From: Pedro Alves <pedro@codesourcery.com>
To: gdb-patches@sourceware.org
Cc: Thiago Jung Bauermann <bauerman@br.ibm.com>,
Philippe Waroquiers <philippe.waroquiers@skynet.be>,
yao@codesourcery.com
Subject: software watchpoints bug (was: Re: x86 watchpoints bug)
Date: Tue, 26 Jul 2011 20:02:00 -0000 [thread overview]
Message-ID: <201107262046.19451.pedro@codesourcery.com> (raw)
In-Reply-To: <1311388735.3205.29.camel@hactar>
On Saturday 23 July 2011 03:38:55, Thiago Jung Bauermann wrote:
> On Fri, 2011-07-22 at 17:40 +0100, Pedro Alves wrote:
> > Maybe better would be to change works_in_software_mode_watchpoint to:
> >
> > int
> > works_in_software_mode_watchpoint (const struct breakpoint *b)
> > {
> > - return b->type == bp_hardware_watchpoint;
> > + return (b->type == bp_watchpoint || b->type == bp_hardware_watchpoint);
> > }
>
> Agreed. I would only comment that the parenthesis are not necessary. :-)
Alright! :-) I've applied the patch below, which does that, and
adds a new test.
>
> Theoretically resources_needed_watchpoint would have to be adapted for
> software watchpoints too, but in practice that function is only called
> in hw_watchpoint_used_count, which is never called with bp_watchpoint as
> an argument.
>
> FWIW, my local branch with my rework of debug registers accounting
> doesn't have hw_watchpoint_used_count anymore.
Oooh! Is that branch somewhere public?
>
> > The error string could also be enhanced to include the real
> > watchpoint type (so a user of masked watchpoints doesn't get
> > confused).
>
> I tried to keep that code agnostic to the type of watchpoint at hand
> (hence the breakpoint_ops methods), so what about a more generic error
> message, like "There is no hardware debug support for this watchpoint."
> or "Expression cannot be implemented with hardware debug resources."?
...
> error (_("Expression cannot be implemented with this type of watchpoint."));
Yeah, I'd go the path of making the error string more generic. Perhaps
s/implemented/watched/.
--
Pedro Alves
2011-07-26 Pedro Alves <pedro@codesourcery.com>
gdb/
* breakpoint.c (works_in_software_mode_watchpoint): Also return
true for software watchpoints.
gdb/testsuite/
* gdb.base/watchpoint.exp
(test_disable_enable_software_watchpoint): New procedure.
(top level): Run it.
---
gdb/breakpoint.c | 3 ++-
gdb/testsuite/gdb.base/watchpoint.exp | 19 +++++++++++++++++++
2 files changed, 21 insertions(+), 1 deletion(-)
Index: src/gdb/testsuite/gdb.base/watchpoint.exp
===================================================================
--- src.orig/gdb/testsuite/gdb.base/watchpoint.exp 2011-07-26 20:34:05.271448492 +0100
+++ src/gdb/testsuite/gdb.base/watchpoint.exp 2011-07-26 20:38:49.661448542 +0100
@@ -630,6 +630,23 @@ proc test_constant_watchpoint {} {
gdb_test_no_output "delete \$bpnum" "delete watchpoint `7 + count'"
}
+proc test_disable_enable_software_watchpoint {} {
+ # This is regression test for a bug that caused `enable' to fail
+ # for software watchpoints.
+
+ # Watch something not memory to force a software watchpoint.
+ gdb_test {watch $pc} ".*atchpoint \[0-9\]+: .pc"
+
+ gdb_test_no_output "disable \$bpnum" "disable watchpoint `\$pc'"
+ gdb_test_no_output "enable \$bpnum" "reenable watchpoint `\$pc'"
+
+ gdb_test "info watchpoint \$bpnum" \
+ ".*watchpoint\[ \t\]+keep\[ \t\]+y\[ \t\]+.pc.*" \
+ "watchpoint `\$pc' is enabled"
+
+ gdb_test_no_output "delete \$bpnum" "delete watchpoint `\$pc'"
+}
+
proc test_watch_location {} {
gdb_breakpoint [gdb_get_line_number "func5 breakpoint here"]
gdb_continue_to_breakpoint "func5 breakpoint here"
@@ -903,6 +920,8 @@ if [initialize] then {
test_constant_watchpoint
+ test_disable_enable_software_watchpoint
+
test_watch_location
}
Index: src/gdb/breakpoint.c
===================================================================
--- src.orig/gdb/breakpoint.c 2011-07-26 20:34:05.271448492 +0100
+++ src/gdb/breakpoint.c 2011-07-26 20:36:48.211448520 +0100
@@ -8637,7 +8637,8 @@ resources_needed_watchpoint (const struc
static int
works_in_software_mode_watchpoint (const struct breakpoint *b)
{
- return b->type == bp_hardware_watchpoint;
+ /* Read and access watchpoints only work with hardware support. */
+ return b->type == bp_watchpoint || b->type == bp_hardware_watchpoint;
}
static enum print_stop_action
next prev parent reply other threads:[~2011-07-26 19:46 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-05-21 22:20 ping: Re: PATCH : allow to set length of hw watchpoints (e.g. for Valgrind gdbserver) Philippe Waroquiers
2011-05-26 19:02 ` Tom Tromey
2011-05-29 13:01 ` Philippe Waroquiers
2011-05-30 15:26 ` Joel Brobecker
2011-05-31 19:07 ` x86 watchpoints bug (Re: ping: Re: PATCH : allow to set length of hw watchpoints (e.g. for Valgrind gdbserver)) Pedro Alves
2011-05-31 20:25 ` Philippe Waroquiers
2011-05-31 20:53 ` Pedro Alves
2011-05-31 21:29 ` Pedro Alves
2011-05-31 22:15 ` Philippe Waroquiers
2011-05-31 23:04 ` Pedro Alves
2011-06-01 14:35 ` Pedro Alves
2011-06-08 22:55 ` Philippe Waroquiers
2011-06-09 0:00 ` Pedro Alves
2011-06-09 22:16 ` Philippe Waroquiers
2011-07-21 17:20 ` Pedro Alves
2011-07-22 16:40 ` Philippe Waroquiers
2011-07-22 16:43 ` Pedro Alves
2011-07-23 16:28 ` Thiago Jung Bauermann
2011-07-26 20:02 ` Pedro Alves [this message]
2011-07-27 3:45 ` software watchpoints bug (was: Re: x86 watchpoints bug) Thiago Jung Bauermann
2011-07-22 17:19 ` x86 watchpoints bug (Re: ping: Re: PATCH : allow to set length of hw watchpoints (e.g. for Valgrind gdbserver)) Pedro Alves
2011-05-27 3:25 ` ping: Re: PATCH : allow to set length of hw watchpoints (e.g. for Valgrind gdbserver) Yao Qi
2011-05-27 17:53 ` Tom Tromey
2011-05-27 17:59 ` Pedro Alves
2011-05-30 4:06 ` Yao Qi
2011-05-30 5:34 ` Philippe Waroquiers
2011-05-30 5:48 ` Yao Qi
2011-05-30 6:31 ` Philippe Waroquiers
2011-05-31 17:31 ` Pedro Alves
2011-05-31 18:06 ` Philippe Waroquiers
2011-06-01 15:15 ` Pedro Alves
2011-06-05 20:55 ` Philippe Waroquiers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=201107262046.19451.pedro@codesourcery.com \
--to=pedro@codesourcery.com \
--cc=bauerman@br.ibm.com \
--cc=gdb-patches@sourceware.org \
--cc=philippe.waroquiers@skynet.be \
--cc=yao@codesourcery.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox