Hi,


Gentle ping.  I sent this patch a couple of weeks ago and wanted to
check whether someone has had a chance to look at it.


The patch addresses PR gdb/21221 and introduces a heuristic that ends
a source-level step after one iteration of a tight empty loop.  I
would appreciate any feedback.


Thanks,
Bai




         原始邮件
         
       
发件人：白 <597498687@qq.com&gt;
发件时间：2026年8月17日 18:07
收件人：gdb-patches <gdb-patches@sourceware.org&gt;
主题：[PATCH] infrun: Introduce empty-loop step heuristic



       Hi,

This&nbsp;patch&nbsp;addresses&nbsp;PR&nbsp;gdb/21221.&nbsp;&nbsp;When&nbsp;a&nbsp;user&nbsp;single-steps&nbsp;a&nbsp;source
line&nbsp;containing&nbsp;a&nbsp;tight&nbsp;empty&nbsp;loop,&nbsp;GDB&nbsp;currently&nbsp;steps&nbsp;every
iteration&nbsp;and&nbsp;only&nbsp;stops&nbsp;once&nbsp;the&nbsp;loop&nbsp;finishes.&nbsp;&nbsp;The&nbsp;patch&nbsp;introduces&nbsp;a
heuristic&nbsp;that&nbsp;ends&nbsp;the&nbsp;step&nbsp;after&nbsp;one&nbsp;iteration&nbsp;by&nbsp;detecting&nbsp;when&nbsp;the
inferior&nbsp;returns&nbsp;to&nbsp;the&nbsp;first&nbsp;stop&nbsp;PC&nbsp;observed&nbsp;on&nbsp;the&nbsp;same&nbsp;source&nbsp;line.

The&nbsp;heuristic&nbsp;is&nbsp;only&nbsp;active&nbsp;in&nbsp;forward&nbsp;execution&nbsp;and&nbsp;includes&nbsp;a
safety&nbsp;threshold&nbsp;to&nbsp;avoid&nbsp;runaway&nbsp;steps.&nbsp;&nbsp;A&nbsp;ui_out&nbsp;message&nbsp;is&nbsp;emitted
so&nbsp;front-ends&nbsp;can&nbsp;parse&nbsp;the&nbsp;stop&nbsp;reason.

Tested&nbsp;by&nbsp;gdb.base/empty-loop-step.exp.

Thanks,
Bai