This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFC] GDB patches for hw watchpoints - revised
- From: Wu Zhou <woodzltc at cn dot ibm dot com>
- To: anton at au1 dot ibm dot com
- Cc: drow at false dot org, gdb-patches at sources dot redhat dot com, mark at xs4all dot nl, bje at au1 dot ibm dot com
- Date: Thu, 22 Dec 2005 11:48:42 +0800 (CST)
- Subject: Re: [RFC] GDB patches for hw watchpoints - revised
> Yeah, we cant assume we have only one data breakpoint register - I think
> some of the 32bit cpus have multiple ones.
Can you recall which cpus have multiple ones? I am now reading related
processor document and find that Book E seems to use a different debug
facility. It has three debug control registers, one debug status
register, two insruction address compare registers and two data address
compare registers (maybe the same as DABR in other POWER/PowerPC arch).
That is in <<Book E: Enhanced PowerPC Architecture>>. So maybe the "at
most one DABR" assertion still hold for most ppc arches. And provided
that the current kernel only support at most one DABR, so this patch still
make sense for GDB, right? Any objection? :-)
What is more, in function ppc_linux_check_watch_resources, there is a run
time check to see whether the command PTRACE_SET_DEBUGREG of ptrace can
succeed. I believe that will make these archs which don't have DABR not
impacted by this patch.
So I am thinking this patch still make sense. What is your thought?
Very happy to know your comments.
Regards
- Wu Zhou