FS#4770 - pata drives not detected on promise fasttrak sata/pata controllers

Attached to Project: Arch Linux
Opened by Robert Howard (iBertus) - Saturday, 03 June 2006, 01:42 GMT
Last edited by Judd Vinet (judd) - Tuesday, 25 July 2006, 22:17 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Architecture not specified
Severity Medium
Priority Normal
Reported Version 0.7.2 Gimmick
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

I've noticed that with the stock Arch kernel (2.6.19.15 when last checked) Promise controllers with both PATA and SATA ports do not correctly detect PATA drives. This is fixed with a small patch to the kernel source. I'm not sure if this is fixed in 2.6.17 kernels but I don't think so. I'll attach the patch taken from the RHEL mailing list which corrects this issue.
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Wednesday, 26 July 2006, 08:54 GMT
Reason for closing:  Fixed
Comment by Robert Howard (iBertus) - Saturday, 03 June 2006, 01:44 GMT
Just a note - I've built a kernel with this patch and it does work without issue so far.
Comment by Tobias Powalowski (tpowa) - Saturday, 03 June 2006, 07:19 GMT
if you are not scared of using testing, i wuld like to hear if it's already fixed in 2.6.17.
Comment by Tobias Powalowski (tpowa) - Saturday, 03 June 2006, 07:38 GMT
hi an other option, not so invasive:
http://www.archlinux.org/~tpowa/2.6.17/2.6.17rc5git8.iso
try to burn this as cd and then boot with arch root=<eg. /dev/sda1 yourrootdevic> init s
and then check if your harddrive is working correct.
thanks
greetings
tpowa
Comment by Robert Howard (iBertus) - Saturday, 03 June 2006, 23:22 GMT
I get "unable to open root device" error when trying it from the live cd.
Comment by Tobias Powalowski (tpowa) - Sunday, 04 June 2006, 06:50 GMT
did you add the root= parameter on prompt? else i is no wonder that it is not booting ;)
Comment by Robert Howard (iBertus) - Sunday, 04 June 2006, 17:20 GMT
Yes, I added root=/dev/sda1 but it didn't work. I'll try the testing kernel tonight and let you know how it works.
Comment by Robert Howard (iBertus) - Tuesday, 06 June 2006, 05:19 GMT
The newest kernel from testing doesn't seem to provide PATA support either. Device nodes were not created for the devices, which are created when using the above patch with 2.6.16 kernels. I haven't tried applying it to the 2.6.17 kernel series. I'll give that a try when time permits.
Comment by Tobias Powalowski (tpowa) - Tuesday, 06 June 2006, 08:18 GMT
ok added the patch to 17-rc6 and 16.20
Comment by James Rayner (iphitus) - Tuesday, 06 June 2006, 09:30 GMT
could you test out kernel26beyond for me and see if this is needed?
Comment by Robert Howard (iBertus) - Tuesday, 06 June 2006, 11:53 GMT
I'll test it later today. Also, if someone with a mixture of SATA/PATA drives could connect try this out that would be nice too.
Comment by Robert Howard (iBertus) - Tuesday, 06 June 2006, 13:36 GMT
This patch seems to be already included with beyond. By greping the file I found most of this patch inside the beyond source and it also works with my drives. Perhaps part of the libata update this is talked about on the beyond site?
Comment by Robert Howard (iBertus) - Tuesday, 06 June 2006, 13:49 GMT
The fix is supplied by promise-sata-controller-pata-port-fix.diff in beyond. I'm not sure where it comes from, but it's pretty much the same patch.
Comment by Judd Vinet (judd) - Tuesday, 25 July 2006, 22:17 GMT
Re-opened by Bjørn Lindeijer:

This patch is causing a kernel oops on startup for me. I finally got around to compiling my own kernel without it and now things work fine again. I am using a promise PCI extension card and I only use a SATA port.
Comment by James Rayner (iphitus) - Wednesday, 26 July 2006, 01:40 GMT
please post the full oops.

tpowa: maybe try the equivalent patch from genpatches, it's probably a bit better synced to 2.6.17.
Comment by Thorbjørn Lindeijer (bjorn) - Wednesday, 26 July 2006, 08:29 GMT
Is there any way to post the full oops without copying it down on paper first? It's quite a bunch of numbers.
Comment by Tobias Powalowski (tpowa) - Wednesday, 26 July 2006, 08:30 GMT
could you please try kernel 17.7? i changed the patch to a 2.6.18 backport
Comment by Thorbjørn Lindeijer (bjorn) - Wednesday, 26 July 2006, 08:53 GMT
Alright I tried, and 2.6.17.7 doesn't oops for me!

Loading...