#746426 gcc: Enable -fasynchronous-unwind-tables on more arches.

Package:
gcc-11
Source:
gcc-11
Description:
GNU C compiler
Submitter:
Kurt Roeckx
Date:
2021-04-28 07:57:03 UTC
Severity:
wishlist
Tags:
#746426#5
Date:
2014-04-29 21:50:27 UTC
From:
To:
Hi,

It seems that not all arches have -fasynchronous-unwind-tables
enabled.  I see it enabled on amd64, i386, s390x, but disabled
on armel/armhf, powerpc.

Could this enable this on more architectures?

I'm currently seeing elfutils test failures on powerpc because
the unwind information is not available.  I'll probably also get
this problem on arm64 if it's not enabled there.


Kurt
#746426#10
Date:
2014-04-29 22:04:14 UTC
From:
To:
So reading about this, this might break glibc when you enable it
and they might need to build some files without it.


Kurt
#746426#19
Date:
2014-05-18 14:37:47 UTC
From:
To:
Am 30.04.2014 00:04, schrieb Kurt Roeckx:

Is this enabled by any other distributions?
#746426#26
Date:
2014-05-18 17:37:01 UTC
From:
To:
On fedora and RHEL -fasynchronous-unwind-tables is enabled on all
architectures for gcc (otherwise rpmbuild will add it as standard flag).
So all packages are always build with it so you will always get reliable
backtraces.
#746426#31
Date:
2014-05-18 22:03:32 UTC
From:
To:
I assume this is enabled on redhat/fedora, but I have no idea.


Kurt
#746426#36
Date:
2014-05-18 22:45:27 UTC
From:
To:
It is enabled on all primary and secondary arches.

What is the glibc breakage you are seeing?
#746426#41
Date:
2014-05-19 18:03:25 UTC
From:
To:
I can't remember, I just found some reference to something that
was broken in libc when it was enabled.  I assume that since
redhat / fedora does this on all primary and secondary arches
that it has been fixed for all arches.


Kurt
#746426#50
Date:
2015-12-28 11:28:14 UTC
From:
To:
So looking at this again, it seems glibc needs to exclude the init
and fini files, and I can find various old patches to fix that.

Is there any reason why gcc doesn't make this the default?


Kurt
#746426#55
Date:
2019-01-23 10:00:21 UTC
From:
To:
Dear submitter,

as the package gcc-5 has just been removed from the Debian archive
unstable we hereby close the associated bug reports.  We are sorry
that we couldn't deal with your issue properly.

For details on the removal, please see https://bugs.debian.org/920166

The version of this package that was in Debian prior to this removal
can still be found using http://snapshot.debian.org/.

This message was generated automatically; if you believe that there is
a problem with it please contact the archive administrators by mailing
ftpmaster@ftp-master.debian.org.

Debian distribution maintenance software
pp.
Scott Kitterman (the ftpmaster behind the curtain)
#746426#60
Date:
2019-01-23 18:54:57 UTC
From:
To:
reopen 746426
reassign 746426 gcc-8
thanks

I've reassign it to gcc-8 now. I can also clone it to gcc-9 if
needed.
is not supported.

These work: amd64, x32

These have at least 1 test that says unwinding is not supported:
arm64, armel, armhf, i386, mips, mips64el, mipsel, ppc64el, s390x,
alpha, hppa, ia64, powerpc, ppc64el, riscv64, sparc64

But those pass at least run-backtrace-native: arm64, armel, armhf,
i386, ppc64el, s390x, powerpc, ppc64, sparc64

This means that for the arches we release for, mips* is the only
one that does not pass run-backtrace-native.

Unknown: powerpcspe, m68k, sh4


Kurt
#746426#71
Date:
2019-02-17 18:40:51 UTC
From:
To:
So at least the test suite failures on riscv64 are caused by gcc
not enabling this.

I've been told that "Unwinding information is not optional on
GNU".

Please just enable this.


Kurt
#746426#76
Date:
2021-02-13 13:07:12 UTC
From:
To:
Dear submitter,

as the package gcc-8 has just been removed from the Debian archive
unstable we hereby close the associated bug reports.  We are sorry
that we couldn't deal with your issue properly.

For details on the removal, please see https://bugs.debian.org/954831

The version of this package that was in Debian prior to this removal
can still be found using http://snapshot.debian.org/.

Please note that the changes have been done on the master archive and
will not propagate to any mirrors until the next dinstall run at the
earliest.

This message was generated automatically; if you believe that there is
a problem with it please contact the archive administrators by mailing
ftpmaster@ftp-master.debian.org.

Debian distribution maintenance software
pp.
Joerg Jaspert (the ftpmaster behind the curtain)
#746426#87
Date:
2021-04-28 07:52:30 UTC
From:
To:
gcc-11 enables this on

  amd64 arm64 i386 ppc64 ppc64el s390x x32

Are there other architectures where it's safe to enable this?