recent rwhod woes

recent rwhod woes

[J Sloan]

Hi, just a reminder that rwhod is broken in recent 2.5.67-xx kernels.

I had earlier reported it incorrectly, saying I first saw the problem
in 2.5.67-mm3, but going back to verify, I see that 2.5.67-mm1 is OK
while 2.5.67-mm2 through -mm4 are broken.

I noticed that rwhod is OK in 2.5.67 + linux patchset 1.1177, but
adding the sk-allocation patch (from -mm2) breaks rwhod.

The symptom here is that running the "ruptime" command on an -mm2
box shows all hosts down - however the other hosts on the subnet are
getting the rwho broadcasts from the -mm2 box, but the -mm2 box is
unable to process rwho broadcasts, including it's own -

I'd hoped I could narrow it down to that patch, but backing out that
single patch does not fix -mm2 through -mm4, so it appears that while
the one patch will break rwhod, there are apparently other changes
that also break rwhod independently of the sk-allocation patch, and
the answer is more involved than I'd hoped.

I've been hammered at work lately so time is limited, but I did want
to raise the rwhod issue again - if I get some time I'll check the
plain -bk tree, or if someone has any ideas I'll be happy to act as
a patch & test monkey in my spare time.

Best Regards,

Joe