From 90cf7c80506afb15067a018dffdac6aaf4a3bad4 Mon Sep 17 00:00:00 2001 From: Aubrey.Li Date: Wed, 17 Jan 2007 16:54:15 +0800 Subject: [PATCH] Add an interface to limit total vfs page cache. The default value is using 90% memory for page cache Signed-off-by: Aubrey.Li Signed-off-by: Thomas Chou --- linux-2.6.x/include/linux/gfp.h | 1 + linux-2.6.x/include/linux/pagemap.h | 2 +- linux-2.6.x/include/linux/sysctl.h | 2 ++ linux-2.6.x/kernel/sysctl.c | 15 +++++++++++++++ linux-2.6.x/mm/page_alloc.c | 17 +++++++++++++++-- 5 files changed, 34 insertions(+), 3 deletions(-) diff --git a/linux-2.6.x/include/linux/gfp.h b/linux-2.6.x/include/linux/gfp.h index bf2b6bc..e118944 100644 --- a/linux-2.6.x/include/linux/gfp.h +++ b/linux-2.6.x/include/linux/gfp.h @@ -46,6 +46,7 @@ struct vm_area_struct; #define __GFP_NOMEMALLOC ((__force gfp_t)0x10000u) /* Don't use emergency reserves */ #define __GFP_HARDWALL ((__force gfp_t)0x20000u) /* Enforce hardwall cpuset memory allocs */ #define __GFP_THISNODE ((__force gfp_t)0x40000u)/* No fallback, no policies */ +#define __GFP_PAGECACHE ((__force gfp_t)0x80000u) /* Is page cache allocation ? */ #define __GFP_BITS_SHIFT 20 /* Room for 20 __GFP_FOO bits */ #define __GFP_BITS_MASK ((__force gfp_t)((1 << __GFP_BITS_SHIFT) - 1)) diff --git a/linux-2.6.x/include/linux/pagemap.h b/linux-2.6.x/include/linux/pagemap.h index c3e255b..890bb23 100644 --- a/linux-2.6.x/include/linux/pagemap.h +++ b/linux-2.6.x/include/linux/pagemap.h @@ -62,7 +62,7 @@ static inline struct page *__page_cache_alloc(gfp_t gfp) static inline struct page *page_cache_alloc(struct address_space *x) { - return __page_cache_alloc(mapping_gfp_mask(x)); + return __page_cache_alloc(mapping_gfp_mask(x)|__GFP_PAGECACHE); } static inline struct page *page_cache_alloc_cold(struct address_space *x) diff --git a/linux-2.6.x/include/linux/sysctl.h b/linux-2.6.x/include/linux/sysctl.h index d98562f..0fe2cb1 100644 --- a/linux-2.6.x/include/linux/sysctl.h +++ b/linux-2.6.x/include/linux/sysctl.h @@ -202,6 +202,7 @@ enum VM_PANIC_ON_OOM=33, /* panic at out-of-memory */ VM_VDSO_ENABLED=34, /* map VDSO into new processes? */ VM_MIN_SLAB=35, /* Percent pages ignored by zone reclaim */ + VM_PAGECACHE_RATIO=36, /* percent of RAM to use as page cache */ }; @@ -964,6 +965,7 @@ extern ctl_handler sysctl_string; extern ctl_handler sysctl_intvec; extern ctl_handler sysctl_jiffies; extern ctl_handler sysctl_ms_jiffies; +extern int sysctl_pagecache_ratio; /* diff --git a/linux-2.6.x/kernel/sysctl.c b/linux-2.6.x/kernel/sysctl.c index 09e569f..ff445d2 100644 --- a/linux-2.6.x/kernel/sysctl.c +++ b/linux-2.6.x/kernel/sysctl.c @@ -1034,6 +1034,21 @@ static ctl_table vm_table[] = { .extra1 = &zero, }, #endif +/* + * NOTE: do not add new entries to this table unless you have read + * Documentation/sysctl/ctl_unnumbered.txt + */ + { + .ctl_name = VM_PAGECACHE_RATIO, + .procname = "pagecache_ratio", + .data = &sysctl_pagecache_ratio, + .maxlen = sizeof(sysctl_pagecache_ratio), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .strategy = &sysctl_intvec, + .extra1 = &zero, + .extra2 = &one_hundred, + }, { .ctl_name = 0 } }; diff --git a/linux-2.6.x/mm/page_alloc.c b/linux-2.6.x/mm/page_alloc.c index 7fbd098..3a0ac4d 100644 --- a/linux-2.6.x/mm/page_alloc.c +++ b/linux-2.6.x/mm/page_alloc.c @@ -81,6 +81,8 @@ int sysctl_lowmem_reserve_ratio[MAX_NR_ZONES-1] = { #endif }; +int sysctl_pagecache_ratio = 10; + EXPORT_SYMBOL(totalram_pages); /* @@ -892,6 +894,7 @@ failed: #define ALLOC_HARDER 0x10 /* try to alloc harder */ #define ALLOC_HIGH 0x20 /* __GFP_HIGH set */ #define ALLOC_CPUSET 0x40 /* check for correct cpuset */ +#define ALLOC_PAGECACHE 0x80 /* __GFP_PAGECACHE set */ /* * Return 1 if free pages are above 'mark'. This takes into account the order @@ -910,6 +913,9 @@ int zone_watermark_ok(struct zone *z, int order, unsigned long mark, if (alloc_flags & ALLOC_HARDER) min -= min / 4; + if (alloc_flags & ALLOC_PAGECACHE) + min = min + (sysctl_pagecache_ratio * z->present_pages) / 100; + if (free_pages <= min + z->lowmem_reserve[classzone_idx]) return 0; for (o = 0; o < order; o++) { @@ -1000,8 +1006,12 @@ restart: return NULL; } - page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, - zonelist, ALLOC_WMARK_LOW|ALLOC_CPUSET); + if (gfp_mask & __GFP_PAGECACHE) + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_WMARK_LOW|ALLOC_CPUSET|ALLOC_PAGECACHE); + else + page = get_page_from_freelist(gfp_mask|__GFP_HARDWALL, order, + zonelist, ALLOC_WMARK_LOW|ALLOC_CPUSET); if (page) goto got_pg; @@ -1027,6 +1037,9 @@ restart: if (wait) alloc_flags |= ALLOC_CPUSET; + if (gfp_mask & __GFP_PAGECACHE) + alloc_flags |= ALLOC_PAGECACHE; + /* * Go through the zonelist again. Let __GFP_HIGH and allocations * coming from realtime tasks go deeper into reserves. -- 1.5.3.3