Package: gnupg2 Version: 2.1.19-3 Control: -1 tags + help On Tue 2017-03-21 17:32:27 -0400, Daniel Kahn Gillmor wrote: > but since the upgrade to 2.1.19-3 in experimental, we're seeing the the > segfault on s390x. > > I'm not even sure how to interpret the error message: is it saying that > make itself Segfaulted, or that gpgscm segfaulted? > > I've connected to zelenka to try to reproduce the problem directly on > s390x but i haven't been able to get there yet.
OK, i've managed to reproduce some of the failures (intermittently) on a
sid chroot (named "gnupg2-xcheck-segfault") in zelenka.
With the source of 2.1.19-3 in /home/dkg/src/gnupg2/gnupg2, and having
built the package out-of-tree in /home/dkg/src/gnupg2/gnupg2/build (via
debian/rules), i repeatedly run just the first two simple test suites
From /home/dkg/src/gnupg2/gnupg2/:
LC_ALL=C EXEEXT=
PATH=../gpgscm:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games
srcdir=/home/dkg/src/gnupg2/gnupg2/build/../tests/openpgp
objdir=/home/dkg/src/gnupg2/gnupg2/build
GPGSCM_PATH=/home/dkg/src/gnupg2/gnupg2/build/../tests/gpgscm:/home/dkg/src/gnupg2/gnupg2/build/../tests/openpgp
/home/dkg/src/gnupg2/gnupg2/build/tests/gpgscm/gpgscm run-tests.scm
version.scm
The failures are intermittent -- sometimes no failures, sometimes both
run-tests.scm and version.scm fail, sometimes only one of them.
In one example, i got this result:
PASS: setup.scm
gpgscm: error running 'version.scm': terminated
FAIL: version.scm
1 tests run, 0 succeeded, 1 failed, 0 skipped.
Failed tests: version.scm
and then looking in dmesg, i see the following log:
[1099865.715125] User process fault: interruption code 0x4003B in
gpgscm[2aae2582000+37000]
[1099865.715135] failing address: 00000000196d1000 TEID: 00000000196d1800
[1099865.715138] Fault in primary space mode while using user ASCE.
[1099865.715141] AS:000000007feec1c7 R3:0000000000000024
[1099865.715145] CPU: 0 PID: 7395 Comm: gpgscm Not tainted 3.16.0-4-s390x #1
Debian 3.16.39-1+deb8u2
[1099865.715148] task: 00000000fb6ea108 ti: 000000003d600000 task.ti:
000000003d600000
[1099865.715151] User PSW : 0705200180000000 000002aae259450a
[1099865.715154] R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:1 AS:0 CC:2 PM:0
EA:3
User GPRS: 000002ab196c6888 000000000000003f 00000000196d1588
0000000000000007
[1099865.715160] 000002aae2596876 0000000000000001 0000000000000008
000002ab196c6888
[1099865.715163] 000002ab196c6868 0000000000000008 0000000000000007
000002ab196c6740
[1099865.715166] 000003fffd59c000 000002ab196c6888 000002aae25968ce
000003ffff893098
[1099865.715175] User Code: 000002aae25944fe: 0707 bcr 0,%r7
000002aae2594500: a718003f lhi %r1,63
#000002aae2594504: ebeff0700024 stmg
%r14,%r15,112(%r15)
>000002aae259450a: 54102000 n
%r1,0(%r2)
000002aae259450e: a7fbff60 aghi
%r15,-160
000002aae2594512: a71e0017 chi %r1,23
000002aae2594516: a7740016 brc
7,2aae2594542
000002aae259451a: e33020080021 clg
%r3,8(%r2)
[1099865.715286] Last Breaking-Event-Address:
[1099865.715301] [<000002aae25968c8>] 0x2aae25968c8
I have no idea what this means :(
Any suggestions? The failures are intermittent enough that i imagine
just a series of repeated givebacks *might* result in one successful
test, but that isn't very satisfying.
I'm reporting this to the BTS so that we have someplace specific to
track the problem. Please drop [email protected] from any
followup!
Hoping for some help,
--dkg
signature.asc
Description: PGP signature
