This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: ping: Re: [patch] Forbid "run" etc. for use_gdb_stub targets
- From: Jan Kratochvil <jan dot kratochvil at redhat dot com>
- To: Doug Evans <dje at google dot com>
- Cc: Pedro Alves <pedro at codesourcery dot com>, gdb-patches at sourceware dot org, Kevin Pouget <kevin dot pouget at gmail dot com>, Tom Tromey <tromey at redhat dot com>
- Date: Sun, 4 Dec 2011 02:29:35 +0100
- Subject: Re: ping: Re: [patch] Forbid "run" etc. for use_gdb_stub targets
- References: <BANLkTinM3j98XgLoXeAk8TLxKSgh2QzYHw@mail.gmail.com> <201110271909.37066.pedro@codesourcery.com> <20111028174014.GB18885@host1.jankratochvil.net> <201110281917.44839.pedro@codesourcery.com> <20111029194745.GA12850@host1.jankratochvil.net> <20111203183613.GA15496@host2.jankratochvil.net> <CADPb22Qt0qnaymAoWGWU48boOKPaYfzCWMEyn5BDkGYNcdfKkg@mail.gmail.com>
On Sun, 04 Dec 2011 02:02:06 +0100, Doug Evans wrote:
> I'd like to get away from conflating anything related to gdb stubs and
> valgrind / gdbserver.
Currently $use_gdb_stub means the startup has been already executed when the
testsuite gets into GDB control.
IIUC in reality currently the real stubs are not in use but gdbserver in
non-extended mode behaves as native GDB with the stubs in the past.
Maybe $use_gdb_stub name is not fortunate for what it does, real stubs are
only subset of $use_gdb_stub cases.
wrt valgrind, you may be referring to:
http://sourceware.org/ml/gdb-patches/2011-10/msg00768.html
# Therefore shouldn't be the gdb.server/ext-*.exp testfiles skipped if the
# target board defined use_gdb_stub? (ext-attach.exp already is skipped, for
# the [is_remote target] reason)
[unreplied]
Currently I did not change on the GDB testsuite behavior what was not clearly
wrong. Therefore:
* If a testcase can run only in some its specific mode/board of run:
* Some testcases skip themselves if their mode/board is not satisfied.
--- if [is_remote target] then { return 0 }
* Some testcases enforce their mode/board no matter what mode has been
specified by user for the testsuite.
--- gdbserver_start_extended
This creates duplicate runs of a testcase in the same mode/board if one
runs the testsuite in multiple modes.
* If a testcase is compatible with any mode/board it uses the mode/board
specified by user for the testsuite.
So far I just fixed those cases there were clearly wrong - such as spawning a
gdbserver and then running new inferior locally in GDB.
In the case of gdb.base/valgrind-db-attach.exp it enforces its mode and always
runs.
Thanks,
Jan