[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] gdbstub: Allow target CPUs to specify watchpoin
From: |
Max Filippov |
Subject: |
Re: [Qemu-devel] [PATCH] gdbstub: Allow target CPUs to specify watchpoint STOP_BEFORE_ACCESS flag |
Date: |
Mon, 29 Sep 2014 22:58:55 +0400 |
On Tue, Sep 16, 2014 at 9:16 AM, Max Filippov <address@hidden> wrote:
> On Mon, Sep 15, 2014 at 9:15 PM, Peter Maydell <address@hidden> wrote:
>> On 15 September 2014 20:59, Max Filippov <address@hidden> wrote:
>>> I've tested xtensa part and have noticed no difference with or without
>>> this patch:
>>> gdb connected to qemu gdbstub always stops right after the watched
>>> instruction.
>>> I guess that without this patch I should have seen gdb stopping not right
>>> after
>>> the watched instruction, but one instruction later, right?
>>
>> Yes, that is the behaviour I saw on ARM. Perhaps for
>> Xtensa gdb thinks it has no watchpoint support at all
>> and is using its "singlestep-and-test" fallback?
>
> Hmmm, interesting. Yes, I see gdb single-stepping and reading watched buffer.
> With the patch it first stops at the store instruction, without it --
> at the next one.
> In both cases the following single-step request does nothing. Need to look at
> it some more.
Looks like I won't be able to look at it in a reasonable time. So, here's my
Tested-by: Max Filippov <address@hidden>
to signify that it at least doesn't have visible negative effect on
target-xtensa.
I'll still try to figure out why xtensa doesn't exhibit the issue that
the patch is
supposed to fix.
--
Thanks.
-- Max