[6232] in linux-scsi channel archive
Re: new ncr/sym53c8xx driver tree at tux
daemon@ATHENA.MIT.EDU (Matthew Jacob)
Wed Mar 31 14:54:06 1999
Date: Wed, 31 Mar 1999 11:27:24 -0800 (PST)
From: Matthew Jacob <mjacob@feral.com>
Reply-To: mjacob@feral.com
To: Gerard Roudier <groudier@club-internet.fr>
cc: linux-scsi@vger.rutgers.edu, ncr53c810@Colorado.EDU
In-Reply-To: <Pine.LNX.3.95.990331193703.399A-100000@localhost>
>
> > scsi : aborting command due to timeout :
> > pid 54, scsi2, channel 0, id 0, lun 0 Test Unit Ready 00 00 00 00 00
>
> The failure occurs on the first TEST_UNIT_READY sent to a device of this
> SCSI BUS. That is simply strange, since the device should be quick at
> responding, and this command does not need any DATA PHASE.
> At this step, the driver does not try to negotiate anything and the SCSI
> BUS has just been reset, so the failure seems double-strange to me.
> And The fact that the first 875 just succeeds makes your problem appear
> triple-strange to me. ;-)
It's auto-request sense that's probably failing for post TUR?
>
> > ncr53c876-1: rev=0x14, base=0x1ff80014000,
> > io_port=0xfffff9fe02010c00, irq=6,7dc
> > ncr53c876-2: rev=0x14, base=0x1ff80016000,
> > io_port=0xfffff9fe02011000, irq=6,7dc
>
> The both functions of the 876 report same IRQ thing that probably means
> that they are sharing this thing.
Right- it's a Symbios 2-chip board- instead of being bridged like say an
Adaptec, they report as 2 separate device functions.
>
> In the Solaris boot-up messages, I donnot see what PCI ressources have
> been assigned to the PCI devices. Were they the same ones ?
I'll check- it'll be a bit hard to determine because the specific values
are all buried in prom range goop.
>
> If I had to guess, I would think that some interrupt that should go
> to the driver has just been lost prior to calling the driver interrupt
> routine.
Possibly. I'll check.
>
> Could you for example, boot your system with an option that excludes one
> of the 876 functions, a follow:
>
> ncr53c8xx=excl:#io_address
>
> Replace #io_address by the corresponding 0xffffff################ of
> one function that seems as real an io address as Windows/NT is an O/S ;-)
>
> You can exclude more than 1 controller, using
>
> ncr53c8xx=excl:#io1,excl:#io2
>
> You also could give a try with the other driver (sym53c8xx) that can be
> configured independantly as either dynamically loadable module or
> statically linked to the kernel, but only shares some CONFIG defines with
> the other driver.
>
> There is also some README.ncr53c8xx file that is not excellent english but
> contains some informations, especially on boot options, as debug ones.
> This file is also intended to help users refining diagnostics before
> reporting problems.
Ah! I'll try and look at that before reporting again.
>
> With the informations you provide, I am not able to tell you more about
> your problem.
>
Will do! I really need this to work as it's a shared configuration that I
need to leave running so I can remotely reboot linux or solaris as
needed.. I'll also evaluate a Ultra2 LVD 53c895 too and let you know how
it goes...
Thanks for the response- I'll try and get you more complete information.
-matt
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
