Gentoo Websites Logo
Go to: Gentoo Home Documentation Forums Lists Bugs Planet Store Wiki Get Gentoo!
Bug 196555 - [4.2/bad code] dev-libs/openssl-0.9.8g RC5 tests broken w/ -ftree-vectorize
Summary: [4.2/bad code] dev-libs/openssl-0.9.8g RC5 tests broken w/ -ftree-vectorize
Status: RESOLVED WORKSFORME
Alias: None
Product: Gentoo Linux
Classification: Unclassified
Component: [OLD] Core system (show other bugs)
Hardware: AMD64 Linux
: High normal (vote)
Assignee: Gentoo Toolchain Maintainers
URL:
Whiteboard:
Keywords:
: 197150 (view as bug list)
Depends on:
Blocks:
 
Reported: 2007-10-20 22:06 UTC by Duncan
Modified: 2008-06-27 12:57 UTC (History)
3 users (show)

See Also:
Package list:
Runtime testing required: ---


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Duncan 2007-10-20 22:06:18 UTC
At least on ~amd64, with gcc-4.2.2 and binutils-2.18.50.0.2, openssl-0.9.8g doesn't like CFLAGS including -ftree-vectorize .  It causes the RC5 encryption tests to fail with random numbers (the "got" result changes every time the ebuild is run). Perhaps a filter-flags is in order?

With CFLAGS=${CFLAGS/-ftree-vectorize /} in /etc/portage/env/dev-libs/openssl ,  (and with MAKEOPTS=-j1 but I just filed that bug), the ebuild completes the tests and merges normally, so it's definitely that flag.

Duncan
Comment 1 Pierrot Rey 2007-10-21 06:49:13 UTC
Thanks, it works great now.

Pierrot
Comment 2 Santiago Gala 2007-10-21 10:27:42 UTC
/me too, I'll emerge it removing the flag. Something worth to report upstream. Is upstream taken care of, or should I report?
Comment 3 SpanKY gentoo-dev 2007-10-21 16:55:45 UTC
we do not want users reporting GCC bugs upstream, we will take care of that step

first we need to isolate and reduce the testcase
Comment 4 Santiago Gala 2007-10-26 10:39:42 UTC
> first we need to isolate and reduce the testcase

as or "isolate and reduce", the funny thing is that the error is actually a test case failing, it looks difficult to miss it. Correct me if I'm wrong, but any test case failing should be reported "as is", this is precisely the purpose of test cases. :)
Comment 5 Jakub Moc (RETIRED) gentoo-dev 2007-10-26 18:21:49 UTC
*** Bug 197150 has been marked as a duplicate of this bug. ***
Comment 6 Luis Diaz 2007-10-28 19:22:42 UTC
Full error details:

test multi-call ....................done
bulk test ok
../util/shlib_wrap.sh ./rc5test
ecb RC5 error encrypting (1)
got     :C4 BA 6A 9C 55 CA 10 F5
expected:21 A5 DB EE 15 4B 8F 6D
ecb RC5 error encrypting (2)
got     :F5 FC 76 70 78 AA 17 C2
expected:F7 C0 13 AC 5B 2B 89 52
ecb RC5 error encrypting (3)
got     :9C 56 00 58 45 16 E5 2B
expected:2F 42 B3 B7 03 69 FC 92
ecb RC5 error encrypting (4)
got     :7A 24 F4 F7 7E C0 06 F4
expected:65 C1 78 B2 84 D1 97 CC
ecb RC5 error encrypting (5)
got     :55 C9 F3 60 46 F1 93 07
expected:EB 44 E4 15 DA 31 98 24
cbc RC5 error encrypting (9)
got     :22 6E 5B 84 93 FD 9F 4B
expected:DC FE 09 85 77 EC A5 FF
cbc RC5 error encrypting (10)
got     :84 B7 40 4D F9 08 0A F5
expected:96 46 FB 77 63 8F 9C A8
cbc RC5 error encrypting (11)
got     :6A C8 4D E6 74 FE E2 FB
expected:B2 B3 20 9D B6 59 4D A4
cbc RC5 error encrypting (12)
got     :F9 6C 11 DB B9 59 A9 30
expected:54 5F 7F 32 A5 FC 38 36
cbc RC5 error encrypting (13)
got     :EE 9D 6E 4F BD 34 95 D1
expected:82 85 E7 C1 B5 BC 74 02
cbc RC5 error encrypting (14)
got     :41 56 E4 2C D7 DA 72 D2
expected:FC 58 6F 92 F7 08 09 34
cbc RC5 error encrypting (15)
got     :2E 2E 82 95 F3 40 8C 39
expected:CF 27 0E F9 71 7F F7 C4
cbc RC5 error encrypting (16)
got     :0E 8C B1 C4 23 C3 80 BC
expected:E4 93 F1 C1 BB 4D 6E 8C
cbc RC5 error encrypting (17)
got     :CE DF 86 18 F8 81 11 DD
expected:5C 4C 04 1E 0F 21 7A C3
cbc RC5 error encrypting (18)
got     :32 B8 7B B1 72 6E 4E FB
expected:92 1F 12 48 53 73 B4 F7
cbc RC5 error encrypting (19)
got     :D7 66 F7 87 A6 97 73 36
expected:5B A0 CA 6B BE 7F 5F AD
cbc RC5 error encrypting (20)
got     :8A 11 F9 C6 95 F5 9D 1C
expected:C5 33 77 1C D0 11 0E 63
cbc RC5 error encrypting (21)
got     :37 C1 07 46 74 4E 92 0F
expected:29 4D DB 46 B3 27 8D 60
cbc RC5 error encrypting (22)
got     :83 7D C1 78 18 C7 B2 5D
expected:DA D6 BD A9 DF E8 F7 E8
cbc RC5 error encrypting (23)
got     :13 AE 76 16 05 50 D6 0C
expected:97 E0 78 78 37 ED 31 7F
cbc RC5 error encrypting (24)
got     :91 74 10 E7 86 6A 44 E7
expected:78 75 DB F6 73 8C 64 78
cbc RC5 error encrypting (25)
got     :FE 1B 45 42 86 27 DF 37
expected:8F 34 C3 C6 81 C9 96 95
cbc RC5 error encrypting (26)
got     :42 F3 F3 9D 5E BD AA 5F
expected:7C B3 F1 DF 34 F9 48 11
cbc RC5 error encrypting (27)
got     :D6 92 D3 EA 75 61 C4 EE
expected:7F D1 A0 23 A5 BB A2 17
make[1]: *** [test_rc5] Error 30
make[1]: Leaving directory `/var/tmp/portage/dev-libs/openssl-0.9.8g/work/openssl-0.9.8g/test'
make: *** [tests] Error 2
 *
 * ERROR: dev-libs/openssl-0.9.8g failed.
 * Call stack:
 *               ebuild.sh, line 1695:  Called dyn_compile
 *               ebuild.sh, line 1033:  Called qa_call 'src_compile'
 *               ebuild.sh, line   44:  Called src_compile
 *   openssl-0.9.8g.ebuild, line  124:  Called src_test
 *   openssl-0.9.8g.ebuild, line  132:  Called die
 * The specific snippet of code:
 *      make test || die "make test failed"
 *  The die message:
 *   make test failed
 *
 * If you need support, post the topmost build error, and the call stack if relevant.
 * A complete build log is located at '/var/tmp/portage/dev-libs/openssl-0.9.8g/temp/build.log'.
 *
Comment 7 Mike Doty (RETIRED) gentoo-dev 2007-11-07 18:54:18 UTC
amd64 emul chroots fail in the same way
Comment 8 SpanKY gentoo-dev 2008-06-21 04:54:39 UTC
assuming gcc-4.3 has the issue fixed
Comment 9 Ryan Hill (RETIRED) gentoo-dev 2008-06-25 02:53:20 UTC
it does.
Comment 10 Duncan 2008-06-27 12:57:29 UTC
Confirmed by original filer as working now, too.

Thanks! =8^)