This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
RE: [PATCH] Add the -catch-load and -catch-unload MI commands.
- From: Marc Khouzam <marc dot khouzam at ericsson dot com>
- To: "'Mircea Gherzan'" <mircea dot gherzan at intel dot com>, "'Pedro Alves'" <palves at redhat dot com>
- Cc: "'gdb-patches at sourceware dot org'" <gdb-patches at sourceware dot org>, "'mgherzan at gmail dot com'" <mgherzan at gmail dot com>
- Date: Mon, 10 Sep 2012 11:03:38 -0400
- Subject: RE: [PATCH] Add the -catch-load and -catch-unload MI commands.
- References: <1346320317-1599-1-git-send-email-mircea.gherzan@intel.com> <F7CE05678329534C957159168FA70DEC5E116C3855@EUSAACMS0703.eamcs.ericsson.se> <503FA714.4020206@intel.com> <F7CE05678329534C957159168FA70DEC5E116C3AD4@EUSAACMS0703.eamcs.ericsson.se> <504DF49D.7070700@redhat.com> <504DFDB8.7090005@intel.com>
> -----Original Message-----
> From: Mircea Gherzan [mailto:mircea.gherzan@intel.com]
> Sent: Monday, September 10, 2012 10:48 AM
> To: Pedro Alves
> Cc: Marc Khouzam; 'gdb-patches@sourceware.org'; 'mgherzan@gmail.com'
> Subject: Re: [PATCH] Add the -catch-load and -catch-unload MI
> commands.
>
> Hi,
>
> Thank you for your feedback.
>
> On 10.09.2012 16:09, Pedro Alves wrote:
> > [...]
> > I don't this would be a good idea compared to separate
> commands. Each
> > of those catch variants takes different parameters/options. I'm not
> > seeing what putting them all under the same roof would buy.
>
> I am still in favour of having them as separate commands, but, if the
> powers that be decide otherwise, I am willing to rework the patch.
>
> > In the CLI, although "catch exec", "catch fork", "catch
> syscall", etc.
> > have the same prefix, they're really implemented as
> separate commands
> > as well.
> >
>
> Please let me know if there are any issues that prevent this
> patch from getting merged.
An enhancement that would be nice (but may not be blocking) is
to support a '-d' flag like -break-insert does. This would
create the catchpoint in a disabled state.
Currently, a frontend that wants to create a disabled catchpoint
must do:
catch load
-break-disabled 1 (assuming the catchpoint number is 1)
in non-stop mode, this can cause a race condition where the
catchpoint could hit before it is disabled. It would be better
to do a single command:
-catch-load -d
Note that I've never actually seen this happen, but I believe
the risk is there.
BTW, we have the same potential problem with watchpoints
which use -break-watch.
Thanks for these proposed new MI commands, they will be useful
for Eclipse.
Marc
>
> Thanks,
> Mircea
>
> --
>
> Intel GmbH
> Dornacher Strasse 1
> 85622 Feldkirchen/Muenchen, Deutschland
> Sitz der Gesellschaft: Feldkirchen bei Muenchen
> Geschaeftsfuehrer: Peter Gleissner, Christian Lamprechter,
> Hannes Schwaderer, Douglas Lusk
> Registergericht: Muenchen HRB 47456
> Ust.-IdNr./VAT Registration No.: DE129385895
> Citibank Frankfurt a.M. (BLZ 502 109 00) 600119052
>
>