Re: [Outreachy kernel] [PATCH] staging: ade7754: Replace mlock with buf_lock and refactor code

[Alison Schofield <amsfield22@gmail.com>]

On Tue, Mar 21, 2017 at 11:48:27PM +0530, Gargi Sharma wrote:
> On Tue, Mar 21, 2017 at 11:42 PM, Alison Schofield <amsfield22@gmail.com> wrote:
> > On Tue, Mar 21, 2017 at 11:18:23PM +0530, Gargi Sharma wrote:
> >> On Tue, Mar 21, 2017 at 10:54 PM, Alison Schofield <amsfield22@gmail.com> wrote:
> >> > On Mon, Mar 20, 2017 at 01:09:21PM +0530, Gargi Sharma wrote:
> >> >> The IIO subsystem is redefining iio_dev->mlock to be used by
> >> >> the IIO core only for protecting device operating mode changes.
> >> >> ie. Changes between INDIO_DIRECT_MODE, INDIO_BUFFER_* modes.
> >> >>
> >> >> In this driver, mlock was being used to protect hardware state
> >> >> changes. Replace it with buf_lock in the devices global data.
> >> >>
> >> >> As buf_lock already protects operations in ade7754_write_frequency,
> >> >> there isn't a need to acquire the lock inside ade7754_spi_write_reg_8
> >> >> when writing to the register.
> >> >
> >> > Hi Gargi,
> >> >
> >> > Looks like something went wrong in your patch below.  It doesn't do what
> >> > you say it'll do...Instead of removing the lock from _write_reg_8()
> >> > it inserts a bunch of code.  Anyway, it seems that w_rite_reg_8() is used
> >> > in multiple places, so removing that lock doesn't appear to be an
> >> > option.
> >> >
> >> > See below...
> >> >
> >> > alisons
> >> >
> >> >>
> >> >> Signed-off-by: Gargi Sharma <gs051095@gmail.com>
> >> >> ---
> >> >>  drivers/staging/iio/meter/ade7754.c | 13 +++++++++++--
> >> >>  1 file changed, 11 insertions(+), 2 deletions(-)
> >> >>
> >> >> diff --git a/drivers/staging/iio/meter/ade7754.c b/drivers/staging/iio/meter/ade7754.c
> >> >> index 024463a..eb03469 100644
> >> >> --- a/drivers/staging/iio/meter/ade7754.c
> >> >> +++ b/drivers/staging/iio/meter/ade7754.c
> >> >> @@ -29,6 +29,15 @@ static int ade7754_spi_write_reg_8(struct device *dev, u8 reg_address, u8 val)
> >> >>       struct iio_dev *indio_dev = dev_to_iio_dev(dev);
> >> >>       struct ade7754_state *st = iio_priv(indio_dev);
> >> >>
> >> >> +     if (reg_address == ADE7754_WAVMODE) {
> >> >> +             st->tx[0] = ADE7754_WRITE_REG(reg_address);
> >> >> +             st->tx[1] = val;
> >> >> +
> >> >> +             ret = spi_write(st->us, st->tx, 2);
> >> >> +
> >> >> +             return ret;
> >> >> +     }
> >> >> +
> >> > What's this?
> >>
> >> When the function ade_spi_write_reg_8() is called inside
> >> ade7754_write_frequency(), we are writing to this( ADE7754_WAVMODE)
> >> register. When writing to this register we don't need to hold the
> >> buf_lock since ade7754_write_frequency() already takes care of that.
> >
> > Oh! I see it now. You created a special 'no lock needed' case
> > inside of --write_reg_8 for  writing frequency.   That works,
> > but it's...ummm...sneaky ;)  Let's see if there's another way.
> >
> > Look back at Lars suggestion on a similar patch.  Maybe that
> > will apply here.
> > http://marc.info/?l=linux-kernel&m=148940648615743&w=2
> >
> 
> I did look at the patch you suggested for inspiration :) What I could
> not understand was the part where Lars wrote about "doing a
> read-modify-write cycle in a protected section." I can write a
> separate function for --write_reg_8 that does not take the lock, but
> do not know how to do a "read-modify-write cycle in a protected
> section" :(
> 
> Gargi

Gargi - That's a great, specific question! Let's pop that back to Lars.

Lars - Can you explain further or perhaps point to a driver that does
something similar?

alisons

> > alisons
> >
> >
> >>
> >> >
> >> >>       mutex_lock(&st->buf_lock);
> >> >>       st->tx[0] = ADE7754_WRITE_REG(reg_address);
> >> >>       st->tx[1] = val;
> >> >> @@ -430,7 +439,7 @@ static ssize_t ade7754_write_frequency(struct device *dev,
> >> >>       if (!val)
> >> >>               return -EINVAL;
> >> >>
> >> >> -     mutex_lock(&indio_dev->mlock);
> >> >> +     mutex_lock(&st->buf_lock);
> >> >>
> >> >>       t = 26000 / val;
> >> >>       if (t > 0)
> >> >> @@ -451,7 +460,7 @@ static ssize_t ade7754_write_frequency(struct device *dev,
> >> >>       ret = ade7754_spi_write_reg_8(dev, ADE7754_WAVMODE, reg);
> >> >>
> >> >>  out:
> >> >> -     mutex_unlock(&indio_dev->mlock);
> >> >> +     mutex_unlock(&st->buf_lock);
> >> >>
> >> >>       return ret ? ret : len;
> >> >>  }
> >>
> >> The buf_lock inside ade7754_write_frequency() takes into account that
> >> when using the function ade7754_spi_write_reg_8, lock is already held
> >> and locking is no longer required inside the ade7754_spi_write_reg_8()
> >> function.
> >>
> >> Let me know if this sounds okay, I can perhaps edit the commit log to
> >> make this clearer.
> >>
> >> Thanks,
> >> Gargi
> >>
> >> >> --
> >> >> 2.7.4
> >> >>
> >> >> --
> >> >> You received this message because you are subscribed to the Google Groups "outreachy-kernel" group.
> >> >> To unsubscribe from this group and stop receiving emails from it, send an email to outreachy-kernel+unsubscribe@googlegroups.com.
> >> >> To post to this group, send email to outreachy-kernel@googlegroups.com.
> >> >> To view this discussion on the web visit https://groups.google.com/d/msgid/outreachy-kernel/1489995561-6988-1-git-send-email-gs051095%40gmail.com.
> >> >> For more options, visit https://groups.google.com/d/optout.
> >> >
> >> > --
> >> > You received this message because you are subscribed to the Google Groups "outreachy-kernel" group.
> >> > To unsubscribe from this group and stop receiving emails from it, send an email to outreachy-kernel+unsubscribe@googlegroups.com.
> >> > To post to this group, send email to outreachy-kernel@googlegroups.com.
> >> > To view this discussion on the web visit https://groups.google.com/d/msgid/outreachy-kernel/20170321172438.GC2793%40d830.WORKGROUP.
> >> > For more options, visit https://groups.google.com/d/optout.
> >
> > --
> > You received this message because you are subscribed to the Google Groups "outreachy-kernel" group.
> > To unsubscribe from this group and stop receiving emails from it, send an email to outreachy-kernel+unsubscribe@googlegroups.com.
> > To post to this group, send email to outreachy-kernel@googlegroups.com.
> > To view this discussion on the web visit https://groups.google.com/d/msgid/outreachy-kernel/20170321181207.GA10699%40d830.WORKGROUP.
> > For more options, visit https://groups.google.com/d/optout.