Re: Slowwwwwwwwwwww NFS read performance....

[Ram Pai]

On Thu, 2004-01-15 at 11:00, Trond Myklebust wrote:
> PÃ on , 14/01/2004 klokka 21:35, skreiv Trond Myklebust:
> 
> > Err.. no...
> > 
> > I didn't have a 2.6.1-mm3 machine ready to go in our GigE testbed (I'm
> > busy compiling one up right now). However I did run a quick test on
> > 2.6.0-test11. Iozone rather than bonnie, but the results should be
> > comparable:
> 
> Hah.... They turned out not to be...
> 
> The changeset with key 
> 
> akpm@xxxxxxxx[torvalds]|ChangeSet|20031230234945|63435
> 
> # ChangeSet
> #   2003/12/30 15:49:45-08:00 akpm@xxxxxxxx
> #   [PATCH] readahead: multiple performance fixes
> #
> #   From: Ram Pai <linuxram@xxxxxxxxxx>
> 
Yes this problem has been reported earlier. Attaching Andrew's patch
that reverts a change. This should work. Please confirm.

Thanks,
RP


diff -puN mm/filemap.c~readahead-partial-backout mm/filemap.c
--- 25/mm/filemap.c~readahead-partial-backout	2004-01-09 17:41:14.000000000 -0800
+++ 25-akpm/mm/filemap.c	2004-01-09 17:41:14.000000000 -0800
@@ -587,22 +587,13 @@ void do_generic_mapping_read(struct addr
 			     read_actor_t actor)
 {
 	struct inode *inode = mapping->host;
-	unsigned long index, offset, last;
+	unsigned long index, offset;
 	struct page *cached_page;
 	int error;
 
 	cached_page = NULL;
 	index = *ppos >> PAGE_CACHE_SHIFT;
 	offset = *ppos & ~PAGE_CACHE_MASK;
-	last = (*ppos + desc->count) >> PAGE_CACHE_SHIFT;
-
-	/*
-	 * Let the readahead logic know upfront about all
-	 * the pages we'll need to satisfy this request
-	 */
-	for (; index < last; index++)
-		page_cache_readahead(mapping, ra, filp, index);
-	index = *ppos >> PAGE_CACHE_SHIFT;
 
 	for (;;) {
 		struct page *page;
@@ -621,6 +612,7 @@ void do_generic_mapping_read(struct addr
 		}
 
 		cond_resched();
+		page_cache_readahead(mapping, ra, filp, index);
 
 		nr = nr - offset;
 find_page:

_