[2345] in linux-scsi channel archive
Re: "read defect list" with 2.0.30-pre7 and patch Aug19
daemon@ATHENA.MIT.EDU (Ulrich Windl)
Tue Aug 26 02:56:50 1997
From: "Ulrich Windl" <ulrich.windl@rz.uni-regensburg.de>
To: Gerard Roudier <groudier@club-internet.fr>
Date: Tue, 26 Aug 1997 08:28:13 +0200
Cc: linux-scsi@vger.rutgers.edu
In-reply-to: <Pine.LNX.3.95.970825191253.1005A-100000@localhost>
On 25 Aug 97 at 19:28, Gerard Roudier wrote:
>
> I donnot use to be sarcastic, or perhaps I does, I donnot know ...
> If you are able to write 50 full lines of text that explains a
> trivial change, you probably will put the Net on the knees the day
> you will explain a complex one. ;-)
> (Sorry Doug, I could'nt resist.)
I'm not sure what the above staement was about, but if it was to my
suggestion to comment the change, I'd just have the following
reasoning:
As those lines were unchanged for a very long time, people might
think these being stable (in the sense of "correct"). If somebody now
changes them, somebody else might think "what are they playing around
there" (Linus sometimes forced people to test new things that way)?
The other think is about the issue whether "EINVAL" is the
most correct thing to return (EPERM "you are not allowed to crash
the kernel" ;-) ).
>
> [ switch to serious mode ]
>
> > > Aug 22 23:57:52 elf kernel: SMalloc: 4096 0000a000
> > > Aug 22 23:57:52 elf kernel: scsi_do_cmd (host = 0, channel = 0 target = 0, buffer =0000a000, bufflen = 4096, done = 001ab754, timeout = 1000, retries = 5)
> > > Aug 22 23:57:52 elf kernel: command : 37 00 14 00 00 00 00 20 00 00
> > > Aug 22 23:57:52 elf kernel: (scsi0:0:0) Data overrun of 16773218 bytes detected in Data-In phase, tag 7; forcing a retry.
>
> A data overrun is generally an application error and so, no error
> recovery should be attempted at driver level, in my opinion.
Right; as soon as the real reson has been found (data overrun) it's
clear that retrying the operation won't help much. In that special
case the kernel is already rather damaged during the first retry...
Ulrich
