NetBSD-Bugs archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
port-sh3/49253: NetBSD/sh3 7.0_BETA binaries made by gcc48 strange somewhere.
>Number: 49253
>Category: port-sh3
>Synopsis: NetBSD/sh3 7.0_BETA binaries made by gcc48 strange somewhere.
>Confidential: no
>Severity: critical
>Priority: high
>Responsible: port-sh3-maintainer
>State: open
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sat Oct 04 06:50:00 +0000 2014
>Originator: Yasushi Oshima
>Release: NetBSD 7.0_BETA
>Organization:
>Environment:
NetBSD usl5p2 7.0_BETA NetBSD 7.0_BETA (GENERIC.201410031710Z) landisk
>Description:
In NetBSD/sh3-port (landisk, hpcsh, ...) 7.0_BETA, some binaries does not work correctly.
1) For building pkgsrc, the gcc occurs SIGSEGV.
see http://mail-index.netbsd.org/port-sh3/2014/08/16/msg000112.html
2) When atf-run in /usr/tests/fs/psshfs, it fails.
the mount_psshfs got signal 4(SIGILL). it is very strange.(*1)
For comparison, I build NetBSD/landisk binary with GCC45 by build.sh -V HAVE_GCC45 and run,
1) 'make pkgsrc/pkgtools/digest' success.
2) atf-run psshfs passed, SIGILL does not occur.
and,
2') in GCC4.8 (Normal release binary) environment,
I copy libpuffs.so.2 binary built by GCC4.5,
atf-run psshfs passed, successful.
In this case, the problem is in libpuffs.so.2 built by gcc4.8 only.
I think these problems are caused by GCC48 for netbsd/sh3, it may generate worse binary.
(Now GCC48 is the default for NetBSD/sh3 in netbsd-7 and current)
*1: results of all atf-run in /usr/tests is here:
http://www.yagoto-urayama.jp/~oshimaya/netbsd/files/atf/netbsd-7/landisk/20141002/
gcc48/ is by the normal release binaries in NetBSD7.0_BETA,
gcc45/ is by binarries built with -V HAVE_GCC=45 in netbsd-7 tree.
>How-To-Repeat:
1) install NetBSD 7.0_BETA to NetBSD/sh3 port
2) build pkgsrc/pkgtools/digest
or
2') run atf-run in /usr/tests/fs/psshfs
>Fix:
no idea.
Home |
Main Index |
Thread Index |
Old Index