This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [patch 1/2] Convert hardware watchpoints to use breakpoint_ops
- From: Joel Brobecker <brobecker at adacore dot com>
- To: Thiago Jung Bauermann <bauerman at br dot ibm dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Mon, 8 Nov 2010 10:43:35 -0800
- Subject: Re: [patch 1/2] Convert hardware watchpoints to use breakpoint_ops
- References: <1282074071.2606.702.camel@hactar> <201010161843.43062.pedro@codesourcery.com> <1287534691.2686.17.camel@hactar>
Just for the record, I've started looking at both patches, and already
spent a good hour on them. But I think that I should give myself some
time to to let every piece sink in, before I make an official review.
I urge another one of us to look at this change, because it's not
completely obvious. In particular, I think it was my suggestion
that prompted the move to using breakpoint_ops for watchpoints.
But it's not obvious that there is a gain. I think the gain shows up
in the second patch, where we avoid having:
1. To define new kinds for range and mask watchpoints
2. To program dispatching manually, Eg:
if (bp->print_one)
bp->print_one (...)
else if (bp->type == hw_watchpoint)
print_one_normal_watchpoint (...)
else if (bp->type == range_watchpoint)
print_one_range_watchpoint (...)
else if (bp->type == mask_watchpoint)
[...]
--
Joel