Re: [PATCH] target/openrisc: Do not reset delay slot flag on early tb exit

[Stafford Horne]

On Wed, May 11, 2022 at 02:56:37PM -0700, Richard Henderson wrote:
> On 5/11/22 14:43, Stafford Horne wrote:
> > At this point how would tb->flags have the right value?  Would it always be 
> > set
> > correctly by `cpu_get_tb_cpu_state` in the `lookup_tb_ptr`call?
> 
> Well, it would be set by cpu_get_tb_cpu_state in cpu_exec, which is then
> passed to tb_gen_code.  If we go around a loop and look it up a second time,
> we'll find a tb with a matching set of flags.

Right, cpu_get_tb_cpu_state called in lookup_tb_ptr will not update tb->flags.

What you mention, that is for when we have to generate a new TB, the tb->flags
get set right before tb_gen_code.

But for the case where we exit the delay-slot TB due to a pending exception I
think the flow would go.

 TB chain:
     -> branch-TB     : set env->flag 1
     -> delay-slot-TB : exit_tb due to condition
 Exit:
     -> return to cpu_tb_exec
       -> tcg_ops->synchronize_from_tb

In this case I don't see how the tb->flag would be updated, ooh, I guess it
would have been set earlier when the TB was generated.  Maybe that is what I am
missing.

-Stafford