public inbox for elfutils@sourceware.org
 help / color / mirror / Atom feed
* Elfutils 0.176 and riscv
@ 2019-02-16 17:09 Kurt Roeckx
  2019-02-16 19:22 ` Mark Wielaard
  0 siblings, 1 reply; 3+ messages in thread
From: Kurt Roeckx @ 2019-02-16 17:09 UTC (permalink / raw)
  To: elfutils-devel

On Debian's riscv64 port, this is a log of the build:
https://buildd.debian.org/status/fetch.php?pkg=elfutils&arch=riscv64&ver=0.176-1&stamp=1550335976&raw=0

It failed with:
FAIL: run-backtrace-native.sh
=============================

0x2000000000	0x2000019000	/lib/riscv64-linux-gnu/ld-2.28.so
0x200001a000	0x200001c000	[vdso: 10935]
0x2000022000	0x2000038000	/lib/riscv64-linux-gnu/libpthread-2.28.so
0x200003c000	0x2000146000	/lib/riscv64-linux-gnu/libc-2.28.so
0x2aaaaaa000	0x2aaaaae000	/<<PKGBUILDDIR>>/tests/backtrace-child
TID 10935:
# 0 0x2000031278    	raise
TID 10938:
# 0 0x2000031278    	raise
/<<PKGBUILDDIR>>/tests/backtrace: dwfl_thread_getframes: No DWARF information found
/<<PKGBUILDDIR>>/tests/backtrace: dwfl_thread_getframes: No DWARF information found
backtrace: backtrace.c:81: callback_verify: Assertion `seen_main' failed.
./test-subr.sh: line 84: 10933 Aborted                 LD_LIBRARY_PATH="${built_library_path}${LD_LIBRARY_PATH:+:}$LD_LIBRARY_PATH" $VALGRIND_CMD "$@"
backtrace-child: no main
FAIL run-backtrace-native.sh (exit status: 1)

FAIL: run-backtrace-dwarf.sh
============================

0x20000a43c6	raise
/<<PKGBUILDDIR>>/tests/backtrace-dwarf: dwfl_thread_getframes: No DWARF information found
dwarf: no main
FAIL run-backtrace-dwarf.sh (exit status: 1)

FAIL: run-backtrace-native-core.sh
==================================

0x200001a000	0x200001b000	linux-vdso.so.1
0x2000000000	0x2000019158	ld-linux-riscv64-lp64d.so.1
0x200003c000	0x2000149220	libc.so.6
0x2000022000	0x200003b478	libpthread.so.0
0x2aaaaaa000	0x2aaaaad0d8	backtrace-child
TID 10994:
# 0 0x2000031278    	raise
TID 10992:
# 0 0x2000029e20    	__GI___pthread_timedjoin_ex
/<<PKGBUILDDIR>>/tests/backtrace: dwfl_thread_getframes: No DWARF information found
/<<PKGBUILDDIR>>/tests/backtrace: dwfl_thread_getframes: No DWARF information found
backtrace: backtrace.c:81: callback_verify: Assertion `seen_main' failed.
./test-subr.sh: line 84: 10998 Aborted                 LD_LIBRARY_PATH="${built_library_path}${LD_LIBRARY_PATH:+:}$LD_LIBRARY_PATH" $VALGRIND_CMD "$@"
backtrace-child-core.10992: no main
rmdir: failed to remove 'test-10977': Directory not empty
FAIL run-backtrace-native-core.sh (exit status: 1)



Kurt

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Elfutils 0.176 and riscv
  2019-02-16 17:09 Elfutils 0.176 and riscv Kurt Roeckx
@ 2019-02-16 19:22 ` Mark Wielaard
  2019-02-17 17:42   ` Florian Weimer
  0 siblings, 1 reply; 3+ messages in thread
From: Mark Wielaard @ 2019-02-16 19:22 UTC (permalink / raw)
  To: Kurt Roeckx; +Cc: elfutils-devel

On Sat, Feb 16, 2019 at 06:09:06PM +0100, Kurt Roeckx wrote:
> On Debian's riscv64 port, this is a log of the build:
> https://buildd.debian.org/status/fetch.php?pkg=elfutils&arch=riscv64&ver=0.176-1&stamp=1550335976&raw=0

You are missing .eh_frame CFI. Without that backtraces
will be unreliable. Make sure that gcc for riscv64 has
-fasynchronous-unwind-tables on by default and these
failures should disappear. It is the default on Fedora
riscv64 and there we see zero-fail.

Apart from that you might want to make sure your glibc has
the following patch:
https://sourceware.org/git/?p=glibc.git;a=commit;h=85bd1ddbdfdfd13cfd06f7c367519b6ed3360843

Cheers,

Mark

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: Elfutils 0.176 and riscv
  2019-02-16 19:22 ` Mark Wielaard
@ 2019-02-17 17:42   ` Florian Weimer
  0 siblings, 0 replies; 3+ messages in thread
From: Florian Weimer @ 2019-02-17 17:42 UTC (permalink / raw)
  To: Mark Wielaard; +Cc: Kurt Roeckx, elfutils-devel

* Mark Wielaard:

> On Sat, Feb 16, 2019 at 06:09:06PM +0100, Kurt Roeckx wrote:
>> On Debian's riscv64 port, this is a log of the build:
>> https://buildd.debian.org/status/fetch.php?pkg=elfutils&arch=riscv64&ver=0.176-1&stamp=1550335976&raw=0
>
> You are missing .eh_frame CFI. Without that backtraces
> will be unreliable. Make sure that gcc for riscv64 has
> -fasynchronous-unwind-tables on by default and these
> failures should disappear. It is the default on Fedora
> riscv64 and there we see zero-fail.

And keep in mind that you have to patch GCC for
-fasynchronous-unwind-tables on other architectures, too.  Unwinding
information is not optional on GNU.

Thanks,
Florian

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2019-02-17 17:42 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-02-16 17:09 Elfutils 0.176 and riscv Kurt Roeckx
2019-02-16 19:22 ` Mark Wielaard
2019-02-17 17:42   ` Florian Weimer

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).