Applying InnoDB Plugin 1.0.5 snapshot, part 6

From revision r5748 to r5783

Detailed revision comments:

r5748 | marko | 2009-09-03 06:05:44 -0500 (Thu, 03 Sep 2009) | 1 line
branches/zip: MLOG_MULTI_REC_END: Correct the comment.
r5751 | marko | 2009-09-03 09:36:15 -0500 (Thu, 03 Sep 2009) | 7 lines
branches/zip: row_merge(): Remove a bogus debug assertion
that was triggered when creating an index on an empty table.

row_merge_sort(): Add debug assertions and comments that justify
the loop termination condition.

The bogus assertion ut_ad(ihalf > 0) was reported by Michael.
r5752 | marko | 2009-09-03 09:55:51 -0500 (Thu, 03 Sep 2009) | 10 lines
branches/zip: recv_recover_page_func(): Write the log sequence number
to the compressed page, if there is one.  Previously, the function only
wrote the LSN to the uncompressed page.

It is not clear why recv_recover_page_func() is updating FIL_PAGE_LSN
in the buffer pool.  The log sequence number will be stamped on the
page when it is flushed to disk, in buf_flush_init_for_writing().
I noticed this inconsistency when analyzing Issue #313, but this patch
does not fix it.  That is no surprise, since FIL_PAGE_LSN should only
matter on disk files, not in the buffer pool.
r5775 | calvin | 2009-09-07 16:15:05 -0500 (Mon, 07 Sep 2009) | 13 lines
branches/zip: Build InnoDB on Windows with UNIV_HOTBACKUP

The changes are non-functional changes for normal InnoDB,
but needed for building the Hot Backup on Windows (with
UNIV_HOTBACKUP defined).

- Define os_aio_use_native_aio for HB.
- Do not acquire seek mutexes for backup since HB is single threaded.
- Do not use srv_flush_log_at_trx_commit for HB build

rb://155

Approved by: Marko
r5777 | marko | 2009-09-08 10:50:25 -0500 (Tue, 08 Sep 2009) | 2 lines
branches/zip: Remove BUF_LRU_INITIAL_RATIO, which should have been removed
together with buf_LRU_get_recent_limit().
r5779 | marko | 2009-09-09 01:17:19 -0500 (Wed, 09 Sep 2009) | 2 lines
branches/zip: buf_page_peek_if_too_old(): Make the bitmasking work when
buf_pool->freed_page_clock is wider than 32 bits.
r5780 | marko | 2009-09-09 01:50:50 -0500 (Wed, 09 Sep 2009) | 1 line
branches/zip: ut_time_ms(): Return ulint, not uint.
r5782 | marko | 2009-09-09 02:00:59 -0500 (Wed, 09 Sep 2009) | 2 lines
branches/zip: buf_page_peek_if_too_old(): Silence a compiler warning
that was introduced in r5779 on 32-bit systems.
r5783 | marko | 2009-09-09 02:25:00 -0500 (Wed, 09 Sep 2009) | 1 line
branches/zip: buf_page_is_accessed(): Correct the function comment.

storage/innodb_plugin/buf/buf0lru.c

@@ -66,11 +66,6 @@ allowed to point to either end of the LRU list. */
 # error "BUF_LRU_NON_OLD_MIN_LEN >= BUF_LRU_OLD_MIN_LEN"
 #endif
 
-/** The whole LRU list length is divided by this number to determine an
-initial segment in buf_LRU_get_recent_limit */
-
-#define BUF_LRU_INITIAL_RATIO	8
-
 /** When dropping the search hash index entries before deleting an ibd
 file, we build a local array of pages belonging to that tablespace
 in the buffer pool. Following is the size of that array. */

storage/innodb_plugin/include/buf0buf.h

@@ -807,8 +807,8 @@ buf_page_set_old(
 	buf_page_t*	bpage,	/*!< in/out: control block */
 	ibool		old);	/*!< in: old */
 /*********************************************************************//**
-Determine the time of last access a block in the buffer pool.
-@return	ut_time_ms() at the time of last access, 0 if not accessed */
+Determine the time of first access of a block in the buffer pool.
+@return	ut_time_ms() at the time of first access, 0 if not accessed */
 UNIV_INLINE
 unsigned
 buf_page_is_accessed(

storage/innodb_plugin/include/buf0buf.ic

@@ -84,7 +84,7 @@ buf_page_peek_if_too_old(
 		return(FALSE);
 	} else {
 		/* FIXME: bpage->freed_page_clock is 31 bits */
-		return((buf_pool->freed_page_clock & ~(1 << 31))
+		return((buf_pool->freed_page_clock & ((1UL << 31) - 1))
 		       > bpage->freed_page_clock
 		       + (buf_pool->curr_size
 			  * (BUF_LRU_OLD_RATIO_DIV - buf_LRU_old_ratio)
@@ -471,8 +471,8 @@ buf_page_set_old(
 }
 
 /*********************************************************************//**
-Determine the time of last access a block in the buffer pool.
-@return	ut_time_ms() at the time of last access, 0 if not accessed */
+Determine the time of first access of a block in the buffer pool.
+@return	ut_time_ms() at the time of first access, 0 if not accessed */
 UNIV_INLINE
 unsigned
 buf_page_is_accessed(

storage/innodb_plugin/include/mtr0mtr.h

@@ -118,7 +118,7 @@ For 1 - 8 bytes, the flag value must give the length also! @{ */
 #define MLOG_WRITE_STRING	((byte)30)	/*!< write a string to
 						a page */
 #define	MLOG_MULTI_REC_END	((byte)31)	/*!< if a single mtr writes
-						log records for several pages,
+						several log records,
 						this log record ends the
 						sequence of these records */
 #define MLOG_DUMMY_RECORD	((byte)32)	/*!< dummy log record used to

storage/innodb_plugin/include/ut0ut.h

@@ -246,7 +246,7 @@ value may wrap around.  It should only be used for heuristic
 purposes.
 @return	ms since epoch */
 UNIV_INTERN
-uint
+ulint
 ut_time_ms(void);
 /*============*/
 #endif /* !UNIV_HOTBACKUP */

storage/innodb_plugin/log/log0recv.c

@@ -1327,6 +1327,7 @@ recv_recover_page_func(
 	buf_block_t*	block)	/*!< in/out: buffer block */
 {
 	page_t*		page;
+	page_zip_des_t*	page_zip;
 	recv_addr_t*	recv_addr;
 	recv_t*		recv;
 	byte*		buf;
@@ -1376,6 +1377,7 @@ recv_recover_page_func(
 	mtr_set_log_mode(&mtr, MTR_LOG_NONE);
 
 	page = block->frame;
+	page_zip = buf_block_get_page_zip(block);
 
 #ifndef UNIV_HOTBACKUP
 	if (just_read_in) {
@@ -1436,13 +1438,19 @@ recv_recover_page_func(
 		if (recv->type == MLOG_INIT_FILE_PAGE) {
 			page_lsn = page_newest_lsn;
 
-			mach_write_ull(page + UNIV_PAGE_SIZE
-				       - FIL_PAGE_END_LSN_OLD_CHKSUM, 0);
-			mach_write_ull(page + FIL_PAGE_LSN, 0);
+			memset(FIL_PAGE_LSN + page, 0, 8);
+			memset(UNIV_PAGE_SIZE - FIL_PAGE_END_LSN_OLD_CHKSUM
+			       + page, 0, 8);
+
+			if (page_zip) {
+				memset(FIL_PAGE_LSN + page_zip->data, 0, 8);
+			}
 		}
 
 		if (recv->start_lsn >= page_lsn) {
 
+			ib_uint64_t	end_lsn;
+
 			if (!modification_to_page) {
 
 				modification_to_page = TRUE;
@@ -1464,11 +1472,17 @@ recv_recover_page_func(
 			recv_parse_or_apply_log_rec_body(recv->type, buf,
 							 buf + recv->len,
 							 block, &mtr);
-			mach_write_ull(page + UNIV_PAGE_SIZE
-				       - FIL_PAGE_END_LSN_OLD_CHKSUM,
-				       recv->start_lsn + recv->len);
-			mach_write_ull(page + FIL_PAGE_LSN,
-				       recv->start_lsn + recv->len);
+
+			end_lsn = recv->start_lsn + recv->len;
+			mach_write_ull(FIL_PAGE_LSN + page, end_lsn);
+			mach_write_ull(UNIV_PAGE_SIZE
+				       - FIL_PAGE_END_LSN_OLD_CHKSUM
+				       + page, end_lsn);
+
+			if (page_zip) {
+				mach_write_ull(FIL_PAGE_LSN
+					       + page_zip->data, end_lsn);
+			}
 		}
 
 		if (recv->len > RECV_DATA_BLOCK_SIZE) {

storage/innodb_plugin/os/os0file.c

@@ -88,7 +88,9 @@ UNIV_INTERN ibool	os_do_not_call_flush_at_each_write	= FALSE;
 /* We do not call os_file_flush in every os_file_write. */
 #endif /* UNIV_DO_FLUSH */
 
-#ifndef UNIV_HOTBACKUP
+#ifdef UNIV_HOTBACKUP
+# define os_aio_use_native_aio	FALSE
+#else /* UNIV_HOTBACKUP */
 /* We use these mutexes to protect lseek + file i/o operation, if the
 OS does not provide an atomic pread or pwrite, or similar */
 #define OS_FILE_N_SEEK_MUTEXES	16
@@ -198,7 +200,7 @@ static ulint	os_aio_n_segments	= ULINT_UNDEFINED;
 /** If the following is TRUE, read i/o handler threads try to
 wait until a batch of new read requests have been posted */
 static ibool	os_aio_recommend_sleep_for_read_threads	= FALSE;
-#endif /* !UNIV_HOTBACKUP */
+#endif /* UNIV_HOTBACKUP */
 
 UNIV_INTERN ulint	os_n_file_reads		= 0;
 UNIV_INTERN ulint	os_bytes_read_since_printout = 0;
@@ -1245,6 +1247,7 @@ os_file_create(
 		}
 #endif
 #ifdef UNIV_NON_BUFFERED_IO
+# ifndef UNIV_HOTBACKUP
 		if (type == OS_LOG_FILE && srv_flush_log_at_trx_commit == 2) {
 			/* Do not use unbuffered i/o to log files because
 			value 2 denotes that we do not flush the log at every
@@ -1253,10 +1256,14 @@ os_file_create(
 			   == SRV_WIN_IO_UNBUFFERED) {
 			attributes = attributes | FILE_FLAG_NO_BUFFERING;
 		}
-#endif
+# else /* !UNIV_HOTBACKUP */
+		attributes = attributes | FILE_FLAG_NO_BUFFERING;
+# endif /* !UNIV_HOTBACKUP */
+#endif /* UNIV_NON_BUFFERED_IO */
 	} else if (purpose == OS_FILE_NORMAL) {
 		attributes = 0;
 #ifdef UNIV_NON_BUFFERED_IO
+# ifndef UNIV_HOTBACKUP
 		if (type == OS_LOG_FILE && srv_flush_log_at_trx_commit == 2) {
 			/* Do not use unbuffered i/o to log files because
 			value 2 denotes that we do not flush the log at every
@@ -1265,7 +1272,10 @@ os_file_create(
 			   == SRV_WIN_IO_UNBUFFERED) {
 			attributes = attributes | FILE_FLAG_NO_BUFFERING;
 		}
-#endif
+# else /* !UNIV_HOTBACKUP */
+		attributes = attributes | FILE_FLAG_NO_BUFFERING;
+# endif /* !UNIV_HOTBACKUP */
+#endif /* UNIV_NON_BUFFERED_IO */
 	} else {
 		attributes = 0;
 		ut_error;
@@ -2227,7 +2237,9 @@ os_file_read(
 	DWORD		low;
 	DWORD		high;
 	ibool		retry;
+#ifndef UNIV_HOTBACKUP
 	ulint		i;
+#endif /* !UNIV_HOTBACKUP */
 
 	ut_a((offset & 0xFFFFFFFFUL) == offset);
 
@@ -2246,16 +2258,20 @@ os_file_read(
 	os_n_pending_reads++;
 	os_mutex_exit(os_file_count_mutex);
 
+#ifndef UNIV_HOTBACKUP
 	/* Protect the seek / read operation with a mutex */
 	i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES;
 
 	os_mutex_enter(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	ret2 = SetFilePointer(file, low, &high, FILE_BEGIN);
 
 	if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) {
 
+#ifndef UNIV_HOTBACKUP
 		os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 		os_mutex_enter(os_file_count_mutex);
 		os_n_pending_reads--;
@@ -2266,7 +2282,9 @@ os_file_read(
 
 	ret = ReadFile(file, buf, (DWORD) n, &len, NULL);
 
+#ifndef UNIV_HOTBACKUP
 	os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	os_mutex_enter(os_file_count_mutex);
 	os_n_pending_reads--;
@@ -2275,7 +2293,7 @@ os_file_read(
 	if (ret && len == n) {
 		return(TRUE);
 	}
-#else
+#else /* __WIN__ */
 	ibool	retry;
 	ssize_t	ret;
 
@@ -2294,7 +2312,7 @@ os_file_read(
 		"InnoDB: Was only able to read %ld.\n",
 		(ulong)n, (ulong)offset_high,
 		(ulong)offset, (long)ret);
-#endif
+#endif /* __WIN__ */
 #ifdef __WIN__
 error_handling:
 #endif
@@ -2343,7 +2361,9 @@ os_file_read_no_error_handling(
 	DWORD		low;
 	DWORD		high;
 	ibool		retry;
+#ifndef UNIV_HOTBACKUP
 	ulint		i;
+#endif /* !UNIV_HOTBACKUP */
 
 	ut_a((offset & 0xFFFFFFFFUL) == offset);
 
@@ -2362,16 +2382,20 @@ os_file_read_no_error_handling(
 	os_n_pending_reads++;
 	os_mutex_exit(os_file_count_mutex);
 
+#ifndef UNIV_HOTBACKUP
 	/* Protect the seek / read operation with a mutex */
 	i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES;
 
 	os_mutex_enter(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	ret2 = SetFilePointer(file, low, &high, FILE_BEGIN);
 
 	if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) {
 
+#ifndef UNIV_HOTBACKUP
 		os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 		os_mutex_enter(os_file_count_mutex);
 		os_n_pending_reads--;
@@ -2382,7 +2406,9 @@ os_file_read_no_error_handling(
 
 	ret = ReadFile(file, buf, (DWORD) n, &len, NULL);
 
+#ifndef UNIV_HOTBACKUP
 	os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	os_mutex_enter(os_file_count_mutex);
 	os_n_pending_reads--;
@@ -2391,7 +2417,7 @@ os_file_read_no_error_handling(
 	if (ret && len == n) {
 		return(TRUE);
 	}
-#else
+#else /* __WIN__ */
 	ibool	retry;
 	ssize_t	ret;
 
@@ -2404,7 +2430,7 @@ os_file_read_no_error_handling(
 
 		return(TRUE);
 	}
-#endif
+#endif /* __WIN__ */
 #ifdef __WIN__
 error_handling:
 #endif
@@ -2463,9 +2489,11 @@ os_file_write(
 	DWORD		ret2;
 	DWORD		low;
 	DWORD		high;
-	ulint		i;
 	ulint		n_retries	= 0;
 	ulint		err;
+#ifndef UNIV_HOTBACKUP
+	ulint		i;
+#endif /* !UNIV_HOTBACKUP */
 
 	ut_a((offset & 0xFFFFFFFF) == offset);
 
@@ -2482,16 +2510,20 @@ os_file_write(
 	os_n_pending_writes++;
 	os_mutex_exit(os_file_count_mutex);
 
+#ifndef UNIV_HOTBACKUP
 	/* Protect the seek / write operation with a mutex */
 	i = ((ulint) file) % OS_FILE_N_SEEK_MUTEXES;
 
 	os_mutex_enter(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	ret2 = SetFilePointer(file, low, &high, FILE_BEGIN);
 
 	if (ret2 == 0xFFFFFFFF && GetLastError() != NO_ERROR) {
 
+#ifndef UNIV_HOTBACKUP
 		os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 		os_mutex_enter(os_file_count_mutex);
 		os_n_pending_writes--;
@@ -2525,7 +2557,9 @@ os_file_write(
 	}
 # endif /* UNIV_DO_FLUSH */
 
+#ifndef UNIV_HOTBACKUP
 	os_mutex_exit(os_file_seek_mutexes[i]);
+#endif /* !UNIV_HOTBACKUP */
 
 	os_mutex_enter(os_file_count_mutex);
 	os_n_pending_writes--;

storage/innodb_plugin/row/row0merge.c

@@ -1576,7 +1576,6 @@ row_merge(
 	ulint		ohalf;	/*!< half the output file */
 
 	UNIV_MEM_ASSERT_W(block[0], 3 * sizeof block[0]);
-	ut_ad(ihalf > 0);
 	ut_ad(ihalf < file->offset);
 
 	of.fd = *tmpfd;
@@ -1665,6 +1664,10 @@ row_merge_sort(
 {
 	ulint	half = file->offset / 2;
 
+	/* The file should always contain at least one byte (the end
+	of file marker).  Thus, it must be at least one block. */
+	ut_ad(file->offset > 0);
+
 	do {
 		ulint	error;
 
@@ -1673,6 +1676,10 @@ row_merge_sort(
 		if (error != DB_SUCCESS) {
 			return(error);
 		}
+
+		/* half > 0 should hold except when the file consists
+		of one block.  No need to merge further then. */
+		ut_ad(half > 0 || file->offset == 1);
 	} while (half < file->offset && half > 0);
 
 	return(DB_SUCCESS);

storage/innodb_plugin/ut/ut0ut.c

@@ -206,7 +206,7 @@ value may wrap around.  It should only be used for heuristic
 purposes.
 @return	ms since epoch */
 UNIV_INTERN
-uint
+ulint
 ut_time_ms(void)
 /*============*/
 {
@@ -214,7 +214,7 @@ ut_time_ms(void)
 
 	ut_gettimeofday(&tv, NULL);
 
-	return((uint) tv.tv_sec * 1000 + tv.tv_usec / 1000);
+	return((ulint) tv.tv_sec * 1000 + tv.tv_usec / 1000);
 }
 #endif /* !UNIV_HOTBACKUP */