Interesting. From my read it looks like it should not be
necessary to restore b_data because getpbuf() calls initpbuf()
which unconditionally reassigns the field. But, that said,
it's probably a good idea to fix it so I'll commit the patch
with a suitable log message.
-Matt
Matthew Dillon
<dillon@backplane.com>
:While cleaning out my source tree I came across this one-line local
:modification. Presumably someone who knows more about the kernel than I
:do recommended I make this change when I was having problems with
:something, but darned if I can remember who or why.
:
:It's probably meaningless, but in case not, I thought I'd just toss it
:up here before deleting it.
:
:-Chris
:
:Index: kern_physio.c
:===================================================================
:RCS file: /home/dcvs/src/sys/kern/kern_physio.c,v
:retrieving revision 1.8
:diff -u -r1.8 kern_physio.c
:--- kern_physio.c 16 Feb 2004 20:11:20 -0000 1.8
:+++ kern_physio.c 15 Jun 2004 03:45:52 -0000
:@@ -143,6 +143,7 @@
: }
: }
: doerror:
:+ bp->b_data = sa;
: relpbuf(bp, NULL);
: return (error);
: }