From: Hugh Dickins Date: Tue, 5 Feb 2008 06:28:54 +0000 (-0800) Subject: tmpfs: radix_tree_preloading X-Git-Tag: v2.6.25-rc1~961 X-Git-Url: http://pilppa.com/gitweb/?a=commitdiff_plain;h=b409f9fcf04692c0f603d28c73d2e3dfed27bf54;p=linux-2.6-omap-h63xx.git tmpfs: radix_tree_preloading Nick has observed that shmem.c still uses GFP_ATOMIC when adding to page cache or swap cache, without any radix tree preload: so tending to deplete emergency reserves of memory. GFP_ATOMIC remains appropriate in shmem_writepage's add_to_swap_cache: it's being called under memory pressure, so must not wait for more memory to become available. But shmem_unuse_inode now has a window in which it can and should preload with GFP_KERNEL, and say GFP_NOWAIT instead of GFP_ATOMIC in its add_to_page_cache. shmem_getpage is not so straightforward: its filepage/swappage integrity relies upon exchanging between caches under spinlock, and it would need a lot of restructuring to place the preloads correctly. Instead, follow its pattern of retrying on races: use GFP_NOWAIT instead of GFP_ATOMIC in add_to_page_cache, and begin each circuit of the repeat loop with a sleeping radix_tree_preload, followed immediately by radix_tree_preload_end - that won't guarantee success in the next add_to_page_cache, but doesn't need to. And we can then remove that bothersome congestion_wait: when needed, it'll automatically get done in the course of the radix_tree_preload. Signed-off-by: Hugh Dickins Looks-good-to: Nick Piggin Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/mm/shmem.c b/mm/shmem.c index a0126c43710..530c5033d02 100644 --- a/mm/shmem.c +++ b/mm/shmem.c @@ -901,12 +901,16 @@ found: error = 1; if (!inode) goto out; + error = radix_tree_preload(GFP_KERNEL); + if (error) + goto out; + error = 1; spin_lock(&info->lock); ptr = shmem_swp_entry(info, idx, NULL); if (ptr && ptr->val == entry.val) error = add_to_page_cache(page, inode->i_mapping, - idx, GFP_ATOMIC); + idx, GFP_NOWAIT); if (error == -EEXIST) { struct page *filepage = find_get_page(inode->i_mapping, idx); error = 1; @@ -931,6 +935,7 @@ found: if (ptr) shmem_swp_unmap(ptr); spin_unlock(&info->lock); + radix_tree_preload_end(); out: unlock_page(page); page_cache_release(page); @@ -1185,6 +1190,16 @@ repeat: goto done; error = 0; gfp = mapping_gfp_mask(mapping); + if (!filepage) { + /* + * Try to preload while we can wait, to not make a habit of + * draining atomic reserves; but don't latch on to this cpu. + */ + error = radix_tree_preload(gfp & ~__GFP_HIGHMEM); + if (error) + goto failed; + radix_tree_preload_end(); + } spin_lock(&info->lock); shmem_recalc_inode(inode); @@ -1266,7 +1281,7 @@ repeat: set_page_dirty(filepage); swap_free(swap); } else if (!(error = add_to_page_cache( - swappage, mapping, idx, GFP_ATOMIC))) { + swappage, mapping, idx, GFP_NOWAIT))) { info->flags |= SHMEM_PAGEIN; shmem_swp_set(info, entry, 0); shmem_swp_unmap(entry); @@ -1280,10 +1295,6 @@ repeat: spin_unlock(&info->lock); unlock_page(swappage); page_cache_release(swappage); - if (error == -ENOMEM) { - /* let kswapd refresh zone for GFP_ATOMICs */ - congestion_wait(WRITE, HZ/50); - } goto repeat; } } else if (sgp == SGP_READ && !filepage) { @@ -1338,7 +1349,7 @@ repeat: shmem_swp_unmap(entry); } if (error || swap.val || 0 != add_to_page_cache_lru( - filepage, mapping, idx, GFP_ATOMIC)) { + filepage, mapping, idx, GFP_NOWAIT)) { spin_unlock(&info->lock); page_cache_release(filepage); shmem_unacct_blocks(info->flags, 1);