[Intel-gfx] [PATCH v2] drm/i915: BXT DDI PHY sequence BUN

[Imre Deak]

On to, 2016-03-31 at 19:47 +0300, Kannan, Vandana wrote:
> > -----Original Message-----
> > From: Deak, Imre
> > Sent: Thursday, March 31, 2016 9:45 AM
> > To: Kannan, Vandana <vandana.kannan at intel.com>; intel-
> > gfx at lists.freedesktop.org
> > Cc: Nikula, Jani <jani.nikula at intel.com>
> > Subject: Re: [PATCH v2] drm/i915: BXT DDI PHY sequence BUN
> > 
> > On to, 2016-03-31 at 22:40 +0530, Vandana Kannan wrote:
> > > According to the BSpec update, bit 7 of PORT_CL1CM_DW0 register
> > needs
> > > to be checked to ensure that the register is in accessible state.
> > > Also, based on a BSpec update, changing the timeout value to check
> > > iphypwrgood, from 10ms to wait for up to 100us.
> > > 
> > > v2: use wait_for_us instead of the atomic call.
> > > 
> > > Signed-off-by: Vandana Kannan <vandana.kannan at intel.com>
> > > Reported-by: Philippe Lecluse <Philippe.Lecluse at intel.com>
> > > Cc: Deak, Imre <imre.deak at intel.com>
> > > Cc: Nikula, Jani <jani.nikula at intel.com>
> > > ---
> > >  drivers/gpu/drm/i915/i915_reg.h  |  1 +
> > >  drivers/gpu/drm/i915/intel_ddi.c | 15 +++++++++++++--
> > >  2 files changed, 14 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/i915_reg.h
> > > b/drivers/gpu/drm/i915/i915_reg.h index 7dfc400..9a02bfc 100644
> > > --- a/drivers/gpu/drm/i915/i915_reg.h
> > > +++ b/drivers/gpu/drm/i915/i915_reg.h
> > > @@ -1318,6 +1318,7 @@ enum skl_disp_power_wells {
> > >  #define _PORT_CL1CM_DW0_A		0x162000
> > >  #define _PORT_CL1CM_DW0_BC		0x6C000
> > >  #define   PHY_POWER_GOOD		(1 << 16)
> > > +#define   PHY_RESERVED			(1 << 7)
> > >  #define BXT_PORT_CL1CM_DW0(phy)		_BXT_PHY((phy),
> > _PORT_CL1CM_DW0_BC, \
> > >  							_PORT_CL1CM_DW0_A)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/intel_ddi.c
> > > b/drivers/gpu/drm/i915/intel_ddi.c
> > > index 62de9f4..05a5b3a 100644
> > > --- a/drivers/gpu/drm/i915/intel_ddi.c
> > > +++ b/drivers/gpu/drm/i915/intel_ddi.c
> > > @@ -2669,9 +2669,20 @@ static void broxton_phy_init(struct
> > drm_i915_private *dev_priv,
> > >  	val |= GT_DISPLAY_POWER_ON(phy);
> > >  	I915_WRITE(BXT_P_CR_GT_DISP_PWRON, val);
> > > 
> > > -	/* Considering 10ms timeout until BSpec is updated */
> > > -	if (wait_for(I915_READ(BXT_PORT_CL1CM_DW0(phy)) &
> > PHY_POWER_GOOD, 10))
> > > +	/*
> > > +	 * The PHY registers start out inaccessible and respond to reads
> > with
> > > +	 * all 1s.  Eventually they become accessible as they power up, then
> > > +	 * the reserved bit will give the default 0.  Poll on the reserved bit
> > > +	 * becoming 0 to find when the PHY is accessible.
> > > +	 * HW team confirmed that the time to reach phypowergood status
> > is
> > > +	 * anywhere between 50 us and 100us.
> > > +	 */
> > > +	if (wait_for_us(((!(I915_READ(BXT_PORT_CL1CM_DW0(phy)) &
> > > +				PHY_RESERVED)) &&
> > > +				((I915_READ(BXT_PORT_CL1CM_DW0(phy)) &
> > > +				  PHY_POWER_GOOD) ==
> > PHY_POWER_GOOD)), 100)) {
> > >  		DRM_ERROR("timeout during PHY%d power on\n", phy);
> > 
> > Thanks. This is still missing Jani's comment, that is reading out the register
> > only once and applying a mask before comparing against
> > PHY_POWER_GOOD.
> > 
> > --Imre
> > 
> [Vandana] 
> Oh! Thought he said its fine on an updated mail.

Hm, if you check below,

> 
> "
> > > +	if (wait_for_atomic_us(((!(I915_READ(BXT_PORT_CL1CM_DW0(phy)) &
> > > +				PHY_RESERVED)) &&
> > > +				((I915_READ(BXT_PORT_CL1CM_DW0(phy)) &
> > > +				  PHY_POWER_GOOD) == PHY_POWER_GOOD)), 100)) {
> > 
> > Is there any reason why you'd need to do the read twice? Why not just 
> > write it as:
> > 
> > (I915_READ(BXT_PORT_CL1CM_DW0(phy)) & (PHY_RESERVED | PHY_POWER_GOOD)) 
> > == PHY_POWER_GOOD
> 
> AFAICT this should be fine.

This just means what I also asked. I'd prefer this, since reading out the
register twice is redundant and makes the reader wonder why it was needed
in the first place.

--Imre