2022-03-31 05:09:26

by Trond Myklebust

Subject: [PATCH] NFS: Replace readdir's use of xxhash() with hash_64()

From: Trond Myklebust <[email protected]>

Both xxhash() and hash_64() appear to give similarly low collision
rates with a standard linearly increasing readdir offset. They both give
similarly higher collision rates when applied to ext4's offsets.

So switch to using the standard hash_64().

Signed-off-by: Trond Myklebust <[email protected]>
---
fs/nfs/Kconfig | 4 ----
fs/nfs/dir.c | 9 +++------
2 files changed, 3 insertions(+), 10 deletions(-)

diff --git a/fs/nfs/Kconfig b/fs/nfs/Kconfig
index 47a53b3362b6..14a72224b657 100644
--- a/fs/nfs/Kconfig
+++ b/fs/nfs/Kconfig
@@ -4,10 +4,6 @@ config NFS_FS
depends on INET && FILE_LOCKING && MULTIUSER
select LOCKD
select SUNRPC
- select CRYPTO
- select CRYPTO_HASH
- select XXHASH
- select CRYPTO_XXHASH
select NFS_ACL_SUPPORT if NFS_V3_ACL
help
Choose Y here if you want to access files residing on other
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 0365063b85a2..c6b263b5faf1 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -39,7 +39,7 @@
#include <linux/sched.h>
#include <linux/kmemleak.h>
#include <linux/xattr.h>
-#include <linux/xxhash.h>
+#include <linux/hash.h>

#include "delegation.h"
#include "iostat.h"
@@ -350,10 +350,7 @@ static int nfs_readdir_page_array_append(struct page *page,
* of directory cookies. Content is addressed by the value of the
* cookie index of the first readdir entry in a page.
*
- * The xxhash algorithm is chosen because it is fast, and is supposed
- * to result in a decent flat distribution of hashes.
- *
- * We then select only the first 18 bits to avoid issues with excessive
+ * We select only the first 18 bits to avoid issues with excessive
* memory use for the page cache XArray. 18 bits should allow the caching
* of 262144 pages of sequences of readdir entries. Since each page holds
* 127 readdir entries for a typical 64-bit system, that works out to a
@@ -363,7 +360,7 @@ static pgoff_t nfs_readdir_page_cookie_hash(u64 cookie)
{
if (cookie == 0)
return 0;
- return xxhash(&cookie, sizeof(cookie), 0) & NFS_READDIR_COOKIE_MASK;
+ return hash_64(cookie, 18);
}

static bool nfs_readdir_page_validate(struct page *page, u64 last_cookie,
--
2.35.1

2022-04-05 09:51:58

by Geert Uytterhoeven

Subject: Re: [PATCH] NFS: Replace readdir's use of xxhash() with hash_64()

Hi Trond,

On Wed, 30 Mar 2022, [email protected] wrote:
> From: Trond Myklebust <[email protected]>
>
> Both xxhash() and hash_64() appear to give similarly low collision
> rates with a standard linearly increasing readdir offset. They both give
> similarly higher collision rates when applied to ext4's offsets.
>
> So switch to using the standard hash_64().
>
> Signed-off-by: Trond Myklebust <[email protected]>

Thanks for your patch!

> --- a/fs/nfs/Kconfig
> +++ b/fs/nfs/Kconfig
> @@ -4,10 +4,6 @@ config NFS_FS
> depends on INET && FILE_LOCKING && MULTIUSER
> select LOCKD
> select SUNRPC
> - select CRYPTO

I'm extremely grateful for this part! ;-)

While updating shmobile_defconfig, I was wondering why I was suddenly
asked about all those crypto drivers...

> - select CRYPTO_HASH
> - select XXHASH
> - select CRYPTO_XXHASH
> select NFS_ACL_SUPPORT if NFS_V3_ACL
> help
> Choose Y here if you want to access files residing on other

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds