>SHORT STORY:
>
>Apology for making fuss about mem=8M. It was pilot error.
After reading your message I tried to boot with mem=4m arca-52...
PID TTY STAT TIME COMMAND
1 ? SW 0:04 (init)
2 ? SW 0:03 (kflushd)
3 ? SW< 0:02 (kswapd)
12 ? D 0:00 update
99 ? SW 0:00 (syslogd)
101 ? SW 0:00 (klogd)
107 ? SW 0:00 (inetd)
122 ? SW 0:00 (gpm)
129 ? SW 0:00 (lpd)
157 ? S 0:00 /usr/sbin/cron
164 2 S 0:00 -bash
165 3 SW 0:00 (getty)
166 4 SW 0:00 (getty)
167 5 SW 0:00 (getty)
168 6 SW 0:00 (getty)
210 ? S 0:04 X :0 -auth /home/andrea/.Xauthority
223 1 SW 0:00 (xterm)
234 ? SW 0:00 (nmbd)
248 2 R 0:00 ps xa
51 ? SW 0:00 (portmap)
163 1 SW 0:00 (bash)
205 1 SW 0:00 (xinit)
212 1 S 0:00 /usr/X11R6/bin/WindowMaker
225 1 SW 0:01 (gnomecal)
227 1 SW 0:00 (xconsole)
229 p0 SW 0:00 (bash)
total used free shared buffers cached
Mem: 3776 3704 72 516 128 1472
-/+ buffers/cache: 2104 1672
Swap: 72256 8448 63808
Everything was fine (except a bit slowed down but not too bad :). I was
running a gnome app + wmaker + X ;). I also reinserted the vmscan changes
from Stephen in my tree (since it seems that the new shrink_mmap() in
do_try_to_free_pages() handle the freeing of the not really freed yet
pages fine). Now the only big mm difference between arca-52 and ac6 is
kswapd. It would be nice of somebody would do some benchmark. I had a
report where a kernel compile is 2 sec faster in arca-52 then ac6, but
that is probably due the idle not rescheduling bug (I extracted the
fix from my tree and I send it to AC some days ago btw). I' ll continue
the testing this evening because now I _have_ to shutdown :(.
Andrea Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/