This is the mail archive of the gdb-patches@sourceware.org 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: [PATCH 0/2] Honour software single step in fallback of displaced stepping


On 04/07/2015 04:52 PM, Yao Qi wrote:
> From: Yao Qi <yao.qi@linaro.org>
> 
> Hi,
> This patch set fixes many fails I've seen in
> gdb.threads/non-stop-fair-events.exp on arm-linux target.  I don't see
> they are covered by Pedro's "All-stop on top of non-stop" V2, so I
> post them out for the review.

They actually are.  :-)  I tested v2 against x86 software single-step,
and it caught issues like that.

This patch:

 [PATCH v2 07/23] Embed the pending step-over chain in thread_info objects
 https://sourceware.org/ml/gdb-patches/2015-04/msg00218.html

splits that code you're touching to a separate "start_step_over_inferior"
function.

And then this patch:

 [PATCH v2 11/23] Use keep_going in proceed and start_step_over too
 https://sourceware.org/ml/gdb-patches/2015-04/msg00203.html

rewrites that whole function to defer to keep_going instead.  keep_going
already handles the case of the breakpoint disappearing (thread_still_needs_step_over
returns false).  And in case the breakpoint is still around, it ends
in 'resume' again, which is then the only place that knows
how to start a displaced step.

I don't mind if you push your patch in first.  I'll just
end up deleting that code again when I rebase it.

Thanks,
Pedro Alves


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