[PATCH mesa v3] i965/gen8+: bo in state base address must be in 32-bit address range

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

 



Gen8+ supports 48-bit virtual addresses, but some objects must always be
allocated inside the 32-bit address range.

In specific, any resource used with flat/heapless (0x00000000-0xfffff000)
General State Heap or Intruction State Heap must be in a 32-bit range
(GSH / ISH), because the General State Offset and Instruction State Offset
are limited to 32-bits.

Use drm_intel_bo_emit_reloc_48bit when the 4GB limit is not necessary, and
the bo can be in the full address space.

This commit introduces a dependency of libdrm 2.4.63, which introduces the
drm_intel_bo_emit_reloc_48bit function.

v2: s/48baddress/48b_address/,
    Only use in OUT_RELOC64 cases, OUT_RELOC implies a 32-bit address offset
    is needed (Ben)
v3: Added OUT_RELOC64_INSIDE_4G, so it stands out when a 64-bit relocation
    needs the 32-bit workaround (Chris)

References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html
Cc: Ben Widawsky <ben@xxxxxxxxxxxx>
Cc: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Michel Thierry <michel.thierry@xxxxxxxxx>
---
 configure.ac                                  |  2 +-
 src/mesa/drivers/dri/i965/gen8_misc_state.c   | 19 +++++++++++--------
 src/mesa/drivers/dri/i965/intel_batchbuffer.c | 20 ++++++++++++++++----
 src/mesa/drivers/dri/i965/intel_batchbuffer.h | 10 ++++++++--
 4 files changed, 36 insertions(+), 15 deletions(-)

diff --git a/configure.ac b/configure.ac
index af61aa2..c92ca44 100644
--- a/configure.ac
+++ b/configure.ac
@@ -68,7 +68,7 @@ AC_SUBST([OSMESA_VERSION])
 dnl Versions for external dependencies
 LIBDRM_REQUIRED=2.4.38
 LIBDRM_RADEON_REQUIRED=2.4.56
-LIBDRM_INTEL_REQUIRED=2.4.60
+LIBDRM_INTEL_REQUIRED=2.4.63
 LIBDRM_NVVIEUX_REQUIRED=2.4.33
 LIBDRM_NOUVEAU_REQUIRED="2.4.33 libdrm >= 2.4.41"
 LIBDRM_FREEDRENO_REQUIRED=2.4.57
diff --git a/src/mesa/drivers/dri/i965/gen8_misc_state.c b/src/mesa/drivers/dri/i965/gen8_misc_state.c
index b20038e..73eba06 100644
--- a/src/mesa/drivers/dri/i965/gen8_misc_state.c
+++ b/src/mesa/drivers/dri/i965/gen8_misc_state.c
@@ -28,6 +28,10 @@
 
 /**
  * Define the base addresses which some state is referenced from.
+ *
+ * Use OUT_RELOC64_INSIDE_4G instead of OUT_RELOC64, the General State
+ * Offset and Instruction State Offset are limited to 32-bits by hardware,
+ * and must be located in the first 4GBs (32-bit offset).
  */
 void gen8_upload_state_base_address(struct brw_context *brw)
 {
@@ -41,19 +45,18 @@ void gen8_upload_state_base_address(struct brw_context *brw)
    OUT_BATCH(0);
    OUT_BATCH(mocs_wb << 16);
    /* Surface state base address: */
-   OUT_RELOC64(brw->batch.bo, I915_GEM_DOMAIN_SAMPLER, 0,
-               mocs_wb << 4 | 1);
+   OUT_RELOC64_INSIDE_4G(brw->batch.bo, I915_GEM_DOMAIN_SAMPLER, 0,
+                         mocs_wb << 4 | 1);
    /* Dynamic state base address: */
-   OUT_RELOC64(brw->batch.bo,
-               I915_GEM_DOMAIN_RENDER | I915_GEM_DOMAIN_INSTRUCTION, 0,
-               mocs_wb << 4 | 1);
+   OUT_RELOC64_INSIDE_4G(brw->batch.bo,
+                         I915_GEM_DOMAIN_RENDER | I915_GEM_DOMAIN_INSTRUCTION, 0,
+                         mocs_wb << 4 | 1);
    /* Indirect object base address: MEDIA_OBJECT data */
    OUT_BATCH(mocs_wb << 4 | 1);
    OUT_BATCH(0);
    /* Instruction base address: shader kernels (incl. SIP) */
-   OUT_RELOC64(brw->cache.bo, I915_GEM_DOMAIN_INSTRUCTION, 0,
-               mocs_wb << 4 | 1);
-
+   OUT_RELOC64_INSIDE_4G(brw->cache.bo, I915_GEM_DOMAIN_INSTRUCTION, 0,
+                         mocs_wb << 4 | 1);
    /* General state buffer size */
    OUT_BATCH(0xfffff001);
    /* Dynamic state buffer size */
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.c b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
index 54081a1..ca90784 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.c
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.c
@@ -409,11 +409,23 @@ bool
 intel_batchbuffer_emit_reloc64(struct brw_context *brw,
                                drm_intel_bo *buffer,
                                uint32_t read_domains, uint32_t write_domain,
-			       uint32_t delta)
+                               uint32_t delta,
+                               bool support_48bit_offset)
 {
-   int ret = drm_intel_bo_emit_reloc(brw->batch.bo, 4*brw->batch.used,
-                                     buffer, delta,
-                                     read_domains, write_domain);
+   int ret;
+
+   /* Not all buffers can be allocated outside the first 4GB, and
+    * offset must be limited to 32-bits.
+    */
+   if (support_48bit_offset)
+      drm_intel_bo_emit_reloc_48bit(brw->batch.bo, 4*brw->batch.used,
+                                    buffer, delta,
+                                    read_domains, write_domain);
+   else
+      drm_intel_bo_emit_reloc(brw->batch.bo, 4*brw->batch.used,
+                              buffer, delta,
+                              read_domains, write_domain);
+
    assert(ret == 0);
    (void) ret;
 
diff --git a/src/mesa/drivers/dri/i965/intel_batchbuffer.h b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
index ef8a6ff..de5023b 100644
--- a/src/mesa/drivers/dri/i965/intel_batchbuffer.h
+++ b/src/mesa/drivers/dri/i965/intel_batchbuffer.h
@@ -62,7 +62,8 @@ bool intel_batchbuffer_emit_reloc64(struct brw_context *brw,
                                     drm_intel_bo *buffer,
                                     uint32_t read_domains,
                                     uint32_t write_domain,
-                                    uint32_t offset);
+                                    uint32_t offset,
+                                    bool support_48bit_offset);
 static inline uint32_t float_as_int(float f)
 {
    union {
@@ -167,7 +168,12 @@ intel_batchbuffer_advance(struct brw_context *brw)
 
 /* Handle 48-bit address relocations for Gen8+ */
 #define OUT_RELOC64(buf, read_domains, write_domain, delta) do { \
-   intel_batchbuffer_emit_reloc64(brw, buf, read_domains, write_domain, delta);	\
+   intel_batchbuffer_emit_reloc64(brw, buf, read_domains, write_domain, delta, 1);	\
+} while (0)
+
+/* Handle 48-bit address relocations for Gen8+, requesting 32-bit offset */
+#define OUT_RELOC64_INSIDE_4G(buf, read_domains, write_domain, delta) do { \
+   intel_batchbuffer_emit_reloc64(brw, buf, read_domains, write_domain, delta, 0);	\
 } while (0)
 
 #define ADVANCE_BATCH() intel_batchbuffer_advance(brw);
-- 
2.5.0

_______________________________________________
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