[3106] in linux-scsi channel archive
Re: Problem with NCR 53C810 an Linux 2.0.29
daemon@ATHENA.MIT.EDU (Kurt Garloff)
Wed Jan 21 05:26:38 1998
Date: Wed, 21 Jan 1998 11:18:06 +0100 (CET)
From: Kurt Garloff <garloff@kg1.ping.de>
To: Stephan Voit <voit@wi-inf.uni-essen.de>
cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <3.0.1.16.19980118225228.2c375334@wi-inf.uni-essen.de>
On Sun, 18 Jan 1998, Stephan Voit wrote:
> My on-board NCR-SCSI-Controller (NCR 53C810) isnīt detected with Linux 2.0.29
> (german distribution from Delix (DLD 5.2)). With Linux 1.2.x (DLD 2.1) there
> are no problems.
>
> My mother board is a SA486P AIO-U (STD) with Phoenix BIOS, bought in 08/94).
> Its PCI bridge is jumpered to IRQ 15 and INT A.
> ...
> >From boot with Linux 2.0.19:
>
> pcibios_init : PCI BIOS revision 2.00 entry at 0xf0100
> ...
> Linux version 2.0.29 (root@erdbeere.delix.de) (gcc version 2.7.2.1) #1 Tue May 6 17:40:22 MET DST 1997
> scsi-ncr53c7,8xx : at PCI bus 0, device 1, function 0
> scsi-ncr53c7,8xx : NCR53c810 at memory 0xd8000000, io 0xc000, irq 15
> ...
> scsi0 : test 1 started
> scsi0 : driver test 1 timed out due to lost interrupt.
> Please verify that the correct IRQ is being used for your board,
> and that the motherboard IRQ jumpering matches the PCI setup on
> PCI systems.
> If you are using a NCR53c810 board in a PCI system, you should
> also verify that the board is jumpered to use PCI INTA, since
> most PCI motherboards lack support for INTB, INTC, and INTD.
>
> --------------------------
> ncr53c810-0: using memory mapped IO at virtual address 0x181b000
> ncr53c810-0: assuming 40MHz clock
> ncr53c810-0: initial value of SCNTL3 = 03, final = 13
> ncr53c810-0: requesting shared irq 15 (dev_id=0xbb8068)
^^^^^^^^^^^^^
> ncr53c810-0: request irq 15 failure
> ncr53c810-0: releasing memory mapped IO region 181b000[128]
> ncr53c810-0: releasing IO region c000[128]
Did you try to disable the second IDE interface in your BIOS setup? It
will probably use IRQ 15 otherwise. Did you set the PCI IRQ reporting
scheme to Level triggered? This is necessary to be able to share IRQs.
There should also be a possibility to disable the second IDE port by
passing a parameter to the kernel. Hmm, let me have a look. ...
(linux/drivers/block/ide.c: 3000+) ide1=noprobe should work.
Kurt Garloff, Dortmund
<K.Garloff@ping.de>
PGP key on http://www.student.physik.uni-dortmund.de/homepages/garloff
