if_re locking patch...
pepper at reppep.com
Mon Sep 6 21:42:00 PDT 2004
At 8:47 PM -0700 2004/09/06, John-Mark Gurney wrote:
>John-Mark Gurney wrote this message on Mon, Sep 06, 2004 at 20:40 -0700:
>> Chris Pepper wrote this message on Mon, Sep 06, 2004 at 23:32 -0400:
>> > Unfortunately "make buildkernel" fails on if_re with
>> > DEVICE_POLLING, which should work according to polling.4.
>> > http://www.reppep.com/~pepper/guest/make-buildkernel-re-failure-2.txt
>> ru committed a fix for this already... but it appears to have failed
>> to be MFC'd... I didn't expect the change to be MFC'd so quickly...
>> I'll make sure that the fix gets MT5..
>Let me correct the record, this change has not been MFC'd yet, and ru
>commited the fix to -current. Basicly add semicolons at the end of the
>prototypes that are ifdef'd DEVICE_POLLING out..
>I have updated the patch, or you can grab HEAD's if_re.c.
Sounds good. Is there a standard way of updating a source
directory to a different CVS tag? I use cvsup to track RELENG_5, but
thought I'd seen someone mention a make incantation here to update a
single directory (/usr/src/sys/dev/re) to HEAD or another tag.
More generally (and even more tangentially), is there an SOP
for keeping patches to cvsupped source? This seems like the kind of
thing lots of FreeBSD folks do for local patches, but I haven't had
to do it myself previously. I could add a couple patch commands to my
cvsup.sh script easily enough, but wonder if there's a better way...
Chris Pepper: <http://www.reppep.com/~pepper/>
Rockefeller University: <http://www.rockefeller.edu/>
More information about the freebsd-current