Red Hat Bugzilla – Bug 174973
Kernel PCI-X breaks on kernel >=2.6.14 if APCI MCFG is enabled.
Last modified: 2015-01-04 17:23:31 EST
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.8) Gecko/20051204 Fedora/1.5-1 Firefox/1.5
Description of problem:
I'm just reboot my FC4 x86-64 machine after updating to the latest kernel (1644).
The machine fails to boot. (attempted to kill init panic)
I've modified the 1644 initrd image, adding some "sleep 10s" so I can see what's broken.
Seems that the the aic7xxx module loads just fine, but does not scan the SCSI bus for devices and as a result, my root MD5/LVM RAID doesn't get detected; no root device; panic.
I should further add that machine was rock solid up until (and including) 2.6.13-1.1532smp.
Tyan Thunder K8WE.
BIOS version v1.01.
2 x Opteron 248.
4 x 512MB PC3200/ECC/REG.
Adatec 29160/64 in a PCI-X 100Mhz slot.
4 x 10K RPM, 36GB, 68p drives in software RAID5.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Install 1644 kernel.
2. Reboot machine.
Actual Results: 1. SCSI driver does not scan bus and does not detect devices.
Expected Results: 1. Scan SCSI bus.
2. Detect 4 SCSI drives on bus.
3. Detect software RAID5.
4. Detect LVM.
5. Mount root.
Might relate to (or duplicate):
Just tried -1653.
Driver loads, AFAIK does not detect the SCSI device (and as a result, does not
scan the SCSI bus), init dies.
After contacting Tyan, It was suggested that I update the BIOS and disable the
ACPI MCFG table by selecting (and here comes the duh!) Linux instead of other
in the OS type.
I'm leaving the bugzilla.kernel.org bug open (just in-case it ACPI MCFG parsing
is broken). Should we leave this one open too?
We'll pick it up when upstream gets it within short time, so there's little
point leaving this one open. Thanks for the update though.