ata_piix problem with Intel ICH7 chipset

[José M. López-Cepero]

Binary package hint: linux-386

I have an Acer Aspire 9410 Centrino Duo laptop and there seems to be a problem with the SATA controller. Although I have found no major flaws yet (ie. no data corruption), every few minutes (maybe 2, maybe 30) the system seems to hang for a while (about 10-30 sec).

Apparently, the problem is that the ata driver is not fully compatible with the chipset. Examining dmesg's output, every time a hang happens the following lines appear:

[17186864.708000] ata1 is slow to respond, please be patient
[17186889.664000] ata1: command 0xa0 timeout, stat 0xd0 host_stat 0x60
[17186889.664000] ata1: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[17186889.664000] ATA: abnormal status 0xD0 on port 0x1F7
[17186891.016000] ata1: status=0x50 { DriveReady SeekComplete }
[17186891.016000] sda: Current: sense key: No Sense
[17186891.016000] Additional sense: No additional sense information
[17186891.020000] sr0: CDROM (ioctl) error, command: <6>Read TOC/PMA/ATIP 43 00 00 00 00 00 00 00 0c 40
[17186891.020000] sr: Current [descriptor]: sense key: Aborted Command
[17186891.020000] Additional sense: Scsi parity error
[17186891.020000] ATA: abnormal status 0xD0 on port 0x1F7
[17186891.020000] ATA: abnormal status 0xD0 on port 0x1F7

The main scheme is always the same: there is an 'ata1 is slow to respond, please be patient', followed by a timeout and a few errors. The CD error message may appear or not; I remember having seen a few different variations (examining /var/log/messages, the "sense key" may be "No sense", "Scsi parity error" or "Aborted command", and there are another slight variations, and the error may come from sda or sr0). The "abnormal status" messages are always there; I have looked around for a reference to what status 0xD0 means, but to no avail. I think what causes the hangs is that the system blocks until a 30 second timeout passes from the "slow to respond" message.

I'm using 2.6.15-26-386 (the 686 version caused the laptop to hang at seemingly random times); as revealed by lspci, the controler is:

0000:00:1f.2 IDE interface: Intel Corporation 82801GBM/GHM (ICH7 Family) Serial ATA Storage Controllers cc=IDE (rev 02)

(numeric code 8086:27c4), which apparently should be well supported by ata_piix. I have not yet tried with another Linux distro, but Windows XP does not exhibit that problem. Also, I exchanged the laptop for another because of an unrelated problem and the hangs kept happening.

I assume this is due to some kind of bug on the sata controllers and that's why I'm reporting the bug here. Feel free to direct me to any other source and I'll repost. Alternatively, if there was some way to set the 'slow to respond' timeout to a shorter time (5 seconds or so), that would probably be a good enough workaround.

I will immediately attach the output of dmesg, lspci and lsmod.

Thanks in advance for your help.

[José M. López-Cepero (cepe)]

dmesg

[José M. López-Cepero (cepe)]

lspci -vvvv

[José M. López-Cepero (cepe)]

/var/log/messages. Search for "ata1 is slow to respond" to locate hangs.

[José M. López-Cepero (cepe)]

lsmod

[José M. López-Cepero (cepe)]

Sorry - I forgot to mention. I'm using Ubuntu Dapper, freshly installed and with all the updates applied.

[José M. López-Cepero (cepe)]

And sorry (again). When the CD error occurs, it is always (or at least the times I have noticed) a "read TOC" error such as the one above. But there is no disc in the drive to read the TOC from. I don't know if it is usual to issue read TOC commands in empty drives...

[José M. López-Cepero (cepe)]

Changing to Confirmed, since it has happened to me in two different laptops (thus nearly excluding hardware problems) and on a stock Ubuntu 6.06 installation.

[José M. López-Cepero (cepe)]

Just wanted to let you know that this bug has a _really_ nasty interaction with the wireless chip on the laptop (a fairly common ipw3945). When one of these hanges occurs, the module may crash with a PREEMPT, which essentially means that wireless is lost unless the computer is rebooted. The probabilities of this occurring are seemingly related with the network activity at the time of the hang. For instance, if I run aMule and have some downloads going on (nothing too speedy, maybe 30 kb/s between upload and download) the wireless connection will rarely be up for more than 1 hour. (No exaggeration - it actually just went out as I was writing this report - total uptime is 12 _minutes_).

I'm a little frustrated at not being able to use my favourite distro effectively on my computer. I understand that maybe this is not the correct place to report the bug, or that I should provide more information, or that you are just very busy at the moment, and that's why there are no advances on the issue. In that case, I'd really like a pointer to what to do next to help you, or to the correct libata mantainer. As I said before, I like to think that I'm a rather savvy Linux user, but if it comes to low-level kernel debugging, I'm totally at a loss.

[Ben Collins (ben-collins)]

Have you tried the latest kernel from dapper-security, as of Aug 3?

[José M. López-Cepero (cepe)]

Hi Ben,

Thanks for your reply. I'm rather confident I'm current with all updates. The kernel package I'm using is

linux-image-2.6.15-26-386
Version 2.6.15-26.46

It is indeed from dapper-security; I can't seem to find a way for synaptic or apt-cache to show me the date, though.

I have seen that Edgy has a slightly more advanced kernel version, and that building custom kernels in Dapper and below is rather discouraged. Do you think updating to Edgy would help? (I have quickly glanced through 2.6.17 changelogs, but found nothing libata related). I have run Breezy and Dapper already before release, but not this early in the release cycle.

Thanks again for your reply. If there is any other test I can carry out for you, please let me know.

[Ben Collins (ben-collins)]

Please try one of the Edgy Knot CD's. You can boot liveCD and check if it operates your hardware correctly.

[chantra (chantra)]

I can report the same problem here.
I've got an aspire 9411
here are some attach files:

[José M. López-Cepero (cepe)]

Edgy Eft LiveCD Knot 1 won't boot. The process hangs rather early in the bootup sequence: in the initrd, while trying to mount the root filesystem. After a few seconds, it bails out to a BusyBox prompt complaining that it does not have access to the tty because job control is turned off (or something similar). The MD5sum of the CD is correct, so it is not a recording problem on my side.

Any other ideas? :(

[chantra (chantra)]

same here again.
2.6.17-5 (if I remember well) don't boot at all, it uncompress the kernel, load essential modules and then hang on "mounting root filesystem".
the timeout could be changed in include/linux/libata.h
bu modifying ATA_TMOUT_BOOT, it is then used in
drivers/scsi/libata-core.c

check this out, might be a solution.
https://launchpad.net/distros/ubuntu/+source/linux-source-2.6.15/+bug/37382
will try it later

[José M. López-Cepero (cepe)]

A small update. Today I have struck a lucky chord: it's already almost 5h uptime and not a single hang (and, therefore, my wireless connection is working happily). If I do a dmesg, the following suspicious lines appear:

[17179641.092000] cdrom: sr0: mrw address space DMA selected
[17179641.116000] Assertion failed! qc->n_elem > 0,drivers/scsi/libata-core.c,ata_fill_sg,line=2531
[17179641.376000] cdrom: sr0: mrw address space DMA selected
[17179641.380000] Assertion failed! qc->n_elem > 0,drivers/scsi/libata-core.c,ata_fill_sg,line=2531

Whatever that means, I have not seen them before (and I don't remember updating the kernel from yesterday to today). The system seems to work perfectly fine, with no evident negative consequences of the failed assertion, and with no hangs. The CD works seemingly OK, and copying large files out of it doesn't clog the CPU, but the maximum transfer speed seems to be around 2 MB/s. I'd rather endure that than the hangs. If any of you know of any way to force the "mrw address space DMA", that might be a good enough workaround for the moment...

I'll leave the computer running overnight to be sure, and then reboot and see if it magically works again...

[chantra (chantra)]

A new dump if it can help.

Well, the dump start with a hibernation, he computer came back from hibernation fine.
Did some reboot to try out alternatives and then let my computer turned on overnight. Even though the computer was not doing much task (simply turn on), there were a "slow to respond" issue about once every hours :s.

[chantra (chantra)]

lately i tried an older kernel, 2.6.15-23-686 that I could download from ftp://ftp.port80.se/ubuntu/pool/main/l/linux-source-2.6.15/ .

It seems to be stable on my edgy system (I did'nt get any "ata1 is slow to respond").
The same kernel on dapper kept on freezing up.

[chantra (chantra)]

well, kernel 2.6.15-23-686 is definitely stable on edgy, but not on dapper.
This is exactely the same .deb package so I start to believe that something else interact with the sata drive and create the slowing down.
Maybe some power management software. I couldn't point this out so far.

[chantra (chantra)]

well, lately i updated my edgy box and the stable kernel wasn't stable anymore.

I finally spotted (i guess) which software messed it up:
*hal
*libhal-storage1
*hal-device-manager
*libhal1

stables version for me are:
hal_0.5.7.1-0ubuntu5_i386.deb
hal-device-manager_0.5.7.1-0ubuntu5_all.deb
libhal-storage1_0.5.7.1-0ubuntu5_i386.deb
libhal1_0.5.7.1-0ubuntu5_i386.deb

josé, you might be able to get thos from ftp://ftp.port80.se/ubuntu

hope this helps

[Prinz Igor (pigor)]

Exactly the same here.

I have a new laptop: Wortmann Terra MOBILE-BUSINESS M 8400 iC-T2500/15.4" GLARE

Is there any hope that it will be fixed soon?

[José M. López-Cepero (cepe)]

Sorry for not answering earlier - I have been without internet connection.

I have found what seems to be a workaround. Don't ask me how it is supposed to work, though :D

Apparently, when I boot up using the Edgy CD I downloaded and choosing "Boot from first hard disk" from the menu, the CD gets put in the "mrw address space DMA" (whatever that is) and the system may go on happily for days (very rarely I get a hangup, but it would be like one in a week or so. I don't know why that happens (note that the only thing from the CD that gets executed is the bootloader), but it seems to be reliable...

Cheers

[teomatto (teomatto)]

hi all, i HAD the same problem with kernel 2.6.15-26-386 and my maxdata centrino duo laptop (with ipw3945).

I solved by upgrading kernel to 2.6.15-26-686
(apt-get install linux-i686)

hope this help
teo

[Prinz Igor (pigor)]

hi. i have just installed edgy knot2. my system does not freeze anymore but the message are all the same (appear just once(!) in dmesg):

ata2 is slow to respond, please be patient
ata2 failed to respond (30 secs)
ata2: command 0xa0 timeout, stat 0xd0 host_stat 0x20
ata2: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
sr0: CDROM (ioctl) error, command: <6>Read TOC/PMA/ATIP 43 00 00 00 00 00 00 00 0c 40
sr: Current [descriptor]: sense key: Aborted Command
   Additional sense: Scsi parity error

What is the difference between Dapper and Edgy? What makes it work in Edgy?

[Christian Mayrhuber (chrismayrhuber)]

This bug seems to be a duplicate of
   [Bug 38760] Re: ata timeout with sony vgn-s580 laptop

I don't know why but those SATA timeout messages occur less frequently for me if
I add the ahci module to /etc/modules.

Acer Aspire 9412WLMI
Kubuntu Dapper with latest security updates.
ii linux-image-2.6.15-26-686 2.6.15-26.46
ii linux-restricted-modules-686 2.6.15.24
ii nvidia-kernel-common 20051028+1
ii nvidia-glx 1.0.8762+2.6.15.11-3

I had to add the boot parameters "video=vga16fb:off video=vesafb:off"
to prevent hard lockups due to bad interaction of nvidia and framebuffer
drivers.

Hibernation, neither suspend to disk, nor suspend to RAM, works for me at all.
The laptop just freezes with a black screen.
I tried KPowersafe and KLaptop with the same results.

[bonsiware (bonsiware-deactivatedaccount)]

Acer Aspire 9411 AWSMi

Same timeout errors happens only if the cdrom drive is empty!

If I have a cd or a dvd in the drive all works fine!!!

Can you confirm?

[Christian Mayrhuber (chrismayrhuber)]

Hi bonsiware,

I cannot confirm, but after installing a new firmware on the TS-632D
DVDRW drive which I found on
 http://www.toshibaer.com/firmware/index.php?path=TS-L632D/
those ata timeouts completely disappeared on my Aspire 9412AWSMi, so
I suppose this is not a kernel bug, but rather was a firmware bug.

I found out that my DVDRW drive had revision AC00 on it, the bin file in
the
zip file is revison AC01. I used TS-L632D_AC01_Acer_OEM.zip

You need Windows to flash the firmware, though. Be aware that flashing a
new
firmware on your DVDRW drive may destroy it if flashing fails or you flash
a
wrong firmware.

Cheers, Christian

[chantra (chantra)]

hi christian,

unfortunately, I cannot confirm that solution :(
I did update the firmware, but still, I get those error :
[17180760.212000] ata1 is slow to respond, please be patient
[17180785.188000] ata1 failed to respond (30 secs)
[17180785.200000] ata1: command 0xa0 timeout, stat 0xd0 host_stat 0x60
[17180785.200000] ata1: translated ATA stat/err 0xd0/00 to SCSI SK/ASC/ASCQ 0xb/47/00
[17180785.200000] sr0: CDROM (ioctl) error, command: <6>Read TOC/PMA/ATIP 43 00 00 00 00 00 00 00 0c 40
[17180785.200000] sr: Current [descriptor]: sense key: Aborted Command
[17180785.200000] Additional sense: Scsi parity error

:(

[bonsiware (bonsiware-deactivatedaccount)]

I've upgraded the firmware too, but I still have the same errors...

In a previous post I told about to have a cd in the drive to avoid timeout errors... well it wasn't exact... but if I BOOT with a cd in the drive and I keep it in everything works fine!!! If the cd is ejected errors come back! It doesn't mean that I keep the same cd in for the entire session... I can change cd, but I can't stay without!

Try and let me know?

[Christian Mayrhuber (chrismayrhuber)]

Hi,

this is strange. The errors are completely gone for me with or without
a CD in the drive after the firmware update.

I'm running Ubuntu Edgy with the generic kernel:
ii linux-image-2.6.17-10-generic 2.6.17-10.33

I did that upgrade because the stalls on Dapper made me crazy.
Unfortunately
the upgrade to Edgy didn't help a bit, the firmware upgrade seems to
fix it, at least for me.

Sad that it doesn't work for you.

Cheers, Christian

On Sun, 19 Nov 2006 22:57:37 +0100, bonsiware <email address hidden> wrote:

> I've upgraded the firmware too, but I still have the same errors...
>
> In a previous post I told about to have a cd in the drive to avoid
> timeout errors... well it wasn't exact... but if I BOOT with a cd in the
> drive and I keep it in everything works fine!!! If the cd is ejected
> errors come back! It doesn't mean that I keep the same cd in for the
> entire session... I can change cd, but I can't stay without!
>
> Try and let me know?
>

[chantra (chantra)]

it seems that as bonsiware said, having a cd in the drive stops the freezes :)

weird, but so far, it works :D

[live_linux_secure (francois-vaningelgom)]

I also have a Acer Aspire 9412AWLMI and i tried much distros with all the same result (freeze...). Someone (witch i thank a lot!) at the fosdem at Brussels showed me this post. Since i putted a cd in the cd drive no more freeze.
I'll try to upgrade firmware of the cd drive and i'll say you what'd the result.
hoping this bug will be soon fixed...
thanks

[Christian Mayrhuber (chrismayrhuber)]

Hi,

I have to report that I still have freezes from time to time (during
normal desktop work):
 - ca. 2 a day without a CD in the drive
 - ca. 1 a day with a CD in the drive

However prior to the firmware update I had around 5 times more freezes per
day.
I'm currently running 2.6.19 on feisty. I first thought that this freezes
were gone, because
I had none during the first days after applying the firmware update, don't
know why.
The kernel versions from 2.6.17 on till 2.6.19 didn't make much of a
difference.

Sorry, that I cannot say anything more positive.

Note that a firmware upgrade may destroy your DVD drive and void the
warranty.

Cheers, Christian

On Mon, 26 Feb 2007 13:39:43 +0100, live_linux_secure
<email address hidden> wrote:

> I also have a Acer Aspire 9412AWLMI and i tried much distros with all
> the same result (freeze...). Someone (witch i thank a lot!) at the
> fosdem at Brussels showed me this post. Since i putted a cd in the cd
> drive no more freeze.
> I'll try to upgrade firmware of the cd drive and i'll say you what'd the
> result.
> hoping this bug will be soon fixed...
> thanks
>

[chantra (chantra)]

using a disc, I don't have any freezes.

I believe since 2.6.19, freezes happens more rarely than it used to be on edgy. If I remember well, on edgy, freezes were happening almost every 30min-1hours.

Hope to see this fixed in the coming release. In the meanwhile, I keep a CD in the tray.

[live_linux_secure (francois-vaningelgom)]

I confirm that it still freezes for me too. A little bit less and not the first day but it still freezes.
I also tried firmware upgrade and cd in the drive with same poor results

François

[Paul Sladen (sladen)]

François: I'm glad the link was helpful! :)

I don't think we know the root cause of this bug; I know that a set of Lenovo ThinkPads suffered something similar to do with the PATA-SATA bridge---I have one of the affected laptops. The workaround for this was that they patched the firmware *on the hard-drive*. (Which also means that if you swap the hard-drive it doesn't make it very far through the boot-process)...

Not sure how to solve this one; the fact that upgrading the firmware on the optical drive has an effect gives the impression that there is something fairly wrong with the hardware itself. It's likely that this was worked around in any MS Windows drivers shipped with the machine, and that is what we'd have to do under Linux ...when the actual problem can be tracked down.

Are there any kernel versions that /don't/ have this issue? I think somebody mentioned a recent Gentoo kernel that may have had some extra IDE patches.

[José M. López-Cepero (cepe)]

I have recently begun using the same laptop I had the problem as my main computer (for the past 3 months I didn't use it much), and apparently updating to Feisty solves the problem. If the involved computers are not mission critical I'd reccommend you to update to Feisty (there's the ocassional breakage, but overall it is rather stable).

I do have a CD in the drive. I haven't yet researched what happens if I remove it.

Best regards, Jose

[Christian Mayrhuber (chrismayrhuber)]

Hi, again!

I expirienced what I assume is a harddrive reset unter MS Windows.
The harddisk makes a noise like it is parking it's heads at arbitrary
points in time. Somtimes, I notice a short delay in mouse movement during
the noise period. I think the windows driver "works around" the system
freeze
by using a smaller timeout than Linux. In theory the same should be
possible
on linux through sysfs.
Issuing as root user (or adding it to /etc/rc.local):

 echo "1" > /sys/bus/scsi/devices/0:0:0:0/timeout

will set the timeout from 30s to 1s.
I tried suspend to ram (and wakeup) with this setting and it still worked..

I'll report back if there are any sideeffects for me.

Cheers, Christian

PS: Is there any tool like sysctl, which sets parameters in /proc, to set
parameters in sysfs?

[chantra (chantra)]

nice one.
concerning sysfs, you have sysfsutils. Install it and add:
bus/scsi/devices/0:0:0:0/timeout = 1
in /etc/sysfs.conf

I'm going to try it out with a timeout of 10, will leave it running and update with the results.

[chantra (chantra)]

well, after 2 hours running, I still get these:
[ 6680.772000] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
[ 6680.772000] ata1.01: cmd a0/00:00:00:00:20/00:00:00:00:00/b0 tag 0 cdb 0x1e data 0
[ 6680.772000] res 40/00:03:00:00:00/00:00:00:00:00/b0 Emask 0x4 (timeout)
[ 6687.772000] ata1: port is slow to respond, please be patient (Status 0xd0)
[ 6710.788000] ata1: port failed to respond (30 secs, Status 0xd0)
[ 6710.788000] ata1: soft resetting port
[ 6711.140000] ata1.00: configured for UDMA/100
[ 6711.320000] ata1.01: configured for UDMA/33
[ 6711.320000] ata1: EH complete
[ 6711.328000] SCSI device sda: 156301488 512-byte hdwr sectors (80026 MB)
[ 6711.328000] sda: Write Protect is off
[ 6711.328000] sda: Mode Sense: 00 3a 00 00
[ 6711.672000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 6711.692000] SCSI device sda: 156301488 512-byte hdwr sectors (80026 MB)
[ 6711.692000] sda: Write Protect is off
[ 6711.692000] sda: Mode Sense: 00 3a 00 00
[ 6711.692000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA

still the "30 secs" even though:
cat '/sys/bus/scsi/devices/0:0:0:0/timeout'
10

[José M. López-Cepero (cepe)]

Using 2.6.20-9-generic SMP on Feisty.

Without a CD in the drive, there are hangs and the dmesg messages are almost exactly the same chantra has reported:

[ 9575.732000] ata1.01: cmd a0/01:00:00:00:00/00:00:00:00:00/b0 tag 0 cdb 0x25 data 8 in
[ 9575.732000] res 50/00:00:00:00:00/00:00:00:00:00/b0 Emask 0x2 (HSM violation)
[ 9575.732000] ata1: soft resetting port
[ 9576.084000] ata1.00: configured for UDMA/100
[ 9576.264000] ata1.01: configured for UDMA/33
[ 9576.264000] ata1: EH complete
[ 9576.280000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[ 9576.296000] sda: Write Protect is off
[ 9576.296000] sda: Mode Sense: 00 3a 00 00
[ 9576.304000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 9576.336000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[ 9576.340000] sda: Write Protect is off
[ 9576.340000] sda: Mode Sense: 00 3a 00 00
[ 9576.340000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[10195.476000] ipw3945: association process canceled
[10531.292000] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
[10531.292000] ata1.01: cmd a0/01:00:00:00:00/00:00:00:00:00/b0 tag 0 cdb 0x43 data 12 in
[10531.292000] res 40/00:03:00:00:00/00:00:00:00:00/b0 Emask 0x4 (timeout)
[10538.292000] ata1: port is slow to respond, please be patient (Status 0xd0)
[10561.308000] ata1: port failed to respond (30 secs, Status 0xd0)
[10561.308000] ata1: soft resetting port
[10561.660000] ata1.00: configured for UDMA/100
[10561.848000] ata1.01: configured for UDMA/33
[10561.848000] ata1: EH complete
[10561.864000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[10561.872000] sda: Write Protect is off
[10561.872000] sda: Mode Sense: 00 3a 00 00
[10561.888000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[10561.912000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[10561.912000] sda: Write Protect is off
[10561.912000] sda: Mode Sense: 00 3a 00 00
[10562.288000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA

I had a few more errors of the "timeout" kind after this one, and they all corresponded to hangs. They seemed to happen very frequently, like every 20 minutes or less. I did not get any other "HSM violation" errors.

The good part: With a CD in the drive, the computer works perfectly. No hangs, no dmesg messages, all good and well.

[Christian Mayrhuber (chrismayrhuber)]

Hi,

thanks for the tip with sysfsutils.

Sorry that setting the timeout doesn't work. I got no timeout until today
since
a week and it seems that the number "30" in the timeout setting is just a
coincidence.
I don't know what this timeout setting is for, then. The dvd-rw drive has
a timeout of 0.

Regards, Christian

On Wed, 07 Mar 2007 22:07:51 +0100, chantra <email address hidden> wrote:

> well, after 2 hours running, I still get these:
> [ 6680.772000] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2
> frozen
> [ 6680.772000] ata1.01: cmd a0/00:00:00:00:20/00:00:00:00:00/b0 tag 0
> cdb 0x1e data 0
> [ 6680.772000] res 40/00:03:00:00:00/00:00:00:00:00/b0 Emask
> 0x4 (timeout)
> [ 6687.772000] ata1: port is slow to respond, please be patient (Status
> 0xd0)
> [ 6710.788000] ata1: port failed to respond (30 secs, Status 0xd0)
> [ 6710.788000] ata1: soft resetting port
> [ 6711.140000] ata1.00: configured for UDMA/100
> [ 6711.320000] ata1.01: configured for UDMA/33
> [ 6711.320000] ata1: EH complete
> [ 6711.328000] SCSI device sda: 156301488 512-byte hdwr sectors (80026
> MB)
> [ 6711.328000] sda: Write Protect is off
> [ 6711.328000] sda: Mode Sense: 00 3a 00 00
> [ 6711.672000] SCSI device sda: write cache: enabled, read cache:
> enabled, doesn't support DPO or FUA
> [ 6711.692000] SCSI device sda: 156301488 512-byte hdwr sectors (80026
> MB)
> [ 6711.692000] sda: Write Protect is off
> [ 6711.692000] sda: Mode Sense: 00 3a 00 00
> [ 6711.692000] SCSI device sda: write cache: enabled, read cache:
> enabled, doesn't support DPO or FUA
>
>
> still the "30 secs" even though:
> cat '/sys/bus/scsi/devices/0:0:0:0/timeout'
> 10
>

[live_linux_secure (francois-vaningelgom)]

Hi all i didn't try the "echo "1" > /sys/bus/scsi/devices/0:0:0:0/timeout" but i tried to search why it was freezing and i saw somewhere (maybe is it stupid?) that it was due to the ata_piix driver which control both SATA and PATA. From this i thought that i could try to add a boot option like "noata_piix". I thought that it wouldn't boot but it booted...
Sadly it was still freezing. But i tried with a cd in the drive, i'm running for 20hours now and no freeze... What's your method to detect freeze you also use dmesg?
I'm running ubuntu feisty beta with the generic kernel 2.6.20-12-generic and with Acer AWSMI (i think).

Francois

[chantra (chantra)]

>>What's your method to detect freeze you also use dmesg?
yop, if not "feeling the freeze" by being in front of my comp, this is the way i detect it.

[Travis Reitter (treitter-dev)]

I'm getting the same problems on:

Feisty (fully updated as of this post, and regularly updated in general)
Linux 2.6.20-13-generic
Core 2 Duo
00:1f.1 IDE interface: Intel Corporation 82801G (ICH7 Family) IDE Controller (rev 02)

Every once and a while, some things will hang for 30 seconds (until the driver times out and it does a soft reset).

A few kernel versions back, my Intel WiFi hardware would die sporadically, and I believe I got kernel messages about hard drive issues at the same time. So, these timeouts are better than forced reboots, but it's still really irritating.

Nobody else seems to have mentioned another strange thing I noticed: 'df' lists my data partitions as "hdaN", while only /dev/sda* exist. This could just be because I've got my partitions mounted by UUID, and 'df' assumes PATA if they aren't mounted by device name, or something.

[Christian Mayrhuber (chrismayrhuber)]

Sorry, this timeout setting doesn't work.
The freezes are pretty obvious, everything except the mousepointer seems
to stall.
After that you can see timeout errors in dmesg.

Cheers, Christian

On Mon, 26 Mar 2007 09:02:15 +0200, live_linux_secure
<email address hidden> wrote:

> Hi all i didn't try the "echo "1" >
> /sys/bus/scsi/devices/0:0:0:0/timeout" but i tried to search why it was
> freezing and i saw somewhere (maybe is it stupid?) that it was due to
> the ata_piix driver which control both SATA and PATA. From this i
> thought that i could try to add a boot option like "noata_piix". I
> thought that it wouldn't boot but it booted...
> Sadly it was still freezing. But i tried with a cd in the drive, i'm
> running for 20hours now and no freeze... What's your method to detect
> freeze you also use dmesg?
> I'm running ubuntu feisty beta with the generic kernel 2.6.20-12-generic
> and with Acer AWSMI (i think).
>
> Francois
>

[live_linux_secure (francois-vaningelgom)]

New dmesg output with feisty without cd in the drive;

[ 8364.784000] ata1.01: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x2 frozen
[ 8364.784000] ata1.01: cmd a0/00:00:00:00:20/00:00:00:00:00/b0 tag 0 cdb 0x1e data 0
[ 8364.784000] res 40/00:03:00:00:00/00:00:00:00:00/b0 Emask 0x4 (timeout)
[ 8371.784000] ata1: port is slow to respond, please be patient (Status 0xd0)
[ 8394.800000] ata1: port failed to respond (30 secs, Status 0xd0)
[ 8394.800000] ata1: soft resetting port
[ 8395.144000] ata1.00: ata_hpa_resize 1: sectors = 234441648, hpa_sectors = 234441648
[ 8395.152000] ata1.00: ata_hpa_resize 1: sectors = 234441648, hpa_sectors = 234441648
[ 8395.152000] ata1.00: configured for UDMA/100
[ 8395.332000] ata1.01: configured for UDMA/33
[ 8395.332000] ata1: EH complete
[ 8395.340000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[ 8395.340000] sda: Write Protect is off
[ 8395.340000] sda: Mode Sense: 00 3a 00 00
[ 8395.340000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 8395.752000] SCSI device sda: 234441648 512-byte hdwr sectors (120034 MB)
[ 8395.752000] sda: Write Protect is off
[ 8395.752000] sda: Mode Sense: 00 3a 00 00
[ 8395.752000] SCSI device sda: write cache: enabled, read cache: enabled, doesn't support DPO or FUA

Maybe it could help....

[Santos (tacio-santos)]

I got this problem in xubuntu desktop feisty 7.0.4. Although I know it should not be a long run solution installing debian etch kernel worked for me.

The debian etch kernel packages I've installed were:

kernel-image-2.6-amd64-generic_2.6.18+6_amd64.deb
linux-image-2.6.18-4-amd64_2.6.18.dfsg.1-12_amd64.deb
linux-image-2.6-amd64_2.6.18+6_amd64.deb

[chantra (chantra)]

@Santos:
>>Although I know it should not be a long run solution installing debian etch kernel worked for me

To you mean that debian etch kernel's kernel does not trigger the error?

If so, this is going to be a matter on patching the actual kernel, I would like to confirm this behaviour.

[Santos (tacio-santos)]

Yes, since then (yesterday) feisty 7.0.4 does not freeze anymore. By the way, my partitions are now recognized as /dev/hdaN instead of /dev/sdaN.

[live_linux_secure (francois-vaningelgom)]

I'm sorry but are you using Debian Etch or Ubuntu Edgy?
My disk is a PATA but is actually recognized as SATA.
I hope there is at least or debian or ubuntu which doesn(t freeze.
Thanks for the reply
François

[Santos (tacio-santos)]

>I'm sorry but are you using Debian Etch or Ubuntu Edgy?
>My disk is a PATA but is actually recognized as SATA.
>I hope there is at least or debian or ubuntu which doesn(t freeze.
>Thanks for the reply

I'm using Ubuntu _Feisty 7.0.4_ with debian etch kernel (linux-image-2.6.18-4-amd64_2.6.18.dfsg.1-12_amd64.deb) and my disk is recognized as an ide (i.e /dev/hdaN ).

See:

http://librarian.launchpad.net/7565042/lspci_lsmod.txt
http://librarian.launchpad.net/7560394/dmesg.txt
https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.15/+bug/53754

[live_linux_secure (francois-vaningelgom)]

I don't know how to install a debian kernel on Ubuntu but i think that if i install a Debian Etch, it could lead to the same resut isn't it?
Thank for your reply
François

[chantra (chantra)]

This make more sense now.

It seems that it works better as IDE rather than SATA.

A fix seem to be in the process: https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/84603/comments/30

People who understand french can give a look at http://forums.archlinux.fr/post3592.html#3592

[Santos (tacio-santos)]

@chantra:
thanks a lot, it's much better having the ubuntu kernel working. By the way, the black list method described in
 https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/84603/comments/30
didn't work for me. I've simply run:

rm /lib/modules/2.6.20-15-generic/kernel/drivers/ata/ata_piix.ko
update-initramfs -u -k 2.6.20-15-generic"

@live_linux_secure
you can manually install a kernel downloading the *.deb files and running "dpkg -i file.deb", but the solution above is much better.

cheers

[chantra (chantra)]

@santos:
If I remove ata_piix I can't get my kernel to boot, it hangs on "mounting root filesystem" or something.

Would recommend people to back up their ata_piix.ko ;) and to have another kernel ready to fall back onto just in case.

[live_linux_secure (francois-vaningelgom)]

I also have the same problem as chantra.
Impossible to boot without ata_piix.ko :-(
I'll try debian etch and i'll tell you what
Francois

[chantra (chantra)]

Hi guys,

Refering to bug#84603 : https://bugs.launchpad.net/bugs/84603

There is yet another way to flash your drive.

Note that you need windows.
Got to http://www.toshibaer.com/firmware/index.php?path=TS-L632D/

get the latest package TS-L632D_AC01_Acer_OEM.zip , also get TS-L632D_SC03.zip

Now, we are going to cross flash the drive with the .BIN from TS-L632D_SC03.zip .
In order to get the flasher to accept croos flashing, you need to launch it from the command line with -nocheck switch like:
sfdnwin.exe -nocheck
qnd choose the TS... .BIN file

[Alan Doyle (alandoyle)]

Hey there,
I've an Acer 9303WSMi with a TSST-L632D Drive which is hanging my system. Have to do the Alt-SysRq-o to recover (Nasty forced head parking clunk on my 160Gb SATA drive :( Does anybody have TS-L632D_AC01_Acer_OEM.zip and TS-L632D_SC03.zip because http://www.toshibaer.com/firmware/index.php?path=TS-L632D/ is inaccessible and I really need to stop my system hanging and save my SATA drive.
Alan

[Tomas Šiaulys (tosi)]

Hey,
I didn't have the original file, but I had it extracted, so I zipped it up for you, together with flashing utility for dos. I used those 2 and everything is fine. I don't have acer firmware though, as I have Asus myself. But you don't need it for cross-flashing. Here's the link:
http://twin.lt/~tomas/sc03.zip
I hope it will help.
Tomas

[Alan Doyle (alandoyle)]

No joy. I tried SFDNDOS SC03.BIN sm and it found my drive but said the firmware was incompatible :( Guess I'll have to keep on looking.

[chantra (chantra)]

@Alan

You might want to give a go to the -nocheck switch:
sfdnwin_3.25.4.exe -nocheck

from the command line

[Alan Doyle (alandoyle)]

Finally got around to upgrading my firmware to SC04 with -nocheck option and SUCCESS :)
No more lockups. No more 100% CPU usage. And no more coasters being burnt :)
Fantastic.
Many thanks for all your help.
Sorry I took so long to reply.

[marcologne (marcologne)]

yes, it worked with me on 2.6.22-14-generic, too

I updated firmware SS03 march 2007 using Samsung tool under xp. The firmware upgrade fixed lockups and the 'ata1: port is slow to respond'-problem'. Hardware: Medion MD 98000 CoreDuo T2300 notebook with Intel 82801GBM/GHM SATA AHCI controller, drive to flash:
Samsung SN-S082D SLIM DVD-RW

Still open is that know issue SATA drive on. SATA initializes quite often, every 5 min for some seconds without lockup. I can't find any dmesg event coresponding to that behaviour. See also https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.20/+bug/84603

If someone has a solution for that, please let me know.

http://samsungoms.com/KorLib/File/sfdnwin.exe
http://www.opendrivers.com/driver/237193/samsung-sn-s082d-slim-firmware-ss03-windows-98se-me-2000-xp-free-download.html

[Launchpad Janitor (janitor)]

Beginning with the Hardy Heron 8.04 development cycle, all open Ubuntu kernel bugs need to be reported against the "linux" kernel package. We are automatically migrating this linux-source-2.6.15 kernel bug to the new "linux" package. We appreciate your patience and understanding as we make this transition. Also, if you would be interested in testing the upcoming Intrepid Ibex 8.10 release, it is available at http://www.ubuntu.com/testing . Please let us know your results. Thanks!