[7717] in linux-scsi channel archive
SCSI queueing status.
daemon@ATHENA.MIT.EDU (Eric Youngdale)
Sat Dec 18 00:09:34 1999
Date: Sat, 18 Dec 1999 00:05:39 -0500 (EST)
From: Eric Youngdale <eric@andante.org>
To: Jens Axboe <axboe@image.dk>, Alan Cox <alan@lxorguk.ukuu.org.uk>,
Douglas Gilbert <dgilbert@interlog.com>, linux-scsi@vger.rutgers.edu
Message-ID: <Pine.LNX.4.10.9912171849590.30229-100000@gwyn.tux.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
I just sent off the last patcheset to Linus to be included in the
next release. As I mentioned before, I am going to be on vacation for the
next 10 days, and while I will have a laptop, I won't be checking it quite
as regularly. If serious issues arise, I will be in a position to try and
investigate to some degree (as best I can without real SCSI hardware). I
do have a source tree on my laptop.
The patchset that I sent out corrects a couple of race conditions,
and solves a DMA pool exhaustion issue that was pointed out.
There are a couple of open issues that I know of:
1) There is still a slight problem of DMA exhaustion, at this
point it the symptom is just a lot of spurious messages about running low,
but the system should work it's way through the problem. I probably need
to tweak the heuristics a bit more. This will only affect people who have
ISA SCSI cards that do DMA, and who also have more than 16M (I remember
when that was a rarity...). At the moment (with the patch), this should
be no more than a nuisance.
2) I have broken the support for > 16 disks. I have a workaround
for this one - a slight modification to scsi_get_request_dev in scsi_lib.c
should get these at least partly working. The person who originally
posted the report is having a different problem now, and I am still
waiting for details before I can figure out what exactly the problem is. I
am enclosing the workaround for the first problem - I am reluctant to
submit workarounds to Linus, as workarounds have this nasty tendancy to
gain a life of their own.
3) There was a report on linux-kernel about a scatter-gather table
segment counting problem, which led to a panic. I ask for, but have yet
to receive details on this one. As far as I know, there is only one
person who has reported this issue.
4) There was one report of being unable to write a CDROM. I have
asked some questions, but at this point I don't know the nature of the
problem, nor do I know how many people would be affected.
-Eric
"The world was a library, and its books were the stones, leaves,
brooks, grass, and the birds of the earth. We learned to do what only
a student of nature ever learns, and that was to feel beauty."
Chief Luther Standing Bear - Teton Sioux
Workaround for >16 disk problem.
*** drivers/scsi/scsi_lib.c Wed Dec 15 21:32:32 1999
--- drivers/scsi/scsi_lib.c Sat Dec 18 00:34:36 1999
***************
*** 599,604 ****
--- 599,614 ----
ASSERT_LOCK(&io_request_lock, 1);
+ /*
+ * FIXME(eric) - we need to fix this so that we recognize that
+ * a single driver might support a number of different major
+ * numbers. For now, this workaround should do the trick.
+ * This is *ONLY* a workaround.
+ */
+ if( major >= SCSI_DISK1_MAJOR && major <= SCSI_DISK7_MAJOR ) {
+ major = SCSI_DISK0_MAJOR;
+ };
+
for (spnt = scsi_devicelist; spnt; spnt = spnt->next) {
/*
* Search for a block device driver that supports this
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
