On Fri, Jan 02, 2004 at 06:01:28PM +0100, Lennart Augustsson wrote:
> Yes, I just experienced this as well.  And if you include doing
> `make depend' it gets even worse, I think.  This code cannot
> stay as it is, I mean a 3-fold slowdown is pretty awful.
> 
> 	-- Lennart

I'm seeing the same problem, too. The kernel is taking up so much
memory and swap is running short, neither of which has never seen on my
512M machine. Current top(1) output is as follows: 

load averages:  3.05,  3.25,  3.24    02:05:49
69 processes:  68 sleeping, 1 on processor

Memory: 183M Act, 103M Inact, 1056K Wired, 45M Exec, 172M File, 22M Free
Swap: 129M Total, 108M Used, 21M Free


  PID USERNAME PRI NICE   SIZE   RES STATE      TIME   WCPU    CPU COMMAND
15314 junyoung   2    0  7268K   13M select     3:11  0.63%  0.63% kdeinit
12481 junyoung   2    0    23M   30M select     9:22  0.15%  0.15% XFree86
29199 junyoung  10    0   728K 1328K wait       0:00  0.44%  0.10% make
    9 root     -18    0     0K  180M reaper     1:57  0.05%  0.05% [reaper]
26254 junyoung   2    0    30M   35M select     4:36  0.00%  0.00% kdeinit
19251 junyoung   2    0  3120K   11M select     4:33  0.00%  0.00% kdeinit
28045 junyoung   2    0    30M   38M select     1:48  0.00%  0.00% kdeinit
 1059 junyoung   2    0  5332K   14M select     1:35  0.00%  0.00% kdeinit
  635 junyoung   2    0  4912K   12M select     1:26  0.00%  0.00% kdeinit
   10 root      18    0     0K  180M syncer     1:10  0.00%  0.00% [ioflush]
  542 junyoung   2    0  4780K   11M select     0:52  0.00%  0.00% kdeinit
23019 junyoung   2    0  5308K   14M select     0:40  0.00%  0.00% kdeinit
  476 junyoung   2    0    12M   14M select     0:29  0.00%  0.00% kget
 6971 junyoung   2    0  5608K 6704K select     0:26  0.00%  0.00% kdeinit
 4512 junyoung   2    0    11M 2240K select     0:19  0.00%  0.00% kdeinit
20822 junyoung   2    0  2728K 4928K select     0:14  0.00%  0.00% artsd
 6575 junyoung   2    0  6320K 6468K select     0:11  0.00%  0.00% kdeinit
13975 junyoung   2    0  5608K 3752K poll       0:10  0.00%  0.00% nabi

And I noticed time took on building XFree86-current got tremendously
longer (2 or 3 fold) with the new bufcache scheme.

Jun-Young

-- 
Bang Jun-Young <junyoung@NetBSD.org>