diff mbox

[V6,2/2,RESEND] ksm: replace jhash2 with faster hash

Message ID 20180525011657.4qxrosmm3xjzo24w@xakep.localdomain (mailing list archive)
State New, archived
Headers show

Commit Message

Pavel Tatashin May 25, 2018, 1:16 a.m. UTC
Hi Timofey,

> > Do you have performance numbers of crc32c without acceleration?
> Yes, https://lkml.org/lkml/2017/12/30/222
> 
> The experimental results (the experimental value is the average of the
> measured values)
> crc32c_intel: 1084.10ns
> crc32c (no hardware acceleration): 7012.51ns
> xxhash32: 2227.75ns
> xxhash64: 1413.16ns
> jhash2: 5128.30ns

Excellent, thank you for this data.

> > I understand that losing half of the hash result might be acceptable in
> > this case, but I am not really sure how XOirng one more time can possibly
> > make hash function worse, could you please elaborate?
> 
> IIRC, because of xor are symmetric
> i.e. shift:
> 0b01011010 >> 4 = 0b0101
> and xor:
> 0b0101 ^ 0b1010 = 0b1111
> Xor will decrease randomness/entropy and will lead to hash collisions.

Makes perfect sense. Yes, XORing two random numbers reduces entropy.

> That possible to move decision from lazy load, to ksm_thread,
> that will allow us to start bench and not slowdown boot.
> 
> But for that to works, ksm must start later, after init of crypto.

After studying this dependency some more I agree, it is OK to choose hash
function where it is, but I still disagree that we must measure the
performance at runtime.

> crc32c with no hw, are slower in compare to jhash2 on x86, so i think on
> other arches result will be same.

Agreed.

Below, is your patch updated with my suggested changes.

1. Removes dependency on crc32c, use it only when it is available.
2. Do not spend time measuring the performance, choose only if there is HW
optimized implementation of crc32c is available.
3. Replace the logic with static branches.
4. Fix a couple minor bugs: 
   fastest_hash_setup() and crc32c_available() were marked as __init
   functions. Thus could be unmapped by the time they are run for the
   first time. I think section mismatch would catch those
   Removed dead code:  "desc.flags = 0", and also replaced desc with sash.
   Removed unnecessary local global "static struct shash_desc desc" this
   removes it from data page.
   Fixed few spelling errors, and other minor changes to pass
   ./scripts/checkpatch.pl

The patch is untested, but should work. Please let me know if you agree
with the changes. If so, you can test and resubmit the series.

Thank you,
Pavel

Patch:

Comments

kernel test robot May 26, 2018, 8:25 p.m. UTC | #1
Hi Pavel,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on mmotm/master]
[also build test ERROR on v4.17-rc6]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Pavel-Tatashin/ksm-replace-jhash2-with-faster-hash/20180527-032512
base:   git://git.cmpxchg.org/linux-mmotm.git master
config: i386-randconfig-x078-201821 (attached as .config)
compiler: gcc-7 (Debian 7.3.0-16) 7.3.0
reproduce:
        # save the attached .config to linux build tree
        make ARCH=i386 

All errors (new ones prefixed by >>):

   mm/ksm.c: In function 'fasthash':
>> mm/ksm.c:338:15: error: implicit declaration of function 'xxhash'; did you mean 'jhash'? [-Werror=implicit-function-declaration]
      return (u32)xxhash(input, length, 0);
                  ^~~~~~
                  jhash
   cc1: some warnings being treated as errors

vim +338 mm/ksm.c

   331	
   332	static u32 fasthash(const void *input, size_t length)
   333	{
   334		if (static_branch_likely(&ksm_use_crc32c))
   335			return crc32c(0, input, length);
   336	
   337		if (static_branch_likely(&ksm_use_xxhash))
 > 338			return (u32)xxhash(input, length, 0);
   339	
   340		/* Is done only once on the first call of fasthash() */
   341		fasthash_setup();
   342	
   343		/* Now, that we know the hash alg., calculate checksum for zero page */
   344		zero_checksum = fasthash(ZERO_PAGE(0), PAGE_SIZE);
   345	
   346		return fasthash(input, length);
   347	}
   348	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation
kernel test robot May 26, 2018, 9:06 p.m. UTC | #2
Hi Pavel,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on mmotm/master]
[also build test ERROR on v4.17-rc6]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Pavel-Tatashin/ksm-replace-jhash2-with-faster-hash/20180527-032512
base:   git://git.cmpxchg.org/linux-mmotm.git master
config: i386-randconfig-s0-201821 (attached as .config)
compiler: gcc-6 (Debian 6.4.0-9) 6.4.0 20171026
reproduce:
        # save the attached .config to linux build tree
        make ARCH=i386 

All errors (new ones prefixed by >>):

   mm/ksm.c: In function 'fasthash':
>> mm/ksm.c:338:15: error: implicit declaration of function 'xxhash' [-Werror=implicit-function-declaration]
      return (u32)xxhash(input, length, 0);
                  ^~~~~~
   cc1: some warnings being treated as errors

vim +/xxhash +338 mm/ksm.c

   331	
   332	static u32 fasthash(const void *input, size_t length)
   333	{
   334		if (static_branch_likely(&ksm_use_crc32c))
   335			return crc32c(0, input, length);
   336	
   337		if (static_branch_likely(&ksm_use_xxhash))
 > 338			return (u32)xxhash(input, length, 0);
   339	
   340		/* Is done only once on the first call of fasthash() */
   341		fasthash_setup();
   342	
   343		/* Now, that we know the hash alg., calculate checksum for zero page */
   344		zero_checksum = fasthash(ZERO_PAGE(0), PAGE_SIZE);
   345	
   346		return fasthash(input, length);
   347	}
   348	

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation
diff mbox

Patch

==========================================================================
From d5f9ecb89ac5de7467fe587b6ccdead39ee00049 Mon Sep 17 00:00:00 2001
From: Timofey Titovets <nefelim4ag@gmail.com>
Date: Wed, 18 Apr 2018 22:32:20 +0300
Subject: [PATCH] ksm: replace jhash2 with faster hash

1. Pickup, Sioh Lee crc32 patch, after some long conversation
2. Merge with my work on xxhash
3. Add autoselect code to choice fastest hash helper.

Base idea are same, replace jhash2 with something faster.

Perf numbers:
Intel(R) Xeon(R) CPU E5-2420 v2 @ 2.20GHz
ksm: crc32c   hash() 12081 MB/s
ksm: xxh64    hash()  8770 MB/s
ksm: xxh32    hash()  4529 MB/s
ksm: jhash2   hash()  1569 MB/s

As jhash2 always will be slower (for data size like PAGE_SIZE),
just drop it from choice.

Add function to autoselect hash algo during first page merging run.

Move init of zero_checksum from init, to first call of fasthash():
  1. KSM Init run on early kernel init,
     run perf testing stuff on main kernel boot thread looks bad to me.
  2. Crypto subsystem not available at that early booting,
     so crc32c even, compiled in, not available
     As crypto and ksm init, run at subsys_initcall() (4) kernel level of
     init, all possible consumers will run later at 5+ levels

Output after first try of KSM to hash page:
ksm: using crc32c as hash function

Thanks.

Changes:
  v1 -> v2:
    - Move xxhash() to xxhash.h/c and separate patches
  v2 -> v3:
    - Move xxhash() xxhash.c -> xxhash.h
    - replace xxhash_t with 'unsigned long'
    - update kerneldoc above xxhash()
  v3 -> v4:
    - Merge xxhash/crc32 patches
    - Replace crc32 with crc32c (crc32 have same as jhash2 speed)
    - Add auto speed test and auto choice of fastest hash function
  v4 -> v5:
    - Pickup missed xxhash patch
    - Update code with compile time choicen xxhash
    - Add more macros to make code more readable
    - As now that only possible use xxhash or crc32c,
      on crc32c allocation error, skip speed test and fallback to xxhash
    - For workaround too early init problem (crc32c not available),
      move zero_checksum init to first call of fastcall()
    - Don't alloc page for hash testing, use arch zero pages for that
  v5 -> v6:
    - Use libcrc32c instead of CRYPTO API, mainly for
      code/Kconfig deps Simplification
    - Add crc32c_available():
      libcrc32c will BUG_ON on crc32c problems,
      so test crc32c available by crc32c_available()
    - Simplify choice_fastest_hash()
    - Simplify fasthash()
    - struct rmap_item && stable_node have sizeof == 64 on x86_64,
      that makes them cache friendly. As we don't suffer from hash
      collisions,  change hash type from unsigned long back to u32.
    - Fix kbuild robot warning, make all local functions static

Signed-off-by: Timofey Titovets <nefelim4ag@gmail.com>
Signed-off-by: leesioh <solee@os.korea.ac.kr>
Reviewed-by: Pavel Tatashin <pasha.tatashin@oracle.com>

CC: Andrea Arcangeli <aarcange@redhat.com>
CC: linux-mm@kvack.org
CC: kvm@vger.kernel.org
---
 mm/Kconfig |  1 +
 mm/ksm.c   | 49 +++++++++++++++++++++++++++++++++++++++++++++----
 2 files changed, 46 insertions(+), 4 deletions(-)

diff --git a/mm/Kconfig b/mm/Kconfig
index e14c01513bfd..6f46cda5c8ed 100644
--- a/mm/Kconfig
+++ b/mm/Kconfig
@@ -298,6 +298,7 @@  config MMU_NOTIFIER
 config KSM
 	bool "Enable KSM for page merging"
 	depends on MMU
+	select XXHASH
 	help
 	  Enable Kernel Samepage Merging: KSM periodically scans those areas
 	  of an application's address space that an app has advised may be
diff --git a/mm/ksm.c b/mm/ksm.c
index e3cbf9a92f3c..1ce4ce6dc313 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -25,7 +25,6 @@ 
 #include <linux/pagemap.h>
 #include <linux/rmap.h>
 #include <linux/spinlock.h>
-#include <linux/jhash.h>
 #include <linux/delay.h>
 #include <linux/kthread.h>
 #include <linux/wait.h>
@@ -39,6 +38,9 @@ 
 #include <linux/freezer.h>
 #include <linux/oom.h>
 #include <linux/numa.h>
+#include <crypto/hash.h>
+#include <linux/crc32c.h>
+#include <linux/xxhash.h>
 
 #include <asm/tlbflush.h>
 #include "internal.h"
@@ -284,6 +286,47 @@  static DEFINE_SPINLOCK(ksm_mmlist_lock);
 		sizeof(struct __struct), __alignof__(struct __struct),\
 		(__flags), NULL)
 
+static DEFINE_STATIC_KEY_FALSE(ksm_use_crc32c);
+static DEFINE_STATIC_KEY_FALSE(ksm_use_xxhash);
+
+static void fasthash_setup(void)
+{
+	struct crypto_shash *shash = crypto_alloc_shash("crc32c", 0, 0);
+
+	if (!IS_ERR(shash)) {
+		/* Use crc32c if any non-generic version is available.
+		 * Generic crypto algorithms have priority 100.
+		 */
+		if (crypto_tfm_alg_priority(&shash->base) > 100) {
+			static_branch_enable(&ksm_use_crc32c);
+			pr_info("ksm: using crc32c as hash function");
+		}
+		crypto_free_shash(shash);
+	}
+
+	if (!static_branch_likely(&ksm_use_crc32c)) {
+		static_branch_enable(&ksm_use_xxhash);
+		pr_info("ksm: using xxhash as hash function");
+	}
+}
+
+static u32 fasthash(const void *input, size_t length)
+{
+	if (static_branch_likely(&ksm_use_crc32c))
+		return crc32c(0, input, length);
+
+	if (static_branch_likely(&ksm_use_xxhash))
+		return (u32)xxhash(input, length, 0);
+
+	/* Is done only once on the first call of fasthash() */
+	fasthash_setup();
+
+	/* Now, that we know the hash alg., calculate checksum for zero page */
+	zero_checksum = fasthash(ZERO_PAGE(0), PAGE_SIZE);
+
+	return fasthash(input, length);
+}
+
 static int __init ksm_slab_init(void)
 {
 	rmap_item_cache = KSM_KMEM_CACHE(rmap_item, 0);
@@ -979,7 +1022,7 @@  static u32 calc_checksum(struct page *page)
 {
 	u32 checksum;
 	void *addr = kmap_atomic(page);
-	checksum = jhash2(addr, PAGE_SIZE / 4, 17);
+	checksum = fasthash(addr, PAGE_SIZE);
 	kunmap_atomic(addr);
 	return checksum;
 }
@@ -3100,8 +3143,6 @@  static int __init ksm_init(void)
 	struct task_struct *ksm_thread;
 	int err;
 
-	/* The correct value depends on page size and endianness */
-	zero_checksum = calc_checksum(ZERO_PAGE(0));
 	/* Default to false for backwards compatibility */
 	ksm_use_zero_pages = false;