Hi!
>>>>> "knielsen" == knielsen <kniel...@knielsen-hq.org> writes:
knielsen> #At lp:maria
knielsen> 2703 kniel...@knielsen-hq.org 2009-06-10
knielsen> Fix XtraDB to build with atomic operations, for good
performance.
knielsen> The root of the problem is that ./configure mixed together two
different things. One is the
knielsen> availability of GCC atomic operation intrinsics. The other is
the selection of which
knielsen> primitives to use for my_atomic implementation.
knielsen> Then at some point a hack was made to not use GCC intrinsics in
my_atomic to work around
knielsen> some test failures. But because the two things are mixed in
./configure, this as a side
knielsen> effect also makes GCC intrinsics unavailable for XtraDB.
knielsen> Fixed by splitting this in two in configure, so that we have
HAVE_GCC_ATOMIC_BUILTINS for
knielsen> GCC intrinsics availability and MY_ATOMIC_MODE_GCC_BUILTINS for
use in my_atomic.
knielsen> modified:
knielsen> configure.in
knielsen> include/atomic/nolock.h
knielsen> === modified file 'configure.in'
knielsen> --- a/configure.in 2009-04-23 13:06:16 +0000
knielsen> +++ b/configure.in 2009-06-10 09:13:53 +0000
knielsen> @@ -1726,6 +1726,30 @@ then
knielsen> fi
knielsen> fi
knielsen> +AC_CACHE_CHECK([whether the compiler provides atomic builtins],
knielsen> + [mysql_cv_gcc_atomic_builtins], [AC_TRY_RUN([
knielsen> + int main()
knielsen> + {
knielsen> + int foo= -10; int bar= 10;
knielsen> + if (!__sync_fetch_and_add(&foo, bar) || foo)
knielsen> + return -1;
knielsen> + bar= __sync_lock_test_and_set(&foo, bar);
knielsen> + if (bar || foo != 10)
knielsen> + return -1;
knielsen> + bar= __sync_val_compare_and_swap(&bar, foo, 15);
knielsen> + if (bar)
knielsen> + return -1;
knielsen> + return 0;
knielsen> + }
knielsen> +], [mysql_cv_gcc_atomic_builtins=yes],
knielsen> + [mysql_cv_gcc_atomic_builtins=no],
knielsen> + [mysql_cv_gcc_atomic_builtins=no])])
knielsen> +
knielsen> +if test "x$mysql_cv_gcc_atomic_builtins" = xyes; then
knielsen> + AC_DEFINE(HAVE_GCC_ATOMIC_BUILTINS, 1,
knielsen> + [Define to 1 if compiler provides atomic builtins.])
knielsen> +fi
knielsen> +
knielsen> AC_ARG_WITH([atomic-ops],
knielsen> AC_HELP_STRING([--with-atomic-ops=rwlocks|smp|up],
knielsen> [Implement atomic operations using pthread rwlocks or
atomic CPU
knielsen> @@ -1739,28 +1763,9 @@ case "$with_atomic_ops" in
knielsen> [Use pthread rwlocks for atomic ops]) ;;
knielsen> "smp") ;;
knielsen> "")
knielsen> - AC_CACHE_CHECK([whether the compiler provides atomic builtins],
knielsen> - [mysql_cv_gcc_atomic_builtins], [AC_TRY_RUN([
knielsen> - int main()
knielsen> - {
knielsen> - int foo= -10; int bar= 10;
knielsen> - if (!__sync_fetch_and_add(&foo, bar) || foo)
knielsen> - return -1;
knielsen> - bar= __sync_lock_test_and_set(&foo, bar);
knielsen> - if (bar || foo != 10)
knielsen> - return -1;
knielsen> - bar= __sync_val_compare_and_swap(&bar, foo, 15);
knielsen> - if (bar)
knielsen> - return -1;
knielsen> - return 0;
knielsen> - }
knielsen> - ], [mysql_cv_gcc_atomic_builtins=yes_but_disabled],
knielsen> - [mysql_cv_gcc_atomic_builtins=no],
knielsen> - [mysql_cv_gcc_atomic_builtins=no])])
knielsen> -
knielsen> - if test "x$mysql_cv_gcc_atomic_builtins" = xyes; then
knielsen> - AC_DEFINE(HAVE_GCC_ATOMIC_BUILTINS, 1,
knielsen> - [Define to 1 if compiler provides atomic builtins.])
knielsen> + if test "x$mysql_cv_gcc_atomic_builtins" = xyes_but_disabled;
then
knielsen> + AC_DEFINE([MY_ATOMIC_MODE_GCC_BUILTINS], [1],
knielsen> + [Use GCC atomic builtins for atomic ops])
<cut>
Under which circumstances could the last 'if' be true? The new code
don't seem to be able to set mysql_cv_gcc_atomic_builtins to
yes_but_disabled which makes the last if always false and we never
define MY_ATOMIC_MODE_GCC_BUILTINS as 1.
Regards,
Monty
_______________________________________________
Mailing list: https://launchpad.net/~maria-developers
Post to : maria-developers@lists.launchpad.net
Unsubscribe : https://launchpad.net/~maria-developers
More help : https://help.launchpad.net/ListHelp
