Re: [9fans] /lib/rfc

[erik quanstrom <quanstro@coraid.com>]

> > Yes, /lib/rfc/grabfc.  Uncomment this line:
> >
> > /cron/sys/cron:#30 9 * * *	local	/lib/rfc/grabrfc
>
> thank you.  unfortunately, it turns out that this script causes
> something that looks like deadlock between ramfs and ftpfs.
> i didn't get a bead on ramfs before ftpfs exited but it was in
> state Pread as was ramfs.
>
> here's where ftpfs was
>
> tyty# cd /proc/389580
> tyty# cat fd
> /lib/rfc
>   0 r  c    0 (000000000000000a 0 00)     0        0 /dev/null
>   1 w  c    0 (000000000000000a 0 00)     0        0 /dev/null
>   2 w  M 256525 (0000000000000001 0 00)  8192   337626 /mnt/term/dev/cons
>   3 rw |    0 (000000000023e801 0 00) 65536 17056533 #|/data
>   4 rw M 258155 (0000000000000141 0 00)  8192     7835 /tmp/ftpo0000389580
>   5 r  c    0 (0000000000000001 0 00)     0    75912 /dev/bintime
>   6 rw M 258155 (0000000000000142 0 00)  8192    83648 /tmp/ftp00000389580
>   7 rw I    1 (00000000000201ad 0 00)     0    86141 /net.alt/tcp/13/data
>   8 rw M 258155 (0000000000000143 0 00)  8192   156468 /tmp/ftpa0000389580
>   9 rw M 258155 (0000000000000144 0 00)  8192    24637 /tmp/ftpb0000389580
[...]

actually, i tracked this a bit further.  the passive connection
gets crossed up.  if killed in this state, ftpfs will exit with
part of the downloading file in its exit status.  this only happens
with larger files.

- erik