linux-mips-fnet
[Top] [All Lists]

Re: current_SC == penguin within interrupt!

To: linux-mips@fnet.fr
Subject: Re: current_SC == penguin within interrupt!
From: Florian Lohoff <flo@mini.gt.owl.de>
Date: Wed, 17 Jun 1998 02:14:24 +0200
In-reply-to: <19980617012442.56315@alpha.franken.de>; from Thomas Bogendoerfer on Wed, Jun 17, 1998 at 01:24:42AM +0200
References: <19980617010049.58002@mini> <19980617012442.56315@alpha.franken.de>
On Wed, Jun 17, 1998 at 01:24:42AM +0200, Thomas Bogendoerfer wrote:
> On Wed, Jun 17, 1998 at 01:00:49AM +0200, Florian Lohoff wrote:
> > My initialiation of the interrupt currently looks like this
> > (grepped from jazz_esp.c)
> > 
> > Whereas i is my current TURBOChannel .....
> > ------------------------------------------------------------------
> > esp->irq = TC0+i;
> > request_irq(TC0+i, esp_intr, SA_INTERRUPT, "PMAZ-AA SCSI", NULL
> > ------------------------------------------------------------------
> > 
> > So ... what did i wrong :) I know ... nearly everything ... 
> 
> looks sane to me. How are the DMA addresses mapped to CPU addresses ?
> Maybe there something fishy in this area, which causes the esp to do

Hmmm ... odd ... i cant reproduce the above ... currently i havnt set
the dma ... all function are nil as the PMAZ-AA is only able
to DMA on the board itself. I enabled all debugging output and
all "functions" should give some output but it stops immediatly when
it should start doing an scsi disk scan.

-------------------------------------------------------------
PMAZ-AA esp_detect called
SCSI ID 7  Clock 40 MHz CCF=0 Time-Out 167 NCR53C9XF(espfast) detected
esp0: FAST chip is fas100a (family=0, version=0)
SCSI ID 7  Clock 40 MHz CCF=0 Time-Out 167 NCR53C9XF(espfast) detected
esp1: FAST chip is fas100a (family=0, version=0)
scsi0 : ESP100A-FAST
scsi1 : ESP100A-FAST
scsi : 2 hosts.
------------------------------------------------------------

Thats it ... nothing more ... 

> strange things. I have to admit, that I didn't see the pengiun panic
> while working on jazz_esp (and there were alot of lockups and panics:-)).

I mostly just copied the jazz_esp things and uncommented all
the vdma things, and got that far :) I think of an irq problem,
the irq things seem not to be so clear on the decstation.

BTW: This is the second time i have a look at the kernel ... :)

Flo
-- 
Florian Lohoff          flo@mini.gt.owl.de              +49-5241-470566

<Prev in Thread] Current Thread [Next in Thread>