>On Sun, 11 Oct 1998, Alan Cox wrote:
>
>> 2.1.10x swap deadlock
>> 2.1.10x can seize up looping through swap_out_vma, shrink_mmap
>> and get_swap_page.
>It includes both Andrea's patch, my patch and a few checks
>to see if we're really out of memory.
When __get_free_pages() return 0 we must know that we are OOM. If we are
not OOM and __get_free_pages() return 0, it means that it' s buggy (at
least in _not_ GFP_ATOMIC context).
The point is that an OOM killer could be _eventually_ a config option for
2.2 and it' s a new _feature_ (I don' t go in the details if it' s
something we want or not in 2.2 here). It has nothing to do with fixing
the bug Alan pointed out. My patch make 2.1 usable as/or better than 2.0
and I think it' s the right/good fix for the 2.1 OOM problem (and
obviously I am happily using it all the time without one problem... from
when Linus pointed out the __GFP_WAIT bug ;-).
>The only problem is that my patch still needs some cleaning
>up and I have a math test this friday ;(
I know what does it mean ;-).
Rik I 100% agree that an OOM killer can be useful, but here I am pointing
out that this is a totoally different issue.
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/