Hi Aaron, On Tue, Jan 23, 2018 at 1:26 AM, Aaron M. Ucko <ucko@debian.org> wrote: > Source: pocl > Version: 0.14-5 > Severity: important > Tags: upstream > Justification: fails to build from source > User: debian-arm@lists.debian.org > Usertags: arm64 > > Builds of pocl for arm64 have been failing, per the below excerpts from > https://buildd.debian.org/status/fetch.php?pkg=pocl&arch=arm64&ver=1.0-2&stamp=1516324314&raw=0 > > I acknowledge that the output I've quoted doesn't make the cause of > failure clear in these cases, but that's all the information I have. > Perhaps you can reproduce the errors on a porter box. > > Could you please take a look? There are two issues causing the test failure. I have a patch which fixes one of the issues. More below... [...] > 7/121 Test #22: kernel/test_shuffle_char ............................................................***Failed Required regular expression found.Regex=[OK The "Regex=[OK]" indicates that the test ran successfully and passed. But... > > ] 29.66 sec > OK ... upstream expects a subset of test_shuffle_* tests to fail. They are marked as WILL_FAIL in the below excerpt from tests/kernel/CMakeLists.txt # interestingly, long/ulong/double do not fail if(ARM64) set_tests_properties("kernel/test_shuffle_char" "kernel/test_shuffle_short" "kernel/test_shuffle_ushort" "kernel/test_shuffle_uint" "kernel/test_shuffle_int" "kernel/test_shuffle_float" ${HALF_TEST} PROPERTIES WILL_FAIL 1) endif() The attached patch (on top of the package repository) fixes it for me. While investigating this failure, I noticed an unrelated issue with the debian build for arm64 - instead of choosing a generic armv8 as the target cpu, the rules file chooses to build with cortex-a53 (value picked from debian/supported-arch). Just thought I'd mention it. [...] > 29/121 Test #16: kernel/test_fabs ....................................................................***Failed Error regular expression found in output. Regex=[FAIL] 4.82 sec > CMake Error at /<<PKGBUILDDIR>>/cmake/run_test.cmake:22 (message): > FAIL: Test exited with nonzero code: > /<<PKGBUILDDIR>>/obj-aarch64-linux-gnu/tests/kernel/kernel;test_fabs > > > STDOUT: > > > > STDERR: > > > > -- OK This one seems to be an actual failure but the test code is a little hard to parse. I'll try and see if I can figure out why it is failing. In the meanwhile, please apply the attached patch as a partial fix. Thanks, Punit
Attachment:
0001-Enable-all-shuffle-tests-on-arm64.patch
Description: Binary data