Bug 110543

Summary: regression: kernel fails to detect ICH5 SATA controller
Product: [Retired] Red Hat Raw Hide Reporter: Kaj J. Niemi <kajtzu>
Component: kernelAssignee: Jeff Garzik <jgarzik>
Status: CLOSED UPSTREAM QA Contact: Brian Brock <bbrock>
Severity: high Docs Contact:
Priority: medium    
Version: 1.0CC: peterm
Target Milestone: ---   
Target Release: ---   
Hardware: All   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2004-03-03 08:22:37 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
Output of lspci while running 2.4.20-20.9
none
Output of lspci -vvv under 2.4.20-20.9
none
dmesg output under 2.4.20-20.9
none
Output of lspci while running 2.4.22-1.2115.nptlsmp
none
Output of lspci -vvv while running 2.4.22-1.2115.nptlsmp
none
Output of dmesg while running 2.4.22-1.2115.nptlsmp (ICH5 lines missing)
none
dmesg while running 2.4.22-1.2138.nptlsmp none

Description Kaj J. Niemi 2003-11-20 23:54:35 UTC 
Description of problem:
rawhide kernel 2.4.22-1.2115.nptlsmp fails to detect ICH5 SATA
controller resulting in horrible performance.
2.4.21-20.1.2024.2.1.nptlsmp seems to exhibit the same behavior.. I
didn't have any interim 2.4.22-1.19xx/2xxx kernels anymore to try with.

Using an older RH9 kernel (2.4.20-20.9) works just fine.

Attached is lspci and dmesg while running under 2.4.20-20.9.

Version-Release number of selected component (if applicable):
2.4.22-1.2115.nptlsmp, 2.4.21-20.1.2024.2.1.nptlsmp, 2.4.20-20.9smp

How reproducible:
Always


Regression -> Severity: high
Comment 1 Kaj J. Niemi 2003-11-20 23:55:31 UTC 
Created attachment 96106 [details]
Output of lspci while running 2.4.20-20.9
Comment 2 Kaj J. Niemi 2003-11-20 23:56:15 UTC 
Created attachment 96107 [details]
Output of lspci -vvv under 2.4.20-20.9
Comment 3 Kaj J. Niemi 2003-11-20 23:57:19 UTC 
Created attachment 96108 [details]
dmesg output under 2.4.20-20.9
Comment 4 Kaj J. Niemi 2003-11-21 00:07:21 UTC 
Created attachment 96109 [details]
Output of lspci while running 2.4.22-1.2115.nptlsmp
Comment 5 Kaj J. Niemi 2003-11-21 00:08:33 UTC 
Created attachment 96110 [details]
Output of lspci -vvv while running 2.4.22-1.2115.nptlsmp
Comment 6 Kaj J. Niemi 2003-11-21 00:10:52 UTC 
Created attachment 96111 [details]
Output of dmesg while running 2.4.22-1.2115.nptlsmp (ICH5 lines missing)
Comment 7 Kaj J. Niemi 2003-11-21 14:36:09 UTC 
Hmmm might the problem be related to the fact that in 2.4.22 SATA is
part of the ata_piix module (drivers/ide/pci/pci.c, line 880)
while in 2.4.20 they're still part of the kernel?

Attempting to modprobe ata_piix locks the system up. ;-)
Comment 8 Kaj J. Niemi 2003-12-16 20:47:44 UTC 
The problem is still there with 2.4.22-1.2135.nptlsmp
Comment 9 Jeff Garzik 2004-01-12 18:55:55 UTC 
You need to set your BIOS to enhanced mode, before loading ata_piix
driver.  Does that solve the lockups?
Comment 10 Kaj J. Niemi 2004-01-14 13:19:15 UTC 
Created attachment 96968 [details]
dmesg while running 2.4.22-1.2138.nptlsmp

Jeff,

Yes and no. ;-) I had time to kick around with this particular problem so
here's what I found out when installing from scratch:

When kickstaring FC1 installations (kernel-2.4.20-1.2115) in enhanced mode the
first thing happening is that anaconda attempts to load ata_piix and the system
freezes.

Changing to AUTO mode (I guess this becomes combined?) loads the ata_piix
module, the drive shows up as /dev/sda and the installation succeeds.
Read/write speeds are also satisfactory

After installation if I change from auto/combined to enhanced mode everything
works and there are no lockups.

I'll wait for FC2 betas, try installing in enhanced mode and see what happens.