From 0c23a5624a80903fba7e635e8c44f38a79caf223 Mon Sep 17 00:00:00 2001
From: Sage Weil <sage@inktank.com>
Date: Mon, 9 Sep 2013 18:18:26 -0700
Subject: [PATCH] common/buffer: cache crcs in buffer::raw
Signed-off-by: Samuel Just <sam.just@inktank.com>
---
src/common/buffer.cc | 69 +++++++++++++++++++++++++++++++++++++++---
src/test/bufferlist.cc | 17 +++++++++++
2 files changed, 81 insertions(+), 5 deletions(-)
diff --git a/src/common/buffer.cc b/src/common/buffer.cc
index 8da4c106d1b..09dac01db17 100644
--- a/src/common/buffer.cc
+++ b/src/common/buffer.cc
@@ -21,6 +21,7 @@
#include "include/atomic.h"
#include "include/types.h"
#include "include/compat.h"
+#include "include/Spinlock.h"
#include <errno.h>
#include <fstream>
@@ -30,6 +31,9 @@
namespace ceph {
+
+static unsigned char zbuf[128];
+
#ifdef BUFFER_DEBUG
static uint32_t simple_spinlock_t buffer_debug_lock = SIMPLE_SPINLOCK_INITIALIZER;
# define bdout { simple_spin_lock(&buffer_debug_lock); std::cout
@@ -60,9 +64,14 @@ bool buffer_track_alloc = get_env_bool("CEPH_BUFFER_TRACK");
unsigned len;
atomic_t nref;
- raw(unsigned l) : data(NULL), len(l), nref(0)
+ Spinlock crc_lock;
+ map<pair<off_t, off_t>, pair<int64_t, int64_t> > crc_map;
+ int64_t crc_in; ///< cached crc base; -1 if invalid
+ int64_t crc_out; ///< cached crc value; -1 if invalid
+
+ raw(unsigned l) : data(NULL), len(l), nref(0), crc_in(-1), crc_out(-1)
{ }
- raw(char *c, unsigned l) : data(c), len(l), nref(0)
+ raw(char *c, unsigned l) : data(c), len(l), nref(0), crc_in(-1), crc_out(-1)
{ }
virtual ~raw() {};
@@ -77,12 +86,31 @@ bool buffer_track_alloc = get_env_bool("CEPH_BUFFER_TRACK");
return c;
}
+ unsigned length() const {
+ return len;
+ }
+
bool is_page_aligned() {
return ((long)data & ~CEPH_PAGE_MASK) == 0;
}
bool is_n_page_sized() {
return (len & ~CEPH_PAGE_MASK) == 0;
}
+ bool get_crc(const pair<off_t, off_t> &fromto,
+ pair<int64_t, int64_t> *crc) const {
+ Spinlock::Locker l(crc_lock);
+ map<pair<off_t, off_t>, pair<int64_t, int64_t> >::const_iterator i =
+ crc_map.find(fromto);
+ if (i == crc_map.end())
+ return false;
+ *crc = i->second;
+ return true;
+ }
+ void set_crc(const pair<off_t, off_t> &fromto,
+ const pair<uint64_t, uint64_t> &crc) {
+ Spinlock::Locker l(crc_lock);
+ crc_map[fromto] = crc;
+ }
};
class buffer::raw_malloc : public buffer::raw {
@@ -1274,9 +1302,40 @@ __u32 buffer::list::crc32c(__u32 crc) const
{
for (std::list<ptr>::const_iterator it = _buffers.begin();
it != _buffers.end();
- ++it)
- if (it->length())
- crc = ceph_crc32c(crc, (unsigned char*)it->c_str(), it->length());
+ ++it) {
+ if (it->length()) {
+ raw *r = it->get_raw();
+ pair<off_t, off_t> ofs(it->offset(), it->offset() + it->length());
+ pair<int64_t, int64_t> ccrc;
+ if (r->get_crc(ofs, &ccrc)) {
+ if (ccrc.first == crc) {
+ // got it already
+ crc = ccrc.second;
+ } else {
+ /* If we have cached crc32c(buf, v) for initial value v,
+ * we can convert this to a different initial value v' by:
+ * crc32c(buf, v') = crc32c(buf, v) ^ adjustment
+ * where adjustment = crc32c(0*len(buf), v ^ v')
+ *
+ * http://crcutil.googlecode.com/files/crc-doc.1.0.pdf
+ * note, u for our crc32c implementation is 0
+ */
+ int64_t adjustment = ccrc.first ^ crc;
+ size_t remaining = it->length();
+ for (; remaining > sizeof(zbuf); remaining -= sizeof(zbuf)) {
+ adjustment = ceph_crc32c(adjustment, zbuf, sizeof(zbuf));
+ }
+ if (remaining)
+ adjustment = ceph_crc32c(adjustment, zbuf, remaining);
+ crc = ccrc.second ^ adjustment;
+ }
+ } else {
+ uint32_t base = crc;
+ crc = ceph_crc32c(crc, (unsigned char*)it->c_str(), it->length());
+ r->set_crc(ofs, make_pair(base, crc));
+ }
+ }
+ }
return crc;
}
diff --git a/src/test/bufferlist.cc b/src/test/bufferlist.cc
index b23bd33e55a..beae868c0da 100644
--- a/src/test/bufferlist.cc
+++ b/src/test/bufferlist.cc
@@ -1649,6 +1649,23 @@ TEST(BufferList, crc32c) {
EXPECT_EQ((unsigned)0x5FA5C0CC, crc);
}
+TEST(BufferList, crc32cappend) {
+ bufferlist bl1;
+ bufferlist bl2;
+
+ for (int j = 0; j < 200; ++j) {
+ bufferlist bl;
+ for (int i = 0; i < 200; ++i) {
+ char x = rand();
+ bl.append(x);
+ bl1.append(x);
+ }
+ bl.crc32c(rand()); // mess with the cached bufferptr crc values
+ bl2.append(bl);
+ }
+ ASSERT_EQ(bl1.crc32c(0), bl2.crc32c(0));
+}
+
TEST(BufferList, compare) {
bufferlist a;
a.append("A");