> From: Paolo Bonzini [mailto:pbonzini@redhat.com]
> On 24/07/19 10:44, Pavel Dovgalyuk wrote:
> > From: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
> >
> > Most of IO instructions can be executed only at the end of the block in
> > icount mode. Therefore translator can set cpu_can_io flag when translating
> > the last instruction.
> > But when the blocks are chained, then this flag is not reset and may
> > remain set at the beginning of the next block.
> > This patch resets the flag before "chaining" the translation blocks.
> >
> > Signed-off-by: Pavel Dovgalyuk <Pavel.Dovgaluk@ispras.ru>
> > ---
> > accel/tcg/tcg-runtime.c | 2 ++
> > 1 file changed, 2 insertions(+)
> >
> > diff --git a/accel/tcg/tcg-runtime.c b/accel/tcg/tcg-runtime.c
> > index 8a1e408e31..fe6b83d0fc 100644
> > --- a/accel/tcg/tcg-runtime.c
> > +++ b/accel/tcg/tcg-runtime.c
> > @@ -151,6 +151,8 @@ void *HELPER(lookup_tb_ptr)(CPUArchState *env)
> > target_ulong cs_base, pc;
> > uint32_t flags;
> >
> > + /* We are going to jump to the next block. can_do_io should be reset */
> > + cpu->can_do_io = !use_icount;
> > tb = tb_lookup__cpu_state(cpu, &pc, &cs_base, &flags, curr_cflags());
> > if (tb == NULL) {
> > return tcg_ctx->code_gen_epilogue;
> >
>
> This only fixes indirect jumps though.
>
> I think you do not need this patch if you remove the assignment in
> cpu_tb_exec, and compile a "move 0 to cpu->can_do_io" in gen_tb_start
> instead.
"move 0 to cpu->can_do_io" only for icount mode?
And we'll also need to set can_do_io to 1 somewhere, because it
is checked in non-icount mode too.
Pavel Dovgalyuk