Re: 2.6.18-rc7-mm1

From: Martin Bligh
Date: Tue Sep 19 2006 - 13:01:09 EST


Andrew Morton wrote:
On Tue, 19 Sep 2006 07:45:06 -0700
"Martin J. Bligh" <mbligh@xxxxxxxxxx> wrote:


- It took maybe ten hours solid work to get this dogpile vaguely
compiling and limping to a login prompt on x86, x86_64 and powerpc. I guess it's worth briefly testing if you're keen.

PPC64 blades shit themselves in a strange way. Possibly the udev
breakage you mentioned? Hard to tell really if people are going to
go around breaking userspace compatibility ;-(


What version of udev is it running?


http://test.kernel.org/abat/48127/debug/console.log

..

sda: Write Protect is off
sda: cache data unavailable
sda: assuming drive cache: write through
SCSI device sda: 143374000 512-byte hdwr sectors (73407 MB)
sda: Write Protect is off
sda: cache data unavailable
sda: assuming drive cache: write through
sda: sda1 sda2 sda3 sda4 < sda5 sda6 sda7 sda8 >
sd 0:0:1:0: Attached scsi disk sda
creating device nodes .[: [0-9]*: bad number
0:0:1:0: sg_io failed status 0x8 0x0 0x0 0x2
0:0:1:0: sense key 0x5 ASC 0x24 ASCQ 0x0
[: [0-9]*: bad number
0:0:1:0: sg_io failed status 0x8 0x0 0x0 0x2
0:0:1:0: sense key 0x5 ASC 0x24 ASCQ 0x0
[: [0-9]*: bad number
0:0:1:0: sg_io failed status 0x8 0x0 0x0 0x2
0:0:1:0: sense key 0x5 ASC 0x24 ASCQ 0x0
[: [0-9]*: bad number
0:0:1:0: sg_io failed status 0x8 0x0 0x0 0x2
0:0:1:0: sense key 0x5 ASC 0x24 ASCQ 0x0
[: [0-9]*: bad number




That all looks rather bad.


ReiserFS: sda2: Using r5 hash to sort names
looking for init ...
found /sbin/init
/init: cannot open .//dev//console: no such file


Bizarrely-formed pathname. Does it always do that?


Working one (-git3): http://test.kernel.org/abat/48064/debug/console.log

Same sgio shit. no mention of /dev/console, but it's an error message,
so not unexpected.

Has udev actually attempted to do anything by this stage?

Buggered if I know. I always just turn it off on my machines.

I wasn't seeing anything that spectacular. It used to be the case that
udev simply hung. But in rc7-mm1 the symptoms are that incoming ssh
sessions hang, but most other things work OK.

Oh well - Greg has split that tree apart and I shall not be pulling the
more problematic bits henceforth.

OK, may not be that at all ... could be something entirely different.
Just seemed co-incicental to your comments.

M.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/