[PATCH 1/1] lib: sbi_hsm: Restor hart state to stop when fails to start

[Joshua Yeong]

On Mon, Feb 19, 2024 at 05:39:26PM +0800, Joshua Yeong <joshua.yeong at starfivetech.com> wrote:
> 在 2024-02-19星期一的 15:14 +0800，Joshua Yeong写道：
> > Hart state should change back to hart stop when
> > hsm_device_hart_start() or sbi_ipi_raw_send() fails to perform hart start.
> >
> > Signed-off-by: Joshua Yeong <joshua.yeong at starfivetech.com>
> > ---
> >  lib/sbi/sbi_hsm.c | 4 ++++
> >  1 file changed, 4 insertions(+)
> >
> > diff --git a/lib/sbi/sbi_hsm.c b/lib/sbi/sbi_hsm.c index
> > 3d60ceb..be48d64 100644
> > --- a/lib/sbi/sbi_hsm.c
> > +++ b/lib/sbi/sbi_hsm.c
> > @@ -360,6 +360,10 @@ int sbi_hsm_hart_start(struct sbi_scratch
> > *scratch,
> >
> >  	if (!rc)
> >  		return 0;
> > +
> > +	/* If it fails to start, change hart state back to stop */
> > +	__sbi_hsm_hart_change_state(hdata,
> SBI_HSM_STATE_START_PENDING,
> > +				    SBI_HSM_STATE_STOPPED);
> 
> This is not needed. Because if hdata->state is not equal to
> SBI_HSM_STATE_STOPPED, SBI_HSM_STATE_START_PENDING will not be
> written.
> 
> Regards,
> Xiang W

The scenario is that the Hart is in SBI_HSM_STATE_STOPPED and then it switches to SBI_HSM_STATE_START_PENDING,
but it got stuck in SBI_HSM_STATE_START_PENDING if it fails to start. The hart's state has impact on
other process where hart's state is taken into consideration, for example system suspend.

Regards,
Joshua 

> 
> >  err:
> >  	hsm_start_ticket_release(hdata);
> >  	return rc;
> > --
> > 2.25.1
> >
> >