I have an 8M tmpfs filesystem whose available blocks are exhausted over
time.  There seems to be a memory leak, because I cannot account for more
than 3M of the use with fstat(1) (for deleted files & dirs) and du(1),

# du -ks /mfs
2724    /mfs
# fstat -f /etc | sort -u -n -k 6,6
USER     CMD          PID   FD MOUNT       INUM MODE         SZ|DV R/W
root     sh         18746   11 /etc          23 -rw-r--r--    2108 r 
root     sh          2244   wd /etc          63 drwxr-xr-x      40 r 
root     fstat       4581    6 /etc         191 -rw-r--r--   40960 r 
# fstat -f /home | sort -u -n -k 6,6
USER     CMD          PID   FD MOUNT       INUM MODE         SZ|DV R/W
root     sh         11753   wd /home        259 drwxr-xr-x     120 r 
# fstat -f /tmp | sort -u -n -k 6,6
USER     CMD          PID   FD MOUNT       INUM MODE         SZ|DV R/W
# fstat -f /var | sort -u -n -k 6,6
USER     CMD          PID   FD MOUNT       INUM MODE         SZ|DV R/W
root     cron        2407   wd /var         402 drwxr-xr-x      20 r 
root     syslogd     2005    9 /var         410 -rw-------     189 w 
root     syslogd     2005   10 /var         411 -rw-------    8154 w 
root     syslogd     2005    7 /var         415 -rw-r--r--   12286 w 
root     syslogd     2005   11 /var         419 -rw-------       0 w 
nobody   thttpd        96   wd /var         432 drwxr-xr-x     400 r 
quagga   zebra       1932    6 /var         478 -rw-------       5 rw
root     sleep      23925    2 /var         479 -rw-r--r--   45008 w 
root     sleep       7094    2 /var         489 -rw-r--r--    4079 w 
root     sleep       4377    2 /var         494 -rw-r--r--   53026 w 
root     sleep      26527    2 /var         497 -rw-r--r--  190547 w 
nobody   thttpd        96    4 /var         499 -rw-r--r--       0 w 
root     sh         16884    4 /var         510 -rw-------     330 w 
root     cron        2407    3 /var         512 -rw-r--r--       5 rw
root     syslogd     2005   12 /var        1372 -rw-r--r--       0 w 
root     sh          1250    1 /var        1553 -rw-r--r--   12287 w 
root     sh          2737    1 /var        1734 -rw-r--r--       0 w 

Nevertheless, my tmpfs is full:

# df -hi
Filesystem        Size       Used      Avail %Cap    iUsed   iAvail %iCap Mounted on
/dev/wd0e         230M        18M       200M   8%     1088   116670   0% /
tmpfs             496K       196K       300K  39%     1211       53  95% /dev
tmpfs             8.0M       8.0M         0B 100%     1856        0 100% /mfs
/etc              230M        18M       200M   8%     1088   116670   0% /permanent/etc
/home             230M        18M       200M   8%     1088   116670   0% /permanent/home
/tmp              230M        18M       200M   8%     1088   116670   0% /permanent/tmp
/var              230M        18M       200M   8%     1088   116670   0% /permanent/var
/mfs/etc          8.0M       8.0M         0B 100%     1856        0 100% /etc
/mfs/home         8.0M       8.0M         0B 100%     1856        0 100% /home
/mfs/tmp          8.0M       8.0M         0B 100%     1856        0 100% /tmp
/mfs/var          8.0M       8.0M         0B 100%     1856        0 100% /var

Does no one else see this?  My application may be a bit unusual, both
in that I use null mounts, and in that I have no swap activated.

Could the cause of the leak be an interaction between my null mounts
and tmpfs?  Also, I am dimly aware of some reference-counting bug in
tmpfs; it was mentioned in one of ad@'s commits to the vmlocking branch.
(I do not run the vmlocking branch.)

Any pointers appreciated.

Dave

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Urbana, IL * (217) 278-3933 ext 24