This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH 8/9] ENABLE_CHECKING refactoring: target-specific parts
- From: Jeff Law <law at redhat dot com>
- To: Mikhail Maltsev <maltsevm at gmail dot com>, gcc-patches mailing list <gcc-patches at gcc dot gnu dot org>, Richard Biener <richard dot guenther at gmail dot com>
- Date: Thu, 29 Oct 2015 22:13:48 -0600
- Subject: Re: [PATCH 8/9] ENABLE_CHECKING refactoring: target-specific parts
- Authentication-results: sourceware.org; auth=none
- References: <56130763 dot 4070206 at gmail dot com> <56130A77 dot 4030108 at gmail dot com> <5632EDD2 dot 7020406 at redhat dot com>
On 10/29/2015 10:10 PM, Jeff Law wrote:
On 10/05/2015 05:40 PM, Mikhail Maltsev wrote:
gcc/ChangeLog:
2015-10-05 Mikhail Maltsev <maltsevm@gmail.com>
* config/alpha/alpha.c (alpha_function_arg): Use gcc_checking_assert.
* config/arm/arm.c (arm_unwind_emit_sequence): Adjust to use
CHECKING_P.
* config/bfin/bfin.c (hwloop_optimize): Likewise.
* config/i386/i386.c (ix86_print_operand_address,
output_387_binary_op): Likewise.
* config/ia64/ia64.c (ia64_sched_init, bundling): Likewise.
* config/m68k/m68k.c (m68k_sched_md_init_global): Likewise.
* config/rs6000/rs6000.c (htm_expand_builtin, rs6000_emit_prologue):
Likewise.
* config/rs6000/rs6000.h: Likewise.
* config/visium/visium.c (visium_setup_incoming_varargs): Likewise.
One more note Mikhail. If you could do another pass on cleaning up
ENABLE_CHECKING macros it'd be appreciated.
They're still in:
find . -type f -print | xargs grep ENABLE_CHECKING | grep -v ChangeLog
./config.in:#undef ENABLE_CHECKING
./c-family/c-omp.c:#ifdef ENABLE_CHECKING
./tree-ssa-loop-manip.c:#ifdef ENABLE_CHECKING
./genconditions.c:#undef ENABLE_CHECKING\n\
./configure.ac: AC_DEFINE(ENABLE_CHECKING, 1,
./configure:$as_echo "#define ENABLE_CHECKING 1" >>confdefs.h
./cp/tree.c:#ifdef ENABLE_CHECKING
./cp/parser.c:#ifdef ENABLE_CHECKING
./cp/mangle.c:#if ENABLE_CHECKING
./cp/mangle.c:#endif /* ENABLE_CHECKING */
./cp/method.c:#ifdef ENABLE_CHECKING
./cp/typeck.c:#ifdef ENABLE_CHECKING
./cp/constexpr.c:#ifdef ENABLE_CHECKING
./cp/constexpr.c:#ifdef ENABLE_CHECKING
./cp/call.c:#ifdef ENABLE_CHECKING
./cp/call.c:#endif /* ENABLE_CHECKING */
./cp/decl2.c:#ifdef ENABLE_CHECKING
./cp/decl2.c:#endif /* ENABLE_CHECKING */
./cp/typeck2.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c: /* We shouldn't get here, but keep going if
!ENABLE_CHECKING. */
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/pt.c:#ifdef ENABLE_CHECKING
./cp/cp-tree.h:#ifdef ENABLE_CHECKING
./cp/cp-tree.h:#ifdef ENABLE_CHECKING
./cp/cp-tree.h:#endif /* ENABLE_CHECKING */
./sese.h:#ifdef ENABLE_CHECKING
./cfganal.c:#if ENABLE_CHECKING
Then we can go through the #if CHECKING_P stuff and see what can/should
be converted to a runtime test. Then sort out the right form of the
flag_checking test we want to use to get the __builtin_expect bits the
way we want.
Thanks again for your effort on this work. It is greatly appreciated.
jeff