Hmm ... I'm don't think so ... maybe the problem is triggered by
the changes in fs/buffer.c but these changes are not the cause for
the problem.
We heave at S.u.S.E. some support mails and support phone calls for
2.0.25 and 2.0.29 with the same problems now often found in
2.0.30. Maybe some of the problems are caused by broken or
misconfigured hardware ... but the rest of these phenomenons?
No, I can reproduce it on my 2.1.x Sparc systems these days every once
in a while, this is why I initially suspected the buffer.c changes
since those were now in both 2.0.30 and current 2.1.x sources.
Now I am starting to suspect a strange race in the page cache, all of
the debugging I've seen on this shows a process stuck on I/O to a page
in the page cache. I will be looking more deeply into this.
---------------------------------------------////
Yow! 11.26 MB/s remote host TCP bandwidth & ////
199 usec remote TCP latency over 100Mb/s ////
ethernet. Beat that! ////
-----------------------------------------////__________ o
David S. Miller, davem@caip.rutgers.edu /_____________/ / // /_/ ><