pata_sis problem?

Dr.U · 2007-06-20 16:37:26

I recently installed Arch 2007.05 Duke Linux Tag whatever and am having a problem with a very slow boot process. I did a

pacman -Syu

but that did not change the behavior.

I searched the forum but only found one similar posting about pata_<something-else> rather than pata_sis, but the symptoms seem quite similar: long pauses in the area "... retrying in 5 secs" as given below (from dmesg):

Freeing unused kernel memory: 268k freed
libata version 2.20 loaded.
pata_sis 0000:00:02.5: version 0.5.0
ata1: PATA max UDMA/100 cmd 0x000101f0 ctl 0x000103f6 bmdma 0x0001ff00 irq 14
ata2: PATA max UDMA/100 cmd 0x00010170 ctl 0x00010376 bmdma 0x0001ff08 irq 15
scsi0 : pata_sis
ata1.00: ATA-7: SAMSUNG SP0802N, TK200-04, max UDMA/133
ata1.00: 156368016 sectors, multi 16: LBA48 
ata1.01: ATA-6: HDS722580VLAT20, V32OA60A, max UDMA/100
ata1.01: 160836480 sectors, multi 16: LBA48 
ata1.00: configured for UDMA/100
ata1.01: configured for UDMA/100
scsi1 : pata_sis
ata2.00: ATAPI, max UDMA/33
ata2.00: qc timeout (cmd 0xef)
ata2.00: failed to set xfermode (err_mask=0x4)
ata2: failed to recover some devices, retrying in 5 secs
ata2.00: qc timeout (cmd 0xef)
ata2.00: failed to set xfermode (err_mask=0x4)
ata2.00: limiting speed to UDMA/33:PIO3
ata2: failed to recover some devices, retrying in 5 secs
ata2.00: qc timeout (cmd 0xef)
ata2.00: failed to set xfermode (err_mask=0x4)
ata2.00: disabled
scsi 0:0:0:0: Direct-Access     ATA      SAMSUNG SP0802N  TK20 PQ: 0 ANSI: 5
scsi 0:0:1:0: Direct-Access     ATA      HDS722580VLAT20  V32O PQ: 0 ANSI: 5

A developer replied to the poster and suggested including a boot parameter like this

legacy_ide earlymodules=sis

and to remove the ro, which I did. For the poster the suggestion was for something other than sis since he had something different. I tried substituting the sis in the hopes that it would help but the relevant part of dmesg did not change and the boot process takes just as long as ever.

Does anyone have a suggestion? I really don't know beans about whatever pata_sis might be so provide me a pretty clear answer about what I can try as a solution (meaning something an idiot could probably understand and carry out successfully).

Thanks in advance,

-- Dr.U

-------------------------------------------------------------------------------
"A child of five would understand this! Send someone to fetch a
child of five..." -- Groucho Marx

Dr.U · 2007-06-21 09:19:57

I found the posting from justin in this same section (title: DVD burner won't mount) and thought that maybe the solution that was suggested there would work for my problem after adjusting for the fact that rather than having pata_amd I have pata_sis. The only problem is that where it was suggested for justin to use amd74xx I am unsure about what I should use ...

Does anyone have an idea? Am I on the right track here or am I looking at a possible trashing of my installation?

Thanks in advance,

-- Dr.U

------------------------------------------------------------------------------------
"I've got the brain of a four year old. I'll bet he was glad to be rid of it."
-- Groucho Marx

meh · 2007-06-24 16:44:23

Well, I hope someone responds, because I have the same problem. It's not a major problem of course, just an annoyance to wait through the errors. I haven't figured out if it's affecting anything. Brand new install, as of last night

Jun 24 07:25:41 rolypoly scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
Jun 24 07:25:41 rolypoly <Adaptec aic7890/91 Ultra2 SCSI adapter>
Jun 24 07:25:41 rolypoly aic7890/91: Ultra2 Wide Channel A, SCSI Id=7, 32/253 SCBs
Jun 24 07:25:41 rolypoly
Jun 24 07:25:41 rolypoly scsi 0:0:0:0: Direct-Access     QUANTUM  ATLAS V  9 WLS   0201 PQ: 0 ANSI: 3
Jun 24 07:25:41 rolypoly scsi0:A:0:0: Tagged Queuing enabled.  Depth 32
Jun 24 07:25:41 rolypoly target0:0:0: Beginning Domain Validation
Jun 24 07:25:41 rolypoly target0:0:0: wide asynchronous
Jun 24 07:25:41 rolypoly target0:0:0: FAST-40 WIDE SCSI 80.0 MB/s ST (25 ns, offset 63)
Jun 24 07:25:41 rolypoly target0:0:0: Domain Validation skipping write tests
Jun 24 07:25:41 rolypoly target0:0:0: Ending Domain Validation
Jun 24 07:25:41 rolypoly scsi 0:0:8:0: Direct-Access     QUANTUM  ATLAS IV 9 WLS   0707 PQ: 0 ANSI: 3
Jun 24 07:25:41 rolypoly scsi0:A:8:0: Tagged Queuing enabled.  Depth 32
Jun 24 07:25:41 rolypoly target0:0:8: Beginning Domain Validation
Jun 24 07:25:41 rolypoly target0:0:8: wide asynchronous
Jun 24 07:25:41 rolypoly target0:0:8: FAST-40 WIDE SCSI 80.0 MB/s ST (25 ns, offset 31)
Jun 24 07:25:41 rolypoly target0:0:8: Domain Validation skipping write tests
Jun 24 07:25:41 rolypoly target0:0:8: Ending Domain Validation
Jun 24 07:25:41 rolypoly ata_piix 0000:00:04.1: version 2.10ac1
Jun 24 07:25:41 rolypoly ata1: PATA max UDMA/33 cmd 0x000101f0 ctl 0x000103f6 bmdma 0x0001d800 irq 14
Jun 24 07:25:41 rolypoly ata2: PATA max UDMA/33 cmd 0x00010170 ctl 0x00010376 bmdma 0x0001d808 irq 15
Jun 24 07:25:41 rolypoly scsi1 : ata_piix
Jun 24 07:25:41 rolypoly ata1.00: ATAPI, max UDMA/33
Jun 24 07:25:41 rolypoly ata1.01: ATAPI, max UDMA/66
Jun 24 07:25:41 rolypoly ata1.00: qc timeout (cmd 0xef)
Jun 24 07:25:41 rolypoly ata1.00: failed to set xfermode (err_mask=0x4)
Jun 24 07:25:41 rolypoly ata1: failed to recover some devices, retrying in 5 secs
Jun 24 07:25:41 rolypoly ata1.00: qc timeout (cmd 0xef)
Jun 24 07:25:41 rolypoly ata1.00: failed to set xfermode (err_mask=0x4)
Jun 24 07:25:41 rolypoly ata1.00: limiting speed to UDMA/33:PIO3
Jun 24 07:25:41 rolypoly ata1: failed to recover some devices, retrying in 5 secs
Jun 24 07:25:41 rolypoly ata1.00: qc timeout (cmd 0xef)
Jun 24 07:25:41 rolypoly ata1.00: failed to set xfermode (err_mask=0x4)
Jun 24 07:25:41 rolypoly ata1.00: disabled
Jun 24 07:25:41 rolypoly ata1: failed to recover some devices, retrying in 5 secs
Jun 24 07:25:41 rolypoly ata1.01: failed to set xfermode (err_mask=0x40)
Jun 24 07:25:41 rolypoly ata1: failed to recover some devices, retrying in 5 secs
Jun 24 07:25:41 rolypoly ata1.01: configured for UDMA/33
Jun 24 07:25:41 rolypoly scsi2 : ata_piix
Jun 24 07:25:41 rolypoly scsi 1:0:1:0: CD-ROM            ATAPI    CD-ROM MAX 56X   1.12 PQ: 0 ANSI: 5

meh · 2007-06-24 17:06:39

In this thread http://bbs.archlinux.org/viewtopic.php?id=25431&p=2, I found suggestions for adding earlymodules=piix to the kernel line. I modified /boot/grub/menu.lst kernel line and added it at the end. Couple reboots later, no more hangs. I don't know if this is going to give me troubles later, but for now it's a much quicker boot.

Also, I ran:
# lsmod | grep ata
# mkinitcpio -M
to compare to their output and look at the config files referenced. My results were very, very similar to the others in that thread. I'm also using a Dell Optiplex.

A thread specific to pata_sis that might help you is: http://bbs.archlinux.org/viewtopic.php?id=29352.

Last edited by meh (2007-06-24 17:23:08)

Dr.U · 2007-06-27 10:19:49

Thanks for the tips, meh. However, I already had found the threads you mentioned and had tried the fixes listed, the results were negative: same problems occurring at boot-up. I have since given up and deleted Arch since this is obviously something beyond my capabilities to solve -- and apparently no one else seems capable either

Unfortunate, as Arch used to be my favorite distro, but it seems that in the last year or so I have experienced many difficult-to-remedy problems after updating. I no longer have interest in farting around trying to get things running again or running smoothly (such as with my recent boot problems). I just want a system that is relatively up-to-date with software versions, boots and runs quickly, and doesn't require a Linux guru to keep it that way

Regards,

-- Dr.U

Cotton · 2007-07-12 21:14:28

I've also experienced the same long boot times described in this thread using pata_sis drivers.

Just tried the 2.6.22 kernel in testing in the hope it would improve but it stalls completely now. It says its trying "soft resets" but just gets stuck in a loop accessing ata1.

brain0 · 2007-07-13 07:43:24

Dr.U wrote:

A developer replied to the poster and suggested including a boot parameter like this
legacy_ide earlymodules=sis
and to remove the ro, which I did.

He had obviously no idea what he was talking about. The "legacy-ide" option only has an effect in the installer (which is, the ide drivers are used instead of the pata drivers), and for the "earlymodules" option to work, the specified module (which is called sis5513 btw) has to be present in initramfs, which it isn't in your case. On an already installed system, you have to modify the initramfs image and rebuild it, for which the system has to be booted. I have also experienced instabilities in pata_sis, so here is what you do:

Installation:
* If you do a fresh install on a sis chipset, run the installer with the legacy-ide option.

If you have already installed the system, do this:
* Get the system to boot somehow (that part may be tricky if your system doesn't boot at all, but we can discuss that later)
* Edit /etc/mkinitcpio.conf: Change the MODULES= line from MODULES="ata_generic pata_sis" to MODULES="sis5513". In the HOOKS line, replace "pata" with "ide".
* Edit /etc/fstab and /boot/grub/menu.lst: Replace all occurences of /dev/sdXY with /dev/hdXY.
* Run "mkinitcpio -p kernel26"
* Reboot

Now you use the old and more stable ide driver.

BTW, someone should report these pata_sis instabilities as bugs or regressions to the kernel mailing list.

Arch Linux

#1 2007-06-20 16:37:26

pata_sis problem?

#2 2007-06-21 09:19:57

Re: pata_sis problem?

#3 2007-06-24 16:44:23

Re: pata_sis problem?

#4 2007-06-24 17:06:39

Re: pata_sis problem?

#5 2007-06-27 10:19:49

Re: pata_sis problem?

#6 2007-07-12 21:14:28

Re: pata_sis problem?

#7 2007-07-13 07:43:24

Re: pata_sis problem?

Board footer