Skip to contentSkip to conversation

drm/i915: Move a bunch of workaround-related code to its own file

PATCH In review Oscar Mateo 3 participants 7 messages Apr 10, 2018 – Apr 12, 2018

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH 1/2 drm/i915: Move a bunch of workaround-related code to its own file

This has grown to be a sizable amount of code, so move it to its own file before we try to refactor anything. For the moment, we are leaving behind the WA BB code and the WAs that get applied (incorrectly) in init_clock_gating, but we will deal with it later.

v2: Use intel_ prefix for code that deals with the hardware (Chris)
v3: Rebased
v4:
- Rebased
- New license header
v5:
- Rebased
- Added some organisational notes to the file (Chris)
v6: Include DOC section in the documentation build (Jani)

---
 Documentation/gpu/i915.rst               |   6 +
 drivers/gpu/drm/i915/Makefile            |   3 +-
 drivers/gpu/drm/i915/intel_engine_cs.c   | 634 ----------------------------
 drivers/gpu/drm/i915/intel_lrc.c         |   1 +
 drivers/gpu/drm/i915/intel_ringbuffer.c  |   1 +
 drivers/gpu/drm/i915/intel_ringbuffer.h  |   3 -
 drivers/gpu/drm/i915/intel_workarounds.c | 684 +++++++++++++++++++++++++++++++
 drivers/gpu/drm/i915/intel_workarounds.h |  13 +
 8 files changed, 707 insertions(+), 638 deletions(-)
 create mode 100644 drivers/gpu/drm/i915/intel_workarounds.c
 create mode 100644 drivers/gpu/drm/i915/intel_workarounds.h
Documentation/gpu/i915.rst +6 -0
index 34d22f2..055df45 100644
@@ -58,6 +58,12 @@ Intel GVT-g Host Support(vGPU device model)
 .. kernel-doc:: drivers/gpu/drm/i915/intel_gvt.c
    :internal:
 
+Workarounds
+-----------
+
+.. kernel-doc:: drivers/gpu/drm/i915/intel_workarounds.c
+   :doc: Hardware workarounds
+
 Display Hardware Handling
 =========================
 
drivers/gpu/drm/i915/Makefile +2 -1
index 0c79c19..9bee52a9 100644
@@ -43,7 +43,8 @@ i915-y := i915_drv.o \
 	  intel_csr.o \
 	  intel_device_info.o \
 	  intel_pm.o \
-	  intel_runtime_pm.o
+	  intel_runtime_pm.o \
+	  intel_workarounds.o
 
 i915-$(CONFIG_COMPAT)   += i915_ioc32.o
 i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o intel_pipe_crc.o
drivers/gpu/drm/i915/intel_engine_cs.c +0 -634

640 lines not on this page · raw message

drivers/gpu/drm/i915/intel_lrc.c +1 -0
index 02b25bf..c250c8b 100644
@@ -139,6 +139,7 @@
 #include "i915_gem_render_state.h"
 #include "intel_lrc_reg.h"
 #include "intel_mocs.h"
+#include "intel_workarounds.h"
 
 #define RING_EXECLIST_QFULL		(1 << 0x2)
 #define RING_EXECLIST1_VALID		(1 << 0x3)
drivers/gpu/drm/i915/intel_ringbuffer.c +1 -0
index 04d9d9a..36acc32 100644
@@ -36,6 +36,7 @@
 #include "i915_gem_render_state.h"
 #include "i915_trace.h"
 #include "intel_drv.h"
+#include "intel_workarounds.h"
 
 /* Rough estimate of the typical request size, performing a flush,
  * set-context and then emitting the batch.
drivers/gpu/drm/i915/intel_ringbuffer.h +0 -3
index 256d584..7170416 100644
@@ -885,9 +885,6 @@ static inline u32 intel_engine_last_submit(struct intel_engine_cs *engine)
 	return READ_ONCE(engine->timeline->seqno);
 }
 
-int init_workarounds_ring(struct intel_engine_cs *engine);
-int intel_ring_workarounds_emit(struct i915_request *rq);
-
 void intel_engine_get_instdone(struct intel_engine_cs *engine,
 			       struct intel_instdone *instdone);
 
drivers/gpu/drm/i915/intel_workarounds.c new file +684 -0

685 lines not on this page · raw message

drivers/gpu/drm/i915/intel_workarounds.h new file +13 -0
new file mode 100644
index 0000000..2afea73
@@ -0,0 +1,13 @@
+/*
+ * SPDX-License-Identifier: MIT
+ *
+ * Copyright © 2014-2018 Intel Corporation
+ */
+
+#ifndef _I915_WORKAROUNDS_H_
+#define _I915_WORKAROUNDS_H_
+
+int init_workarounds_ring(struct intel_engine_cs *engine);
+int intel_ring_workarounds_emit(struct i915_request *rq);
+
+#endif
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw

PATCH 2/2 drm/i915: Split out functions for different kinds of workarounds

There are different kind of workarounds (those that modify registers that live in the context image, those that modify global registers, those that whitelist registers, etc...) and they have different requirements in terms of where they are applied and how. Also, by splitting them apart, it should be easier to decide where a new workaround should go.

v2:
- Add multiple MISSING_CASE
- Rebased

v3:
- Rename mmio_workarounds to gt_workarounds (Chris, Mika)
- Create empty placeholders for BDW and CHV GT WAs
- Rebased

v4: Rebased

v5:
- Rebased
- FORCE_TO_NONPRIV register exists since BDW, so make a path for it to achieve universality, even if empty (Chris)

---
 drivers/gpu/drm/i915/i915_gem.c          |   3 +
 drivers/gpu/drm/i915/i915_gem_context.c  |   6 +
 drivers/gpu/drm/i915/intel_lrc.c         |  14 +-
 drivers/gpu/drm/i915/intel_ringbuffer.c  |   8 +-
 drivers/gpu/drm/i915/intel_workarounds.c | 636 +++++++++++++++++++------------
 drivers/gpu/drm/i915/intel_workarounds.h |   8 +-
 6 files changed, 433 insertions(+), 242 deletions(-)
drivers/gpu/drm/i915/i915_gem.c +3 -0
index 28ab0be..ab846f8 100644
@@ -35,6 +35,7 @@
 #include "intel_drv.h"
 #include "intel_frontbuffer.h"
 #include "intel_mocs.h"
+#include "intel_workarounds.h"
 #include "i915_gemfs.h"
 #include <linux/dma-fence-array.h>
 #include <linux/kthread.h>
@@ -5198,6 +5199,8 @@ int i915_gem_init_hw(struct drm_i915_private *dev_priv)
 		}
 	}
 
+	intel_gt_workarounds_apply(dev_priv);
+
 	i915_gem_init_swizzling(dev_priv);
 
 	/*
drivers/gpu/drm/i915/i915_gem_context.c +6 -0
index 5cfac02..9b3834a 100644
@@ -90,6 +90,7 @@
 #include <drm/i915_drm.h>
 #include "i915_drv.h"
 #include "i915_trace.h"
+#include "intel_workarounds.h"
 
 #define ALL_L3_SLICES(dev) (1 << NUM_L3_SLICES(dev)) - 1
 
@@ -459,11 +460,16 @@ static bool needs_preempt_context(struct drm_i915_private *i915)
 int i915_gem_contexts_init(struct drm_i915_private *dev_priv)
 {
 	struct i915_gem_context *ctx;
+	int ret;
 
 	/* Reassure ourselves we are only called once */
 	GEM_BUG_ON(dev_priv->kernel_context);
 	GEM_BUG_ON(dev_priv->preempt_context);
 
+	ret = intel_ctx_workarounds_init(dev_priv);
+	if (ret)
+		return ret;
+
 	INIT_LIST_HEAD(&dev_priv->contexts.list);
 	INIT_WORK(&dev_priv->contexts.free_work, contexts_free_worker);
 	init_llist_head(&dev_priv->contexts.free_list);
drivers/gpu/drm/i915/intel_lrc.c +11 -3
index c250c8b..568a59b 100644
@@ -1725,6 +1725,10 @@ static int gen8_init_render_ring(struct intel_engine_cs *engine)
 	if (ret)
 		return ret;
 
+	ret = intel_whitelist_workarounds_apply(engine);
+	if (ret)
+		return ret;
+
 	/* We need to disable the AsyncFlip performance optimisations in order
 	 * to use MI_WAIT_FOR_EVENT within the CS. It should already be
 	 * programmed to '1' on all products.
@@ -1735,7 +1739,7 @@ static int gen8_init_render_ring(struct intel_engine_cs *engine)
 
 	I915_WRITE(INSTPM, _MASKED_BIT_ENABLE(INSTPM_FORCE_ORDERING));
 
-	return init_workarounds_ring(engine);
+	return 0;
 }
 
 static int gen9_init_render_ring(struct intel_engine_cs *engine)
@@ -1746,7 +1750,11 @@ static int gen9_init_render_ring(struct intel_engine_cs *engine)
 	if (ret)
 		return ret;
 
-	return init_workarounds_ring(engine);
+	ret = intel_whitelist_workarounds_apply(engine);
+	if (ret)
+		return ret;
+
+	return 0;
 }
 
 static void reset_common_ring(struct intel_engine_cs *engine,
@@ -2071,7 +2079,7 @@ static int gen8_init_rcs_context(struct i915_request *rq)
 {
 	int ret;
 
-	ret = intel_ring_workarounds_emit(rq);
+	ret = intel_ctx_workarounds_emit(rq);
 	if (ret)
 		return ret;
 
drivers/gpu/drm/i915/intel_ringbuffer.c +6 -2
index 36acc32..757bb09 100644
@@ -600,7 +600,7 @@ static int intel_rcs_ctx_init(struct i915_request *rq)
 {
 	int ret;
 
-	ret = intel_ring_workarounds_emit(rq);
+	ret = intel_ctx_workarounds_emit(rq);
 	if (ret != 0)
 		return ret;
 
@@ -618,6 +618,10 @@ static int init_render_ring(struct intel_engine_cs *engine)
 	if (ret)
 		return ret;
 
+	ret = intel_whitelist_workarounds_apply(engine);
+	if (ret)
+		return ret;
+
 	/* WaTimedSingleVertexDispatch:cl,bw,ctg,elk,ilk,snb */
 	if (IS_GEN(dev_priv, 4, 6))
 		I915_WRITE(MI_MODE, _MASKED_BIT_ENABLE(VS_TIMER_DISPATCH));
@@ -659,7 +663,7 @@ static int init_render_ring(struct intel_engine_cs *engine)
 	if (INTEL_GEN(dev_priv) >= 6)
 		I915_WRITE_IMR(engine, ~engine->irq_keep_mask);
 
-	return init_workarounds_ring(engine);
+	return 0;
 }
 
 static u32 *gen6_signal(struct i915_request *rq, u32 *cs)
drivers/gpu/drm/i915/intel_workarounds.c +401 -235

824 lines not on this page · raw message

drivers/gpu/drm/i915/intel_workarounds.h +6 -2
index 2afea73..d9b0cc5 100644
@@ -7,7 +7,11 @@
 #ifndef _I915_WORKAROUNDS_H_
 #define _I915_WORKAROUNDS_H_
 
-int init_workarounds_ring(struct intel_engine_cs *engine);
-int intel_ring_workarounds_emit(struct i915_request *rq);
+int intel_ctx_workarounds_init(struct drm_i915_private *dev_priv);
+int intel_ctx_workarounds_emit(struct i915_request *rq);
+
+void intel_gt_workarounds_apply(struct drm_i915_private *dev_priv);
+
+int intel_whitelist_workarounds_apply(struct intel_engine_cs *engine);
 
 #endif
Signature · 6 lines
1.9.1

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
#2 lore raw
In reply to Oscar Mateo

Quoting Oscar Mateo (2018-04-10 17:12:46)

Quoted 15 lines

This has grown to be a sizable amount of code, so move it to its own file before we try to refactor anything. For the moment, we are leaving behind the WA BB code and the WAs that get applied (incorrectly) in init_clock_gating, but we will deal with it later.

v2: Use intel_ prefix for code that deals with the hardware (Chris)
v3: Rebased
v4:
- Rebased
- New license header
v5:
- Rebased
- Added some organisational notes to the file (Chris)
v6: Include DOC section in the documentation build (Jani)

Quoted 4 lines

I presume this impacts your work. Would you rather have this patch in place or land your w/a series first? (Assuming all goes well!) -Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#3 lore raw
Quoted 16 lines · Chris Wilson

Quoting Oscar Mateo (2018-04-10 17:12:46)

Quoted 15 lines

This has grown to be a sizable amount of code, so move it to its own file before we try to refactor anything. For the moment, we are leaving behind the WA BB code and the WAs that get applied (incorrectly) in init_clock_gating, but we will deal with it later.

v2: Use intel_ prefix for code that deals with the hardware (Chris)
v3: Rebased
v4:
- Rebased
- New license header
v5:
- Rebased
- Added some organisational notes to the file (Chris)
v6: Include DOC section in the documentation build (Jani)

Quoted 7 lines
Quoted 4 lines

I presume this impacts your work. Would you rather have this patch in place or land your w/a series first? (Assuming all goes well!) -Chris

I can do this patch first and I'll rebase the Gen11 WAs series later, no problem (in fact, this refactoring started because it was quite problematic to decide where each Gen11 WA should go).

Thanks for asking!
Oscar
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#4 lore raw

Quoting Oscar Mateo (2018-04-10 17:24:42)

Quoted 23 lines · Chris Wilson
Quoted 20 lines

Quoting Oscar Mateo (2018-04-10 17:12:46)

Quoted 19 lines

This has grown to be a sizable amount of code, so move it to its own file before we try to refactor anything. For the moment, we are leaving behind the WA BB code and the WAs that get applied (incorrectly) in init_clock_gating, but we will deal with it later.

v2: Use intel_ prefix for code that deals with the hardware (Chris)
v3: Rebased
v4:
- Rebased
- New license header
v5:
- Rebased
- Added some organisational notes to the file (Chris)
v6: Include DOC section in the documentation build (Jani)

Quoted 7 lines
Quoted 3 lines

I presume this impacts your work. Would you rather have this patch in place or land your w/a series first? (Assuming all goes well!) -Chris

I can do this patch first and I'll rebase the Gen11 WAs series later, no problem (in fact, this refactoring started because it was quite problematic to decide where each Gen11 WA should go).

I did a pass to try and make checkpatch happier and pushed. Thanks for the patches, hopefully not causing too much work for you! -Chris
_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#5 lore raw

✗ Fi.CI.CHECKPATCH: warning for series starting with [1/2] drm/i915: Move a bunch of workaround-related code to its own file

In reply to Oscar Mateo
== Series Details ==

== Series Details ==

Series: series starting with [1/2] drm/i915: Move a bunch of workaround-related code to its own file
URL : https://patchwork.freedesktop.org/series/41486/
State : warning

== Summary ==

$ dim checkpatch origin/drm-tip
8044c5a3e8e3 drm/i915: Move a bunch of workaround-related code to its own file
-:742: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating?
#742: 
new file mode 100644

-:791: ERROR:TRAILING_WHITESPACE: trailing whitespace #791: FILE: drivers/gpu/drm/i915/intel_workarounds.c:45:
+ * $

-:816: WARNING:MACRO_WITH_FLOW_CONTROL: Macros with flow control statements should be avoided #816: FILE: drivers/gpu/drm/i915/intel_workarounds.c:70:
+#define WA_REG(addr, mask, val) do { \
+ const int r = wa_add(dev_priv, (addr), (mask), (val)); \
+ if (r) \
+ return r; \
+ } while (0)

-:822: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'mask' - possible side-effects? #822: FILE: drivers/gpu/drm/i915/intel_workarounds.c:76:
+#define WA_SET_BIT_MASKED(addr, mask) \
+ WA_REG(addr, (mask), _MASKED_BIT_ENABLE(mask))

-:825: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'mask' - possible side-effects? #825: FILE: drivers/gpu/drm/i915/intel_workarounds.c:79:
+#define WA_CLR_BIT_MASKED(addr, mask) \
+ WA_REG(addr, (mask), _MASKED_BIT_DISABLE(mask))

-:828: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'mask' - possible side-effects? #828: FILE: drivers/gpu/drm/i915/intel_workarounds.c:82:
+#define WA_SET_FIELD_MASKED(addr, mask, value) \
+ WA_REG(addr, mask, _MASKED_FIELD(mask, value))

-:955: WARNING:LONG_LINE: line over 100 characters #955: FILE: drivers/gpu/drm/i915/intel_workarounds.c:209:
+ I915_WRITE(GEN9_CSFE_CHICKEN1_RCS, _MASKED_BIT_ENABLE(GEN9_PREEMPT_GPGPU_SYNC_SWITCH_DISABLE));

-:1246: ERROR:SPACING: spaces required around that '=' (ctx:VxW) #1246: FILE: drivers/gpu/drm/i915/intel_workarounds.c:500:
+ ret= wa_ring_whitelist_reg(engine, GEN8_CS_CHICKEN1);
^

-:1290: CHECK:OPEN_ENDED_LINE: Lines should not end with a '(' #1290: FILE: drivers/gpu/drm/i915/intel_workarounds.c:544:
+ WA_SET_BIT_MASKED(

-:1350: CHECK:OPEN_ENDED_LINE: Lines should not end with a '(' #1350: FILE: drivers/gpu/drm/i915/intel_workarounds.c:604:
+ WA_SET_BIT_MASKED(

-:1430: WARNING:MISSING_EOF_NEWLINE: adding a line without newline at end of file #1430: FILE: drivers/gpu/drm/i915/intel_workarounds.c:684:
+}

total: 2 errors, 4 warnings, 5 checks, 1381 lines checked e9f5cebfbf56 drm/i915: Split out functions for different kinds of workarounds -:409: CHECK:OPEN_ENDED_LINE: Lines should not end with a '(' #409: FILE: drivers/gpu/drm/i915/intel_workarounds.c:364:
+ WA_SET_BIT_MASKED(

-:456: CHECK:OPEN_ENDED_LINE: Lines should not end with a '(' #456: FILE: drivers/gpu/drm/i915/intel_workarounds.c:399:
+ WA_SET_BIT_MASKED(

-:512: CHECK:BRACES: braces {} should be used on all arms of this statement #512: FILE: drivers/gpu/drm/i915/intel_workarounds.c:450:
+ if (INTEL_GEN(dev_priv) < 8)
[...]
+ else if (IS_BROADWELL(dev_priv))
[...]
+ else if (IS_CHERRYVIEW(dev_priv))
[...]
+ else if (IS_SKYLAKE(dev_priv))
[...]
+ else if (IS_BROXTON(dev_priv))
[...]
+ else if (IS_KABYLAKE(dev_priv))
[...]
+ else if (IS_GEMINILAKE(dev_priv))
[...]
+ else if (IS_COFFEELAKE(dev_priv))
[...]
+ else if (IS_CANNONLAKE(dev_priv))
[...]
+ else {
[...]

-:530: CHECK:BRACES: Unbalanced braces around else statement #530: FILE: drivers/gpu/drm/i915/intel_workarounds.c:468:
+ else {

-:822: WARNING:LINE_SPACING: Missing a blank line after declarations #822: FILE: drivers/gpu/drm/i915/intel_workarounds.c:744:
+ int ret = gen9_whitelist_workarounds_apply(engine);
+ if (ret)

-:851: WARNING:LINE_SPACING: Missing a blank line after declarations #851: FILE: drivers/gpu/drm/i915/intel_workarounds.c:767:
+ int ret = gen9_whitelist_workarounds_apply(engine);
+ if (ret)

-:868: WARNING:LINE_SPACING: Missing a blank line after declarations #868: FILE: drivers/gpu/drm/i915/intel_workarounds.c:781:
+ int ret = gen9_whitelist_workarounds_apply(engine);
+ if (ret)

-:889: WARNING:LINE_SPACING: Missing a blank line after declarations #889: FILE: drivers/gpu/drm/i915/intel_workarounds.c:795:
+ int ret = gen9_whitelist_workarounds_apply(engine);
+ if (ret)

-:933: CHECK:BRACES: braces {} should be used on all arms of this statement #933: FILE: drivers/gpu/drm/i915/intel_workarounds.c:822:
+ if (INTEL_GEN(dev_priv) < 8)
[...]
+ else if (IS_BROADWELL(dev_priv))
[...]
else if (IS_CHERRYVIEW(dev_priv))
[...]
+ err = chv_whitelist_workarounds_apply(engine);
[...]
- err = skl_init_workarounds(engine);
[...]
else if (IS_BROXTON(dev_priv))
[...]
+ err = bxt_whitelist_workarounds_apply(engine);
[...]
- err = kbl_init_workarounds(engine);
[...]
else if (IS_GEMINILAKE(dev_priv))
[...]
+ err = glk_whitelist_workarounds_apply(engine);
[...]

-:959: CHECK:BRACES: Unbalanced braces around else statement #959: FILE: drivers/gpu/drm/i915/intel_workarounds.c:840:
+ else {

total: 0 errors, 4 warnings, 6 checks, 905 lines checked

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

#6 lore raw

✗ Fi.CI.SPARSE: warning for series starting with [1/2] drm/i915: Move a bunch of workaround-related code to its own file

In reply to Oscar Mateo
== Series Details ==

== Series Details ==

Series: series starting with [1/2] drm/i915: Move a bunch of workaround-related code to its own file
URL : https://patchwork.freedesktop.org/series/41486/
State : warning

== Summary ==

$ dim sparse origin/drm-tip
Commit: drm/i915: Move a bunch of workaround-related code to its own file
+drivers/gpu/drm/i915/intel_workarounds.c:684:1: warning: no newline at end of file

Commit: drm/i915: Split out functions for different kinds of workarounds -O:drivers/gpu/drm/i915/intel_workarounds.c:684:1: warning: no newline at end of file +

_______________________________________________
Intel-gfx mailing list
[email protected]
https://lists.freedesktop.org/mailman/listinfo/intel-gfx