V8 5.4 compilation failure on linuxOne/s390x

[mhdawson]

• Version: master
• Platform: s390
• Subsystem: v8

Saw this failure on recent PR run to upgrade to v8 5.4. PPC failures in that run have been seen and we just need to backport a PR from V8 master onto 5.4 and that should be happening now. (5.4 is not yet 'beta' yet and was cut recently so we've not done all backports quite yet)

However, the compilation failure on s390 is not something we've seen before in the builds in our CI or the google CI so opening this issue to track.

https://ci.nodejs.org/job/node-test-commit-v8-linux/285/

make[2]: Entering directory `/data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out'
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/atomicops_internals_x86_gcc.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/accounting-allocator.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/cpu.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/bits.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/division-by-constant.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/debug/stack_trace.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/file-utils.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/functional.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/ieee754.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/logging.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/once.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/platform/time.o
  CXX(target) /data/iojs/build/workspace/node-test-commit-v8-linux/nodes/rhel72-s390x/v8test/v8test/deps/v8/out/s390x.release/obj.target/v8_libbase/src/base/platform/condition-variable.o
../src/base/ieee754.cc: In function ‘int v8::base::ieee754::{anonymous}::_ZN2v84base7ieee75412_GLOBAL__N_117__kernel_rem_pio2EPdS3_iiiPKi.constprop.0(double*, double*, int, int, int, const int32_t*)’:
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
       for (k = 1; iq[jk - k] == 0; k++) {
                            ^
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
../src/base/ieee754.cc:612:28: error: array subscript is below array bounds [-Werror=array-bounds]
cc1plus: all warnings being treated as errors

[jbajwa]

Seems to be a hiccup. The file was modified sometime in june

c781e831 (bmeurer 2016-06-17 08:21:48 -0700 612)       for (k = 1; iq[jk - k] == 0; k++) {

the most recent jenkins job builds cleanly.
https://ci.nodejs.org/job/node-test-commit-v8-linux/nodes=rhel72-s390x,v8test=v8test/289/

[john-yan]

What level of gcc is used?

[jbajwa]

since its a rhel72 box, I'm guessing it should be gcc 4.8.x

[mhdawson]

Note above the referenced ci job was not run on the branch with 5.4 and we still see the compile failure.

This is the version info for the compiler:

[root@test-linuxonecc-rhel72-s390x-1 ~]# gcc --version
gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-4)
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

[john-yan]

waiting for backporting: https://codereview.chromium.org/2299763002/

[john-yan]

backport is completed at https://codereview.chromium.org/2302873002/

[targos]

Closing as the fix is now included in the v5.4 PR

[verbus]

Getting the following error:

CXX(host) /home/firstuser/arangodb/build/3rdParty/V8/v5.7.492.77/x64.release/obj.host/v8_libbase/src/base/bits.o
In file included from ../../../../3rdParty/V8/v5.7.492.77/./src/base/format-macros.h:27:0,
from ../../../../3rdParty/V8/v5.7.492.77/./src/base/macros.h:9,
from ../../../../3rdParty/V8/v5.7.492.77/./src/base/bits.h:11,
from ../../../../3rdParty/V8/v5.7.492.77/src/base/bits.cc:5:
../../../../3rdParty/V8/v5.7.492.77/./src/base/build_config.h:142:2: error: #error Target architecture x64 is only supported on x64 host
#error Target architecture x64 is only supported on x64 host
^~~~~
src/v8_libbase.host.x64.release.mk:182: recipe for target '/home/firstuser/arangodb/build/3rdParty/V8/v5.7.492.77/x64.release/obj.host/v8_libbase/src/base/bits.o' failed
make[3]: *** [/home/firstuser/arangodb/build/3rdParty/V8/v5.7.492.77/x64.release/obj.host/v8_libbase/src/base/bits.o] Error 1

firstuser@etpgxb6:~/arangodb$ cat VERSIONS
STARTER_REV "latest"
SYNCER_REV "latest"
GSEARCH_ID_HTTP "010085642145132923492:fixi4yzeiz8"
GSEARCH_ID_AQL "010085642145132923492:6ymjhhr677k"
GSEARCH_ID_Manual "010085642145132923492:djexw6vlsgo"
GSEARCH_ID_Cookbook "010085642145132923492:nknyifjn7tu"
GCHANGE_FREQ "daily"
GPRIORITY "0.3"
BROWSEABLE_VERSIONS "'devel', '3.3', '3.2', '3.1', '3.0', '2.8'"

sudo dmesg | grep Linux
[ 0.371721] Linux version 4.13.0-19-generic (buildd@z13-013) (gcc version 7.2.0 (Ubuntu 7.2.0-8ubuntu3)) #22-Ubuntu SMP Mon Dec 4 11:58:26 UTC 2017 (Ubuntu 4.13.0-19.22-generic 4.13.13)
[ 0.371727] setup.1a06a7: Linux is running as a z/VM guest operating system in 64-bit mode

cat /proc/version
Linux version 4.13.0-19-generic (buildd@z13-013) (gcc version 7.2.0 (Ubuntu 7.2.0-8ubuntu3)) #22-Ubuntu SMP Mon Dec 4 11:58:26 UTC 2017

[bnoordhuis]

@verbus Wrong bug tracker. Report that to arangodb.