Bug#750996: eglibc FTBFS on Alpha: malloc/malloc.os build failure and testsuite failures.
On Tue, Jun 10, 2014 at 08:10:00PM +0200, Mark Wielaard wrote:
> On Mon, 2014-06-09 at 18:02 +0200, Aurelien Jarno wrote:
> > systemtap-sdt-dev was supposed to be something transparent for the
> > glibc, but in practice it causes build failure on at least on alpha (see
> > above). Looking at the BTS, I see it also causes problems with GCC, so I
> > am a bit concerned on other side effects we might haven't seen yet.
>
> Both issues were GCC bugs now fixed on mainline. See
> https://gcc.gnu.org/PR61231 and http://gcc.gnu.org/PR61336.
> The first one was not really related to sys/sdt.h at all. The second was
> indeed a bug in GCC on alpha triggered by the usage of the "i"
> constrained in the sys/sdt.h asm, now fixed.
I doesn't seems to be the case. PR61336 is about the sys/sdt.h code
triggering an ICE, and it has indeed been fixed. That said it now emits
an error instead of an ICE, so sys/sdt.h is still not usable on alpha,
as the last comment says:
| Richard Henderson 2014-06-02 16:47:20 UTC
| The ICE has been resolved.
|
| Note that the asm in question comes from system tap, which has not been
| ported to alpha. So you're probably better off disabling that in your
| (e)glibc build too.
I am therefore re-asking if you can provide a list of architectures
where sys/sdt.h is known to work and doesn't have any issues.
--
Aurelien Jarno GPG: 4096R/1DDD8C9B
aurelien@aurel32.net http://www.aurel32.net
Reply to: