2002-06-23 19:45:31

by Marc Lefranc

[permalink] [raw]
Subject: aha152x driver broken in 2.4.19-pre10


Hi,

it looks like the aha152x driver was broken between 2.4.19-pre8 and
2.4.19-pre10. Apparently there is a problem with a lost interrupt. I
am afraid I cannot do much to fix this problem, but I am willing to
perform any test that would help.

All the best,
Marc

P.S. As I'm not subscribed to the mailing list, thanks in advance for
cc'ing the comments to me.

Jun 23 20:48:38 socrate kernel: Linux version 2.4.19-pre10 ([email protected]) (gcc version 2.96 20000731 (Red Hat Linux 7.3 2.96-110)) #5 Wed Jun 19 21:44:39 CEST 2002
[...]
Jun 23 20:49:17 socrate kernel: SCSI subsystem driver Revision: 1.00
Jun 23 20:49:18 socrate kernel: aha152x: BIOS test: passed, detected 1 controller(s)
Jun 23 20:49:18 socrate kernel: aha152x: resetting bus...
Jun 23 20:49:18 socrate kernel: aha152x0: vital data: rev=1, io=0x140 (0x140/0x140), irq=11, scsiid=7, reconnect=enabled, parity=enabled, synchronous=enabled, delay=1000, extended translation=disabled
Jun 23 20:49:18 socrate kernel: aha152x0: trying software interrupt, lost.
Jun 23 20:49:18 socrate kernel: aha152x0: IRQ 11 possibly wrong. Please verify.Jun 23 20:49:18 socrate insmod: /lib/modules/2.4.19-pre10/kernel/drivers/scsi/aha152x.o: init_module: No such device
Jun 23 20:49:18 socrate insmod: Hint: insmod errors can be caused by incorrect module parameters, including invalid IO or IRQ parameters
Jun 23 20:49:18 socrate insmod: /lib/modules/2.4.19-pre10/kernel/drivers/scsi/aha152x.o: insmod scsi_hostadapter failed

And going back to the kernel previously in use on this machine (-pre8
+ Andrea Arcangeli's vm 34 patch applied) :

Jun 23 21:02:41 socrate kernel: Linux version 2.4.19-pre8-vm34 ([email protected]) (gcc version 2.96 20000731 (Red Hat Linux 7.1 2.96-98)) #1 Sun May 5 22:37:18 CEST 2002
[...]
Jun 23 21:03:11 socrate kernel: SCSI subsystem driver Revision: 1.00
Jun 23 21:03:12 socrate kernel: aha152x: BIOS test: passed, detected 1 controller(s)
Jun 23 21:03:12 socrate kernel: aha152x: resetting bus...
Jun 23 21:03:12 socrate kernel: aha152x0: vital data: rev=1, io=0x140 (0x140/0x140), irq=11, scsiid=7, reconnect=enabled, parity=enabled, synchronous=enabled, delay=1000, extended translation=disabled
Jun 23 21:03:12 socrate kernel: aha152x0: trying software interrupt, ok.
Jun 23 21:03:12 socrate kernel: scsi0 : Adaptec 152x SCSI driver; $Revision: 2.4 $



2002-06-24 06:52:17

by Martin Diehl

[permalink] [raw]
Subject: Re: aha152x driver broken in 2.4.19-pre10

On 23 Jun 2002, Marc Lefranc wrote:

> it looks like the aha152x driver was broken between 2.4.19-pre8 and
> 2.4.19-pre10. Apparently there is a problem with a lost interrupt. I
> am afraid I cannot do much to fix this problem, but I am willing to
> perform any test that would help.

Hi,

same for me. AFAICS when looking into the diff which got into -pre10 the
only change around the place where the lost interrupt is detected is a
dropped spin_lock_irq() around a 1sec mdelay() while probing the
interrupt. The interesting observation however is with 2.5.24 which has
exactly the same code as 2.4.19-pre10 and - surprise - does work for me
just fine. So I don't have a good idea what might cause the problem with
2.4.19-pre10...

Martin