RHEL4 and AIC7899

[Rick Stevens]

On Wed, 2005-11-16 at 20:59 +0100, Oliver wrote:
> >> >> >> >> during the first boot from the install CD of RHEL4AS, die system hangs
> >> >> >> >> while trying the module aic7xxx. It's installed an AIC7899 dualchannel
> >> >> >> >> onboard in the server (IBM IntelliPro M [6850]). Ther are no IDE
> >> >> >> >> harddisk
> >> >> >> >> devices in the computer.
> >> >> >> >
> >> >> >> > It sometimes takes a while for the aic7xxx driver to load up on the
> >> >> >> > initial install (it has to do a SCSI bus reset, then a spinup, then a
> >> >> >> > probe of the drives).  This has been known to take over three minutes
> >> >> >> > in some cases.  Did you wait that long?
> >> >> >> 
> >> >> >>
> >> >> >> I've waited more than 30 minutes... Just tested the Server with other
> >> >> >> Linux versions like Knoppix_3.8 and with WindowsXP. Knoppix took 30
> >> >> >> seconds for probing...
> >> >> >> 
> >> >> >> So I don't know, why RHEL4 doesn't recognize the controller correctly.
> >> >> 
> >> >> > Hmmm.  Have you tried installing in text mode and checking the output
> >> >> > on the other three consoles (ALT-F2, ALT-F3, ALT-F4) to see if there's
> >> >> > something displayed there that might give a clue?
> >> >> > You didn't say if the CD was IDE or not (it may well be).  If that's the
> >> >> > case, you might also try installing with DMA disabled for the IDE bus:
> >> >> >         boot: linux ide=nodma
> >> >> 
> >> >> Now I downloaded the RHEL4AS Upd2 CDs and tried to install from them. I've
> >> >> started the installation at the boot prompt with    linux ide=nodma text
> >> >> There is one CD-ROM drives on each IDE channel.
> >> >> 
> >> >> Like before, it seems(?) to hang.
> >> >> Here are the screens...
> >> >> 
> >> >> Installation screen shows:
> >> >>   loading aic7xxx driver
> >> >> Alt-F3 shows:
> >> >>   loaded e100 from /modules/
> >> >>   loaded aic7xxx from /modules
> >> >>   inserted /tmp/mii.ko
> >> >>   inserted /tmp/e100.ko
> >> >> Alt-F4 shows:
> >> >>   (6)e100: eth0: e100_probe: addr ...
> >> >>   (6)ACPI: PCI interrupt 0000:03:01.0[A] -> GSI 10 (level low) -> IRQ 10
> >> >>   (6)ACPI: PCI interrupt 0000:03:01.1[B] -> GSI 10 (level low) -> IRQ 10
> >> >>   (6)scsi0: Adaptec AIC7xxx EISA/VLB/PCI SCSI HBA DRIVER, Rev 6.2.36
> >> >>   (4)       (Adaptec aic7899 Ultra160 SCSI adapter)
> >> >>   (4)       aic7899: Ultra160 Wide Channel A, SCSI id=7, 32/253 SCBs
> >> >>   (4)
> >> >> 
> >> >> Hope you can help me and you have an idea :)!?
> >> 
> >> > Well, it's obvious the driver loaded and recognizes your controller.
> >> > It's NOT seeing the drives, however, and that's where the problem lies.
> >> 
> >> > Stupid question, but are you 100% certain you a) have the SCSI bus
> >> > terminated correctly, and b) have the drives spinning up?  Get into the
> >> > SCSI BIOS and make sure it starts the drives at power up (disable the
> >> > "SCSI START/STOP" stuff).
> >> 
> >> Rick
> >> When  I  wait  for  about  half  an  hour  I  get error messages about not
> >> responding drives. He probes all IDs and at everyone an error appears that
> >> the reset didn't work...
> >> 
> >> But  when  I  boot with a 2.4 kerner version (I think the Knoppix v3.8 had
> >> also  the  old  kernel),  the  recognization  of the drives needs about 10
> >> seconds.
> >> 
> >> The system is terminated as followed
> >> Channel A: Host controller and End of SCSI cable
> >> Channel B (extern): Only Host controller
> >> 
> >> I can remember have read about the external termination at IBM website. It
> >> should be fine if the termination on the backside is only on with external
> >> devices on the bus. Is this correct?
> 
> > Your termination seems OK.  This is a very weird problem.  Since you're
> > using RHEL4, have you contacted Red Hat support regarding this issue?
> > Remember, this list isn't official...we're just a bunch of Red Hat users
> > helping each other out.
> 
> Rick
> My  support  contract  is  just  the  Knowledgebase,  no phone or personal
> support. Because of that I haven't contacted RH yet.
> 
> I've booted the system with the actual Knoppix v4 (Kernel 2.6.12) and this
> version  work  also  very  well.  What  the  hell has RH edited in the AS4
> kernel?

AS4 also uses the 2.6 kernel, but with Red Hat tweaks.  I've heard of
some issues with their aic7xxx driver, but nothing this drastic.

I did find a similar bug in bugzilla:

https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=150548

It looks to be very similar to your situation and the fix was a BIOS
update to the motherboard.  I think you said that it's an IBM machine.
Check the IBM site for BIOS updates and apply them.

----------------------------------------------------------------------
- Rick Stevens, Senior Systems Engineer     rstevens at vitalstream.com -
- VitalStream, Inc.                       http://www.vitalstream.com -
-                                                                    -
-            We look for things.  Things that make us go!            -
----------------------------------------------------------------------