Re: [PATCH v1 3/5] bus: mhi: core: Skip handling BHI irq if MHI reg access is not allowed

[Hemant Kumar]

Hi Mani,

On 5/13/20 12:04 AM, Manivannan Sadhasivam wrote:
> On Tue, May 12, 2020 at 05:28:45PM -0700, Hemant Kumar wrote:
> > Hi Mani,
> >
> > On 5/11/20 11:53 PM, Manivannan Sadhasivam wrote:
> > > On Mon, May 11, 2020 at 07:03:07PM -0700, Hemant Kumar wrote:
> > > > Driver continues handling of BHI interrupt even if MHI register access
> > > > is not allowed. By doing so it calls the status call back and performs
> > > > early notification for the MHI client. This is not needed when MHI
> > > > register access is not allowed. Hence skip the handling in this case and
> > > > return. Also add debug log to print device state, local EE and device EE
> > > > when reg access is valid.
> > > >
> > > > Signed-off-by: Hemant Kumar <hemantk@xxxxxxxxxxxxxx>
> > > > Reviewed-by: Jeffrey Hugo <jhugo@xxxxxxxxxxxxxx>
> > > > ---
> > > >    drivers/bus/mhi/core/main.c | 21 ++++++++++++++-------
> > > >    1 file changed, 14 insertions(+), 7 deletions(-)
> > > >
> > > > diff --git a/drivers/bus/mhi/core/main.c b/drivers/bus/mhi/core/main.c
> > > > index 9ec9b36..467c0ba 100644
> > > > --- a/drivers/bus/mhi/core/main.c
> > > > +++ b/drivers/bus/mhi/core/main.c
> > > > @@ -369,22 +369,29 @@ irqreturn_t mhi_irq_handler(int irq_number, void *dev)
> > > >    	return IRQ_HANDLED;
> > > >    }
> > > > -irqreturn_t mhi_intvec_threaded_handler(int irq_number, void *dev)
> > > > +irqreturn_t mhi_intvec_threaded_handler(int irq_number, void *priv)
> > > >    {
> > > > -	struct mhi_controller *mhi_cntrl = dev;
> > > > +	struct mhi_controller *mhi_cntrl = priv;
> > > > +	struct device *dev = &mhi_cntrl->mhi_dev->dev;
> > > >    	enum mhi_state state = MHI_STATE_MAX;
> > > >    	enum mhi_pm_state pm_state = 0;
> > > >    	enum mhi_ee_type ee = 0;
> > > >    	write_lock_irq(&mhi_cntrl->pm_lock);
> > > > -	if (MHI_REG_ACCESS_VALID(mhi_cntrl->pm_state)) {
> > > > -		state = mhi_get_mhi_state(mhi_cntrl);
> > > > -		ee = mhi_cntrl->ee;
> > > > -		mhi_cntrl->ee = mhi_get_exec_env(mhi_cntrl);
> > > > +	if (!MHI_REG_ACCESS_VALID(mhi_cntrl->pm_state)) {
> > > > +		write_unlock_irq(&mhi_cntrl->pm_lock);
> > >
> > > write_lock is only used for protecting 'mhi_cntrl->ee' but here we are not
> > > updating it if reg access is not valid. So there is no reason to hold this lock.
> > Original code is using write_lock to protect pm_state as well as
> > mhi_cntrl->ee. This patch is keeping the lock same as original code. Just if
> > condition logic is negated here due to that write_unlock_irq is added under
> > if condition.
>
> 'mhi_cntrl->pm_state' is not always protected by 'pm_lock' and that too
> write_lock is used here but 'pm_state' is not modified. So as like in most of
> the places, locks are abused here as well.
>
> I think after 5.8, you should really think about fixing the usage of locks
> throughout the MHI stack.
>
> So I'll take this patch as it is.
Thanks for accepting this patch. Will revisit and review usage of locks 
in MHI stack and identify areas of improvement.
> Thanks,
> Mani
>
> > > > +		goto exit_intvec;
> > > >    	}
> > > > +	state = mhi_get_mhi_state(mhi_cntrl);
> > > > +	ee = mhi_cntrl->ee;
> > > > +	mhi_cntrl->ee = mhi_get_exec_env(mhi_cntrl);
> > >
> > > But it is needed here.
> > >
> > > Thanks,
> > > Mani
> > >
> > > > +	dev_dbg(dev, "local ee:%s device ee:%s dev_state:%s\n",
> > > > +		TO_MHI_EXEC_STR(mhi_cntrl->ee), TO_MHI_EXEC_STR(ee),
> > > > +		TO_MHI_STATE_STR(state));
> > > > +
> > > >    	if (state == MHI_STATE_SYS_ERR) {
> > > > -		dev_dbg(&mhi_cntrl->mhi_dev->dev, "System error detected\n");
> > > > +		dev_dbg(dev, "System error detected\n");
> > > >    		pm_state = mhi_tryset_pm_state(mhi_cntrl,
> > > >    					       MHI_PM_SYS_ERR_DETECT);
> > > >    	}
> > > > --
> > > > The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> > > > a Linux Foundation Collaborative Project
> >
> > --
> > The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
> > a Linux Foundation Collaborative Project

--
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project