</para>
</listitem>
</varlistentry>
+
+ <varlistentry id="guc-bgwriter-flush-after" xreflabel="bgwriter_flush_after">
+ <term><varname>bgwriter_flush_after</varname> (<type>int</type>)
+ <indexterm>
+ <primary><varname>bgwriter_flush_after</> configuration parameter</primary>
+ </indexterm>
+ </term>
+ <listitem>
+ <para>
+ Whenever more than <varname>bgwriter_flush_after</varname> bytes have
+ been written by the bgwriter, hint to OS to flush these writes to the
+ underlying storage. Doing so will limit the amount of dirty data in
+ the kernel's page cache, reducing the likelihood of stalls when fsync
+ is issued at the end of a checkpoint, or when the OS writes out data
+ in larger batches in the background. Often that will result in
+ greatly reduced transaction latency, but there also are some cases,
+ especially with workloads that are bigger than <xref
+ linkend="guc-shared-buffers">, but smaller than the OS's page cache,
+ where performance might degrade. This setting may have no effect on
+ some platforms. <literal>0</literal> disables controlled flushing.
+ The default is <literal>256Kb</> on Linux, <literal>0</> otherwise.
+ This parameter can only be set in the <filename>postgresql.conf</>
+ file or on the server command line.
+ </para>
+ </listitem>
+ </varlistentry>
</variablelist>
<para>
</para>
</listitem>
</varlistentry>
+
+ <varlistentry id="guc-backend-flush-after" xreflabel="backend_flush_after">
+ <term><varname>backend_flush_after</varname> (<type>int</type>)
+ <indexterm>
+ <primary><varname>backend_flush_after</> configuration parameter</primary>
+ </indexterm>
+ </term>
+ <listitem>
+ <para>
+ Whenever more than <varname>backend_flush_after</varname> bytes have
+ been written by a single backend, hint to OS to flush these writes to
+ the underlying storage. Doing so will limit the amount of dirty data
+ in the kernel's page cache, reducing the likelihood of stalls when
+ fsync is issued at the end of a checkpoint, or when the OS writes out
+ data in larger batches in the background. Often that will result in
+ greatly reduced transaction latency, but there also are some cases,
+ especially with workloads that are bigger than <xref
+ linkend="guc-shared-buffers">, but smaller than the OS's page cache,
+ where performance might degrade. Note that because
+ <varname>backend_flush_after</varname> is per-backend, the total
+ amount of dirty data in the kerne's page cache can be considerably
+ bigger than this setting. This setting may have no effect on some
+ platforms. <literal>0</literal> disables controlled flushing. The
+ default is <literal>256Kb</> on Linux, <literal>0</> otherwise. This
+ parameter can only be set in the <filename>postgresql.conf</> file or
+ on the server command line.
+ </para>
+ </listitem>
+ </varlistentry>
</variablelist>
</sect2>
</sect1>
</listitem>
</varlistentry>
+ <varlistentry id="guc-checkpoint-flush-after" xreflabel="checkpoint_flush_after">
+ <term><varname>checkpoint_flush_after</varname> (<type>int</type>)
+ <indexterm>
+ <primary><varname>checkpoint_flush_after</> configuration parameter</primary>
+ </indexterm>
+ </term>
+ <listitem>
+ <para>
+ Whenever more than <varname>checkpoint_flush_after</varname> bytes
+ have been written while performing a checkpoint, hint to OS to flush
+ these writes to the underlying storage. Doing so will limit the
+ amount of dirty data in the kernel's page cache, reducing the
+ likelihood of stalls when fsync is issued at the end of a checkpoint,
+ or when the OS writes out data in larger batches in the background.
+ Often that will result in greatly reduced transaction latency, but
+ there also are some cases, especially with workloads that are bigger
+ than <xref linkend="guc-shared-buffers">, but smaller than the OS's
+ page cache, where performance might degrade. This setting may have no
+ effect on some platforms. <literal>0</literal> disables controlled
+ flushing. The default is <literal>256Kb</> on Linux, <literal>0</>
+ otherwise. This parameter can only be set in the
+ <filename>postgresql.conf</> file or on the server command line.
+ </para>
+ </listitem>
+ </varlistentry>
+
<varlistentry id="guc-checkpoint-warning" xreflabel="checkpoint_warning">
<term><varname>checkpoint_warning</varname> (<type>integer</type>)
<indexterm>
unexpected variation in the number of WAL segments needed.
</para>
+ <para>
+ On Linux and POSIX platforms <xref linkend="guc-checkpoint-flush-after">
+ allows to guide the OS that pages written by the checkpoint should be
+ flushed to disk. Otherwise, these pages may be kept in the OS's page
+ cache, inducing a stall when <literal>fsync</> is called later. This
+ setting helps to reduce transaction latency, but it also can an adverse
+ effect on performance; particularly for workloads that are bigger than
+ <xref linkend="guc-shared-buffers">, but smaller than the OS's page cache.
+ It should be beneficial for high write loads on HDD. This feature probably
+ brings no benefit on SSD, as the I/O write latency is small on such
+ hardware, thus it may be disabled.
+ </para>
+
<para>
The number of WAL segment files in <filename>pg_xlog</> directory depends on
<varname>min_wal_size</>, <varname>max_wal_size</> and
sigjmp_buf local_sigjmp_buf;
MemoryContext bgwriter_context;
bool prev_hibernate;
+ WritebackContext wb_context;
/*
* Properly accept or ignore signals the postmaster might send us.
ALLOCSET_DEFAULT_MAXSIZE);
MemoryContextSwitchTo(bgwriter_context);
+ WritebackContextInit(&wb_context, &bgwriter_flush_after);
+
/*
* If an exception is encountered, processing resumes here.
*
/* Flush any leaked data in the top-level context */
MemoryContextResetAndDeleteChildren(bgwriter_context);
+ /* re-initilialize to avoid repeated errors causing problems */
+ WritebackContextInit(&wb_context, &bgwriter_flush_after);
+
/* Now we can allow interrupts again */
RESUME_INTERRUPTS();
/*
* Do one cycle of dirty-buffer writing.
*/
- can_hibernate = BgBufferSync();
+ can_hibernate = BgBufferSync(&wb_context);
/*
* Send off activity statistics to the stats collector
LWLockMinimallyPadded *BufferIOLWLockArray = NULL;
LWLockTranche BufferIOLWLockTranche;
LWLockTranche BufferContentLWLockTranche;
+WritebackContext BackendWritebackContext;
/*
/* Init other shared buffer-management stuff */
StrategyInitialize(!foundDescs);
+
+ /* Initialize per-backend file flush context */
+ WritebackContextInit(&BackendWritebackContext,
+ &backend_flush_after);
}
/*
bool track_io_timing = false;
int effective_io_concurrency = 0;
+/*
+ * GUC variables about triggering kernel writeback for buffers written; OS
+ * dependant defaults are set via the GUC mechanism.
+ */
+int checkpoint_flush_after = 0;
+int bgwriter_flush_after = 0;
+int backend_flush_after = 0;
+
/*
* How many buffers PrefetchBuffer callers should try to stay ahead of their
* ReadBuffer calls by. This is maintained by the assign hook for
static void PinBuffer_Locked(BufferDesc *buf);
static void UnpinBuffer(BufferDesc *buf, bool fixOwner);
static void BufferSync(int flags);
-static int SyncOneBuffer(int buf_id, bool skip_recently_used);
+static int SyncOneBuffer(int buf_id, bool skip_recently_used, WritebackContext *flush_context);
static void WaitIO(BufferDesc *buf);
static bool StartBufferIO(BufferDesc *buf, bool forInput);
static void TerminateBufferIO(BufferDesc *buf, bool clear_dirty,
static void AtProcExit_Buffers(int code, Datum arg);
static void CheckForBufferLeaks(void);
static int rnode_comparator(const void *p1, const void *p2);
+static int buffertag_comparator(const void *p1, const void *p2);
/*
MemSet((char *) bufBlock, 0, BLCKSZ);
/* don't set checksum for all-zero page */
smgrextend(smgr, forkNum, blockNum, (char *) bufBlock, false);
+
+ /*
+ * XXX: Note that we're *not* doing a ScheduleBufferTagForWriteback
+ * here. At least on linux doing so defeats 'delayed allocation',
+ * leading to more fragmented files.
+ */
}
else
{
FlushBuffer(buf, NULL);
LWLockRelease(BufferDescriptorGetContentLock(buf));
+ ScheduleBufferTagForWriteback(&BackendWritebackContext,
+ &buf->tag);
+
TRACE_POSTGRESQL_BUFFER_WRITE_DIRTY_DONE(forkNum, blockNum,
smgr->smgr_rnode.node.spcNode,
smgr->smgr_rnode.node.dbNode,
int num_to_write;
int num_written;
int mask = BM_DIRTY;
+ WritebackContext wb_context;
/* Make sure we can handle the pin inside SyncOneBuffer */
ResourceOwnerEnlargeBuffers(CurrentResourceOwner);
if (num_to_write == 0)
return; /* nothing to do */
+
+ WritebackContextInit(&wb_context, &checkpoint_flush_after);
+
TRACE_POSTGRESQL_BUFFER_SYNC_START(NBuffers, num_to_write);
/*
*/
if (bufHdr->flags & BM_CHECKPOINT_NEEDED)
{
- if (SyncOneBuffer(buf_id, false) & BUF_WRITTEN)
+ if (SyncOneBuffer(buf_id, false, &wb_context) & BUF_WRITTEN)
{
TRACE_POSTGRESQL_BUFFER_SYNC_WRITTEN(buf_id);
BgWriterStats.m_buf_written_checkpoints++;
* bgwriter_lru_maxpages to 0.)
*/
bool
-BgBufferSync(void)
+BgBufferSync(WritebackContext *wb_context)
{
/* info obtained from freelist.c */
int strategy_buf_id;
/* Execute the LRU scan */
while (num_to_scan > 0 && reusable_buffers < upcoming_alloc_est)
{
- int buffer_state = SyncOneBuffer(next_to_clean, true);
+ int buffer_state = SyncOneBuffer(next_to_clean, true,
+ wb_context);
if (++next_to_clean >= NBuffers)
{
* Note: caller must have done ResourceOwnerEnlargeBuffers.
*/
static int
-SyncOneBuffer(int buf_id, bool skip_recently_used)
+SyncOneBuffer(int buf_id, bool skip_recently_used, WritebackContext *wb_context)
{
BufferDesc *bufHdr = GetBufferDescriptor(buf_id);
int result = 0;
+ BufferTag tag;
ReservePrivateRefCountEntry();
FlushBuffer(bufHdr, NULL);
LWLockRelease(BufferDescriptorGetContentLock(bufHdr));
+
+ tag = bufHdr->tag;
+
UnpinBuffer(bufHdr, true);
+ ScheduleBufferTagForWriteback(wb_context, &tag);
+
return result | BUF_WRITTEN;
}
else
return 0;
}
+
+
+/*
+ * BufferTag comparator.
+ */
+static int
+buffertag_comparator(const void *a, const void *b)
+{
+ const BufferTag *ba = (const BufferTag *) a;
+ const BufferTag *bb = (const BufferTag *) b;
+ int ret;
+
+ ret = rnode_comparator(&ba->rnode, &bb->rnode);
+
+ if (ret != 0)
+ return ret;
+
+ if (ba->forkNum < bb->forkNum)
+ return -1;
+ if (ba->forkNum > bb->forkNum)
+ return 1;
+
+ if (ba->blockNum < bb->blockNum)
+ return -1;
+ if (ba->blockNum > bb->blockNum)
+ return 1;
+
+ return 0;
+}
+
+
+/*
+ * Initialize a writeback context, discarding potential previous state.
+ *
+ * *max_coalesce is a pointer to a variable containing the current maximum
+ * number of writeback requests that will be coalesced into a bigger one. A
+ * value <= 0 means that no writeback control will be performed. max_pending
+ * is a pointer instead of an immediate value, so the coalesce limits can
+ * easily changed by the GUC mechanism, and so calling code does not have to
+ * check the current config variables.
+ */
+void
+WritebackContextInit(WritebackContext *context, int *max_pending)
+{
+ Assert(*max_pending <= WRITEBACK_MAX_PENDING_FLUSHES);
+
+ context->max_pending = max_pending;
+ context->nr_pending = 0;
+}
+
+
+/*
+ * Add buffer to list of pending writeback requests.
+ */
+void
+ScheduleBufferTagForWriteback(WritebackContext *context, BufferTag *tag)
+{
+ PendingWriteback *pending;
+
+ /* nothing to do if flushing is disabled */
+ if (*context->max_pending <= 0 && context->nr_pending <= 0)
+ return;
+
+ Assert(*context->max_pending <= WRITEBACK_MAX_PENDING_FLUSHES);
+
+ pending = &context->pending_writebacks[context->nr_pending++];
+
+ pending->tag = *tag;
+
+ if (context->nr_pending >= *context->max_pending)
+ IssuePendingWritebacks(context);
+}
+
+/*
+ * Issue all pending writeback requests, previously scheduled with
+ * ScheduleBufferTagForWriteback, to the OS.
+ *
+ * Because this is only used to improve the OSs IO scheduling we try to never
+ * error out - it's just a hint.
+ */
+void
+IssuePendingWritebacks(WritebackContext *context)
+{
+ int i;
+
+ if (context->nr_pending == 0)
+ return;
+
+ /*
+ * Executing the writes in-order can make them a lot faster, and allows to
+ * merge writeback requests to consecutive blocks into larger writebacks.
+ */
+ qsort(&context->pending_writebacks, context->nr_pending,
+ sizeof(PendingWriteback), buffertag_comparator);
+
+ /*
+ * Coalesce neighbouring writes, but nothing else. For that we iterate
+ * through the, now sorted, array of pending flushes, and look forward to
+ * find all neighbouring (or identical) writes.
+ */
+ for (i = 0; i < context->nr_pending; i++)
+ {
+ PendingWriteback *cur;
+ PendingWriteback *next;
+ SMgrRelation reln;
+ int ahead;
+ BufferTag tag;
+ Size nblocks = 1;
+
+ cur = &context->pending_writebacks[i];
+ tag = cur->tag;
+
+ /*
+ * Peek ahead, into following writeback requests, to see if they can
+ * be combined with the current one.
+ */
+ for (ahead = 0; i + ahead + 1 < context->nr_pending; ahead++)
+ {
+ next = &context->pending_writebacks[i + ahead + 1];
+
+ /* different file, skip */
+ if (!RelFileNodeEquals(cur->tag.rnode, next->tag.rnode) ||
+ cur->tag.forkNum != cur->tag.forkNum)
+ break;
+
+ /* ok, block flushed twice, skip */
+ if (cur->tag.blockNum == next->tag.blockNum)
+ continue;
+
+ /* only merge consecutive writes */
+ if (cur->tag.blockNum + 1 != next->tag.blockNum)
+ break;
+
+ nblocks++;
+ cur = next;
+ }
+
+ i += ahead;
+
+ /* and finally tell the kernel to write the data to storage */
+ reln = smgropen(tag.rnode, InvalidBackendId);
+ smgrwriteback(reln, tag.forkNum, tag.blockNum, nblocks);
+ }
+
+ context->nr_pending = 0;
+}
/*
* We fsync the files later but first flush them to avoid spamming the
* cache and hopefully get the kernel to start writing them out before
- * the fsync comes. Ignore any error, since it's only a hint.
+ * the fsync comes.
*/
- (void) pg_flush_data(dstfd, offset, nbytes);
+ pg_flush_data(dstfd, offset, nbytes);
}
if (CloseTransientFile(dstfd))
#include <sys/file.h>
#include <sys/param.h>
#include <sys/stat.h>
+#ifndef WIN32
+#include <sys/mman.h>
+#endif
#include <unistd.h>
#include <fcntl.h>
#ifdef HAVE_SYS_RESOURCE_H
/* Define PG_FLUSH_DATA_WORKS if we have an implementation for pg_flush_data */
#if defined(HAVE_SYNC_FILE_RANGE)
#define PG_FLUSH_DATA_WORKS 1
+#elif !defined(WIN32) && defined(MS_ASYNC)
+#define PG_FLUSH_DATA_WORKS 1
#elif defined(USE_POSIX_FADVISE) && defined(POSIX_FADV_DONTNEED)
#define PG_FLUSH_DATA_WORKS 1
#endif
}
/*
- * pg_flush_data --- advise OS that the data described won't be needed soon
+ * pg_flush_data --- advise OS that the described dirty data should be flushed
*
- * Not all platforms have sync_file_range or posix_fadvise; treat as no-op
- * if not available. Also, treat as no-op if enableFsync is off; this is
- * because the call isn't free, and some platforms such as Linux will actually
- * block the requestor until the write is scheduled.
+ * An offset of 0 with an amount of 0 means that the entire file should be
+ * flushed.
*/
-int
-pg_flush_data(int fd, off_t offset, off_t amount)
+void
+pg_flush_data(int fd, off_t offset, off_t nbytes)
{
#ifdef PG_FLUSH_DATA_WORKS
- if (enableFsync)
- {
+
+ /*
+ * Right now file flushing is primarily used to avoid making later
+ * fsync()/fdatasync() calls have a less impact. Thus don't trigger
+ * flushes if fsyncs are disabled - that's a decision we might want to
+ * make configurable at some point.
+ */
+ if (!enableFsync)
+ return;
+
#if defined(HAVE_SYNC_FILE_RANGE)
- return sync_file_range(fd, offset, amount, SYNC_FILE_RANGE_WRITE);
+ {
+ int rc = 0;
+
+ /*
+ * sync_file_range(SYNC_FILE_RANGE_WRITE), currently linux specific,
+ * tells the OS that writeback for the passed in blocks should be
+ * started, but that we don't want to wait for completion. Note that
+ * this call might block if too much dirty data exists in the range.
+ * This is the preferrable method on OSs supporting it, as it works
+ * reliably when available (contrast to msync()) and doesn't flush out
+ * clean data (like FADV_DONTNEED).
+ */
+ rc = sync_file_range(fd, offset, nbytes,
+ SYNC_FILE_RANGE_WRITE);
+
+ /* don't error out, this is just a performance optimization */
+ if (rc != 0)
+ {
+ ereport(WARNING,
+ (errcode_for_file_access(),
+ errmsg("could not flush dirty data: %m")));
+ }
+ }
+#elif !defined(WIN32) && defined(MS_ASYNC)
+ {
+ int rc = 0;
+ void *p;
+
+ /*
+ * On many OSs msync() on a mmap'ed file triggers writeback. On linux
+ * it only does so when MS_SYNC is specified, but then it does the
+ * writeback synchronously. Luckily all common linux systems have
+ * sync_file_range(). This is preferrable over FADV_DONTNEED because
+ * it doesn't flush out clean data.
+ *
+ * We map the file (mmap()), tell the kernel to sync back the contents
+ * (msync()), and then remove the mapping again (munmap()).
+ */
+ p = mmap(NULL, context->nbytes,
+ PROT_READ | PROT_WRITE, MAP_SHARED,
+ context->fd, context->offset);
+ if (p == MAP_FAILED)
+ {
+ ereport(WARNING,
+ (errcode_for_file_access(),
+ errmsg("could not mmap while flushing dirty data in file \"%s\": %m",
+ context->filename ? context->filename : "")));
+ goto out;
+ }
+
+ rc = msync(p, context->nbytes, MS_ASYNC);
+ if (rc != 0)
+ {
+ ereport(WARNING,
+ (errcode_for_file_access(),
+ errmsg("could not flush dirty data in file \"%s\": %m",
+ context->filename ? context->filename : "")));
+ /* NB: need to fall through to munmap()! */
+ }
+
+ rc = munmap(p, context->nbytes);
+ if (rc != 0)
+ {
+ /* FATAL error because mapping would remain */
+ ereport(FATAL,
+ (errcode_for_file_access(),
+ errmsg("could not munmap while flushing blocks in file \"%s\": %m",
+ context->filename ? context->filename : "")));
+ }
+ }
#elif defined(USE_POSIX_FADVISE) && defined(POSIX_FADV_DONTNEED)
- return posix_fadvise(fd, offset, amount, POSIX_FADV_DONTNEED);
+ {
+ int rc = 0;
+
+ /*
+ * Signal the kernel that the passed in range should not be cached
+ * anymore. This has the, desired, side effect of writing out dirty
+ * data, and the, undesired, side effect of likely discarding useful
+ * clean cached blocks. For the latter reason this is the least
+ * preferrable method.
+ */
+
+ rc = posix_fadvise(context->fd, context->offset, context->nbytes,
+ POSIX_FADV_DONTNEED);
+
+ /* don't error out, this is just a performance optimization */
+ if (rc != 0)
+ {
+ ereport(WARNING,
+ (errcode_for_file_access(),
+ errmsg("could not flush dirty data in file \"%s\": %m",
+ context->filename ? context->filename : "")));
+ goto out;
+ }
+ }
#else
#error PG_FLUSH_DATA_WORKS should not have been defined
#endif
- }
-#endif
- return 0;
+
+#endif /* PG_FLUSH_DATA_WORKS */
}
#endif
}
+void
+FileWriteback(File file, off_t offset, int amount)
+{
+ int returnCode;
+
+ Assert(FileIsValid(file));
+
+ DO_DB(elog(LOG, "FileWriteback: %d (%s) " INT64_FORMAT " %d",
+ file, VfdCache[file].fileName,
+ (int64) offset, amount));
+
+ returnCode = FileAccess(file);
+ if (returnCode < 0)
+ return;
+
+ pg_flush_data(VfdCache[file].fd, offset, amount);
+}
+
int
FileRead(File file, char *buffer, int amount)
{
}
/*
- * We ignore errors from pg_flush_data() because this is only a hint.
+ * pg_flush_data() ignores errors, which is ok because this is only a
+ * hint.
*/
- (void) pg_flush_data(fd, 0, 0);
+ pg_flush_data(fd, 0, 0);
(void) CloseTransientFile(fd);
}
#endif /* USE_PREFETCH */
}
+/*
+ * mdwriteback() -- Tell the kernel to write pages back to storage.
+ *
+ * This accepts a rnage of blocks because flushing several pages at once is
+ * considerably more efficient than doing so individually.
+ */
+void
+mdwriteback(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum, int nblocks)
+{
+ off_t seekpos;
+ MdfdVec *v;
+
+ /*
+ * Issue flush requests in as few requests as possible; have to split at
+ * segment boundaries though, since those are actually separate files.
+ */
+ while (nblocks != 0)
+ {
+ int nflush = nblocks;
+ int segnum_start, segnum_end;
+
+ v = _mdfd_getseg(reln, forknum, blocknum, false, EXTENSION_RETURN_NULL);
+
+ /*
+ * We might be flushing buffers of already removed relations, that's
+ * ok, just ignore that case.
+ */
+ if (!v)
+ return;
+
+ /* compute offset inside the current segment */
+ segnum_start = blocknum / RELSEG_SIZE;
+
+ /* compute number of desired writes within the current segment */
+ segnum_end = (blocknum + nblocks - 1) / RELSEG_SIZE;
+ if (segnum_start != segnum_end)
+ nflush = RELSEG_SIZE - (blocknum % ((BlockNumber) RELSEG_SIZE) );
+
+ Assert(nflush >= 1);
+ Assert(nflush <= nblocks);
+
+ seekpos = (off_t) BLCKSZ * (blocknum % ((BlockNumber) RELSEG_SIZE));
+
+ FileWriteback(v->mdfd_vfd, seekpos, BLCKSZ * nflush);
+
+ nblocks -= nflush;
+ blocknum += nflush;
+ }
+}
/*
* mdread() -- Read the specified block from a relation.
BlockNumber blocknum, char *buffer);
void (*smgr_write) (SMgrRelation reln, ForkNumber forknum,
BlockNumber blocknum, char *buffer, bool skipFsync);
+ void (*smgr_writeback) (SMgrRelation reln, ForkNumber forknum,
+ BlockNumber blocknum, int nblocks);
BlockNumber (*smgr_nblocks) (SMgrRelation reln, ForkNumber forknum);
void (*smgr_truncate) (SMgrRelation reln, ForkNumber forknum,
BlockNumber nblocks);
static const f_smgr smgrsw[] = {
/* magnetic disk */
{mdinit, NULL, mdclose, mdcreate, mdexists, mdunlink, mdextend,
- mdprefetch, mdread, mdwrite, mdnblocks, mdtruncate, mdimmedsync,
- mdpreckpt, mdsync, mdpostckpt
+ mdprefetch, mdread, mdwrite, mdwriteback, mdnblocks, mdtruncate,
+ mdimmedsync, mdpreckpt, mdsync, mdpostckpt
}
};
buffer, skipFsync);
}
+
+/*
+ * smgrwriteback() -- Trigger kernel writeback for the supplied range of
+ * blocks.
+ */
+void
+smgrwriteback(SMgrRelation reln, ForkNumber forknum, BlockNumber blocknum,
+ int nblocks)
+{
+ (*(smgrsw[reln->smgr_which].smgr_writeback)) (reln, forknum, blocknum,
+ nblocks);
+}
+
/*
* smgrnblocks() -- Calculate the number of blocks in the
* supplied relation.
check_effective_io_concurrency, assign_effective_io_concurrency, NULL
},
+ {
+ {"checkpoint_flush_after", PGC_SIGHUP, RESOURCES_ASYNCHRONOUS,
+ gettext_noop("Number of pages after which previously performed writes are flushed to disk."),
+ NULL,
+ GUC_UNIT_BLOCKS
+ },
+ &checkpoint_flush_after,
+ /* see bufmgr.h: OS dependant default */
+ DEFAULT_CHECKPOINT_FLUSH_AFTER, 0, WRITEBACK_MAX_PENDING_FLUSHES,
+ NULL, NULL, NULL
+ },
+
+ {
+ {"backend_flush_after", PGC_USERSET, WAL_CHECKPOINTS,
+ gettext_noop("Number of pages after which previously performed writes are flushed to disk."),
+ NULL,
+ GUC_UNIT_BLOCKS
+ },
+ &backend_flush_after,
+ /* see bufmgr.h: OS dependant default */
+ DEFAULT_BACKEND_FLUSH_AFTER, 0, WRITEBACK_MAX_PENDING_FLUSHES,
+ NULL, NULL, NULL
+ },
+
+ {
+ {"bgwriter_flush_after", PGC_SIGHUP, WAL_CHECKPOINTS,
+ gettext_noop("Number of pages after which previously performed writes are flushed to disk."),
+ NULL,
+ GUC_UNIT_BLOCKS
+ },
+ &bgwriter_flush_after,
+ /* see bufmgr.h: 16 on Linux, 0 otherwise */
+ DEFAULT_BGWRITER_FLUSH_AFTER, 0, WRITEBACK_MAX_PENDING_FLUSHES,
+ NULL, NULL, NULL
+ },
+
{
{"max_worker_processes",
PGC_POSTMASTER,
#define BUFMGR_INTERNALS_H
#include "storage/buf.h"
+#include "storage/bufmgr.h"
#include "storage/latch.h"
#include "storage/lwlock.h"
#include "storage/shmem.h"
#define UnlockBufHdr(bufHdr) SpinLockRelease(&(bufHdr)->buf_hdr_lock)
+/*
+ * The PendingWriteback & WritebackContext structure are used to keep
+ * information about pending flush requests to be issued to the OS.
+ */
+typedef struct PendingWriteback
+{
+ /* could store different types of pending flushes here */
+ BufferTag tag;
+} PendingWriteback;
+
+/* typedef forward declared in bufmgr.h */
+typedef struct WritebackContext
+{
+ /* max number of writeback requests to coalesce */
+ int *max_pending;
+
+ /* current number of pending writeback requests */
+ int nr_pending;
+
+ /* pending requests */
+ PendingWriteback pending_writebacks[WRITEBACK_MAX_PENDING_FLUSHES];
+} WritebackContext;
+
/* in buf_init.c */
extern PGDLLIMPORT BufferDescPadded *BufferDescriptors;
+extern PGDLLIMPORT WritebackContext BackendWritebackContext;
/* in localbuf.c */
extern BufferDesc *LocalBufferDescriptors;
/*
- * Internal routines: only called by bufmgr
+ * Internal buffer management routines
*/
+/* bufmgr.c */
+extern void WritebackContextInit(WritebackContext *context, int *max_coalesce);
+extern void IssuePendingWritebacks(WritebackContext *context);
+extern void ScheduleBufferTagForWriteback(WritebackContext *context, BufferTag *tag);
/* freelist.c */
extern BufferDesc *StrategyGetBuffer(BufferAccessStrategy strategy);
* replay; otherwise same as RBM_NORMAL */
} ReadBufferMode;
+/* forward declared, to avoid having to expose buf_internals.h here */
+struct WritebackContext;
+
/* in globals.c ... this duplicates miscadmin.h */
extern PGDLLIMPORT int NBuffers;
/* in bufmgr.c */
+#define WRITEBACK_MAX_PENDING_FLUSHES 128
+
+/* FIXME: Also default to on for mmap && msync(MS_ASYNC)? */
+#ifdef HAVE_SYNC_FILE_RANGE
+#define DEFAULT_CHECKPOINT_FLUSH_AFTER 32
+#define DEFAULT_BACKEND_FLUSH_AFTER 16
+#define DEFAULT_BGWRITER_FLUSH_AFTER 64
+#else
+#define DEFAULT_CHECKPOINT_FLUSH_AFTER 0
+#define DEFAULT_BACKEND_FLUSH_AFTER 0
+#define DEFAULT_BGWRITER_FLUSH_AFTER 0
+#endif /* HAVE_SYNC_FILE_RANGE */
+
extern bool zero_damaged_pages;
extern int bgwriter_lru_maxpages;
extern double bgwriter_lru_multiplier;
extern bool track_io_timing;
extern int target_prefetch_pages;
+extern int checkpoint_flush_after;
+extern int backend_flush_after;
+extern int bgwriter_flush_after;
+
/* in buf_init.c */
extern PGDLLIMPORT char *BufferBlocks;
extern void AbortBufferIO(void);
extern void BufmgrCommit(void);
-extern bool BgBufferSync(void);
+extern bool BgBufferSync(struct WritebackContext *wb_context);
extern void AtProcExit_LocalBuffers(void);
extern off_t FileSeek(File file, off_t offset, int whence);
extern int FileTruncate(File file, off_t offset);
extern char *FilePathName(File file);
+extern void FileWriteback(File file, off_t offset, int amount);
/* Operations that allow use of regular stdio --- USE WITH CAUTION */
extern FILE *AllocateFile(const char *name, const char *mode);
extern int pg_fsync_no_writethrough(int fd);
extern int pg_fsync_writethrough(int fd);
extern int pg_fdatasync(int fd);
-extern int pg_flush_data(int fd, off_t offset, off_t amount);
+extern void pg_flush_data(int fd, off_t offset, off_t amount);
extern void fsync_fname(char *fname, bool isdir);
extern void SyncDataDirectory(void);
BlockNumber blocknum, char *buffer);
extern void smgrwrite(SMgrRelation reln, ForkNumber forknum,
BlockNumber blocknum, char *buffer, bool skipFsync);
+extern void smgrwriteback(SMgrRelation reln, ForkNumber forknum,
+ BlockNumber blocknum, int nblocks);
extern BlockNumber smgrnblocks(SMgrRelation reln, ForkNumber forknum);
extern void smgrtruncate(SMgrRelation reln, ForkNumber forknum,
BlockNumber nblocks);
char *buffer);
extern void mdwrite(SMgrRelation reln, ForkNumber forknum,
BlockNumber blocknum, char *buffer, bool skipFsync);
+extern void mdwriteback(SMgrRelation reln, ForkNumber forknum,
+ BlockNumber blocknum, int nblocks);
extern BlockNumber mdnblocks(SMgrRelation reln, ForkNumber forknum);
extern void mdtruncate(SMgrRelation reln, ForkNumber forknum,
BlockNumber nblocks);
PendingOperationEntry
PendingRelDelete
PendingUnlinkEntry
+PendingWriteback
PerlInterpreter
Perl_ppaddr_t
Permutation
WriteDataPtr
WriteExtraTocPtr
WriteFunc
+WritebackContext
X509
X509_NAME
X509_NAME_ENTRY
projects / users / andresfreund / postgres.git / commitdiff