Successfully identified regression in *gcc* in CI configuration tcwg_bmk_gnu_tx1/gnu-release-aarch64-spec2k6-O2. So far, this commit has regressed CI configurations: - tcwg_bmk_gnu_tx1/gnu-release-aarch64-spec2k6-O2
Culprit: <cut> commit a4dfaad2e5594d871fe00a1116005e28f95d644e Author: Richard Sandiford richard.sandiford@arm.com Date: Mon Sep 30 16:20:44 2019 +0000
Remove global call sets: gcse.c
This is another case in which we can conservatively treat partial kills as full kills. Again this is in principle a bug fix for TARGET_HARD_REGNO_CALL_PART_CLOBBERED targets, but in practice it probably doesn't make a difference.
2019-09-30 Richard Sandiford richard.sandiford@arm.com
gcc/ * gcse.c: Include function-abi.h. (compute_hash_table_work): Use insn_callee_abi to get the ABI of the call insn target. Invalidate partially call-clobbered registers as well as fully call-clobbered ones.
From-SVN: r276323 </cut>
Results regressed to (for first_bad == a4dfaad2e5594d871fe00a1116005e28f95d644e) # reset_artifacts: -10 # build_abe binutils: -9 # build_abe stage1 -- --set gcc_override_configure=--disable-libsanitizer: -8 # build_abe linux: -7 # build_abe glibc: -6 # build_abe stage2 -- --set gcc_override_configure=--disable-libsanitizer: -5 # true: 0 # benchmark -O2 -- artifacts/build-a4dfaad2e5594d871fe00a1116005e28f95d644e/results_id: 1 # 447.dealII,[.] _ZNK12SparseMatrixIdE5vmultI6VectorIdES3_EEvRT regressed by 112
from (for last_good == c1b582720a39c3cb944aaff0298f721cdd034e3f) # reset_artifacts: -10 # build_abe binutils: -9 # build_abe stage1 -- --set gcc_override_configure=--disable-libsanitizer: -8 # build_abe linux: -7 # build_abe glibc: -6 # build_abe stage2 -- --set gcc_override_configure=--disable-libsanitizer: -5 # true: 0 # benchmark -O2 -- artifacts/build-c1b582720a39c3cb944aaff0298f721cdd034e3f/results_id: 1
Artifacts of last_good build: https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... Results ID of last_good: tx1_64/tcwg_bmk_gnu_tx1/bisect-gnu-release-aarch64-spec2k6-O2/1425 Artifacts of first_bad build: https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... Results ID of first_bad: tx1_64/tcwg_bmk_gnu_tx1/bisect-gnu-release-aarch64-spec2k6-O2/1421 Build top page/logs: https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa...
Configuration details:
Reproduce builds: <cut> mkdir investigate-gcc-a4dfaad2e5594d871fe00a1116005e28f95d644e cd investigate-gcc-a4dfaad2e5594d871fe00a1116005e28f95d644e
git clone https://git.linaro.org/toolchain/jenkins-scripts
mkdir -p artifacts/manifests curl -o artifacts/manifests/build-baseline.sh https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... --fail curl -o artifacts/manifests/build-parameters.sh https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... --fail curl -o artifacts/test.sh https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... --fail chmod +x artifacts/test.sh
# Reproduce the baseline build (build all pre-requisites) ./jenkins-scripts/tcwg_bmk-build.sh @@ artifacts/manifests/build-baseline.sh
cd gcc
# Reproduce first_bad build git checkout --detach a4dfaad2e5594d871fe00a1116005e28f95d644e ../artifacts/test.sh
# Reproduce last_good build git checkout --detach c1b582720a39c3cb944aaff0298f721cdd034e3f ../artifacts/test.sh
cd .. </cut>
History of pending regressions and results: https://git.linaro.org/toolchain/ci/base-artifacts.git/log/?h=linaro-local/c...
Artifacts: https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa... Build log: https://ci.linaro.org/job/tcwg_bmk_ci_gnu-bisect-tcwg_bmk_tx1-gnu-release-aa...
Full commit (up to 1000 lines): <cut> commit a4dfaad2e5594d871fe00a1116005e28f95d644e Author: Richard Sandiford richard.sandiford@arm.com Date: Mon Sep 30 16:20:44 2019 +0000
Remove global call sets: gcse.c
This is another case in which we can conservatively treat partial kills as full kills. Again this is in principle a bug fix for TARGET_HARD_REGNO_CALL_PART_CLOBBERED targets, but in practice it probably doesn't make a difference.
2019-09-30 Richard Sandiford richard.sandiford@arm.com
gcc/ * gcse.c: Include function-abi.h. (compute_hash_table_work): Use insn_callee_abi to get the ABI of the call insn target. Invalidate partially call-clobbered registers as well as fully call-clobbered ones.
From-SVN: r276323 --- gcc/ChangeLog | 7 +++++++ gcc/gcse.c | 10 ++++++++-- 2 files changed, 15 insertions(+), 2 deletions(-)
diff --git a/gcc/ChangeLog b/gcc/ChangeLog index 9d8b083d539..1b1c31cf009 100644 --- a/gcc/ChangeLog +++ b/gcc/ChangeLog @@ -1,3 +1,10 @@ +2019-09-30 Richard Sandiford richard.sandiford@arm.com + + * gcse.c: Include function-abi.h. + (compute_hash_table_work): Use insn_callee_abi to get the ABI of + the call insn target. Invalidate partially call-clobbered + registers as well as fully call-clobbered ones. + 2019-09-30 Richard Sandiford richard.sandiford@arm.com
* function.c (aggregate_value_p): Work out which ABI the diff --git a/gcc/gcse.c b/gcc/gcse.c index 9bde8619b7b..aeb59c645e1 100644 --- a/gcc/gcse.c +++ b/gcc/gcse.c @@ -160,6 +160,7 @@ along with GCC; see the file COPYING3. If not see #include "dbgcnt.h" #include "gcse.h" #include "gcse-common.h" +#include "function-abi.h"
/* We support GCSE via Partial Redundancy Elimination. PRE optimizations are a superset of those done by classic GCSE. @@ -1528,8 +1529,13 @@ compute_hash_table_work (struct gcse_hash_table_d *table) if (CALL_P (insn)) { hard_reg_set_iterator hrsi; - EXECUTE_IF_SET_IN_HARD_REG_SET (regs_invalidated_by_call, - 0, regno, hrsi) + + /* We don't track modes of hard registers, so we need + to be conservative and assume that partial kills + are full kills. */ + HARD_REG_SET callee_clobbers + = insn_callee_abi (insn).full_and_partial_reg_clobbers (); + EXECUTE_IF_SET_IN_HARD_REG_SET (callee_clobbers, 0, regno, hrsi) record_last_reg_set_info (insn, regno);
if (! RTL_CONST_OR_PURE_CALL_P (insn) </cut>