* cdrom
@ 1998-08-31 16:56 Ulf Carlsson
1998-08-31 22:19 ` cdrom Thomas Bogendoerfer
0 siblings, 1 reply; 11+ messages in thread
From: Ulf Carlsson @ 1998-08-31 16:56 UTC (permalink / raw)
To: linux
Hi,
Has someone managed to mount a CD yet?
Unable to handle kernel paging request at virtual address 00000000, epc ==
88021bcc, ra == 880f51c4 Oops: 0000
....
epc : 88021bcc
Status: 3004fc02
Cause : 00000008
Segmentition fault
Then i get the prompt back.
The second time I try:
scsi: aborting command due to timeout : pid 665, scsi0, channel 0, id4,
lun 0 Test Unit Ready 00 00 00 00 00
scsi0: Aborting connected command 665 - stopping DMA sending wd33c93 ABORT
command - flushing fifo - asr=00, sr=16, 0 bytes un-transferred
(timeout=-1) - sending wd33c93 DISCONNECT command - asr=00, sr=16.page
fault from irq handler: 0000
...
epc : 88021bcc
Status: 1004fc02
Cause : 00000008
Aiee killing interrupt handler
Kernel panic: Attempted to kill the idle task!
In swapper task - not syncing
Really lots of error messages.
Well, I was thinking if I could boot from the cd and install without
networking, doesn't work very well though.
- Ulf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-08-31 16:56 cdrom Ulf Carlsson
@ 1998-08-31 22:19 ` Thomas Bogendoerfer
1998-08-31 23:26 ` cdrom ralf
1998-09-01 14:55 ` cdrom ralf
0 siblings, 2 replies; 11+ messages in thread
From: Thomas Bogendoerfer @ 1998-08-31 22:19 UTC (permalink / raw)
To: Ulf Carlsson; +Cc: linux
On Mon, Aug 31, 1998 at 06:56:10PM +0200, Ulf Carlsson wrote:
> Hi,
> Has someone managed to mount a CD yet?
my Indy doesn't have a CDrom drive. But it works on my M700. So it's
probably related to the scsi low level driver. Ralf mentioned some
problems with DAT, but that could also be a generic problem.
> scsi: aborting command due to timeout : pid 665, scsi0, channel 0, id4,
> lun 0 Test Unit Ready 00 00 00 00 00
and CDrom drive works with IRIX ?
Thomas.
--
See, you not only have to be a good coder to create a system like Linux,
you have to be a sneaky bastard too ;-)
[Linus Torvalds in <4rikft$7g5@linux.cs.Helsinki.FI>]
^ permalink raw reply [flat|nested] 11+ messages in thread* Re: cdrom
1998-08-31 22:19 ` cdrom Thomas Bogendoerfer
@ 1998-08-31 23:26 ` ralf
1998-09-01 14:55 ` cdrom ralf
1 sibling, 0 replies; 11+ messages in thread
From: ralf @ 1998-08-31 23:26 UTC (permalink / raw)
To: Thomas Bogendoerfer, Ulf Carlsson; +Cc: linux
On Tue, Sep 01, 1998 at 12:19:11AM +0200, Thomas Bogendoerfer wrote:
> On Mon, Aug 31, 1998 at 06:56:10PM +0200, Ulf Carlsson wrote:
> > Hi,
> > Has someone managed to mount a CD yet?
>
> my Indy doesn't have a CDrom drive. But it works on my M700. So it's
> probably related to the scsi low level driver. Ralf mentioned some
> problems with DAT, but that could also be a generic problem.
DAT and a couple more of the wd33c93 driver options available via the
kernel command line. I wouldn't wonder if a CDROM or second disk results
in the same sick effects.
Somebody recently offered to take a look on driver issues, any success
yet?
Ralf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-08-31 22:19 ` cdrom Thomas Bogendoerfer
1998-08-31 23:26 ` cdrom ralf
@ 1998-09-01 14:55 ` ralf
1998-09-01 17:50 ` cdrom Ulf Carlsson
1998-09-01 19:19 ` cdrom Alex deVries
1 sibling, 2 replies; 11+ messages in thread
From: ralf @ 1998-09-01 14:55 UTC (permalink / raw)
To: Thomas Bogendoerfer, Ulf Carlsson; +Cc: linux
On Tue, Sep 01, 1998 at 12:19:11AM +0200, Thomas Bogendoerfer wrote:
> On Mon, Aug 31, 1998 at 06:56:10PM +0200, Ulf Carlsson wrote:
> > Hi,
> > Has someone managed to mount a CD yet?
>
> my Indy doesn't have a CDrom drive. But it works on my M700. So it's
> probably related to the scsi low level driver. Ralf mentioned some
> problems with DAT, but that could also be a generic problem.
>
> > scsi: aborting command due to timeout : pid 665, scsi0, channel 0, id4,
> > lun 0 Test Unit Ready 00 00 00 00 00
>
> and CDrom drive works with IRIX ?
The problems seems to be associated with either sync SCSI or disconnect/
reconnect. I did a two line modification to sgiwd93.c which disables
sync and disconnect/reconnect. I now have since about a hour running:
dd if=/dev/sda of=/dev/zero &
dd if=/dev/sdb of=/dev/zero &
dd if=/dev/sdc of=/dev/zero &
find / /ext -xdev | cpio -o -H crc -F /dev/nst0 --verbose
No problem so far. (Except that SCSI performance even with only single
disk activity is further converging to zero ...)
Ralf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 14:55 ` cdrom ralf
@ 1998-09-01 17:50 ` Ulf Carlsson
1998-09-01 18:05 ` cdrom Alex deVries
1998-09-01 22:08 ` cdrom ralf
1998-09-01 19:19 ` cdrom Alex deVries
1 sibling, 2 replies; 11+ messages in thread
From: Ulf Carlsson @ 1998-09-01 17:50 UTC (permalink / raw)
To: ralf; +Cc: Thomas Bogendoerfer, linux
On Tue, 1 Sep 1998 ralf@uni-koblenz.de wrote:
> On Tue, Sep 01, 1998 at 12:19:11AM +0200, Thomas Bogendoerfer wrote:
>
> The problems seems to be associated with either sync SCSI or disconnect/
> reconnect. I did a two line modification to sgiwd93.c which disables
> sync and disconnect/reconnect. I now have since about a hour running:
>
> dd if=/dev/sda of=/dev/zero &
> dd if=/dev/sdb of=/dev/zero &
> dd if=/dev/sdc of=/dev/zero &
> find / /ext -xdev | cpio -o -H crc -F /dev/nst0 --verbose
>
> No problem so far. (Except that SCSI performance even with only single
> disk activity is further converging to zero ...)
I tried to implement your ideas, I disabled sync, I don't know how you
disable disconnect/reconnect. Anyway, I don't think that's the problem.
The problem is the sigsegv in mount the oops causes, do we have something
like a 'half' mounted device?
Unable to handle kernel paging request at virtual address 00000000, epc ==
88021bcc, ra == 8809414
Oops: 0000
...
epc : 88021bcc
Status: 3004fc02
Cause : 00000008
Segmentation fault
I get those other messages ...
scsi: aborting command due to timeout ...
... when I try to access the directory where I mounted the CDROM drive.
Do you think my SC hack is causing this mess?
- Ulf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 17:50 ` cdrom Ulf Carlsson
@ 1998-09-01 18:05 ` Alex deVries
1998-09-01 22:08 ` cdrom ralf
1 sibling, 0 replies; 11+ messages in thread
From: Alex deVries @ 1998-09-01 18:05 UTC (permalink / raw)
To: Ulf Carlsson; +Cc: ralf, Thomas Bogendoerfer, linux
Not to be a whiner, but I've *always* had a problem with accessing more
than one SCSI device at a time, both with a CDROM and a hard disk.
- Alex
--
Alex deVries, puffin on LinuxNet.
http://www.engsoc.carleton.ca/~adevries/ .
On Tue, 1 Sep 1998, Ulf Carlsson wrote:
> Date: Tue, 1 Sep 1998 19:50:46 +0200 (CEST)
> From: Ulf Carlsson <grim@zigzegv.ml.org>
> To: ralf@uni-koblenz.de
> Cc: Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
> linux@cthulhu.engr.sgi.com
> Subject: Re: cdrom
>
> On Tue, 1 Sep 1998 ralf@uni-koblenz.de wrote:
>
> > On Tue, Sep 01, 1998 at 12:19:11AM +0200, Thomas Bogendoerfer wrote:
> >
> > The problems seems to be associated with either sync SCSI or disconnect/
> > reconnect. I did a two line modification to sgiwd93.c which disables
> > sync and disconnect/reconnect. I now have since about a hour running:
> >
> > dd if=/dev/sda of=/dev/zero &
> > dd if=/dev/sdb of=/dev/zero &
> > dd if=/dev/sdc of=/dev/zero &
> > find / /ext -xdev | cpio -o -H crc -F /dev/nst0 --verbose
> >
> > No problem so far. (Except that SCSI performance even with only single
> > disk activity is further converging to zero ...)
>
> I tried to implement your ideas, I disabled sync, I don't know how you
> disable disconnect/reconnect. Anyway, I don't think that's the problem.
> The problem is the sigsegv in mount the oops causes, do we have something
> like a 'half' mounted device?
>
> Unable to handle kernel paging request at virtual address 00000000, epc ==
> 88021bcc, ra == 8809414
> Oops: 0000
> ...
> epc : 88021bcc
> Status: 3004fc02
> Cause : 00000008
> Segmentation fault
>
> I get those other messages ...
>
> scsi: aborting command due to timeout ...
>
> ... when I try to access the directory where I mounted the CDROM drive.
> Do you think my SC hack is causing this mess?
>
> - Ulf
>
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 17:50 ` cdrom Ulf Carlsson
1998-09-01 18:05 ` cdrom Alex deVries
@ 1998-09-01 22:08 ` ralf
1998-09-02 15:58 ` cdrom Ulf Carlsson
1 sibling, 1 reply; 11+ messages in thread
From: ralf @ 1998-09-01 22:08 UTC (permalink / raw)
To: Ulf Carlsson; +Cc: Thomas Bogendoerfer, linux
On Tue, Sep 01, 1998 at 07:50:46PM +0200, Ulf Carlsson wrote:
> I tried to implement your ideas, I disabled sync, I don't know how you
> disable disconnect/reconnect. Anyway, I don't think that's the problem.
> The problem is the sigsegv in mount the oops causes, do we have something
> like a 'half' mounted device?
The mount operation is assumed to be atomic. If it's not you're in deep
trouble.
> Unable to handle kernel paging request at virtual address 00000000, epc ==
> 88021bcc, ra == 8809414
> Oops: 0000
> ...
> epc : 88021bcc
> Status: 3004fc02
> Cause : 00000008
> Segmentation fault
>
> I get those other messages ...
As I recently told you on IRC - the patch as you've posted it is not
correct. It will misstreat VCEI exceptions.
> scsi: aborting command due to timeout ...
>
> ... when I try to access the directory where I mounted the CDROM drive.
> Do you think my SC hack is causing this mess?
Not the timeouts, they're a genuine wd driver bug.
Ralf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 22:08 ` cdrom ralf
@ 1998-09-02 15:58 ` Ulf Carlsson
1998-09-09 12:42 ` cdrom ralf
0 siblings, 1 reply; 11+ messages in thread
From: Ulf Carlsson @ 1998-09-02 15:58 UTC (permalink / raw)
To: ralf; +Cc: linux
On Wed, 2 Sep 1998 ralf@uni-koblenz.de wrote:
> As I recently told you on IRC - the patch as you've posted it is not
> correct. It will misstreat VCEI exceptions.
My idea is based on that we write the wrong cache line back, and that's
why we receive the invalid instruction errors.
Maybe this is foolish, but anyway: If we have data from main memory cached
in the secondary cache and then overwrite that data line in main memory
with an instruction line and cache the instruction. We receive a VCEI when
we try to access the cached line, and our handler writes the data back
instead of the intstruction and causes the invalid instructions. Well,
this is the only idea I have at the moment.
- Ulf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-02 15:58 ` cdrom Ulf Carlsson
@ 1998-09-09 12:42 ` ralf
0 siblings, 0 replies; 11+ messages in thread
From: ralf @ 1998-09-09 12:42 UTC (permalink / raw)
To: Ulf Carlsson; +Cc: linux
On Wed, Sep 02, 1998 at 05:58:15PM +0200, Ulf Carlsson wrote:
> On Wed, 2 Sep 1998 ralf@uni-koblenz.de wrote:
>
> > As I recently told you on IRC - the patch as you've posted it is not
> > correct. It will misstreat VCEI exceptions.
>
> My idea is based on that we write the wrong cache line back, and that's
> why we receive the invalid instruction errors.
>
> Maybe this is foolish, but anyway: If we have data from main memory cached
> in the secondary cache and then overwrite that data line in main memory
> with an instruction line and cache the instruction. We receive a VCEI when
> we try to access the cached line, and our handler writes the data back
> instead of the intstruction and causes the invalid instructions. Well,
> this is the only idea I have at the moment.
Note that we don't have to care about if the line contains data or
instructions. All we know is that the l2 line corrosponding to the
instruction which threw the vcei exception is in some state other than
invalid. We cannot accidently writeback an instruction line to memory
because the Hit_Writeback_Inv_SD cacheop we're using will only write
dirty lines back.
Ralf
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 14:55 ` cdrom ralf
1998-09-01 17:50 ` cdrom Ulf Carlsson
@ 1998-09-01 19:19 ` Alex deVries
1998-09-01 21:56 ` cdrom ralf
1 sibling, 1 reply; 11+ messages in thread
From: Alex deVries @ 1998-09-01 19:19 UTC (permalink / raw)
To: ralf; +Cc: Thomas Bogendoerfer, Ulf Carlsson, linux
On Tue, 1 Sep 1998 ralf@uni-koblenz.de wrote:
> The problems seems to be associated with either sync SCSI or disconnect/
> reconnect. I did a two line modification to sgiwd93.c which disables
> sync and disconnect/reconnect. I now have since about a hour running:
Could you upload those patches to the cvs?
- Alex
^ permalink raw reply [flat|nested] 11+ messages in thread
* Re: cdrom
1998-09-01 19:19 ` cdrom Alex deVries
@ 1998-09-01 21:56 ` ralf
0 siblings, 0 replies; 11+ messages in thread
From: ralf @ 1998-09-01 21:56 UTC (permalink / raw)
To: Alex deVries; +Cc: Thomas Bogendoerfer, Ulf Carlsson, linux
On Tue, Sep 01, 1998 at 03:19:31PM -0400, Alex deVries wrote:
> On Tue, 1 Sep 1998 ralf@uni-koblenz.de wrote:
> > The problems seems to be associated with either sync SCSI or disconnect/
> > reconnect. I did a two line modification to sgiwd93.c which disables
> > sync and disconnect/reconnect. I now have since about a hour running:
>
> Could you upload those patches to the cvs?
Toxic waste like that hack doesn't belong into a CVS archive.
Ralf
^ permalink raw reply [flat|nested] 11+ messages in thread
end of thread, other threads:[~1998-09-09 13:34 UTC | newest]
Thread overview: 11+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
1998-08-31 16:56 cdrom Ulf Carlsson
1998-08-31 22:19 ` cdrom Thomas Bogendoerfer
1998-08-31 23:26 ` cdrom ralf
1998-09-01 14:55 ` cdrom ralf
1998-09-01 17:50 ` cdrom Ulf Carlsson
1998-09-01 18:05 ` cdrom Alex deVries
1998-09-01 22:08 ` cdrom ralf
1998-09-02 15:58 ` cdrom Ulf Carlsson
1998-09-09 12:42 ` cdrom ralf
1998-09-01 19:19 ` cdrom Alex deVries
1998-09-01 21:56 ` cdrom ralf
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox