[Intel-gfx] [PATCH v2 5/7] drm/i915: sseu: Convert subslice count fields to subslice mask
Imre Deak
imre.deak at intel.com
Wed Aug 31 16:13:05 UTC 2016
In an upcoming patch we'll need the actual mask of subslices in addition
to their count, so convert the subslice_per_slice field to a mask.
Also we can easily calculate subslice_total from the other fields, so
instead of storing a cached version of this, add a helper to calculate
it.
v2:
- Use hweight8() on u8 typed vars instead of hweight32(). (Ben)
Reviewed-by: Robert Bragg <robert at sixbynine.org> (v1)
Reviewed-by: Ben Widawsky <benjamin.widawsky at intel.com> (v1)
Tested-by: Ben Widawsky <benjamin.widawsky at intel.com> (v1)
Signed-off-by: Imre Deak <imre.deak at intel.com>
---
drivers/gpu/drm/i915/i915_debugfs.c | 37 +++++++----------
drivers/gpu/drm/i915/i915_drv.c | 2 +-
drivers/gpu/drm/i915/i915_drv.h | 8 +++-
drivers/gpu/drm/i915/intel_device_info.c | 69 +++++++++++++++-----------------
drivers/gpu/drm/i915/intel_lrc.c | 2 +-
5 files changed, 55 insertions(+), 63 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 0af047d..e4a0495 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -5025,7 +5025,7 @@ static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv,
continue;
sseu->slice_mask = BIT(0);
- sseu->subslice_per_slice++;
+ sseu->subslice_mask |= BIT(ss);
eu_cnt = ((sig1[ss] & CHV_EU08_PG_ENABLE) ? 0 : 2) +
((sig1[ss] & CHV_EU19_PG_ENABLE) ? 0 : 2) +
((sig1[ss] & CHV_EU210_PG_ENABLE) ? 0 : 2) +
@@ -5034,7 +5034,6 @@ static void cherryview_sseu_device_status(struct drm_i915_private *dev_priv,
sseu->eu_per_subslice = max_t(unsigned int,
sseu->eu_per_subslice, eu_cnt);
}
- sseu->subslice_total = sseu->subslice_per_slice;
}
static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
@@ -5066,8 +5065,6 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
GEN9_PGCTL_SSB_EU311_ACK;
for (s = 0; s < s_max; s++) {
- unsigned int ss_cnt = 0;
-
if ((s_reg[s] & GEN9_PGCTL_SLICE_ACK) == 0)
/* skip disabled slice */
continue;
@@ -5075,18 +5072,19 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
sseu->slice_mask |= BIT(s);
if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))
- ss_cnt = INTEL_INFO(dev_priv)->sseu.subslice_per_slice;
+ sseu->subslice_mask =
+ INTEL_INFO(dev_priv)->sseu.subslice_mask;
for (ss = 0; ss < ss_max; ss++) {
unsigned int eu_cnt;
- if (IS_BROXTON(dev_priv) &&
- !(s_reg[s] & (GEN9_PGCTL_SS_ACK(ss))))
- /* skip disabled subslice */
- continue;
+ if (IS_BROXTON(dev_priv)) {
+ if (!(s_reg[s] & (GEN9_PGCTL_SS_ACK(ss))))
+ /* skip disabled subslice */
+ continue;
- if (IS_BROXTON(dev_priv))
- ss_cnt++;
+ sseu->subslice_mask |= BIT(ss);
+ }
eu_cnt = 2 * hweight32(eu_reg[2*s + ss/2] &
eu_mask[ss%2]);
@@ -5095,11 +5093,6 @@ static void gen9_sseu_device_status(struct drm_i915_private *dev_priv,
sseu->eu_per_subslice,
eu_cnt);
}
-
- sseu->subslice_total += ss_cnt;
- sseu->subslice_per_slice = max_t(unsigned int,
- sseu->subslice_per_slice,
- ss_cnt);
}
}
@@ -5112,13 +5105,11 @@ static void broadwell_sseu_device_status(struct drm_i915_private *dev_priv,
sseu->slice_mask = slice_info & GEN8_LSLICESTAT_MASK;
if (sseu->slice_mask) {
- sseu->subslice_per_slice =
- INTEL_INFO(dev_priv)->sseu.subslice_per_slice;
- sseu->subslice_total = hweight8(sseu->slice_mask) *
- sseu->subslice_per_slice;
+ sseu->subslice_mask = INTEL_INFO(dev_priv)->sseu.subslice_mask;
sseu->eu_per_subslice =
INTEL_INFO(dev_priv)->sseu.eu_per_subslice;
- sseu->eu_total = sseu->eu_per_subslice * sseu->subslice_total;
+ sseu->eu_total = sseu->eu_per_subslice *
+ sseu_subslice_total(sseu);
/* subtract fused off EU(s) from enabled slice(s) */
for (s = 0; s < hweight8(sseu->slice_mask); s++) {
@@ -5139,9 +5130,9 @@ static void i915_print_sseu_info(struct seq_file *m, bool is_available_info,
seq_printf(m, " %s Slice Total: %u\n", type,
hweight8(sseu->slice_mask));
seq_printf(m, " %s Subslice Total: %u\n", type,
- sseu->subslice_total);
+ sseu_subslice_total(sseu));
seq_printf(m, " %s Subslice Per Slice: %u\n", type,
- sseu->subslice_per_slice);
+ hweight8(sseu->subslice_mask));
seq_printf(m, " %s EU Total: %u\n", type,
sseu->eu_total);
seq_printf(m, " %s EU Per Subslice: %u\n", type,
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 3c417b2..978107c 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -331,7 +331,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
value = 1;
break;
case I915_PARAM_SUBSLICE_TOTAL:
- value = INTEL_INFO(dev)->sseu.subslice_total;
+ value = sseu_subslice_total(&INTEL_INFO(dev_priv)->sseu);
if (!value)
return -ENODEV;
break;
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 9525762..52d7092 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -673,8 +673,7 @@ struct intel_csr {
struct sseu_dev_info {
u8 slice_mask;
- u8 subslice_total;
- u8 subslice_per_slice;
+ u8 subslice_mask;
u8 eu_total;
u8 eu_per_subslice;
u8 min_eu_in_pool;
@@ -685,6 +684,11 @@ struct sseu_dev_info {
u8 has_eu_pg:1;
};
+static inline unsigned int sseu_subslice_total(const struct sseu_dev_info *sseu)
+{
+ return hweight8(sseu->slice_mask) * hweight8(sseu->subslice_mask);
+}
+
struct intel_device_info {
u32 display_mmio_offset;
u16 device_id;
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
index 139d529..0bdec5c 100644
--- a/drivers/gpu/drm/i915/intel_device_info.c
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -54,26 +54,25 @@ static void cherryview_sseu_info_init(struct drm_i915_private *dev_priv)
sseu->slice_mask = BIT(0);
if (!(fuse & CHV_FGT_DISABLE_SS0)) {
- sseu->subslice_per_slice++;
+ sseu->subslice_mask |= BIT(0);
eu_dis = fuse & (CHV_FGT_EU_DIS_SS0_R0_MASK |
CHV_FGT_EU_DIS_SS0_R1_MASK);
sseu->eu_total += 8 - hweight32(eu_dis);
}
if (!(fuse & CHV_FGT_DISABLE_SS1)) {
- sseu->subslice_per_slice++;
+ sseu->subslice_mask |= BIT(1);
eu_dis = fuse & (CHV_FGT_EU_DIS_SS1_R0_MASK |
CHV_FGT_EU_DIS_SS1_R1_MASK);
sseu->eu_total += 8 - hweight32(eu_dis);
}
- sseu->subslice_total = sseu->subslice_per_slice;
/*
* CHV expected to always have a uniform distribution of EU
* across subslices.
*/
- sseu->eu_per_subslice = sseu->subslice_total ?
- sseu->eu_total / sseu->subslice_total :
+ sseu->eu_per_subslice = sseu_subslice_total(sseu) ?
+ sseu->eu_total / sseu_subslice_total(sseu) :
0;
/*
* CHV supports subslice power gating on devices with more than
@@ -81,7 +80,7 @@ static void cherryview_sseu_info_init(struct drm_i915_private *dev_priv)
* more than one EU pair per subslice.
*/
sseu->has_slice_pg = 0;
- sseu->has_subslice_pg = (sseu->subslice_total > 1);
+ sseu->has_subslice_pg = sseu_subslice_total(sseu) > 1;
sseu->has_eu_pg = (sseu->eu_per_subslice > 2);
}
@@ -91,20 +90,19 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
struct sseu_dev_info *sseu = &info->sseu;
int s_max = 3, ss_max = 4, eu_max = 8;
int s, ss;
- u32 fuse2, ss_disable, eu_disable;
+ u32 fuse2, eu_disable;
u8 eu_mask = 0xff;
fuse2 = I915_READ(GEN8_FUSE2);
sseu->slice_mask = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
- ss_disable = (fuse2 & GEN9_F2_SS_DIS_MASK) >> GEN9_F2_SS_DIS_SHIFT;
/*
* The subslice disable field is global, i.e. it applies
* to each of the enabled slices.
*/
- sseu->subslice_per_slice = ss_max - hweight32(ss_disable);
- sseu->subslice_total = hweight8(sseu->slice_mask) *
- sseu->subslice_per_slice;
+ sseu->subslice_mask = (1 << ss_max) - 1;
+ sseu->subslice_mask &= ~((fuse2 & GEN9_F2_SS_DIS_MASK) >>
+ GEN9_F2_SS_DIS_SHIFT);
/*
* Iterate through enabled slices and subslices to
@@ -119,7 +117,7 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
for (ss = 0; ss < ss_max; ss++) {
int eu_per_ss;
- if (ss_disable & BIT(ss))
+ if (!(sseu->subslice_mask & BIT(ss)))
/* skip disabled subslice */
continue;
@@ -145,9 +143,9 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
* recovery. BXT is expected to be perfectly uniform in EU
* distribution.
*/
- sseu->eu_per_subslice = sseu->subslice_total ?
+ sseu->eu_per_subslice = sseu_subslice_total(sseu) ?
DIV_ROUND_UP(sseu->eu_total,
- sseu->subslice_total) : 0;
+ sseu_subslice_total(sseu)) : 0;
/*
* SKL supports slice power gating on devices with more than
* one slice, and supports EU power gating on devices with
@@ -160,11 +158,11 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
(IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) &&
hweight8(sseu->slice_mask) > 1;
sseu->has_subslice_pg =
- IS_BROXTON(dev_priv) && sseu->subslice_total > 1;
+ IS_BROXTON(dev_priv) && sseu_subslice_total(sseu) > 1;
sseu->has_eu_pg = sseu->eu_per_subslice > 2;
if (IS_BROXTON(dev_priv)) {
-#define IS_SS_DISABLED(_ss_disable, ss) (_ss_disable & BIT(ss))
+#define IS_SS_DISABLED(ss) (!(sseu->subslice_mask & BIT(ss)))
/*
* There is a HW issue in 2x6 fused down parts that requires
* Pooled EU to be enabled as a WA. The pool configuration
@@ -172,16 +170,15 @@ static void gen9_sseu_info_init(struct drm_i915_private *dev_priv)
* doesn't affect if the device has all 3 subslices enabled.
*/
/* WaEnablePooledEuFor2x6:bxt */
- info->has_pooled_eu = ((sseu->subslice_per_slice == 3) ||
- (sseu->subslice_per_slice == 2 &&
+ info->has_pooled_eu = ((hweight8(sseu->subslice_mask) == 3) ||
+ (hweight8(sseu->subslice_mask) == 2 &&
INTEL_REVID(dev_priv) < BXT_REVID_C0));
sseu->min_eu_in_pool = 0;
if (info->has_pooled_eu) {
- if (IS_SS_DISABLED(ss_disable, 0) ||
- IS_SS_DISABLED(ss_disable, 2))
+ if (IS_SS_DISABLED(2) || IS_SS_DISABLED(0))
sseu->min_eu_in_pool = 3;
- else if (IS_SS_DISABLED(ss_disable, 1))
+ else if (IS_SS_DISABLED(1))
sseu->min_eu_in_pool = 6;
else
sseu->min_eu_in_pool = 9;
@@ -195,11 +192,17 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
struct sseu_dev_info *sseu = &mkwrite_device_info(dev_priv)->sseu;
const int s_max = 3, ss_max = 3, eu_max = 8;
int s, ss;
- u32 fuse2, eu_disable[s_max], ss_disable;
+ u32 fuse2, eu_disable[s_max];
fuse2 = I915_READ(GEN8_FUSE2);
sseu->slice_mask = (fuse2 & GEN8_F2_S_ENA_MASK) >> GEN8_F2_S_ENA_SHIFT;
- ss_disable = (fuse2 & GEN8_F2_SS_DIS_MASK) >> GEN8_F2_SS_DIS_SHIFT;
+ /*
+ * The subslice disable field is global, i.e. it applies
+ * to each of the enabled slices.
+ */
+ sseu->subslice_mask = BIT(ss_max) - 1;
+ sseu->subslice_mask &= ~((fuse2 & GEN8_F2_SS_DIS_MASK) >>
+ GEN8_F2_SS_DIS_SHIFT);
eu_disable[0] = I915_READ(GEN8_EU_DISABLE0) & GEN8_EU_DIS0_S0_MASK;
eu_disable[1] = (I915_READ(GEN8_EU_DISABLE0) >> GEN8_EU_DIS0_S1_SHIFT) |
@@ -210,14 +213,6 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
(32 - GEN8_EU_DIS1_S2_SHIFT));
/*
- * The subslice disable field is global, i.e. it applies
- * to each of the enabled slices.
- */
- sseu->subslice_per_slice = ss_max - hweight32(ss_disable);
- sseu->subslice_total = hweight8(sseu->slice_mask) *
- sseu->subslice_per_slice;
-
- /*
* Iterate through enabled slices and subslices to
* count the total enabled EU.
*/
@@ -229,7 +224,7 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
for (ss = 0; ss < ss_max; ss++) {
u32 n_disabled;
- if (ss_disable & (0x1 << ss))
+ if (!(sseu->subslice_mask & BIT(ss)))
/* skip disabled subslice */
continue;
@@ -250,8 +245,9 @@ static void broadwell_sseu_info_init(struct drm_i915_private *dev_priv)
* subslices with the exception that any one EU in any one subslice may
* be fused off for die recovery.
*/
- sseu->eu_per_subslice = sseu->subslice_total ?
- DIV_ROUND_UP(sseu->eu_total, sseu->subslice_total) : 0;
+ sseu->eu_per_subslice = sseu_subslice_total(sseu) ?
+ DIV_ROUND_UP(sseu->eu_total,
+ sseu_subslice_total(sseu)) : 0;
/*
* BDW supports slice power gating on devices with more than
@@ -375,9 +371,10 @@ void intel_device_info_runtime_init(struct drm_i915_private *dev_priv)
info->has_snoop = false;
DRM_DEBUG_DRIVER("slice total: %u\n", hweight8(info->sseu.slice_mask));
- DRM_DEBUG_DRIVER("subslice total: %u\n", info->sseu.subslice_total);
+ DRM_DEBUG_DRIVER("subslice total: %u\n",
+ sseu_subslice_total(&info->sseu));
DRM_DEBUG_DRIVER("subslice per slice: %u\n",
- info->sseu.subslice_per_slice);
+ hweight8(info->sseu.subslice_mask));
DRM_DEBUG_DRIVER("EU total: %u\n", info->sseu.eu_total);
DRM_DEBUG_DRIVER("EU per subslice: %u\n", info->sseu.eu_per_subslice);
DRM_DEBUG_DRIVER("has slice power gating: %s\n",
diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c
index 7209553..92bfe47 100644
--- a/drivers/gpu/drm/i915/intel_lrc.c
+++ b/drivers/gpu/drm/i915/intel_lrc.c
@@ -1905,7 +1905,7 @@ make_rpcs(struct drm_i915_private *dev_priv)
if (INTEL_INFO(dev_priv)->sseu.has_subslice_pg) {
rpcs |= GEN8_RPCS_SS_CNT_ENABLE;
- rpcs |= INTEL_INFO(dev_priv)->sseu.subslice_per_slice <<
+ rpcs |= hweight8(INTEL_INFO(dev_priv)->sseu.subslice_mask) <<
GEN8_RPCS_SS_CNT_SHIFT;
rpcs |= GEN8_RPCS_ENABLE;
}
--
2.5.0
More information about the Intel-gfx
mailing list