suppose what I would have to do is install all relevant kernels from here http://gentoo-portage.com/sys-kernel/vanilla-sources and do "make oldconfig" each time.
Please let me know if you meant something else.

(In reply to comment #9)
> Since the issue appears to be reproducible 100% of the time on your system,
> would it be possible to carry out a bisection of commits between 2.6.27 and
> the
> first failing kernel?

OK - I installed kernel 2.6.27.35 and "echo disk > /sys/power/sate" does what it should (hibernate) and then installed kernel 2.6.28 and "echo disk > /sys/power/sate" gives the error "bash: echo: write error: Input/output error" and does not hibernate, as described above.

Oh, and something else that might be of importance.  With 2.6.28, I don't get the warning messages as in comment #7 (which I get e.g. with kernel 2.6.30.5), but instead I get the following message:  "thinkpad_acpi: acpi_bus_get_device(bay) failed: -19" during startup. So maybe the failing hibernate is not relatedto the warning messages from comment #7 after all...
I am attaching the dmesg with 2.6.28.

Created attachment 23219 [details]
dmesg 2.6.28

what about suspend? Does the system hang?

please run "echo mem > /sys/power/state; dmesg > dmesg-suspend" and see if you can get the log after the failure.

And it would be great if you can verify if it works in 2.6.32.

thanks for your interest in this bug.

suspend does not work:
"echo mem > /sys/power/state; dmesg > dmesg-suspend" shuts down the disk, on the console I get "bash: echo: write error: Input/output error" and the file dmesg-suspend is simply not there after a reboot (I have to do it with the power button).

sorry, I cannot try kernel 2.6.32 for now, since on boot I get an error message, something about the root filesystem not being able to mount read/write. This happens despite the fact that I used make oldconfig from 2.6.31 and left every new option with the default value. I guess this should be another bug report though.....

It's great that kernel bugzilla is back.

can you please verify if the problem still exists in the latest upstream
kernel?

Sorry, I cannot - I have sold the thinkpad... I guess the bug can be closed as it does not seem to be common, maybe it was just my hard disk making problems.

Get it. Bug closed.