On Wed, Feb 03, 1999 at 02:14:40PM -0500, Alexander Viro wrote:
> What??? Wait... it's grow_inodes() and it looks like we never shrink the
> damned icache again. Looks like it was always so with possible exception
Does the explain the NTFS problems I was looking into? I'd noticed that
try_to_free_pages didn't do anything with the icache and in my innocence^W
ignorance, assumed that shrink_dcache_memory did something about it (maybe
I thought inodes weren't cached without a dentry or something).
> of 2.1.36-2.1.42 when we used slab instead of bare pages. Damn. It's not a
> leak per se - we just never shrink icache. Now, the fact that you managed
Out of curiosity, what was the reason behind not using the slab allocator?
I (think) I've noticed an interesting problem -- the slab allocator uses
page blocks of order > 0 for some caches, but get_free_pages only tries to
free single pages. As I can't think of a sane way to 'create' multi-page
blocks in gfp/try_to_free_page, maybe the slab allocator should reduce
the page block size it wants in low memory / fragmented memory situations.
Cheers,
Steve
-- If breastfeeding a pig is sickening to you, what the hell are you doing on usenet? -- Aimee, on rec.music.tori-amos- 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/