[Intel-gfx] [PATCH v3 02/15] drm/i915/rkl: Program BW_BUDDY0 registers instead of BW_BUDDY1/2
Aditya Swarup
aditya.swarup at intel.com
Thu Jun 4 01:18:02 UTC 2020
On Wed, Jun 03, 2020 at 04:12:59PM -0700, Matt Roper wrote:
> On Wed, Jun 03, 2020 at 03:34:32PM -0700, Aditya Swarup wrote:
> > On Wed, Jun 03, 2020 at 02:15:16PM -0700, Matt Roper wrote:
> > > RKL uses the same BW_BUDDY programming table as TGL, but programs the
> > > values into a single set BUDDY0 set of registers rather than the
> > > BUDDY1/BUDDY2 sets used by TGL.
> > >
> > > Bspec: 49218
> > > Cc: Aditya Swarup <aditya.swarup at intel.com>
> > > Signed-off-by: Matt Roper <matthew.d.roper at intel.com>
> > > ---
> > > .../drm/i915/display/intel_display_power.c | 44 +++++++++++--------
> > > drivers/gpu/drm/i915/i915_reg.h | 14 ++++--
> > > 2 files changed, 35 insertions(+), 23 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c
> > > index 72312b67b57a..2c1ce50b572b 100644
> > > --- a/drivers/gpu/drm/i915/display/intel_display_power.c
> > > +++ b/drivers/gpu/drm/i915/display/intel_display_power.c
> > > @@ -5254,7 +5254,7 @@ static void tgl_bw_buddy_init(struct drm_i915_private *dev_priv)
> > > enum intel_dram_type type = dev_priv->dram_info.type;
> > > u8 num_channels = dev_priv->dram_info.num_channels;
> > > const struct buddy_page_mask *table;
> > > - int i;
> > > + int config, min_buddy, max_buddy, i;
> > >
> > > if (IS_TGL_REVID(dev_priv, TGL_REVID_A0, TGL_REVID_B0))
> > > /* Wa_1409767108: tgl */
> > > @@ -5262,29 +5262,35 @@ static void tgl_bw_buddy_init(struct drm_i915_private *dev_priv)
> > > else
> > > table = tgl_buddy_page_masks;
> > >
> > > - for (i = 0; table[i].page_mask != 0; i++)
> > > - if (table[i].num_channels == num_channels &&
> > > - table[i].type == type)
> > > + if (IS_ROCKETLAKE(dev_priv)) {
> > > + min_buddy = max_buddy = 0;
> > > + } else {
> > > + min_buddy = 1;
> > > + max_buddy = 2;
> > > + }
> > > +
> > > + for (config = 0; table[config].page_mask != 0; config++)
> > > + if (table[config].num_channels == num_channels &&
> > > + table[config].type == type)
> > > break;
> > >
> > > - if (table[i].page_mask == 0) {
> > > + if (table[config].page_mask == 0) {
> > > drm_dbg(&dev_priv->drm,
> > > "Unknown memory configuration; disabling address buddy logic.\n");
> > > - intel_de_write(dev_priv, BW_BUDDY1_CTL, BW_BUDDY_DISABLE);
> > > - intel_de_write(dev_priv, BW_BUDDY2_CTL, BW_BUDDY_DISABLE);
> > > + for (i = min_buddy; i <= max_buddy; i++)
> > > + intel_de_write(dev_priv, BW_BUDDY_CTL(i),
> > > + BW_BUDDY_DISABLE);
> > > } else {
> > > - intel_de_write(dev_priv, BW_BUDDY1_PAGE_MASK,
> > > - table[i].page_mask);
> > > - intel_de_write(dev_priv, BW_BUDDY2_PAGE_MASK,
> > > - table[i].page_mask);
> > > -
> > > - /* Wa_22010178259:tgl */
> > > - intel_de_rmw(dev_priv, BW_BUDDY1_CTL,
> > > - BW_BUDDY_TLB_REQ_TIMER_MASK,
> > > - REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK, 0x8));
> > > - intel_de_rmw(dev_priv, BW_BUDDY2_CTL,
> > > - BW_BUDDY_TLB_REQ_TIMER_MASK,
> > > - REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK, 0x8));
> > > + for (i = min_buddy; i <= max_buddy; i++) {
> > > + intel_de_write(dev_priv, BW_BUDDY_PAGE_MASK(i),
> > > + table[config].page_mask);
> > > +
> > > + /* Wa_22010178259:tgl,rkl */
> > > + intel_de_rmw(dev_priv, BW_BUDDY_CTL(i),
> > > + BW_BUDDY_TLB_REQ_TIMER_MASK,
> > > + REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK,
> > > + 0x8));
> > We should be using REG_FIELD_PREP() in i915_reg.h to declare
> > TLB_REQ_TIMER value and then use the value here.
>
> Any specific reason why? The value "8" doesn't have any specific
> hardware meaning that would be meaningful to define in the general
> register definitions. It's just a value that this specific hardware
> workaround asked for in this case. I'm not sure if we want to spread
> the definition of the workaround into the register file if the value
> isn't going to be meaningful to other driver programming or workarounds.
>
>
> Matt
The value 8 is constant and as such should be defined in a header file
for that bitmask. If there was variable used to prepare the value on the
fly, I would have understood this usage.
If you are concerned about 8 being a random value and not spreading to
i915_reg.h, I would prefer a macro in i915_reg.h like:
#define TLB_REQ_TIMER(x) REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK,(x))
Then the value doesn't filter to i915_reg.h and according to me looks
cleaner.
Most of the usage of REG_FIELD_PREP in *.c files is based on creating
bitfields using variable. Here we are using a constant which can easily
be moved to i915_reg.h. It shouldn't matter if it is a WA or not.
Adi
>
> > > + }
> > > }
> > > }
> > >
> > > diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
> > > index 578cfe11cbb9..3e79cefc510a 100644
> > > --- a/drivers/gpu/drm/i915/i915_reg.h
> > > +++ b/drivers/gpu/drm/i915/i915_reg.h
> > > @@ -7837,13 +7837,19 @@ enum {
> > > #define WAIT_FOR_PCH_RESET_ACK (1 << 1)
> > > #define WAIT_FOR_PCH_FLR_ACK (1 << 0)
> > >
> > > -#define BW_BUDDY1_CTL _MMIO(0x45140)
> > > -#define BW_BUDDY2_CTL _MMIO(0x45150)
> > > +#define _BW_BUDDY0_CTL 0x45130
> > > +#define _BW_BUDDY1_CTL 0x45140
> > > +#define BW_BUDDY_CTL(x) _MMIO(_PICK_EVEN(x, \
> > > + _BW_BUDDY0_CTL, \
> > > + _BW_BUDDY1_CTL))
> > > #define BW_BUDDY_DISABLE REG_BIT(31)
> > > #define BW_BUDDY_TLB_REQ_TIMER_MASK REG_GENMASK(21, 16)
> > >
> > > -#define BW_BUDDY1_PAGE_MASK _MMIO(0x45144)
> > > -#define BW_BUDDY2_PAGE_MASK _MMIO(0x45154)
> > > +#define _BW_BUDDY0_PAGE_MASK 0x45134
> > > +#define _BW_BUDDY1_PAGE_MASK 0x45144
> > > +#define BW_BUDDY_PAGE_MASK(x) _MMIO(_PICK_EVEN(x, \
> > > + _BW_BUDDY0_PAGE_MASK, \
> > > + _BW_BUDDY1_PAGE_MASK))
> > >
> > > #define HSW_NDE_RSTWRN_OPT _MMIO(0x46408)
> > > #define RESET_PCH_HANDSHAKE_ENABLE (1 << 4)
> > > --
> > > 2.24.1
> > >
>
> --
> Matt Roper
> Graphics Software Engineer
> VTT-OSGC Platform Enablement
> Intel Corporation
> (916) 356-2795
More information about the Intel-gfx
mailing list