Re: [CI-NOTIFY]: TCWG Bisect tcwg_cross/gnu-master-aarch64-build_cross - Build # 35 - Successful!

Hi Hongtao,

This was due to a bug in our bisection scripts. I've disabled notifications to patch authors and linaro-toolchain@ while I'm testing the fix.

Sorry for the noise.

-- Maxim Kuvyrkov https://www.linaro.org

On Jul 12, 2021, at 10:25 AM, Liu, Hongtao hongtao.liu@intel.com wrote:

...

-----Original Message----- From: ci_notify@linaro.org ci_notify@linaro.org Sent: Sunday, July 11, 2021 9:59 PM To: tcwg-validation@linaro.org; Liu, Hongtao hongtao.liu@intel.com; linaro- toolchain@lists.linaro.org Subject: [CI-NOTIFY]: TCWG Bisect tcwg_cross/gnu-master-aarch64- build_cross - Build # 35 - Successful!

Successfully identified regression in *glibc* in CI configuration tcwg_cross/gnu-master-aarch64-build_cross. So far, this commit has regressed CI configurations:

• tcwg_cross/gnu-master-aarch64-build_cross

Culprit:

<cut> commit 6ec6c77867af4ddfec7323e0ac6ede89effca852 Author: liuhongt <hongtao.liu@intel.com> Date: Wed Jul 7 10:07:59 2021 +0800

soft-fp: Add __extendhfsf2/__extendhfdf2, __truncsfhf2/__truncdfhf2, __eqhf2/__nehf2

1. Add __extendhfdf2/__extendhfsf2 to return an IEEE half converted to

IEEE double/single. 2. Add __truncdfhf2/__extendsfhf2 to truncate IEEE double/single into IEEE half. 3. Add __eqhf2/__nehf2 to return 0 if a == b and a,b are not NAN, otherwise return 1.

These are needed by x86 _Float16:

https://software.intel.com/content/www/us/en/develop/download/intel- avx512-fp16-architecture-specification.html

support in GCC.

</cut>

Results regressed to (for first_bad == 6ec6c77867af4ddfec7323e0ac6ede89effca852) # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # build_abe stage1: 2 # build_abe linux: 3 # First few build errors in logs: # 00:00:14 make[2]: *** [/home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/sysd-rules:483: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/start.o] Error 1 # 00:00:14 make[2]: *** [/home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux- gnu/aarch64-linux-gnu/glibc-glibc.git~master/sysd-rules:499: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/start.os] Error 1 # 00:00:14 make[2]: *** [/home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux- gnu/aarch64-linux-gnu/glibc-glibc.git~master/sysd-rules:483: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/crtn.o] Error 1 # 00:00:14 make[2]: *** [/home/tcwg-buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux- gnu/aarch64-linux-gnu/glibc-glibc.git~master/sysd-rules:483: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/crti.o] Error 1 # 00:00:14 make[2]: *** [../o-iterator.mk:9: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/check_fds.o] Error 1 # 00:00:14 make[2]: *** [../o-iterator.mk:9: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/check_fds.os] Error 1 # 00:00:14 ../sysdeps/aarch64/nptl/tls.h:91:19: error: ‘__builtin_thread_pointer’ is not supported on this target # 00:00:14 make[2]: *** [/home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/sysd-rules:485: /home/tcwg- buildslave/workspace/tcwg_gnu_0/abe/builds/x86_64-pc-linux-gnu/aarch64- linux-gnu/glibc-glibc.git~master/csu/libc-tls.o] Error 1 # 00:00:14 ../sysdeps/unix/sysv/linux/aarch64/sysdep.h:190:17: error: invalid register name for ‘_x0’ # 00:00:14 ../sysdeps/unix/sysv/linux/aarch64/sysdep.h:174:22: error: invalid register name for ‘_x8’

I'm not sure why adding hf soft-fp resulted in such error, anyone can help look into it?

...

from (for last_good == b46cfcef3f7dab8e90729c6616ac1420f2642245) # reset_artifacts: -10 # true: 0 # build_abe binutils: 1 # build_abe stage1: 2 # build_abe linux: 3 # build_abe glibc: 4 # build_abe stage2: 5 # build_abe qemu: 6

Artifacts of last_good build: https://ci.linaro.org/job/tcwg_cross-bisect-gnu- master-aarch64-build_cross/35/artifact/artifacts/build- b46cfcef3f7dab8e90729c6616ac1420f2642245/ Artifacts of first_bad build: https://ci.linaro.org/job/tcwg_cross-bisect-gnu- master-aarch64-build_cross/35/artifact/artifacts/build- 6ec6c77867af4ddfec7323e0ac6ede89effca852/ Build top page/logs: https://ci.linaro.org/job/tcwg_cross-bisect-gnu-master- aarch64-build_cross/35/

Configuration details:

Reproduce builds:

<cut> mkdir investigate-glibc-6ec6c77867af4ddfec7323e0ac6ede89effca852 cd investigate-glibc-6ec6c77867af4ddfec7323e0ac6ede89effca852

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_cross-bisect-gnu-master-aarch64- build_cross/35/artifact/artifacts/manifests/build-baseline.sh --fail curl -o artifacts/manifests/build-parameters.sh https://ci.linaro.org/job/tcwg_cross- bisect-gnu-master-aarch64-build_cross/35/artifact/artifacts/manifests/build- parameters.sh --fail curl -o artifacts/test.sh https://ci.linaro.org/job/tcwg_cross-bisect-gnu-master-aarch64- build_cross/35/artifact/artifacts/test.sh --fail chmod +x artifacts/test.sh

# Reproduce the baseline build (build all pre-requisites) ./jenkins- scripts/tcwg_gnu-build.sh @@ artifacts/manifests/build-baseline.sh

cd glibc

# Reproduce first_bad build git checkout --detach 6ec6c77867af4ddfec7323e0ac6ede89effca852 ../artifacts/test.sh

# Reproduce last_good build git checkout --detach b46cfcef3f7dab8e90729c6616ac1420f2642245 ../artifacts/test.sh

cd ..

</cut>

History of pending regressions and results: https://git.linaro.org/toolchain/ci/base-artifacts.git/log/?h=linaro- local/ci/tcwg_cross/gnu-master-aarch64-build_cross

Artifacts: https://ci.linaro.org/job/tcwg_cross-bisect-gnu-master-aarch64- build_cross/35/artifact/artifacts/ Build log: https://ci.linaro.org/job/tcwg_cross-bisect-gnu-master-aarch64- build_cross/35/consoleText

Full commit (up to 1000 lines):

<cut> commit 6ec6c77867af4ddfec7323e0ac6ede89effca852 Author: liuhongt <hongtao.liu@intel.com> Date: Wed Jul 7 10:07:59 2021 +0800

soft-fp: Add __extendhfsf2/__extendhfdf2, __truncsfhf2/__truncdfhf2, __eqhf2/__nehf2

1. Add __extendhfdf2/__extendhfsf2 to return an IEEE half converted to

IEEE double/single. 2. Add __truncdfhf2/__extendsfhf2 to truncate IEEE double/single into IEEE half. 3. Add __eqhf2/__nehf2 to return 0 if a == b and a,b are not NAN, otherwise return 1.

These are needed by x86 _Float16:

https://software.intel.com/content/www/us/en/develop/download/intel- avx512-fp16-architecture-specification.html

support in GCC.

soft-fp/eqhf2.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++ soft-fp/extendhfdf2.c | 53 +++++++++++++++++++++++++++++++++++++++++++++++++++ soft-fp/extendhfsf2.c | 49 +++++++++++++++++++++++++++++++++++++++++++++++ soft-fp/half.h | 2 ++ soft-fp/truncdfhf2.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++ soft-fp/truncsfhf2.c | 48 ++++++++++++++++++++++++++++++++++++++++++++++ 6 files changed, 253 insertions(+)

diff --git a/soft-fp/eqhf2.c b/soft-fp/eqhf2.c new file mode 100644 index 0000000000..6d6634e5c5 --- /dev/null +++ b/soft-fp/eqhf2.c @@ -0,0 +1,49 @@ +/* Software floating-point emulation.

• Return 0 iff a == b, 1 otherwise
• Copyright (C) 2021 Free Software Foundation, Inc.
• This file is part of the GNU C Library.
• The GNU C Library is free software; you can redistribute it and/or
• modify it under the terms of the GNU Lesser General Public
• License as published by the Free Software Foundation; either
• version 2.1 of the License, or (at your option) any later version.
• In addition to the permissions in the GNU Lesser General Public
• License, the Free Software Foundation gives you unlimited
• permission to link the compiled version of this file into
• combinations with other programs, and to distribute those
• combinations without any restriction coming from the use of this
• file. (The Lesser General Public License restrictions do apply in
• other respects; for example, they cover modification of the file,
• and distribution when not linked into a combine executable.)
• The GNU C Library is distributed in the hope that it will be useful,
• but WITHOUT ANY WARRANTY; without even the implied warranty of
• MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
• Lesser General Public License for more details.
• You should have received a copy of the GNU Lesser General Public
• License along with the GNU C Library; if not, see
• http://www.gnu.org/licenses/. */

+#include "soft-fp.h" +#include "half.h"

+CMPtype +__eqhf2 (HFtype a, HFtype b) +{

• FP_DECL_EX;
• FP_DECL_H (A);
• FP_DECL_H (B);
• CMPtype r;
• FP_INIT_EXCEPTIONS;
• FP_UNPACK_RAW_H (A, a);
• FP_UNPACK_RAW_H (B, b);
• FP_CMP_EQ_H (r, A, B, 1);
• FP_HANDLE_EXCEPTIONS;
• return r;

+}

+strong_alias (__eqhf2, __nehf2); diff --git a/soft-fp/extendhfdf2.c b/soft-fp/extendhfdf2.c new file mode 100644 index 0000000000..337ba791d4 --- /dev/null +++ b/soft-fp/extendhfdf2.c @@ -0,0 +1,53 @@ +/* Software floating-point emulation.

• Return an IEEE half converted to IEEE double
• Copyright (C) 2021 Free Software Foundation, Inc.
• This file is part of the GNU C Library.
• The GNU C Library is free software; you can redistribute it and/or
• modify it under the terms of the GNU Lesser General Public
• License as published by the Free Software Foundation; either
• version 2.1 of the License, or (at your option) any later version.
• In addition to the permissions in the GNU Lesser General Public
• License, the Free Software Foundation gives you unlimited
• permission to link the compiled version of this file into
• combinations with other programs, and to distribute those
• combinations without any restriction coming from the use of this
• file. (The Lesser General Public License restrictions do apply in
• other respects; for example, they cover modification of the file,
• and distribution when not linked into a combine executable.)
• The GNU C Library is distributed in the hope that it will be useful,
• but WITHOUT ANY WARRANTY; without even the implied warranty of
• MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
• Lesser General Public License for more details.
• You should have received a copy of the GNU Lesser General Public
• License along with the GNU C Library; if not, see
• http://www.gnu.org/licenses/. */

+#define FP_NO_EXACT_UNDERFLOW +#include "soft-fp.h" +#include "half.h" +#include "double.h"

+DFtype +__extendhfdf2 (HFtype a) +{

• FP_DECL_EX;
• FP_DECL_H (A);
• FP_DECL_D (R);
• DFtype r;
• FP_INIT_EXCEPTIONS;
• FP_UNPACK_RAW_H (A, a);

+#if _FP_W_TYPE_SIZE < _FP_FRACBITS_D

• FP_EXTEND (D, H, 2, 1, R, A);

+#else

• FP_EXTEND (D, H, 1, 1, R, A);

+#endif

• FP_PACK_RAW_D (r, R);
• FP_HANDLE_EXCEPTIONS;
• return r;

+} diff --git a/soft-fp/extendhfsf2.c b/soft-fp/extendhfsf2.c new file mode 100644 index 0000000000..a02f46d9a9 --- /dev/null +++ b/soft-fp/extendhfsf2.c @@ -0,0 +1,49 @@ +/* Software floating-point emulation.

• Return an IEEE half converted to IEEE single
• Copyright (C) 2021 Free Software Foundation, Inc.
• This file is part of the GNU C Library.
• The GNU C Library is free software; you can redistribute it and/or
• modify it under the terms of the GNU Lesser General Public
• License as published by the Free Software Foundation; either
• version 2.1 of the License, or (at your option) any later version.
• In addition to the permissions in the GNU Lesser General Public
• License, the Free Software Foundation gives you unlimited
• permission to link the compiled version of this file into
• combinations with other programs, and to distribute those
• combinations without any restriction coming from the use of this
• file. (The Lesser General Public License restrictions do apply in
• other respects; for example, they cover modification of the file,
• and distribution when not linked into a combine executable.)
• The GNU C Library is distributed in the hope that it will be useful,
• but WITHOUT ANY WARRANTY; without even the implied warranty of
• MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
• Lesser General Public License for more details.
• You should have received a copy of the GNU Lesser General Public
• License along with the GNU C Library; if not, see
• http://www.gnu.org/licenses/. */

+#define FP_NO_EXACT_UNDERFLOW +#include "soft-fp.h" +#include "half.h" +#include "single.h"

+SFtype +__extendhfsf2 (HFtype a) +{

• FP_DECL_EX;
• FP_DECL_H (A);
• FP_DECL_S (R);
• SFtype r;
• FP_INIT_EXCEPTIONS;
• FP_UNPACK_RAW_H (A, a);
• FP_EXTEND (S, H, 1, 1, R, A);
• FP_PACK_RAW_S (r, R);
• FP_HANDLE_EXCEPTIONS;
• return r;

+} diff --git a/soft-fp/half.h b/soft-fp/half.h index ca2c1fad03..70af2712fc 100644 --- a/soft-fp/half.h +++ b/soft-fp/half.h @@ -167,4 +167,6 @@ union _FP_UNION_H #define _FP_FRAC_HIGH_RAW_H(X) _FP_FRAC_HIGH_1 (X) #define _FP_FRAC_HIGH_DW_H(X) _FP_FRAC_HIGH_1 (X)

+#define FP_CMP_EQ_H(r, X, Y, ex) _FP_CMP_EQ (H, 1, (r), X, Y, (ex))

#endif /* !SOFT_FP_HALF_H */ diff --git a/soft-fp/truncdfhf2.c b/soft-fp/truncdfhf2.c new file mode 100644 index 0000000000..8bcb278769 --- /dev/null +++ b/soft-fp/truncdfhf2.c @@ -0,0 +1,52 @@ +/* Software floating-point emulation.

• Truncate IEEE double into IEEE half.
• Copyright (C) 2021 Free Software Foundation, Inc.
• This file is part of the GNU C Library.
• The GNU C Library is free software; you can redistribute it and/or
• modify it under the terms of the GNU Lesser General Public
• License as published by the Free Software Foundation; either
• version 2.1 of the License, or (at your option) any later version.
• In addition to the permissions in the GNU Lesser General Public
• License, the Free Software Foundation gives you unlimited
• permission to link the compiled version of this file into
• combinations with other programs, and to distribute those
• combinations without any restriction coming from the use of this
• file. (The Lesser General Public License restrictions do apply in
• other respects; for example, they cover modification of the file,
• and distribution when not linked into a combine executable.)
• The GNU C Library is distributed in the hope that it will be useful,
• but WITHOUT ANY WARRANTY; without even the implied warranty of
• MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
• Lesser General Public License for more details.
• You should have received a copy of the GNU Lesser General Public
• License along with the GNU C Library; if not, see
• http://www.gnu.org/licenses/. */

+#include "soft-fp.h" +#include "half.h" +#include "double.h"

+HFtype +__truncdfhf2 (DFtype a) +{

• FP_DECL_EX;
• FP_DECL_D (A);
• FP_DECL_H (R);
• HFtype r;
• FP_INIT_ROUNDMODE;
• FP_UNPACK_SEMIRAW_D (A, a);

+#if _FP_W_TYPE_SIZE < _FP_FRACBITS_D

• FP_TRUNC (H, D, 1, 2, R, A);

+#else

• FP_TRUNC (H, D, 1, 1, R, A);

+#endif

• FP_PACK_SEMIRAW_H (r, R);
• FP_HANDLE_EXCEPTIONS;
• return r;

+} diff --git a/soft-fp/truncsfhf2.c b/soft-fp/truncsfhf2.c new file mode 100644 index 0000000000..25bee29f7f --- /dev/null +++ b/soft-fp/truncsfhf2.c @@ -0,0 +1,48 @@ +/* Software floating-point emulation.

• Truncate IEEE single into IEEE half.
• Copyright (C) 2021 Free Software Foundation, Inc.
• This file is part of the GNU C Library.
• The GNU C Library is free software; you can redistribute it and/or
• modify it under the terms of the GNU Lesser General Public
• License as published by the Free Software Foundation; either
• version 2.1 of the License, or (at your option) any later version.
• In addition to the permissions in the GNU Lesser General Public
• License, the Free Software Foundation gives you unlimited
• permission to link the compiled version of this file into
• combinations with other programs, and to distribute those
• combinations without any restriction coming from the use of this
• file. (The Lesser General Public License restrictions do apply in
• other respects; for example, they cover modification of the file,
• and distribution when not linked into a combine executable.)
• The GNU C Library is distributed in the hope that it will be useful,
• but WITHOUT ANY WARRANTY; without even the implied warranty of
• MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
• Lesser General Public License for more details.
• You should have received a copy of the GNU Lesser General Public
• License along with the GNU C Library; if not, see
• http://www.gnu.org/licenses/. */

+#include "soft-fp.h" +#include "half.h" +#include "single.h"

+HFtype +__truncsfhf2 (SFtype a) +{

• FP_DECL_EX;
• FP_DECL_S (A);
• FP_DECL_H (R);
• HFtype r;
• FP_INIT_ROUNDMODE;
• FP_UNPACK_SEMIRAW_S (A, a);
• FP_TRUNC (H, S, 1, 1, R, A);
• FP_PACK_SEMIRAW_H (r, R);
• FP_HANDLE_EXCEPTIONS;
• return r;

+}

</cut>

---

linaro-toolchain mailing list linaro-toolchain@lists.linaro.org https://lists.linaro.org/mailman/listinfo/linaro-toolchain