mysql8.4/mysql-stack.patch

38 lines
1.6 KiB
Diff

Talk about silly byte-shaving ... the "stock" value of STACK_MIN_SIZE
does not work on s390x, and probably not on some other platforms either.
Knock it up to twice the pre-5.0.33 value to have some margin for future
changes in compiler, glibc, etc. Suspect this number if
"execution_constants" regression test fails.
As of 5.0.45 and F9, it seems 16384 isn't enough anymore either (ppc fails,
and seems to now need something above 24576 --- why?). STACK_MIN_SIZE has
to be enough less than DEFAULT_THREAD_STACK / 8 to provide some headroom,
which means we are now also forced to raise DEFAULT_THREAD_STACK for 32-bit
arches.
diff -Naur mysql-5.0.45.orig/include/my_pthread.h mysql-5.0.45/include/my_pthread.h
--- mysql-5.0.45.orig/include/my_pthread.h 2007-07-04 09:06:05.000000000 -0400
+++ mysql-5.0.45/include/my_pthread.h 2008-01-08 14:51:39.000000000 -0500
@@ -737,7 +737,7 @@
*/
#define DEFAULT_THREAD_STACK (512*1024L)
#else
-#define DEFAULT_THREAD_STACK (192*1024)
+#define DEFAULT_THREAD_STACK (512*1024)
#endif
#endif
diff -Naur mysql-5.0.45.orig/sql/mysql_priv.h mysql-5.0.45/sql/mysql_priv.h
--- mysql-5.0.45.orig/sql/mysql_priv.h 2007-07-04 09:06:41.000000000 -0400
+++ mysql-5.0.45/sql/mysql_priv.h 2008-01-08 14:51:55.000000000 -0500
@@ -173,7 +173,7 @@
Feel free to raise this by the smallest amount you can to get the
"execution_constants" test to pass.
*/
-#define STACK_MIN_SIZE 12000 // Abort if less stack during eval.
+#define STACK_MIN_SIZE 49152 // Abort if less stack during eval.
#define STACK_MIN_SIZE_FOR_OPEN 1024*80
#define STACK_BUFF_ALLOC 352 // For stack overrun checks