[1184] in linux-scsi channel archive
Re: Problems with DPT, scsi and 2.0.27 (fwd)
daemon@ATHENA.MIT.EDU (Dario_Ballabio@milano.europe.dg.co)
Fri Dec 27 06:35:39 1996
Date: Fri, 27 Dec 1996 12:33:01 +0100
From: Dario_Ballabio@milano.europe.dg.com
To: linux-scsi@vger.rutgers.edu
A problem raised by Bob in this mailing list has been fixed using
the version of the EATA driver that I posted here some time ago.
If anybody can reproduce a problem using the EATA driver, feel free
to contact me directly.
Cheers,
db
************************ START OF FORWARDED MESSAGE **********************
From : mt>rmshep
To : dario
Subject : Re: Problems with DPT, scsi and 2.0.27 (fwd)
Date : 12/19/96 05:14
In-Reply-To: <199612181045.LAA20878@milano.europe.dg.com.>
First thanks again for your help.
I first tried compiling the kernel with your driver unmodified. It failed
to recognize the DPT card (see the meessages below.
> I noticed that you have a strange baseio address (0x0510). I would not
> expect a pci card having an address below 0x0fff, so maybe the
I don't know what to say to this. I bought this card more that 18 months
ago. It has alwasy worked flawlessly. I don't recall setting any jumpers
or software when I installed it. I don't know who the baseio address got
set. Maybe I should call DPT for a bios upgrade??
> driver get confused and thinks that you have an ISA board.
> Pls try this quick and dirty fix:
> at line 144 in eata.c replace
> #define ISA_RANGE 0x0fff
> with the following:
> #define ISA_RANGE 0x03ff
This worked!! (see below.) I booted without any problems, there is one
message about address 0x1f0 in use. All of the units on the scsi bus were
correctly recognized.
AND!! ta da everything worked. The backups worked flawlessly.
Interestingly, I was going to back up to a second machine so I created a
200MB tarball of the data files and with the eata_dma driver, ftp dumped
core everytime it got to 70+MB on the tarball with a message about not
reading the drive. Your driver never had a problem with this either.
> An alternative workaround could be to force an EISA address to your board,
> and in this case the driver must work as it is now.
> If you let me know how it works, I can try to put in the driver the
> best solution for your configuration.
Unless, there is some other problem lurking The Machine is now working
correctly as near as I can tell.
Thanks _LOTS_ for your quick help.
bob shepard
