Chuck Silvers <c...@chuq.com> wrote:
> we shouldn't need to change the genfs code to make "soft" work.
> if the underlying RPCs time out and all the retries are exhausted,
> the NFS code should report the error back to the genfs code by doing
> the usual B_ERROR/b_error thing with the buffer, and the genfs code
> should handle that by unlocking pages, etc, just like it would for
> a failed write to a scsi or ata device, and eventually that should
> percolate back up the stack until the cv_wait() returns.
> does this not work currently?
It almost works with a few fixes.
I tried without touching genfs beyond changing cv_wait() into cv_timedwait()
with a debug message, like this:
/* Wait for output to complete. */
if (!wasclean && !async && vp->v_numoutput != 0) {
- while (vp->v_numoutput != 0)
- cv_wait(&vp->v_cv, slock);
+ while (vp->v_numoutput != 0) {
+ int cv_error;
+
+ cv_error = cv_timedwait(&vp->v_cv, slock, 2 * hz);
+ if (cv_error) {
+ printf("%s: failed to complete I/O on %s, "
+ "vp = %p, numoutput = %d, error = %d\n",
+ l->l_name? l->l_name : l->l_proc->p_comm,
+ vp->v_mount->mnt_stat.f_mntonname,
+ vp, vp->v_numoutput, cv_error);
+ }
+
+ }
}
onworklst = (vp->v_iflag & VI_ONWORKLST) != 0;
mutex_exit(slock);
I have also changes in NFS code so that RPC on soft mounts can timeout and set
bp->b_error. What happens here is that we loop in the while (vp->v_numoutput !=
0) block with vp->v_numoutput draining down to 2 and then we loop forever with
this value.
I note we have this in genfs_do_io(), and I suspect this is the same 2 value:
if (iowrite) {
mutex_enter(vp->v_interlock);
vp->v_numoutput += 2;
mutex_exit(vp->v_interlock);
}
mbp = ge
Why the vp->v_numoutput += 2 ?
--
Emmanuel Dreyfus
http://hcpnet.free.fr/pubz
m...@netbsd.org
