linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
* [PATCH]shmem: reduce one time of locking in pagefault
@ 2010-07-07  1:15 Shaohua Li
  2010-07-07  1:32 ` Andrew Morton
  2010-07-09  1:13 ` Hugh Dickins
  0 siblings, 2 replies; 6+ messages in thread
From: Shaohua Li @ 2010-07-07  1:15 UTC (permalink / raw)
  To: lkml, linux-mm; +Cc: Andrew Morton, Andi Kleen, Zhang, Yanmin

[-- Attachment #1: Type: text/plain, Size: 3922 bytes --]

I'm running a shmem pagefault test case (see attached file) under a 64 CPU
system. Profile shows shmem_inode_info->lock is heavily contented and 100%
CPUs time are trying to get the lock. In the pagefault (no swap) case,
shmem_getpage gets the lock twice, the last one is avoidable if we prealloc a
page so we could reduce one time of locking. This is what below patch does.

The result of the test case:
2.6.35-rc3: ~20s
2.6.35-rc3 + patch: ~12s
so this is 40% improvement.

One might argue if we could have better locking for shmem. But even shmem is lockless,
the pagefault will soon have pagecache lock heavily contented because shmem must add
new page to pagecache. So before we have better locking for pagecache, improving shmem
locking doesn't have too much improvement. I did a similar pagefault test against
a ramfs file, the test result is ~10.5s.

Signed-off-by: Shaohua Li <shaohua.li@intel.com>

diff --git a/mm/shmem.c b/mm/shmem.c
index f65f840..c5f2939 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1223,6 +1223,7 @@ static int shmem_getpage(struct inode *inode, unsigned long idx,
 	struct shmem_sb_info *sbinfo;
 	struct page *filepage = *pagep;
 	struct page *swappage;
+	struct page *prealloc_page = NULL;
 	swp_entry_t *entry;
 	swp_entry_t swap;
 	gfp_t gfp;
@@ -1247,7 +1248,6 @@ repeat:
 		filepage = find_lock_page(mapping, idx);
 	if (filepage && PageUptodate(filepage))
 		goto done;
-	error = 0;
 	gfp = mapping_gfp_mask(mapping);
 	if (!filepage) {
 		/*
@@ -1258,7 +1258,19 @@ repeat:
 		if (error)
 			goto failed;
 		radix_tree_preload_end();
+		if (sgp != SGP_READ) {
+			/* don't care if this successes */
+			prealloc_page = shmem_alloc_page(gfp, info, idx);
+			if (prealloc_page) {
+				if (mem_cgroup_cache_charge(prealloc_page,
+				    current->mm, GFP_KERNEL)) {
+					page_cache_release(prealloc_page);
+					prealloc_page = NULL;
+				}
+			}
+		}
 	}
+	error = 0;
 
 	spin_lock(&info->lock);
 	shmem_recalc_inode(inode);
@@ -1407,28 +1419,37 @@ repeat:
 		if (!filepage) {
 			int ret;
 
-			spin_unlock(&info->lock);
-			filepage = shmem_alloc_page(gfp, info, idx);
-			if (!filepage) {
-				shmem_unacct_blocks(info->flags, 1);
-				shmem_free_blocks(inode, 1);
-				error = -ENOMEM;
-				goto failed;
-			}
-			SetPageSwapBacked(filepage);
+			if (!prealloc_page) {
+				spin_unlock(&info->lock);
+				filepage = shmem_alloc_page(gfp, info, idx);
+				if (!filepage) {
+					shmem_unacct_blocks(info->flags, 1);
+					shmem_free_blocks(inode, 1);
+					error = -ENOMEM;
+					goto failed;
+				}
+				SetPageSwapBacked(filepage);
 
-			/* Precharge page while we can wait, compensate after */
-			error = mem_cgroup_cache_charge(filepage, current->mm,
-					GFP_KERNEL);
-			if (error) {
-				page_cache_release(filepage);
-				shmem_unacct_blocks(info->flags, 1);
-				shmem_free_blocks(inode, 1);
-				filepage = NULL;
-				goto failed;
+				/* Precharge page while we can wait, compensate
+				 * after
+				 */
+				error = mem_cgroup_cache_charge(filepage,
+					current->mm, GFP_KERNEL);
+				if (error) {
+					page_cache_release(filepage);
+					shmem_unacct_blocks(info->flags, 1);
+					shmem_free_blocks(inode, 1);
+					filepage = NULL;
+					goto failed;
+				}
+
+				spin_lock(&info->lock);
+			} else {
+				filepage = prealloc_page;
+				prealloc_page = NULL;
+				SetPageSwapBacked(filepage);
 			}
 
-			spin_lock(&info->lock);
 			entry = shmem_swp_alloc(info, idx, sgp);
 			if (IS_ERR(entry))
 				error = PTR_ERR(entry);
@@ -1469,6 +1490,10 @@ repeat:
 	}
 done:
 	*pagep = filepage;
+	if (prealloc_page) {
+		mem_cgroup_uncharge_cache_page(prealloc_page);
+		page_cache_release(prealloc_page);
+	}
 	return 0;
 
 failed:
@@ -1476,6 +1501,10 @@ failed:
 		unlock_page(filepage);
 		page_cache_release(filepage);
 	}
+	if (prealloc_page) {
+		mem_cgroup_uncharge_cache_page(prealloc_page);
+		page_cache_release(prealloc_page);
+	}
 	return error;
 }
 


[-- Attachment #2: shmem-test.c --]
[-- Type: text/x-csrc, Size: 1293 bytes --]

#include <sys/mman.h>
#include <sys/time.h>
#include <unistd.h>
#include <pthread.h>
#include <stdlib.h>
#include <sys/types.h>
#include <sys/stat.h>

#define THREAD_NUM (64L)
#define MEM_SIZE (1024*1024*1024*16L)
#define PER_TH_SIZE (MEM_SIZE/THREAD_NUM)

void *thread_func(void *data)
{
	char *addr = data;
	unsigned long size = PER_TH_SIZE, index = 0;
	int t;

	while (index < size) {
		t = *(addr + index);
		index += 4096;
	}
}

int main(int argc, char *argv[])
{
	int i;
	pthread_t threads[THREAD_NUM];
	pthread_attr_t attr;
	struct timeval start, stop, diff;
	char *mem;

	mem = mmap(NULL, MEM_SIZE, PROT_READ|PROT_WRITE,
		MAP_SHARED|MAP_ANON, 0, 0);
	if (!mem) {
		perror("mmap error");
		exit(1);
	}
	
	pthread_attr_init(&attr);
	pthread_attr_setdetachstate(&attr, PTHREAD_CREATE_JOINABLE);
	gettimeofday(&start, NULL);
	for (i = 0; i < THREAD_NUM; i++)
		if (pthread_create(&threads[i], &attr, thread_func,
			mem + PER_TH_SIZE * i)) {
			perror("thread create error");
			exit(1);
		}

	for (i=0; i< THREAD_NUM; i++)
		pthread_join(threads[i], NULL);

	gettimeofday(&stop, NULL);
	timersub(&stop, &start, &diff);
	printf("Thread %ld Mem %dG time %lu.%03lusec\n",
		THREAD_NUM, MEM_SIZE/1024/1024/1024,
		diff.tv_sec, diff.tv_usec/1000);

	pthread_attr_destroy(&attr);
	return 0;
}

^ permalink raw reply related	[flat|nested] 6+ messages in thread
end of thread, other threads:[~2010-07-09  2:52 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-07-07  1:15 [PATCH]shmem: reduce one time of locking in pagefault Shaohua Li
2010-07-07  1:32 ` Andrew Morton
2010-07-07  1:39   ` Shaohua Li
2010-07-09  1:28     ` Hugh Dickins
2010-07-09  1:13 ` Hugh Dickins
2010-07-09  2:52   ` Shaohua Li

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).