On Thu, 26 Apr 2001, Mike Galbraith wrote:
> On Thu, 26 Apr 2001, Rik van Riel wrote:
>
> > On Thu, 26 Apr 2001, Mike Galbraith wrote:
> >
> > > 1. pagecache is becoming swapcache and must be aged before anything is
> > > done. Meanwhile we're calling refill_inactive_scan() so fast that noone
> > > has a chance to touch a page. Age becomes a simple counter.. I think.
> > > When you hit a big surge, swap pages are at the back of all lists, so all
> > > of your valuable cache gets reclaimed before we write even one swap page.
> >
> > Does the patch I sent to [EMAIL PROTECTED] last night help in
> > this ?
> >
> > I found that the way refill_inactive_scan() and swap_out() are being
> > called from the main loop in refill_inactive() aren't equal and have
> > fixed that in a way which (IMHO) also beautifies the code a bit.
> >
> > (and makes sure background aging doesn't get out of hand with a few
> > simple checks)
>
> That patch livelocked my box with only ~1000 pages on any list.
>
> I can go back and test some more if you want.
I put it back in, the livelock is 100% repeatable (10 repeats). It's
deactivating/laundering itself to death. :) 3mb for my 386-20/0.96.9
would have been enough to frolic (slowly) in.. this box can't live.
procs memory swap io system cpu
r b w swpd free buff cache si so bi bo in cs us sy id
...
37 3 0 988 1940 1912 32368 0 0 40 49 117 293 90 10 0
47 0 0 1348 1940 1320 26816 0 0 80 17 121 415 86 14 0
39 1 0 1364 1972 1076 20728 0 0 184 0 124 294 84 16 0
39 3 2 1456 1940 232 14720 0 572 67 638 159 2225 85 15 0
29 10 2 1456 1612 416 6908 0 0 252 72 235 2253 84 16 0
17 18 2 1292 1420 608 4216 0 60 340 454 279 3289 29 20 51
33 1 2 1296 1408 488 3464 0 4 317 752 295 2432 11 47 42
locked here
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
