Re: [PATCH] drm/i915: Fix single msg gmbus_xfers writes

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Feb 15, 2012 6:48 PM, "Daniel Vetter" <daniel@xxxxxxxx> wrote:
>
> On Thu, Feb 09, 2012 at 12:03:17PM -0800, Benson Leung wrote:
> > gmbus_xfer with a single message (particularly a single message write) would
> > set Bus Cycle Select to 100b, the Gen Stop cycle, instead of 101b,
> > No Index, Stop cycle. This would not start single message i2c transactions.
> >
> > Also, gmbus_xfer done: will disable the interface without checking if
> > it is idle. In the case of writes, there will be no wait on status or delay
> > to ensure the write starts and completes before the interface is turned off.
> >
> > Fixed the former issue by using the same cycle selection as used in the
> > I2C_M_RD for the write case.
> > GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0)
> > Fixed the latter by waiting on GMBUS_ACTIVE to deassert before disable.
> >
> > Signed-off-by: Benson Leung <bleung@xxxxxxxxxxxx>

Reviewed-by: Daniel Kurtz <djkurtz@xxxxxxxxxxxx>

>
> Can you clarify the commit message a bit and say that the first hunk is
> just for optics and the issue is only with the write path (because the
> read path is correct already). Silly me is just to easily confused ;-)
>
> Btw, I've reworked the gmbus -> gpio bit-banging fallback code a bit and
> if that passes review and all I'll reenable gmbus by default again. See
>
> http://cgit.freedesktop.org/~danvet/drm/log/?h=gmbus

If the write case is fixed by Benson's patch, is there any known use
case that still requires i2c bit banging on these pins?  It would be a
nice cleanup to remove it completely.

Also, I can think of at least two further potential performance
improvements that I was wondering if anybody has yet pursued:
  (1) Enabling the i915's gmbus interrupt.  This would eliminate the
need for the (relatively slow) wait_for polling loop.
  (2) Taking advantage of the i915's "INDEX" cycles to combine writing
a (1 or 2 byte) address & reading back an array of bytes into a single
transaction.

Best Regards,
-Daniel

>
> Yours, Daniel
>
> > ---
> >  drivers/gpu/drm/i915/intel_i2c.c |   13 +++++++++----
> >  1 files changed, 9 insertions(+), 4 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
> > index d30cccc..64bb9cd 100644
> > --- a/drivers/gpu/drm/i915/intel_i2c.c
> > +++ b/drivers/gpu/drm/i915/intel_i2c.c
> > @@ -249,7 +249,8 @@ gmbus_xfer(struct i2c_adapter *adapter,
> >
> >               if (msgs[i].flags & I2C_M_RD) {
> >                       I915_WRITE(GMBUS1 + reg_offset,
> > -                                GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
> > +                                GMBUS_CYCLE_WAIT |
> > +                                (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
> >                                  (len << GMBUS_BYTE_COUNT_SHIFT) |
> >                                  (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
> >                                  GMBUS_SLAVE_READ | GMBUS_SW_RDY);
> > @@ -278,7 +279,8 @@ gmbus_xfer(struct i2c_adapter *adapter,
> >
> >                       I915_WRITE(GMBUS3 + reg_offset, val);
> >                       I915_WRITE(GMBUS1 + reg_offset,
> > -                                (i + 1 == num ? GMBUS_CYCLE_STOP : GMBUS_CYCLE_WAIT) |
> > +                                GMBUS_CYCLE_WAIT |
> > +                                (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
> >                                  (msgs[i].len << GMBUS_BYTE_COUNT_SHIFT) |
> >                                  (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
> >                                  GMBUS_SLAVE_WRITE | GMBUS_SW_RDY);
> > @@ -317,9 +319,12 @@ clear_err:
> >       I915_WRITE(GMBUS1 + reg_offset, 0);
> >
> >  done:
> > -     /* Mark the GMBUS interface as disabled. We will re-enable it at the
> > -      * start of the next xfer, till then let it sleep.
> > +     /* Mark the GMBUS interface as disabled after waiting for idle.
> > +      * We will re-enable it at the start of the next xfer,
> > +      * till then let it sleep.
> >        */
> > +     if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0, 10))
> > +             DRM_INFO("GMBUS timed out waiting for idle\n");
> >       I915_WRITE(GMBUS0 + reg_offset, 0);
> >       return i;
> >
> > --
> > 1.7.1
> >
> > _______________________________________________
> > dri-devel mailing list
> > dri-devel@xxxxxxxxxxxxxxxxxxxxx
> > http://lists.freedesktop.org/mailman/listinfo/dri-devel
>
> --
> Daniel Vetter
> Mail: daniel@xxxxxxxx
> Mobile: +41 (0)79 365 57 48
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/dri-devel



[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux