This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [pushed + testcase] Re: [PATCH] Fix GDB hang with remote after error from resume
- From: Edjunior Machado <edjunior at gmail dot com>
- To: Yao Qi <qiyaoltc at gmail dot com>
- Cc: Andreas Arnez <arnez at linux dot vnet dot ibm dot com>, Sergio Durigan Junior <sergiodj at redhat dot com>, Pedro Alves <palves at redhat dot com>, GDB Patches <gdb-patches at sourceware dot org>, David Edelsohn <dje dot gcc at gmail dot com>, Edjunior Machado <emachado at redhat dot com>
- Date: Mon, 15 Jan 2018 18:37:02 +0100
- Subject: Re: [pushed + testcase] Re: [PATCH] Fix GDB hang with remote after error from resume
- Authentication-results: sourceware.org; auth=none
- References: <m3fu7dtzy8.fsf@oc1027705133.ibm.com> <7eccd434-6f47-590e-e53f-32076e99c98b@redhat.com> <m31sir5gik.fsf@oc1027705133.ibm.com> <CAH=s-POSOGChGGzM-iwTceSS8vw8cZbibdWnT2smk-kM32K=Cw@mail.gmail.com>
Hi all,
On Mon, Jan 15, 2018 at 3:52 PM, Yao Qi <qiyaoltc@gmail.com> wrote:
> On Mon, Jan 15, 2018 at 2:43 PM, Andreas Arnez <arnez@linux.vnet.ibm.com>
> wrote:
> >
> > Sergio, now that this is in, we should be able to re-activate the
> > reporting for the s390 (and Power) Buildbots, right?
> >
>
> I am not sure ppc64 issue is resolved. See
> https://sourceware.org/ml/gdb-patches/2017-12/msg00414.html
> Note that I had a patch to fix fails in
> gdb.threads/process-dies-while-handling-bp.exp
> in this thread, but the tests with gdbserver still hang in some watchpoint
> tests.
>
>
Andreas' patch does fix the gdb.base/watchpoint.exp hang on ppc64*
native-gdbserver, but as Yao mentioned, it is still hanging on
gdb.threads/process-dies-while-handling-bp.exp without his fix (
https://sourceware.org/ml/gdb-patches/2017-12/msg00434.html).
That said, just tested upstream testsuite, with Andreas' patch included,
plus Yao's aforementioned patch and native-gdbserver testrun no longer
hangs. These are the results I got with native-gdbserver on Fedora26:
ppc64le:
=== gdb Summary ===
# of expected passes 50410
# of unexpected failures 550
# of unexpected successes 3
# of expected failures 58
# of unknown successes 4
# of known failures 64
# of unresolved testcases 5
# of untested testcases 49
# of unsupported tests 171
/home/fedora/binutils-gdb.git/build/gdb/gdb version 8.1.50.20180115-git
-nw -nx -data-directory
/home/fedora/binutils-gdb.git/build/gdb/testsuite/../data-directory -ex
"set auto-connect-native-target off"
ppc64:
=== gdb Summary ===
# of expected passes 50133
# of unexpected failures 1109
# of unexpected successes 3
# of expected failures 58
# of unknown successes 4
# of known failures 65
# of unresolved testcases 6
# of untested testcases 51
# of unsupported tests 162
/home/fedora/binutils-gdb.git/build/gdb/gdb version 8.1.50.20180115-git
-nw -nx -data-directory
/home/fedora/binutils-gdb.git/build/gdb/testsuite/../data-directory -ex
"set auto-connect-native-target off"
(FWIW, at first glance, gdb.linespec/cpcompletion.exp in ppc64 BE seems to
be one of the main culprits for the whole testrun taking longer (~1h) and
with considerably more failures than in ppc64le, aparently due to the
ppc64's function descriptor.)
Please let me know if you need any additional testing.
Thanks and regards,
Edjunior