Messages in this thread | | | From | Nikita Danilov <> | Date | Thu, 17 Jul 2003 17:50:44 +0400 | Subject | Re: [PATCH] 1/5 VM changes: zone-pressure.patch |
| |
Andrew Morton writes: > Nikita Danilov <Nikita@Namesys.COM> wrote: > > > > > OK, fixes a bug. > > > > What bug? > > Failing to consider mapped pages on the active list until the scanning > priority gets large. > > I ran up your five patches on a 256MB box, running `qsbench -m 350'. It got > all slow then the machine seized up. I'll poke at it some. >
My understanding of this is following:
1. kswapd/balance_pgdat runs through all priorities, but fails to rise zone->free_pages above zone->pages_min. So it exet. Which is strange.
2. nobody is going to wake it up, because all memory allocators are looping indefinitely in that newly introduced do_retry loop inside __alloc_pages.
I think either wakeup_kswapd() should be called in __alloc_pages() before goto rebalance, or kswapd should loop until it restores all zones to the balance (actually, comment before balance_pgdat() says it does), or both.
Nikita. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |