This is the mail archive of the gdb-prs@sources.redhat.com mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: gdb/331: watch is finicky about giving a hardware watchpoint


The following reply was made to PR gdb/331; it has been noted by GNATS.

From: Dru Nelson <dru@redwoodsoft.com>
To: Daniel Jacobowitz <drow@mvista.com>
Cc: dru-gdb@redwoodsoft.com,  <gdb-gnats@sources.redhat.com>
Subject: Re: gdb/331: watch is finicky about giving a hardware watchpoint
Date: Sat, 2 Feb 2002 15:17:26 -0800 (PST)

 One last email.
 
 I'll work on that patch tonight. Shouldn't be too hard.
 
 Thanks,
 
 Dru Nelson
 San Mateo, California
 
 
 On Sat, 2 Feb 2002, Daniel Jacobowitz wrote:
 
 > On Sat, Feb 02, 2002 at 10:41:42PM -0000, dru-gdb@redwoodsoft.com wrote:
 > >
 > > >Number:         331
 > > >Category:       gdb
 > > >Synopsis:       watch is finicky about giving a hardware watchpoint
 > > >Confidential:   no
 > > >Severity:       serious
 > > >Priority:       medium
 > > >Responsible:    unassigned
 > > >State:          open
 > > >Class:          sw-bug
 > > >Submitter-Id:   net
 > > >Arrival-Date:   Sat Feb 02 14:48:01 PST 2002
 > > >Closed-Date:
 > > >Last-Modified:
 > > >Originator:     Dru
 > > >Release:        gdb 5.1.1
 > > >Organization:
 > > >Environment:
 > > linux 2.2.16 (patched with jcrown debug register patches)
 > > x86
 > > >Description:
 > > When I set a watch to a named variable, gdb will create a
 > > 'hardware watchpoint'. However if I issue:
 > > 'watch 0x81d79a8' gdb creates a software watchpoint.
 > > This is annoying since I'm debugging after a segmentation
 > > violation and I need to re-run the program to see who is
 > > writing to that memory.
 > >
 > > Why isn't this a valid expression to create a hardware
 > > watchpoint that doesn't rely on variable scope?
 >
 > That's watching the value of '0x81d79a8'.  A constant.  You want:
 > watch *(int *)0x81d79a8
 >
 > which should create a hardware watchpoint.
 >
 > I agree we should warn about this.
 >
 > --
 > Daniel Jacobowitz                           Carnegie Mellon University
 > MontaVista Software                         Debian GNU/Linux Developer
 >
 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]