[PATCH] replay: fix watchpoint processing for reverse debugging

Pavel Dovgalyuk posted 1 patch 2 years, 11 months ago
Test checkpatch passed
Patches applied successfully (tree, apply log)
git fetch https://github.com/patchew-project/qemu tags/patchew/162072430303.827403.7379783546934958566.stgit@pasha-ThinkPad-X280
Maintainers: Paolo Bonzini <pbonzini@redhat.com>
softmmu/physmem.c |   10 ++++++++++
1 file changed, 10 insertions(+)
[PATCH] replay: fix watchpoint processing for reverse debugging
Posted by Pavel Dovgalyuk 2 years, 11 months ago
This patch enables reverse debugging with watchpoints.
Reverse continue scans the execution to find the breakpoints
and watchpoints that should fire. It uses helper function
replay_breakpoint() for that. But this function needs to access
icount, which can't be correct in the middle of TB.
Therefore, in case of watchpoint, we have to retranslate the block
to allow this access.

Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
---
 softmmu/physmem.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/softmmu/physmem.c b/softmmu/physmem.c
index 5232696571..2826d60cef 100644
--- a/softmmu/physmem.c
+++ b/softmmu/physmem.c
@@ -904,6 +904,16 @@ void cpu_check_watchpoint(CPUState *cpu, vaddr addr, vaddr len,
         if (watchpoint_address_matches(wp, addr, len)
             && (wp->flags & flags)) {
             if (replay_running_debug()) {
+                /*
+                 * replay_breakpoint reads icount.
+                 * Force recompile to succeed, because icount may
+                 * be read only at the end of the block.
+                 */
+                if (!cpu->can_do_io) {
+                    /* Force execution of one insn next time.  */
+                    cpu->cflags_next_tb = 1 | CF_LAST_IO | curr_cflags(cpu);
+                    cpu_loop_exit_restore(cpu, ra);
+                }
                 /*
                  * Don't process the watchpoints when we are
                  * in a reverse debugging operation.


Re: [PATCH] replay: fix watchpoint processing for reverse debugging
Posted by Pavel Dovgalyuk 2 years, 11 months ago
ping

On 11.05.2021 12:11, Pavel Dovgalyuk wrote:
> This patch enables reverse debugging with watchpoints.
> Reverse continue scans the execution to find the breakpoints
> and watchpoints that should fire. It uses helper function
> replay_breakpoint() for that. But this function needs to access
> icount, which can't be correct in the middle of TB.
> Therefore, in case of watchpoint, we have to retranslate the block
> to allow this access.
> 
> Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgalyuk@ispras.ru>
> ---
>   softmmu/physmem.c |   10 ++++++++++
>   1 file changed, 10 insertions(+)
> 
> diff --git a/softmmu/physmem.c b/softmmu/physmem.c
> index 5232696571..2826d60cef 100644
> --- a/softmmu/physmem.c
> +++ b/softmmu/physmem.c
> @@ -904,6 +904,16 @@ void cpu_check_watchpoint(CPUState *cpu, vaddr addr, vaddr len,
>           if (watchpoint_address_matches(wp, addr, len)
>               && (wp->flags & flags)) {
>               if (replay_running_debug()) {
> +                /*
> +                 * replay_breakpoint reads icount.
> +                 * Force recompile to succeed, because icount may
> +                 * be read only at the end of the block.
> +                 */
> +                if (!cpu->can_do_io) {
> +                    /* Force execution of one insn next time.  */
> +                    cpu->cflags_next_tb = 1 | CF_LAST_IO | curr_cflags(cpu);
> +                    cpu_loop_exit_restore(cpu, ra);
> +                }
>                   /*
>                    * Don't process the watchpoints when we are
>                    * in a reverse debugging operation.
> 


Re: [PATCH] replay: fix watchpoint processing for reverse debugging
Posted by Richard Henderson 2 years, 11 months ago
On 5/11/21 2:11 AM, Pavel Dovgalyuk wrote:
> This patch enables reverse debugging with watchpoints.
> Reverse continue scans the execution to find the breakpoints
> and watchpoints that should fire. It uses helper function
> replay_breakpoint() for that. But this function needs to access
> icount, which can't be correct in the middle of TB.
> Therefore, in case of watchpoint, we have to retranslate the block
> to allow this access.
> 
> Signed-off-by: Pavel Dovgalyuk<Pavel.Dovgalyuk@ispras.ru>
> ---
>   softmmu/physmem.c |   10 ++++++++++
>   1 file changed, 10 insertions(+)

Queued to tcg-next, thanks.


r~