[Qemu-devel] [4432] Debugger single step without interrupts (Jason Wessel).

[Qemu-devel] [4432] Debugger single step without interrupts (Jason Wessel).

[Edgar E. Iglesias]

Revision: 4432
          http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=4432
Author:   edgar_igl
Date:     2008-05-12 07:44:23 +0000 (Mon, 12 May 2008)

Log Message:
-----------
Debugger single step without interrupts (Jason Wessel).

Modified Paths:
--------------
    trunk/vl.c

Modified: trunk/vl.c
===================================================================
--- trunk/vl.c	2008-05-11 19:24:10 UTC (rev 4431)
+++ trunk/vl.c	2008-05-12 07:44:23 UTC (rev 4432)
@@ -7032,6 +7032,7 @@
     qemu_aio_poll();
 
     if (vm_running) {
+        if (!(cur_cpu->singlestep_enabled & SSTEP_NOTIMER))
         qemu_run_timers(&active_timers[QEMU_TIMER_VIRTUAL],
                         qemu_get_clock(vm_clock));
         /* run dma transfers, if any */
@@ -7039,6 +7040,7 @@
     }
 
     /* real time timers */
+    if (!(cur_cpu->singlestep_enabled & SSTEP_NOTIMER))
     qemu_run_timers(&active_timers[QEMU_TIMER_REALTIME],
                     qemu_get_clock(rt_clock));
 

Re: [Qemu-devel] [4432] Debugger single step without interrupts (Jason Wessel).

[Fabrice Bellard]

I see no reason why the real time timers should be stopped while single
stepping. These timers have no relation with the emulation.

BTW, it seems that some devices incorrectly use the real time timers...

Regards,

Fabrice.

Edgar E. Iglesias wrote:
> Revision: 4432
>           http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=4432
> Author:   edgar_igl
> Date:     2008-05-12 07:44:23 +0000 (Mon, 12 May 2008)
> 
> Log Message:
> -----------
> Debugger single step without interrupts (Jason Wessel).
> 
> Modified Paths:
> --------------
>     trunk/vl.c
> 
> Modified: trunk/vl.c
> ===================================================================
> --- trunk/vl.c	2008-05-11 19:24:10 UTC (rev 4431)
> +++ trunk/vl.c	2008-05-12 07:44:23 UTC (rev 4432)
> @@ -7032,6 +7032,7 @@
>      qemu_aio_poll();
>  
>      if (vm_running) {
> +        if (!(cur_cpu->singlestep_enabled & SSTEP_NOTIMER))
>          qemu_run_timers(&active_timers[QEMU_TIMER_VIRTUAL],
>                          qemu_get_clock(vm_clock));
>          /* run dma transfers, if any */
> @@ -7039,6 +7040,7 @@
>      }
>  
>      /* real time timers */
> +    if (!(cur_cpu->singlestep_enabled & SSTEP_NOTIMER))
>      qemu_run_timers(&active_timers[QEMU_TIMER_REALTIME],
>                      qemu_get_clock(rt_clock));
>  
> 
> 
> 
> 
> 

Re: [Qemu-devel] [4432] Debugger single step without interrupts (Jason Wessel).

[Edgar E. Iglesias]

On Mon, May 12, 2008 at 12:08:37PM +0200, Fabrice Bellard wrote:
> I see no reason why the real time timers should be stopped while single
> stepping. These timers have no relation with the emulation.

Thanks for noticing.

> BTW, it seems that some devices incorrectly use the real time timers...

OK, not sure if thats the reason for disabling them. Jason?
I tried to keep them on with the etrax machine and the sstep works equally well.

Best regards
-- 
Edgar E. Iglesias
Axis Communications AB

Index: vl.c
===================================================================
--- vl.c	(revision 4432)
+++ vl.c	(working copy)
@@ -7040,7 +7040,6 @@
     }
 
     /* real time timers */
-    if (!(cur_cpu->singlestep_enabled & SSTEP_NOTIMER))
     qemu_run_timers(&active_timers[QEMU_TIMER_REALTIME],
                     qemu_get_clock(rt_clock));
 

Re: [Qemu-devel] [4432] Debugger single step without interrupts (Jason Wessel).

[Jason Wessel]

Edgar E. Iglesias wrote:
> On Mon, May 12, 2008 at 12:08:37PM +0200, Fabrice Bellard wrote:
>   
>> I see no reason why the real time timers should be stopped while single
>> stepping. These timers have no relation with the emulation.
>>     
>
> Thanks for noticing.
>
>   
>> BTW, it seems that some devices incorrectly use the real time timers...
>>     
>
> OK, not sure if thats the reason for disabling them. Jason?
> I tried to keep them on with the etrax machine and the sstep works equally well.
>   

There is no particular need to disable the real time timers at this
point. I had exchanged mail with Fabrice over a year ago on this one.
The patch you applied for the irq stepping fixes had never been formally
re-submitted by myself or had that piece cleaned up. I had posted some
old patches in response because someone asked about them on the mailing
list.

Unless you beat me too it, I will gladly submit a patch to clean up
stepping after the gdb monitor cleanup. :-)

Jason.