On Tue, 28 Mar 2006, rick@snowhite.cis.uoguelph.ca wrote:

> I've run the little test script and it seemed to be fixed after I had
> patched my server to set VA_UTIMES_NULL when setting to server time(s).
>
> So, I think that your diagnosis was correct, but your patch didn't quite
> fix it. Although I haven't tested it, here is the equivalent patch to
> what I did, for OpenBSD3.8. Maybe you can try this?
>
> Good luck with it, rick

I installed a 2.6.16 linux system. The problematic call with my diff
was only setting mtime TOSERVER, so the count never reached 2.

I can confirm your diff fixes things for me. I'm woried about one
thing though: what if the client does atime TOSERVER mtime TOCLIENT?
That should not be allowed.

So here's a revised diff that catches that.

Han, could you try this one?

-Otto

Index: nfsm_subs.h
================================================== =================
RCS file: /cvs/src/sys/nfs/nfsm_subs.h,v
retrieving revision 1.15
diff -u -p -r1.15 nfsm_subs.h
--- nfsm_subs.h 24 Jun 2004 19:35:26 -0000 1.15
+++ nfsm_subs.h 28 Mar 2006 20:11:09 -0000
@@ -471,16 +471,19 @@
fxdr_nfsv3time(tl, &(a)->va_atime); \
break; \
case NFSV3SATTRTIME_TOSERVER: \
+ (a)->va_vaflags |= VA_UTIMES_NULL; \
getnanotime(&(a)->va_atime); \
break; \
}; \
nfsm_dissect(tl, u_int32_t *, NFSX_UNSIGNED); \
switch (fxdr_unsigned(int, *tl)) { \
case NFSV3SATTRTIME_TOCLIENT: \
+ (a)->va_vaflags &= ~VA_UTIMES_NULL; \
nfsm_dissect(tl, u_int32_t *, 2 * NFSX_UNSIGNED); \
fxdr_nfsv3time(tl, &(a)->va_mtime); \
break; \
case NFSV3SATTRTIME_TOSERVER: \
+ (a)->va_vaflags |= VA_UTIMES_NULL; \
getnanotime(&(a)->va_mtime); \
break; \
}; }