On 11/14/2017 02:59 PM, Joseph Myers wrote: > On Tue, 14 Nov 2017, Martin Liška wrote: > >> Hello. >> >> Quite obvious fix for SH target. Joseph can you please continue with testing? >> I don't have a machine to test the patch. > > I don't have SH hardware, but can confirm that, in conjunction with > Martin's glibc patch to work around strncpy warnings (where the proper fix > and another patch for one case are still under discussion), this fixes the > sh4-linux-gnu glibc build with build-many-glibcs.py. (The glibc testsuite > build fails - errors in string/bug-strncat1.c, string/tester.c at least - > but the failures I see look like more fallout from Martin's GCC changes, > probably cases where the testsuite needs to disable the warnings, not > anything SH-specific.) >
Hello. Good, I'm planning to install slightly simplified version of the patch. Martin
>From 8dcdd9c110e39da619f0fa0e73d024186df66515 Mon Sep 17 00:00:00 2001 From: marxin <mli...@suse.cz> Date: Wed, 15 Nov 2017 08:09:18 +0100 Subject: [PATCH] Use proper probability (PR target/82927) gcc/ChangeLog: 2017-11-15 Martin Liska <mli...@suse.cz> PR target/82927 * config/sh/sh-mem.cc: Use proper probability for REG_BR_PROB_NOTE. --- gcc/config/sh/sh-mem.cc | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/gcc/config/sh/sh-mem.cc b/gcc/config/sh/sh-mem.cc index 8fce9799921..4c33260d84e 100644 --- a/gcc/config/sh/sh-mem.cc +++ b/gcc/config/sh/sh-mem.cc @@ -183,8 +183,12 @@ expand_block_move (rtx *operands) return false; } -static const int prob_unlikely = REG_BR_PROB_BASE / 10; -static const int prob_likely = REG_BR_PROB_BASE / 4; +static const int prob_unlikely + = profile_probability::from_reg_br_prob_base (REG_BR_PROB_BASE / 10) + .to_reg_br_prob_note (); +static const int prob_likely + = profile_probability::from_reg_br_prob_base (REG_BR_PROB_BASE / 4) + .to_reg_br_prob_note (); /* Emit code to perform a strcmp. -- 2.14.3
