List:Commits« Previous MessageNext Message »
From:marko.makela Date:October 19 2010 6:45am
Subject:bzr push into mysql-5.5-innodb branch (marko.makela:3187 to 3188)
View as plain text  
 3188 Marko Mäkelä	2010-10-19
      ibuf_set_del_mark(): Do not complain about already delete-marked records.
      
      According to a comment in row_upd_sec_index_entry(), it is a
      legitimate situation that can be caused by a lock wait.

    modified:
      storage/innobase/ibuf/ibuf0ibuf.c
 3187 Marko Mäkelä	2010-10-19 [merge]
      Merge Bug #56680 from mysql-5.1.
      
      Additional fixes in 5.5:
      
      ibuf_set_del_mark(): Add diagnostics when setting a buffered delete-mark fails.
      
      ibuf_delete(): Correct a misleading comment about non-found records.
      
      rec_print(): Add a const qualifier to the index parameter.
      
      Bug #56680 wrong InnoDB results from a case-insensitive covering index
      
      row_search_for_mysql(): When a secondary index record might not be
      visible in the current transaction's read view and we consult the
      clustered index and optionally some undo log records, return the
      relevant columns of the clustered index record to MySQL instead of the
      secondary index record.
      
      ibuf_insert_to_index_page_low(): New function, refactored from
      ibuf_insert_to_index_page().
      
      ibuf_insert_to_index_page(): When we are inserting a record in place
      of a delete-marked record and some fields of the record differ, update
      that record just like row_ins_sec_index_entry_by_modify() would do.
      
      btr_cur_update_alloc_zip(): Make the function public.
      
      mysql_row_templ_t: Add clust_rec_field_no.
      
      row_sel_store_mysql_rec(), row_sel_push_cache_row_for_mysql(): Add the
      flag rec_clust, for returning data at clust_rec_field_no instead of
      rec_field_no. Resurrect the debug assertion that the record not be
      marked for deletion. (Bug #55626)
      
      [UNIV_DEBUG || UNIV_IBUF_DEBUG] ibuf_debug, buf_page_get_gen(),
      buf_flush_page_try():
      Implement innodb_change_buffering_debug=1 for evicting pages from the
      buffer pool, so that change buffering will be attempted more
      frequently.

    added:
      mysql-test/suite/innodb/r/innodb_bug56680.result
      mysql-test/suite/innodb/t/innodb_bug56680.test
    modified:
      storage/innobase/btr/btr0cur.c
      storage/innobase/buf/buf0buf.c
      storage/innobase/buf/buf0flu.c
      storage/innobase/handler/ha_innodb.cc
      storage/innobase/ibuf/ibuf0ibuf.c
      storage/innobase/include/btr0cur.h
      storage/innobase/include/buf0flu.h
      storage/innobase/include/ibuf0ibuf.h
      storage/innobase/include/rem0rec.h
      storage/innobase/include/row0mysql.h
      storage/innobase/include/row0upd.h
      storage/innobase/rem/rem0rec.c
      storage/innobase/row/row0mysql.c
      storage/innobase/row/row0sel.c
      storage/innobase/row/row0upd.c
=== modified file 'storage/innobase/ibuf/ibuf0ibuf.c'
--- a/storage/innobase/ibuf/ibuf0ibuf.c	revid:marko.makela@strippedo0fdd
+++ b/storage/innobase/ibuf/ibuf0ibuf.c	revid:marko.makela@stripped
@@ -3997,21 +3997,22 @@ ibuf_set_del_mark(
 		rec = page_cur_get_rec(&page_cur);
 		page_zip = page_cur_get_page_zip(&page_cur);
 
-		if (UNIV_UNLIKELY
-		    (rec_get_deleted_flag(
+		/* Delete mark the old index record. According to a
+		comment in row_upd_sec_index_entry(), it can already
+		have been delete marked if a lock wait occurred in
+		row_ins_index_entry() in a previous invocation of
+		row_upd_sec_index_entry(). */
+
+		if (UNIV_LIKELY
+		    (!rec_get_deleted_flag(
 			    rec, dict_table_is_comp(index->table)))) {
-			ut_print_timestamp(stderr);
-			fputs("  InnoDB: record is already delete-marked\n",
-			      stderr);
-			goto failure;
+			btr_cur_set_deleted_flag_for_ibuf(rec, page_zip,
+							  TRUE, mtr);
 		}
-
-		btr_cur_set_deleted_flag_for_ibuf(rec, page_zip, TRUE, mtr);
 	} else {
 		ut_print_timestamp(stderr);
 		fputs("  InnoDB: unable to find a record to delete-mark\n",
 		      stderr);
-failure:
 		fputs("InnoDB: tuple ", stderr);
 		dtuple_print(stderr, entry);
 		fputs("\n"

Attachment: [text/bzr-bundle] bzr/marko.makela@oracle.com-20101019064438-gfls1n6fign1brrh.bundle
Thread
bzr push into mysql-5.5-innodb branch (marko.makela:3187 to 3188) marko.makela19 Oct