This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [reverse RFA] no singlestep-over-BP in reverse
- From: Michael Snyder <msnyder at vmware dot com>
- To: teawater <teawater at gmail dot com>
- Cc: "gdb-patches at sourceware dot org" <gdb-patches at sourceware dot org>
- Date: Tue, 16 Sep 2008 11:37:30 -0700
- Subject: Re: [reverse RFA] no singlestep-over-BP in reverse
- References: <48CEAA05.8050006@vmware.com> <20080915184245.GA21388@caradoc.them.org> <daef60380809152110u663350abx76b283d519c5a09d@mail.gmail.com>
teawater wrote:
On Tue, Sep 16, 2008 at 02:42, Daniel Jacobowitz <drow@false.org
<mailto:drow@false.org>> wrote:
If there is a breakpoint on the previous instruction, will you hit it
before or after de-executing that instruction? It seems like this
logic should be somehow still necessary... but I can't put my finger
on when.
I think is before exec the next instruction, the GDB will get breakpoint
trap.
But I found that there is a bug in inside record replay mode, I stop the
GDB after exec the instruction.
Aha, let's figure out a test for this question.
It's rather important. ;-)
Michael, how do you deal with the breakpoint in gdb-freeplay and vmware
record?
Both use simulated hardware breakpoints. That is, they both
accept the Z0 message and emulate hardware breakpoint semantics
by watching the address of execution (one instruction at a time).