Kernel is GENERIC-NODEBUG. No special sysctls or tunables set on my VM: [ngie@fbsd12 /usr/tests]$ sudo kyua -v test_suites.FreeBSD.allow_sysctl_side_effects=1 debug sys/kern/coredump_phnum_test:coredump_phnum kern.coredump: 1 -> 1 kern.corefile: %N.core -> /tmp/kyua.WvTTRC/2/work/coredump_phnum_helper.core Executing command [ /usr/tests/sys/kern/coredump_phnum_helper ] Executing command [ grep Number of program headers: ] Executing command [ grep -1 program headers ] Executing command [ grep -A1 ^ \[ 0\] ] Executing command [ wc -l ] kern.coredump: 1 -> 1 kern.corefile: /tmp/kyua.WvTTRC/2/work/coredump_phnum_helper.core -> %N.core Fail: regexp 65535 \(66169\) not in stdout Number of program headers: 65535 Fail: regexp There are 66169 program headers not in stdout Entry point 0x0 There are 65535 program headers, starting at offset 64 Fail: regexp 0000000000000001 .* 66169 not in stdout [ 0] (null) NULL 0000000000000000 00000000 0000000000000000 0000000000000000 0 0 0 Fail: regexp 66545 not in stdout 66546 Files left in work directory after failure: coredump_phnum_helper.core, coredump_phnum_restore_state.sh sys/kern/coredump_phnum_test:coredump_phnum -> failed: Test case body returned a non-ok exit code, but this is not allowed [ngie@fbsd12 /usr/tests]$ uname -a FreeBSD fbsd12 12.0-CURRENT FreeBSD 12.0-CURRENT #7 r309448: Fri Dec 2 15:39:37 PST 2016 ngie@fbsd12:/usr/obj/usr/src/svn/sys/GENERIC-NODEBUG amd64 [ngie@fbsd12 /usr/tests]$ sh -c 'ulimit -a' cpu time (seconds, -t) unlimited file size (512-blocks, -f) unlimited data seg size (kbytes, -d) 33554432 stack size (kbytes, -s) 524288 core file size (512-blocks, -c) unlimited max memory size (kbytes, -m) unlimited locked memory (kbytes, -l) 64 max user processes (-u) 7578 open files (-n) 87516 virtual mem size (kbytes, -v) unlimited swap limit (kbytes, -w) unlimited socket buffer size (bytes, -b) unlimited pseudo-terminals (-p) unlimited kqueues (-k) unlimited umtx shared locks (-o) unlimited
A commit references this bug: Author: ngie Date: Sat Dec 3 02:17:50 UTC 2016 New revision: 309463 URL: https://svnweb.freebsd.org/changeset/base/309463 Log: Expect :coredump_phnum to fail The number of program headers on my system (GENERIC-NODEBUG) don't match the expected (hardcoded) number in the test PR: 215019 Changes: head/tests/sys/kern/coredump_phnum_test.sh
> Number of program headers: 65535 This piece is a deficiency in base's readelf. Elfutils' version prints the real number of program headers after the first, in the extended case. (The actual number of sections dumped still doesn't match, though.)
https://reviews.freebsd.org/D8701 , https://reviews.freebsd.org/D8702 for the readelf issue.
(In reply to Conrad E. Meyer from comment #2) > Elfutils' version prints the real number of program headers after the first, in the extended case. Whoops, I meant binutils' version (from ports). Elfutils' version also prints the right number but the formatting is more different from the other two.
A related elftoolchain fix for 'readelf -l': https://reviews.freebsd.org/D8703
One last fix needed for elftoolchain readelf: https://reviews.freebsd.org/D8707 I'll work on making the test more flexible in terms of the exact number of sections, but won't mark the test as non-broken until base readelf can be used to run the test.
A commit references this bug: Author: cem Date: Sun Dec 4 03:50:57 UTC 2016 New revision: 309525 URL: https://svnweb.freebsd.org/changeset/base/309525 Log: coredump_phnum_test: Make expected pheader count more flexible Note: this test still requires binutils readelf (from ports) to pass, until a few issues are resolved in elftoolchain. PR: 215019 Reported by: ngie@ Sponsored by: Dell EMC Isilon Changes: head/tests/sys/kern/coredump_phnum_test.sh
Uh... now it's passing with GENERIC-NODEBUG using elftoolchain's readelf and binutil's readelf >_>..
Why did you reopen this if it's passing?
(In reply to Conrad Meyer from comment #9) It's expected to fail on 12.0-CURRENT, and it no longer fails on my machine :/. For some odd reason it fails on GENERIC with ci.freebsd.org though.
A commit references this bug: Author: ngie Date: Wed Jul 19 16:08:08 UTC 2017 New revision: 321210 URL: https://svnweb.freebsd.org/changeset/base/321210 Log: Remove expected failure for :coredump_phnum The testcase no longer fails on ^/head because readelf has established parity with binutils' copy of readelf. This issue is not seen on Jenkins because `test_suites.FreeBSD.allow_sysctl_side_effects` isn't set in kyua.conf on the CI host, i.e., the test is skipped. PR: 215019 Tested with: binutils (amd64-binutils-2.28,1); elftoolchain (r3561M) Changes: head/tests/sys/kern/coredump_phnum_test.sh
(In reply to Ngie Cooper from comment #10) It technically wasn't passing after emaste made the necessary changes to readelf in elftoolchain; please see the commit referenced in comment # 11.
MARKED AS SPAM