linux-scsi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* potential initialisation problem in the aic94xx
@ 2005-11-28 21:52 James Bottomley
  2005-12-06  4:22 ` James Bottomley
  0 siblings, 1 reply; 3+ messages in thread
From: James Bottomley @ 2005-11-28 21:52 UTC (permalink / raw)
  To: Jeff Garzik; +Cc: SCSI Mailing List

I'm not sure if this is a genuine problem, or I just happen to have a
broken card:  Since the card is unusual in being keyed 3.3v only, I
don't actually have any working x86 machine I can try it in.

The symptoms I see on my ia64 box are that as soon as the driver tries
to touch the Region 0 BAR via mmio, the system gives me a MCE.  This is
probably because the card fails to respond on this address and thus
gives some type of PCI bus timeout.  My suspicion is that the card
requires some firmware setup that the on-board BIOS does before the BARs
respond (and since this is an ia64 box, it doesn't run the BIOS).  Could
someone (i.e. Jeff) with access to the Docs have a look and see if this
suspicion can be verified (or someone with a card try it in an ia64 to
see if the same symptoms are observed)?

Thanks,

James



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: potential initialisation problem in the aic94xx
  2005-11-28 21:52 potential initialisation problem in the aic94xx James Bottomley
@ 2005-12-06  4:22 ` James Bottomley
  2005-12-08 20:22   ` Matthew Wilcox
  0 siblings, 1 reply; 3+ messages in thread
From: James Bottomley @ 2005-12-06  4:22 UTC (permalink / raw)
  To: Jeff Garzik; +Cc: SCSI Mailing List

On Mon, 2005-11-28 at 15:52 -0600, James Bottomley wrote:
> I'm not sure if this is a genuine problem, or I just happen to have a
> broken card:  Since the card is unusual in being keyed 3.3v only, I
> don't actually have any working x86 machine I can try it in.

OK, I am sure now.  I've found an x86_64 machine with 3.3v slots and
verified that this card does work.

Therefore, we have a serious bug in the aic94xx driver init code.  It's
clearly relying on something to set it up correctly in the BIOS rather
than doing everything in the driver init routine.  This means (unless
corrected) that the card will only work on x86 like platforms.

James



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: potential initialisation problem in the aic94xx
  2005-12-06  4:22 ` James Bottomley
@ 2005-12-08 20:22   ` Matthew Wilcox
  0 siblings, 0 replies; 3+ messages in thread
From: Matthew Wilcox @ 2005-12-08 20:22 UTC (permalink / raw)
  To: James Bottomley; +Cc: Jeff Garzik, SCSI Mailing List

On Mon, Dec 05, 2005 at 11:22:13PM -0500, James Bottomley wrote:
> Therefore, we have a serious bug in the aic94xx driver init code.  It's
> clearly relying on something to set it up correctly in the BIOS rather
> than doing everything in the driver init routine.  This means (unless
> corrected) that the card will only work on x86 like platforms.

... and it won't work when hotplugged, even on an x86 platform.

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2005-12-08 20:22 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-11-28 21:52 potential initialisation problem in the aic94xx James Bottomley
2005-12-06  4:22 ` James Bottomley
2005-12-08 20:22   ` Matthew Wilcox

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).