[322] in linux-scsi channel archive
Re: NCR 53c400 problems
daemon@ATHENA.MIT.EDU (Richard Gooch)
Wed Jul 5 06:22:47 1995
Date: Wed, 5 Jul 1995 16:05:14 +1000
From: Richard Gooch <rgooch@atnf.CSIRO.AU>
To: Kevin Lentin <kevinl@cs.monash.edu.au>
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <199507050229.MAA28685@fangorn.cs.monash.edu.au>
Kevin Lentin writes:
> Richard Gooch Wrote ...
> >
> > Hi, all. I'm having problems with my new, no-name ISA SCSI card with
> > an NCR 53c400A chip. This board is jumpered to I/O port 0x350 and IRQ 5.
> > When the system boots up, the SCSI BIOS correctly probes my SCSI
> > drive. I can correctly FDISK and FORMAT my drive under DOG (and even
> > boot DOG from the drive), so the controller appears to work.
> >
> > However, when I boot Linux (v1.1.72 with Kevin Lentin's 53c400.patch
> > (version 1(beta)) file), I get the following message on bootup:
> >
> > scsi0: at port 856 irq 5 options CAN_QUEUE=16 CMD_PER_LUN=2 release=1 generic
> > options AUTOPROBE_IRQ AUTOSENSE PSEUDO DMA generic release=6 ncr53c400 release=1
> > scsi0: Generic NCR5380/NCR53C400 Scsi Driver
> > scsi : 1 hosts.
>
> I thought my patch included the change that made that 856 come out in hex.
> Maybe not.
> >
> > and then the system hangs (waits forever:-/). It doesn't matter if
> > there is a SCSI drive on the bus or not: either way it hangs. Does
> > anyone have any idea what may be going wrong? I'm starting to get
> > desperate and considering parting with more money for an Adaptec 1522
> > :-(
>
> Weird. I've never had this on my card.
>
> Is everything terminated right? If all you have is the card and the drive
> then the terminating resisters should be installed in the card and the
> drive should be terminated? Have you tried playing with the terminating
> power settings on the drive? I've found that DOG is willing to accept or
> reject my SCSI setup based on different criteria to my linux system.
Yes, all is terminated correctly. I've also tried other drives. No
luck.
> > BTW: when I type "ncr53c400=0x300,7" at the LILO prompt, the above
> > message (...at port 856 irq 5) does not change: I can't seem to
> > override these parameters (but this is not as important as getting the
> > controller to work under Linux:-)
>
> Check in the Makefile in drivers/scsi. I compile in the 0x350,5 in there,
> although the lilo stuff should still override it. I'll check it out at home
> tonight.
>
> Although if you card is jumpered at 0x350,5 then it should be fine. Have
> you tried a kernel without my patches, using ncr5380=0x358,-1
> See if that works.
I've just done this (with v1.2.8-ELF) and it seemed to work. But
when I tried to read from the drive (dd 5 MBytes in 1 MByte blocks),
it failed with a phase error. I'm about to run off and try an Adaptec
1522 (I have to get this machine going soon).
Regards,
Richard....
