This new version picks the suggestion from Michał Winiarski related to v3 [1].
48-bit virtual address range will be enabled in i915 soon, but some objects must be referenced by 32-bit offsets. These patches use a new kernel flag to specify if this restriction applies or not.
I'm sending these patches to comply with the i915 merge process. Once the kernel patch is merged, I'll make a new libdrm release and address the mesa build dependency.
[1] http://lists.freedesktop.org/archives/dri-devel/2015-August/087837.html
Michel Thierry (2): intel: 48b ppgtt support (EXEC_OBJECT_SUPPORTS_48B_ADDRESS flag) intel: add drm_intel_bo_use_48b_address_range to symbol-check test
include/drm/i915_drm.h | 3 +- intel/intel-symbol-check | 1 + intel/intel_bufmgr.c | 11 ++++++ intel/intel_bufmgr.h | 1 + intel/intel_bufmgr_gem.c | 88 +++++++++++++++++++++++++++++++++++++---------- intel/intel_bufmgr_priv.h | 14 ++++++++ 6 files changed, 98 insertions(+), 20 deletions(-)
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com Signed-off-by: Michel Thierry michel.thierry@intel.com --- include/drm/i915_drm.h | 3 +- intel/intel_bufmgr.c | 11 ++++++ intel/intel_bufmgr.h | 1 + intel/intel_bufmgr_gem.c | 88 +++++++++++++++++++++++++++++++++++++---------- intel/intel_bufmgr_priv.h | 14 ++++++++ 5 files changed, 97 insertions(+), 20 deletions(-)
diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index ded43b1..426b25c 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -680,7 +680,8 @@ struct drm_i915_gem_exec_object2 { #define EXEC_OBJECT_NEEDS_FENCE (1<<0) #define EXEC_OBJECT_NEEDS_GTT (1<<1) #define EXEC_OBJECT_WRITE (1<<2) -#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_WRITE<<1) +#define EXEC_OBJECT_SUPPORTS_48B_ADDRESS (1<<3) +#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_SUPPORTS_48B_ADDRESS<<1) __u64 flags;
__u64 rsvd1; diff --git a/intel/intel_bufmgr.c b/intel/intel_bufmgr.c index 14ea9f9..0856e60 100644 --- a/intel/intel_bufmgr.c +++ b/intel/intel_bufmgr.c @@ -293,6 +293,17 @@ drm_intel_bo_madvise(drm_intel_bo *bo, int madv) }
int +drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{ + if (bo->bufmgr->bo_use_48b_address_range) { + bo->bufmgr->bo_use_48b_address_range(bo, enable); + return 0; + } + + return -ENODEV; +} + +int drm_intel_bo_references(drm_intel_bo *bo, drm_intel_bo *target_bo) { return bo->bufmgr->bo_references(bo, target_bo); diff --git a/intel/intel_bufmgr.h b/intel/intel_bufmgr.h index 95eecb8..a14c78f 100644 --- a/intel/intel_bufmgr.h +++ b/intel/intel_bufmgr.h @@ -164,6 +164,7 @@ int drm_intel_bo_get_tiling(drm_intel_bo *bo, uint32_t * tiling_mode, int drm_intel_bo_flink(drm_intel_bo *bo, uint32_t * name); int drm_intel_bo_busy(drm_intel_bo *bo); int drm_intel_bo_madvise(drm_intel_bo *bo, int madv); +int drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable);
int drm_intel_bo_disable_reuse(drm_intel_bo *bo); int drm_intel_bo_is_reusable(drm_intel_bo *bo); diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c index 2723e21..09d82d2 100644 --- a/intel/intel_bufmgr_gem.c +++ b/intel/intel_bufmgr_gem.c @@ -83,6 +83,22 @@ #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) #define MAX2(A, B) ((A) > (B) ? (A) : (B))
+/** + * upper_32_bits - return bits 32-63 of a number + * @n: the number we're accessing + * + * A basic shift-right of a 64- or 32-bit quantity. Use this to suppress + * the "right shift count >= width of type" warning when that quantity is + * 32-bits. + */ +#define upper_32_bits(n) ((__u32)(((n) >> 16) >> 16)) + +/** + * lower_32_bits - return bits 0-31 of a number + * @n: the number we're accessing + */ +#define lower_32_bits(n) ((__u32)(n)) + typedef struct _drm_intel_bo_gem drm_intel_bo_gem;
struct drm_intel_gem_bo_bucket { @@ -237,6 +253,15 @@ struct _drm_intel_bo_gem { bool is_userptr;
/** + * Boolean of whether this buffer can be placed in the full 48-bit + * address range on gen8+. + * + * By default, buffers will be keep in a 32-bit range, unless this + * flag is explicitly set. + */ + bool use_48b_address_range; + + /** * Size in bytes of this buffer and its relocation descendents. * * Used to avoid costly tree walking in @@ -395,14 +420,16 @@ drm_intel_gem_dump_validation_list(drm_intel_bufmgr_gem *bufmgr_gem) drm_intel_bo_gem *target_gem = (drm_intel_bo_gem *) target_bo;
- DBG("%2d: %d (%s)@0x%08llx -> " - "%d (%s)@0x%08lx + 0x%08x\n", + DBG("%2d: %d (%s)@0x%08x %08x -> " + "%d (%s)@0x%08x %08x + 0x%08x\n", i, bo_gem->gem_handle, bo_gem->name, - (unsigned long long)bo_gem->relocs[j].offset, + upper_32_bits(bo_gem->relocs[j].offset), + lower_32_bits(bo_gem->relocs[j].offset), target_gem->gem_handle, target_gem->name, - target_bo->offset64, + upper_32_bits(target_bo->offset64), + lower_32_bits(target_bo->offset64), bo_gem->relocs[j].delta); } } @@ -468,11 +495,15 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) drm_intel_bufmgr_gem *bufmgr_gem = (drm_intel_bufmgr_gem *)bo->bufmgr; drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *)bo; int index; + int flags = 0; + + if (need_fence) + flags |= EXEC_OBJECT_NEEDS_FENCE; + if (bo_gem->use_48b_address_range) + flags |= EXEC_OBJECT_SUPPORTS_48B_ADDRESS;
if (bo_gem->validate_index != -1) { - if (need_fence) - bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |= - EXEC_OBJECT_NEEDS_FENCE; + bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |= flags; return; }
@@ -501,13 +532,9 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) bufmgr_gem->exec2_objects[index].alignment = bo->align; bufmgr_gem->exec2_objects[index].offset = 0; bufmgr_gem->exec_bos[index] = bo; - bufmgr_gem->exec2_objects[index].flags = 0; + bufmgr_gem->exec2_objects[index].flags = flags; bufmgr_gem->exec2_objects[index].rsvd1 = 0; bufmgr_gem->exec2_objects[index].rsvd2 = 0; - if (need_fence) { - bufmgr_gem->exec2_objects[index].flags |= - EXEC_OBJECT_NEEDS_FENCE; - } bufmgr_gem->exec_count++; }
@@ -780,6 +807,7 @@ retry: bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = true; + bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, alignment);
@@ -926,6 +954,7 @@ drm_intel_gem_bo_alloc_userptr(drm_intel_bufmgr *bufmgr, bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false; + bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);
@@ -1081,6 +1110,7 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr, bo_gem->bo.handle = open_arg.handle; bo_gem->global_name = handle; bo_gem->reusable = false; + bo_gem->use_48b_address_range = false;
memclear(get_tiling); get_tiling.handle = bo_gem->gem_handle; @@ -1930,6 +1960,13 @@ do_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, return 0; }
+static void +drm_intel_gem_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{ + drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *) bo; + bo_gem->use_48b_address_range = enable; +} + static int drm_intel_gem_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, drm_intel_bo *target_bo, uint32_t target_offset, @@ -2073,10 +2110,12 @@ drm_intel_update_buffer_offsets(drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec_objects[i].offset != bo->offset64) { - DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n", - bo_gem->gem_handle, bo_gem->name, bo->offset64, - (unsigned long long)bufmgr_gem->exec_objects[i]. - offset); + DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n", + bo_gem->gem_handle, bo_gem->name, + upper_32_bits(bo->offset64), + lower_32_bits(bo->offset64), + upper_32_bits(bufmgr_gem->exec_objects[i].offset), + lower_32_bits(bufmgr_gem->exec_objects[i].offset)); bo->offset64 = bufmgr_gem->exec_objects[i].offset; bo->offset = bufmgr_gem->exec_objects[i].offset; } @@ -2094,9 +2133,12 @@ drm_intel_update_buffer_offsets2 (drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec2_objects[i].offset != bo->offset64) { - DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n", - bo_gem->gem_handle, bo_gem->name, bo->offset64, - (unsigned long long)bufmgr_gem->exec2_objects[i].offset); + DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n", + bo_gem->gem_handle, bo_gem->name, + upper_32_bits(bo->offset64), + lower_32_bits(bo->offset64), + upper_32_bits(bufmgr_gem->exec2_objects[i].offset), + lower_32_bits(bufmgr_gem->exec2_objects[i].offset)); bo->offset64 = bufmgr_gem->exec2_objects[i].offset; bo->offset = bufmgr_gem->exec2_objects[i].offset; } @@ -2481,6 +2523,7 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false; + bo_gem->use_48b_address_range = false;
DRMINITLISTHEAD(&bo_gem->vma_list); DRMLISTADDTAIL(&bo_gem->name_list, &bufmgr_gem->named); @@ -3278,6 +3321,13 @@ drm_intel_bufmgr_gem_init(int fd, int batch_size) } }
+ if (bufmgr_gem->gen >= 8) { + gp.param = I915_PARAM_HAS_ALIASING_PPGTT; + ret = drmIoctl(bufmgr_gem->fd, DRM_IOCTL_I915_GETPARAM, &gp); + if (ret == 0 && *gp.value == 3) + bufmgr_gem->bufmgr.bo_use_48b_address_range = drm_intel_gem_bo_use_48b_address_range; + } + /* Let's go with one relocation per every 2 dwords (but round down a bit * since a power of two will mean an extra page allocation for the reloc * buffer). diff --git a/intel/intel_bufmgr_priv.h b/intel/intel_bufmgr_priv.h index 59ebd18..5c17ffb 100644 --- a/intel/intel_bufmgr_priv.h +++ b/intel/intel_bufmgr_priv.h @@ -152,6 +152,20 @@ struct _drm_intel_bufmgr { void (*destroy) (drm_intel_bufmgr *bufmgr);
/** + * Indicate if the buffer can be placed anywhere in the full ppgtt + * address range (2^48). + * + * Any resource used with flat/heapless (0x00000000-0xfffff000) + * General State Heap (GSH) or Intructions State Heap (ISH) must + * be in a 32-bit range. 48-bit range will only be used when explicitly + * requested. + * + * \param bo Buffer to set the use_48b_address_range flag. + * \param enable The flag value. + */ + void (*bo_use_48b_address_range) (drm_intel_bo *bo, uint32_t enable); + + /** * Add relocation entry in reloc_buf, which will be updated with the * target buffer's real offset on on command submission. *
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
include/drm/i915_drm.h | 3 +- intel/intel_bufmgr.c | 11 ++++++ intel/intel_bufmgr.h | 1 + intel/intel_bufmgr_gem.c | 88 +++++++++++++++++++++++++++++++++++++---------- intel/intel_bufmgr_priv.h | 14 ++++++++ 5 files changed, 97 insertions(+), 20 deletions(-)
diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index ded43b1..426b25c 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -680,7 +680,8 @@ struct drm_i915_gem_exec_object2 { #define EXEC_OBJECT_NEEDS_FENCE (1<<0) #define EXEC_OBJECT_NEEDS_GTT (1<<1) #define EXEC_OBJECT_WRITE (1<<2) -#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_WRITE<<1) +#define EXEC_OBJECT_SUPPORTS_48B_ADDRESS (1<<3) +#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_SUPPORTS_48B_ADDRESS<<1) __u64 flags;
__u64 rsvd1; diff --git a/intel/intel_bufmgr.c b/intel/intel_bufmgr.c index 14ea9f9..0856e60 100644 --- a/intel/intel_bufmgr.c +++ b/intel/intel_bufmgr.c @@ -293,6 +293,17 @@ drm_intel_bo_madvise(drm_intel_bo *bo, int madv) }
int +drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
- if (bo->bufmgr->bo_use_48b_address_range) {
bo->bufmgr->bo_use_48b_address_range(bo, enable);return 0;- }
- return -ENODEV;
+}
+int drm_intel_bo_references(drm_intel_bo *bo, drm_intel_bo *target_bo) { return bo->bufmgr->bo_references(bo, target_bo); diff --git a/intel/intel_bufmgr.h b/intel/intel_bufmgr.h index 95eecb8..a14c78f 100644 --- a/intel/intel_bufmgr.h +++ b/intel/intel_bufmgr.h @@ -164,6 +164,7 @@ int drm_intel_bo_get_tiling(drm_intel_bo *bo, uint32_t * tiling_mode, int drm_intel_bo_flink(drm_intel_bo *bo, uint32_t * name); int drm_intel_bo_busy(drm_intel_bo *bo); int drm_intel_bo_madvise(drm_intel_bo *bo, int madv); +int drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable);
int drm_intel_bo_disable_reuse(drm_intel_bo *bo); int drm_intel_bo_is_reusable(drm_intel_bo *bo); diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c index 2723e21..09d82d2 100644 --- a/intel/intel_bufmgr_gem.c +++ b/intel/intel_bufmgr_gem.c @@ -83,6 +83,22 @@ #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) #define MAX2(A, B) ((A) > (B) ? (A) : (B))
+/**
- upper_32_bits - return bits 32-63 of a number
- @n: the number we're accessing
- A basic shift-right of a 64- or 32-bit quantity. Use this to suppress
- the "right shift count >= width of type" warning when that quantity is
- 32-bits.
- */
+#define upper_32_bits(n) ((__u32)(((n) >> 16) >> 16))
+/**
- lower_32_bits - return bits 0-31 of a number
- @n: the number we're accessing
- */
+#define lower_32_bits(n) ((__u32)(n))
typedef struct _drm_intel_bo_gem drm_intel_bo_gem;
struct drm_intel_gem_bo_bucket { @@ -237,6 +253,15 @@ struct _drm_intel_bo_gem { bool is_userptr;
/**
* Boolean of whether this buffer can be placed in the full 48-bit* address range on gen8+.** By default, buffers will be keep in a 32-bit range, unless this* flag is explicitly set.*/- bool use_48b_address_range;
- /**
- Size in bytes of this buffer and its relocation descendents.
- Used to avoid costly tree walking in
@@ -395,14 +420,16 @@ drm_intel_gem_dump_validation_list(drm_intel_bufmgr_gem *bufmgr_gem) drm_intel_bo_gem *target_gem = (drm_intel_bo_gem *) target_bo;
DBG("%2d: %d (%s)@0x%08llx -> ""%d (%s)@0x%08lx + 0x%08x\n",
DBG("%2d: %d (%s)@0x%08x %08x -> ""%d (%s)@0x%08x %08x + 0x%08x\n", i, bo_gem->gem_handle, bo_gem->name,
(unsigned long long)bo_gem->relocs[j].offset,
upper_32_bits(bo_gem->relocs[j].offset),lower_32_bits(bo_gem->relocs[j].offset), target_gem->gem_handle, target_gem->name,
target_bo->offset64,
upper_32_bits(target_bo->offset64),lower_32_bits(target_bo->offset64), bo_gem->relocs[j].delta);@@ -468,11 +495,15 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) drm_intel_bufmgr_gem *bufmgr_gem = (drm_intel_bufmgr_gem *)bo->bufmgr; drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *)bo; int index;
int flags = 0;
if (need_fence)
flags |= EXEC_OBJECT_NEEDS_FENCE;if (bo_gem->use_48b_address_range)
flags |= EXEC_OBJECT_SUPPORTS_48B_ADDRESS;if (bo_gem->validate_index != -1) {
if (need_fence)bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |=EXEC_OBJECT_NEEDS_FENCE;
bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |= flags;@@ -501,13 +532,9 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) bufmgr_gem->exec2_objects[index].alignment = bo->align; bufmgr_gem->exec2_objects[index].offset = 0; bufmgr_gem->exec_bos[index] = bo;
- bufmgr_gem->exec2_objects[index].flags = 0;
- bufmgr_gem->exec2_objects[index].flags = flags; bufmgr_gem->exec2_objects[index].rsvd1 = 0; bufmgr_gem->exec2_objects[index].rsvd2 = 0;
- if (need_fence) {
bufmgr_gem->exec2_objects[index].flags |=EXEC_OBJECT_NEEDS_FENCE;- } bufmgr_gem->exec_count++;
}
@@ -780,6 +807,7 @@ retry: bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = true;
bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, alignment);
@@ -926,6 +954,7 @@ drm_intel_gem_bo_alloc_userptr(drm_intel_bufmgr *bufmgr, bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);
@@ -1081,6 +1110,7 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr, bo_gem->bo.handle = open_arg.handle; bo_gem->global_name = handle; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
memclear(get_tiling); get_tiling.handle = bo_gem->gem_handle;
@@ -1930,6 +1960,13 @@ do_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, return 0; }
+static void +drm_intel_gem_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
- drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *) bo;
- bo_gem->use_48b_address_range = enable;
+}
static int drm_intel_gem_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, drm_intel_bo *target_bo, uint32_t target_offset, @@ -2073,10 +2110,12 @@ drm_intel_update_buffer_offsets(drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec_objects[i].offset != bo->offset64) {
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name, bo->offset64,(unsigned long long)bufmgr_gem->exec_objects[i].offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n",bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec_objects[i].offset),lower_32_bits(bufmgr_gem->exec_objects[i].offset)); bo->offset64 = bufmgr_gem->exec_objects[i].offset; bo->offset = bufmgr_gem->exec_objects[i].offset;@@ -2094,9 +2133,12 @@ drm_intel_update_buffer_offsets2 (drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec2_objects[i].offset != bo->offset64) {
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name, bo->offset64,(unsigned long long)bufmgr_gem->exec2_objects[i].offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n",bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec2_objects[i].offset),lower_32_bits(bufmgr_gem->exec2_objects[i].offset)); bo->offset64 = bufmgr_gem->exec2_objects[i].offset; bo->offset = bufmgr_gem->exec2_objects[i].offset;@@ -2481,6 +2523,7 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
DRMINITLISTHEAD(&bo_gem->vma_list); DRMLISTADDTAIL(&bo_gem->name_list, &bufmgr_gem->named);
@@ -3278,6 +3321,13 @@ drm_intel_bufmgr_gem_init(int fd, int batch_size) } }
- if (bufmgr_gem->gen >= 8) {
gp.param = I915_PARAM_HAS_ALIASING_PPGTT;ret = drmIoctl(bufmgr_gem->fd, DRM_IOCTL_I915_GETPARAM, &gp);if (ret == 0 && *gp.value == 3)bufmgr_gem->bufmgr.bo_use_48b_address_range = drm_intel_gem_bo_use_48b_address_range;- }
- /* Let's go with one relocation per every 2 dwords (but round down a bit
- since a power of two will mean an extra page allocation for the reloc
- buffer).
diff --git a/intel/intel_bufmgr_priv.h b/intel/intel_bufmgr_priv.h index 59ebd18..5c17ffb 100644 --- a/intel/intel_bufmgr_priv.h +++ b/intel/intel_bufmgr_priv.h @@ -152,6 +152,20 @@ struct _drm_intel_bufmgr { void (*destroy) (drm_intel_bufmgr *bufmgr);
/**
* Indicate if the buffer can be placed anywhere in the full ppgtt* address range (2^48).** Any resource used with flat/heapless (0x00000000-0xfffff000)* General State Heap (GSH) or Intructions State Heap (ISH) must* be in a 32-bit range. 48-bit range will only be used when explicitly* requested.** \param bo Buffer to set the use_48b_address_range flag.* \param enable The flag value.*/- void (*bo_use_48b_address_range) (drm_intel_bo *bo, uint32_t enable);
- /**
- Add relocation entry in reloc_buf, which will be updated with the
- target buffer's real offset on on command submission.
-- 2.5.0
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
Thanks,
-Michel
include/drm/i915_drm.h | 3 +- intel/intel_bufmgr.c | 11 ++++++ intel/intel_bufmgr.h | 1 + intel/intel_bufmgr_gem.c | 88 +++++++++++++++++++++++++++++++++++++---------- intel/intel_bufmgr_priv.h | 14 ++++++++ 5 files changed, 97 insertions(+), 20 deletions(-)
diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index ded43b1..426b25c 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -680,7 +680,8 @@ struct drm_i915_gem_exec_object2 { #define EXEC_OBJECT_NEEDS_FENCE (1<<0) #define EXEC_OBJECT_NEEDS_GTT (1<<1) #define EXEC_OBJECT_WRITE (1<<2) -#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_WRITE<<1) +#define EXEC_OBJECT_SUPPORTS_48B_ADDRESS (1<<3) +#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_SUPPORTS_48B_ADDRESS<<1) __u64 flags;
__u64 rsvd1; diff --git a/intel/intel_bufmgr.c b/intel/intel_bufmgr.c index 14ea9f9..0856e60 100644 --- a/intel/intel_bufmgr.c +++ b/intel/intel_bufmgr.c @@ -293,6 +293,17 @@ drm_intel_bo_madvise(drm_intel_bo *bo, int madv) }
int +drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
- if (bo->bufmgr->bo_use_48b_address_range) {
bo->bufmgr->bo_use_48b_address_range(bo, enable);return 0;- }
- return -ENODEV;
+}
+int drm_intel_bo_references(drm_intel_bo *bo, drm_intel_bo *target_bo) { return bo->bufmgr->bo_references(bo, target_bo); diff --git a/intel/intel_bufmgr.h b/intel/intel_bufmgr.h index 95eecb8..a14c78f 100644 --- a/intel/intel_bufmgr.h +++ b/intel/intel_bufmgr.h @@ -164,6 +164,7 @@ int drm_intel_bo_get_tiling(drm_intel_bo *bo, uint32_t * tiling_mode, int drm_intel_bo_flink(drm_intel_bo *bo, uint32_t * name); int drm_intel_bo_busy(drm_intel_bo *bo); int drm_intel_bo_madvise(drm_intel_bo *bo, int madv); +int drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable);
int drm_intel_bo_disable_reuse(drm_intel_bo *bo); int drm_intel_bo_is_reusable(drm_intel_bo *bo); diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c index 2723e21..09d82d2 100644 --- a/intel/intel_bufmgr_gem.c +++ b/intel/intel_bufmgr_gem.c @@ -83,6 +83,22 @@ #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) #define MAX2(A, B) ((A) > (B) ? (A) : (B))
+/**
- upper_32_bits - return bits 32-63 of a number
- @n: the number we're accessing
- A basic shift-right of a 64- or 32-bit quantity. Use this to suppress
- the "right shift count >= width of type" warning when that quantity is
- 32-bits.
- */
+#define upper_32_bits(n) ((__u32)(((n) >> 16) >> 16))
+/**
- lower_32_bits - return bits 0-31 of a number
- @n: the number we're accessing
- */
+#define lower_32_bits(n) ((__u32)(n))
typedef struct _drm_intel_bo_gem drm_intel_bo_gem;
struct drm_intel_gem_bo_bucket {
@@ -237,6 +253,15 @@ struct _drm_intel_bo_gem { bool is_userptr;
/**
* Boolean of whether this buffer can be placed in the full 48-bit* address range on gen8+.** By default, buffers will be keep in a 32-bit range, unless this* flag is explicitly set.*/- bool use_48b_address_range;
- /**
- Size in bytes of this buffer and its relocation descendents.
- Used to avoid costly tree walking in
@@ -395,14 +420,16 @@ drm_intel_gem_dump_validation_list(drm_intel_bufmgr_gem *bufmgr_gem) drm_intel_bo_gem *target_gem = (drm_intel_bo_gem *) target_bo;
DBG("%2d: %d (%s)@0x%08llx -> ""%d (%s)@0x%08lx + 0x%08x\n",
DBG("%2d: %d (%s)@0x%08x %08x -> ""%d (%s)@0x%08x %08x + 0x%08x\n", i, bo_gem->gem_handle, bo_gem->name,
(unsigned long long)bo_gem->relocs[j].offset,
upper_32_bits(bo_gem->relocs[j].offset),lower_32_bits(bo_gem->relocs[j].offset), target_gem->gem_handle, target_gem->name,
target_bo->offset64,
upper_32_bits(target_bo->offset64),lower_32_bits(target_bo->offset64), bo_gem->relocs[j].delta);@@ -468,11 +495,15 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) drm_intel_bufmgr_gem *bufmgr_gem = (drm_intel_bufmgr_gem *)bo->bufmgr; drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *)bo; int index;
int flags = 0;
if (need_fence)
flags |= EXEC_OBJECT_NEEDS_FENCE;if (bo_gem->use_48b_address_range)
flags |= EXEC_OBJECT_SUPPORTS_48B_ADDRESS;if (bo_gem->validate_index != -1) {
if (need_fence)bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |=EXEC_OBJECT_NEEDS_FENCE;
bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |= flags;@@ -501,13 +532,9 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) bufmgr_gem->exec2_objects[index].alignment = bo->align; bufmgr_gem->exec2_objects[index].offset = 0; bufmgr_gem->exec_bos[index] = bo;
- bufmgr_gem->exec2_objects[index].flags = 0;
- bufmgr_gem->exec2_objects[index].flags = flags; bufmgr_gem->exec2_objects[index].rsvd1 = 0; bufmgr_gem->exec2_objects[index].rsvd2 = 0;
- if (need_fence) {
bufmgr_gem->exec2_objects[index].flags |=EXEC_OBJECT_NEEDS_FENCE;- } bufmgr_gem->exec_count++; }
@@ -780,6 +807,7 @@ retry: bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = true;
bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, alignment);
@@ -926,6 +954,7 @@ drm_intel_gem_bo_alloc_userptr(drm_intel_bufmgr *bufmgr, bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);
@@ -1081,6 +1110,7 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr, bo_gem->bo.handle = open_arg.handle; bo_gem->global_name = handle; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
memclear(get_tiling); get_tiling.handle = bo_gem->gem_handle;
@@ -1930,6 +1960,13 @@ do_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, return 0; }
+static void +drm_intel_gem_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
- drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *) bo;
- bo_gem->use_48b_address_range = enable;
+}
- static int drm_intel_gem_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, drm_intel_bo *target_bo, uint32_t target_offset,
@@ -2073,10 +2110,12 @@ drm_intel_update_buffer_offsets(drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec_objects[i].offset != bo->offset64) {
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name, bo->offset64,(unsigned long long)bufmgr_gem->exec_objects[i].offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n",bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec_objects[i].offset),lower_32_bits(bufmgr_gem->exec_objects[i].offset)); bo->offset64 = bufmgr_gem->exec_objects[i].offset; bo->offset = bufmgr_gem->exec_objects[i].offset;@@ -2094,9 +2133,12 @@ drm_intel_update_buffer_offsets2 (drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec2_objects[i].offset != bo->offset64) {
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name, bo->offset64,(unsigned long long)bufmgr_gem->exec2_objects[i].offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x %08x\n",bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec2_objects[i].offset),lower_32_bits(bufmgr_gem->exec2_objects[i].offset)); bo->offset64 = bufmgr_gem->exec2_objects[i].offset; bo->offset = bufmgr_gem->exec2_objects[i].offset;@@ -2481,6 +2523,7 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false;
DRMINITLISTHEAD(&bo_gem->vma_list); DRMLISTADDTAIL(&bo_gem->name_list, &bufmgr_gem->named);
@@ -3278,6 +3321,13 @@ drm_intel_bufmgr_gem_init(int fd, int batch_size) } }
- if (bufmgr_gem->gen >= 8) {
gp.param = I915_PARAM_HAS_ALIASING_PPGTT;ret = drmIoctl(bufmgr_gem->fd, DRM_IOCTL_I915_GETPARAM, &gp);if (ret == 0 && *gp.value == 3)bufmgr_gem->bufmgr.bo_use_48b_address_range = drm_intel_gem_bo_use_48b_address_range;- }
- /* Let's go with one relocation per every 2 dwords (but round down a bit
- since a power of two will mean an extra page allocation for the reloc
- buffer).
diff --git a/intel/intel_bufmgr_priv.h b/intel/intel_bufmgr_priv.h index 59ebd18..5c17ffb 100644 --- a/intel/intel_bufmgr_priv.h +++ b/intel/intel_bufmgr_priv.h @@ -152,6 +152,20 @@ struct _drm_intel_bufmgr { void (*destroy) (drm_intel_bufmgr *bufmgr);
/**
* Indicate if the buffer can be placed anywhere in the full ppgtt* address range (2^48).** Any resource used with flat/heapless (0x00000000-0xfffff000)* General State Heap (GSH) or Intructions State Heap (ISH) must* be in a 32-bit range. 48-bit range will only be used when explicitly* requested.** \param bo Buffer to set the use_48b_address_range flag.* \param enable The flag value.*/- void (*bo_use_48b_address_range) (drm_intel_bo *bo, uint32_t enable);
- /**
- Add relocation entry in reloc_buf, which will be updated with the
- target buffer's real offset on on command submission.
-- 2.5.0
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
Kristian
Thanks,
-Michel
include/drm/i915_drm.h | 3 +- intel/intel_bufmgr.c | 11 ++++++ intel/intel_bufmgr.h | 1 + intel/intel_bufmgr_gem.c | 88 +++++++++++++++++++++++++++++++++++++---------- intel/intel_bufmgr_priv.h | 14 ++++++++ 5 files changed, 97 insertions(+), 20 deletions(-)
diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index ded43b1..426b25c 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -680,7 +680,8 @@ struct drm_i915_gem_exec_object2 { #define EXEC_OBJECT_NEEDS_FENCE (1<<0) #define EXEC_OBJECT_NEEDS_GTT (1<<1) #define EXEC_OBJECT_WRITE (1<<2) -#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_WRITE<<1) +#define EXEC_OBJECT_SUPPORTS_48B_ADDRESS (1<<3) +#define __EXEC_OBJECT_UNKNOWN_FLAGS -(EXEC_OBJECT_SUPPORTS_48B_ADDRESS<<1) __u64 flags;
__u64 rsvd1;diff --git a/intel/intel_bufmgr.c b/intel/intel_bufmgr.c index 14ea9f9..0856e60 100644 --- a/intel/intel_bufmgr.c +++ b/intel/intel_bufmgr.c @@ -293,6 +293,17 @@ drm_intel_bo_madvise(drm_intel_bo *bo, int madv) }
int +drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
if (bo->bufmgr->bo_use_48b_address_range) {bo->bufmgr->bo_use_48b_address_range(bo, enable);return 0;}return -ENODEV;+}
+int drm_intel_bo_references(drm_intel_bo *bo, drm_intel_bo *target_bo) { return bo->bufmgr->bo_references(bo, target_bo); diff --git a/intel/intel_bufmgr.h b/intel/intel_bufmgr.h index 95eecb8..a14c78f 100644 --- a/intel/intel_bufmgr.h +++ b/intel/intel_bufmgr.h @@ -164,6 +164,7 @@ int drm_intel_bo_get_tiling(drm_intel_bo *bo, uint32_t * tiling_mode, int drm_intel_bo_flink(drm_intel_bo *bo, uint32_t * name); int drm_intel_bo_busy(drm_intel_bo *bo); int drm_intel_bo_madvise(drm_intel_bo *bo, int madv); +int drm_intel_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable);
int drm_intel_bo_disable_reuse(drm_intel_bo *bo); int drm_intel_bo_is_reusable(drm_intel_bo *bo); diff --git a/intel/intel_bufmgr_gem.c b/intel/intel_bufmgr_gem.c index 2723e21..09d82d2 100644 --- a/intel/intel_bufmgr_gem.c +++ b/intel/intel_bufmgr_gem.c @@ -83,6 +83,22 @@ #define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0])) #define MAX2(A, B) ((A) > (B) ? (A) : (B))
+/**
- upper_32_bits - return bits 32-63 of a number
- @n: the number we're accessing
- A basic shift-right of a 64- or 32-bit quantity. Use this to
suppress
- the "right shift count >= width of type" warning when that quantity
is
- 32-bits.
- */
+#define upper_32_bits(n) ((__u32)(((n) >> 16) >> 16))
+/**
- lower_32_bits - return bits 0-31 of a number
- @n: the number we're accessing
- */
+#define lower_32_bits(n) ((__u32)(n))
typedef struct _drm_intel_bo_gem drm_intel_bo_gem;
struct drm_intel_gem_bo_bucket {
@@ -237,6 +253,15 @@ struct _drm_intel_bo_gem { bool is_userptr;
/**
* Boolean of whether this buffer can be placed in the full48-bit
* address range on gen8+.** By default, buffers will be keep in a 32-bit range, unlessthis
* flag is explicitly set.*/bool use_48b_address_range;/** * Size in bytes of this buffer and its relocation descendents. * * Used to avoid costly tree walking in@@ -395,14 +420,16 @@ drm_intel_gem_dump_validation_list(drm_intel_bufmgr_gem *bufmgr_gem) drm_intel_bo_gem *target_gem = (drm_intel_bo_gem *) target_bo;
DBG("%2d: %d (%s)@0x%08llx -> ""%d (%s)@0x%08lx + 0x%08x\n",
DBG("%2d: %d (%s)@0x%08x %08x -> ""%d (%s)@0x%08x %08x + 0x%08x\n", i, bo_gem->gem_handle, bo_gem->name,
(unsigned long long)bo_gem->relocs[j].offset,
upper_32_bits(bo_gem->relocs[j].offset),lower_32_bits(bo_gem->relocs[j].offset), target_gem->gem_handle, target_gem->name,
target_bo->offset64,
upper_32_bits(target_bo->offset64),lower_32_bits(target_bo->offset64), bo_gem->relocs[j].delta); } }@@ -468,11 +495,15 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) drm_intel_bufmgr_gem *bufmgr_gem = (drm_intel_bufmgr_gem *)bo->bufmgr; drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *)bo; int index;
int flags = 0;if (need_fence)flags |= EXEC_OBJECT_NEEDS_FENCE;if (bo_gem->use_48b_address_range)flags |= EXEC_OBJECT_SUPPORTS_48B_ADDRESS; if (bo_gem->validate_index != -1) {
if (need_fence)bufmgr_gem->exec2_objects[bo_gem->validate_index].flags |=
EXEC_OBJECT_NEEDS_FENCE;
bufmgr_gem->exec2_objects[bo_gem->validate_index].flags|= flags; return; }
@@ -501,13 +532,9 @@ drm_intel_add_validate_buffer2(drm_intel_bo *bo, int need_fence) bufmgr_gem->exec2_objects[index].alignment = bo->align; bufmgr_gem->exec2_objects[index].offset = 0; bufmgr_gem->exec_bos[index] = bo;
bufmgr_gem->exec2_objects[index].flags = 0;
bufmgr_gem->exec2_objects[index].flags = flags; bufmgr_gem->exec2_objects[index].rsvd1 = 0; bufmgr_gem->exec2_objects[index].rsvd2 = 0;
if (need_fence) {bufmgr_gem->exec2_objects[index].flags |=EXEC_OBJECT_NEEDS_FENCE;} bufmgr_gem->exec_count++;@@ -780,6 +807,7 @@ retry: bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = true;
bo_gem->use_48b_address_range = false; drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem,alignment);
@@ -926,6 +954,7 @@ drm_intel_gem_bo_alloc_userptr(drm_intel_bufmgr *bufmgr, bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false; drm_intel_bo_gem_set_in_aperture_size(bufmgr_gem, bo_gem, 0);@@ -1081,6 +1110,7 @@ drm_intel_bo_gem_create_from_name(drm_intel_bufmgr *bufmgr, bo_gem->bo.handle = open_arg.handle; bo_gem->global_name = handle; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false; memclear(get_tiling); get_tiling.handle = bo_gem->gem_handle;@@ -1930,6 +1960,13 @@ do_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, return 0; }
+static void +drm_intel_gem_bo_use_48b_address_range(drm_intel_bo *bo, uint32_t enable) +{
drm_intel_bo_gem *bo_gem = (drm_intel_bo_gem *) bo;bo_gem->use_48b_address_range = enable;+}
- static int drm_intel_gem_bo_emit_reloc(drm_intel_bo *bo, uint32_t offset, drm_intel_bo *target_bo, uint32_t
target_offset, @@ -2073,10 +2110,12 @@ drm_intel_update_buffer_offsets(drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec_objects[i].offset != bo->offset64) {
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name,bo->offset64,
(unsigned longlong)bufmgr_gem->exec_objects[i].
offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x%08x\n",
bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec_objects[i].offset),
lower_32_bits(bufmgr_gem->exec_objects[i].offset)); bo->offset64 = bufmgr_gem->exec_objects[i].offset; bo->offset = bufmgr_gem->exec_objects[i].offset; } @@ -2094,9 +2133,12 @@ drm_intel_update_buffer_offsets2 (drm_intel_bufmgr_gem *bufmgr_gem)
/* Update the buffer offset */ if (bufmgr_gem->exec2_objects[i].offset != bo->offset64){
DBG("BO %d (%s) migrated: 0x%08lx -> 0x%08llx\n",bo_gem->gem_handle, bo_gem->name,bo->offset64,
(unsigned longlong)bufmgr_gem->exec2_objects[i].offset);
DBG("BO %d (%s) migrated: 0x%08x %08x -> 0x%08x%08x\n",
bo_gem->gem_handle, bo_gem->name,upper_32_bits(bo->offset64),lower_32_bits(bo->offset64),upper_32_bits(bufmgr_gem->exec2_objects[i].offset),
lower_32_bits(bufmgr_gem->exec2_objects[i].offset)); bo->offset64 = bufmgr_gem->exec2_objects[i].offset; bo->offset = bufmgr_gem->exec2_objects[i].offset; } @@ -2481,6 +2523,7 @@ drm_intel_bo_gem_create_from_prime(drm_intel_bufmgr *bufmgr, int prime_fd, int s bo_gem->used_as_reloc_target = false; bo_gem->has_error = false; bo_gem->reusable = false;
bo_gem->use_48b_address_range = false; DRMINITLISTHEAD(&bo_gem->vma_list); DRMLISTADDTAIL(&bo_gem->name_list, &bufmgr_gem->named);@@ -3278,6 +3321,13 @@ drm_intel_bufmgr_gem_init(int fd, int batch_size) } }
if (bufmgr_gem->gen >= 8) {gp.param = I915_PARAM_HAS_ALIASING_PPGTT;ret = drmIoctl(bufmgr_gem->fd, DRM_IOCTL_I915_GETPARAM,&gp);
if (ret == 0 && *gp.value == 3)bufmgr_gem->bufmgr.bo_use_48b_address_range =drm_intel_gem_bo_use_48b_address_range;
}/* Let's go with one relocation per every 2 dwords (but rounddown a bit * since a power of two will mean an extra page allocation for the reloc * buffer). diff --git a/intel/intel_bufmgr_priv.h b/intel/intel_bufmgr_priv.h index 59ebd18..5c17ffb 100644 --- a/intel/intel_bufmgr_priv.h +++ b/intel/intel_bufmgr_priv.h @@ -152,6 +152,20 @@ struct _drm_intel_bufmgr { void (*destroy) (drm_intel_bufmgr *bufmgr);
/**
* Indicate if the buffer can be placed anywhere in the fullppgtt
* address range (2^48).** Any resource used with flat/heapless (0x00000000-0xfffff000)* General State Heap (GSH) or Intructions State Heap (ISH) must* be in a 32-bit range. 48-bit range will only be used whenexplicitly
* requested.** \param bo Buffer to set the use_48b_address_range flag.* \param enable The flag value.*/void (*bo_use_48b_address_range) (drm_intel_bo *bo, uint32_tenable);
/** * Add relocation entry in reloc_buf, which will be updated withthe * target buffer's real offset on on command submission. * -- 2.5.0
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
Thanks,
-Michel
[1]http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote:
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 (GSH) or Instruction State Heap (ISH) must be in a 32-bit range, because the General State Offset and Instruction State Offset are limited to 32-bits.
The i915 driver has been modified to provide a flag to set when the 4GB limit is not necessary in a given bo (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). 48-bit range will only be used when explicitly requested.
Callers to the existing drm_intel_bo_emit_reloc function should set the use_48b_address_range flag beforehand, in order to use full ppgtt range.
v2: Make set/clear functions nops on pre-gen8 platforms, and use them internally in emit_reloc functions (Ben) s/48BADDRESS/48B_ADDRESS/ (Dave) v3: Keep set/clear functions internal, no-one needs to use them directly. v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type before enabling set/clear function, print full offsets in debug statements, using port of lower_32_bits and upper_32_bits from linux kernel (Michał)
References: http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html Cc: Ben Widawsky ben@bwidawsk.net Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes
(http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Regards, Emil
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote:
On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: > > > 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 (GSH) or Instruction State Heap (ISH) must be in a > 32-bit range, because the General State Offset and Instruction State > Offset > are limited to 32-bits. > > The i915 driver has been modified to provide a flag to set when the > 4GB > limit is not necessary in a given bo > (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). > 48-bit range will only be used when explicitly requested. > > Callers to the existing drm_intel_bo_emit_reloc function should set > the > use_48b_address_range flag beforehand, in order to use full ppgtt > range. > > v2: Make set/clear functions nops on pre-gen8 platforms, and use them > internally in emit_reloc functions (Ben) > s/48BADDRESS/48B_ADDRESS/ (Dave) > v3: Keep set/clear functions internal, no-one needs to use them > directly. > v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type > before enabling set/clear function, print full offsets in debug > statements, using port of lower_32_bits and upper_32_bits > from linux > kernel (Michał) > > References: > http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html > Cc: Ben Widawsky ben@bwidawsk.net > Cc: Michał Winiarski michal.winiarski@intel.com
+Kristian
LGTM. Acked-by: Michał Winiarski michal.winiarski@intel.com
> Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes
(http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here: https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
Thanks,
-Michel
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote:
On 9/14/2015 2:54 PM, Michał Winiarski wrote: > > > > On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >> >> >> >> 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 (GSH) or Instruction State Heap (ISH) must be in >> a >> 32-bit range, because the General State Offset and Instruction State >> Offset >> are limited to 32-bits. >> >> The i915 driver has been modified to provide a flag to set when the >> 4GB >> limit is not necessary in a given bo >> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >> 48-bit range will only be used when explicitly requested. >> >> Callers to the existing drm_intel_bo_emit_reloc function should set >> the >> use_48b_address_range flag beforehand, in order to use full ppgtt >> range. >> >> v2: Make set/clear functions nops on pre-gen8 platforms, and use >> them >> internally in emit_reloc functions (Ben) >> s/48BADDRESS/48B_ADDRESS/ (Dave) >> v3: Keep set/clear functions internal, no-one needs to use them >> directly. >> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type >> before enabling set/clear function, print full offsets in >> debug >> statements, using port of lower_32_bits and upper_32_bits >> from linux >> kernel (Michał) >> >> References: >> >> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >> Cc: Ben Widawsky ben@bwidawsk.net >> Cc: Michał Winiarski michal.winiarski@intel.com > > > > > +Kristian > > LGTM. > Acked-by: Michał Winiarski michal.winiarski@intel.com > >> Signed-off-by: Michel Thierry michel.thierry@intel.com
Hi Kristian,
More comments on this? I've resent the mesa patch with the last changes
(http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html).
Michał has agreed with the interface and will use it for his work. If mesa doesn't plan to use this for now, it's ok. The kernel changes have been done to prevent any regressions when the support 48-bit flag is not set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here: https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Kristian
Thanks,
-Michel
On Tue, Oct 13, 2015 at 02:51:36PM -0700, Kristian Høgsberg wrote:
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote:
On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry michel.thierry@intel.com wrote: > > > On 9/14/2015 2:54 PM, Michał Winiarski wrote: >> >> >> >> On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >>> >>> >>> >>> 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 (GSH) or Instruction State Heap (ISH) must be in >>> a >>> 32-bit range, because the General State Offset and Instruction State >>> Offset >>> are limited to 32-bits. >>> >>> The i915 driver has been modified to provide a flag to set when the >>> 4GB >>> limit is not necessary in a given bo >>> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >>> 48-bit range will only be used when explicitly requested. >>> >>> Callers to the existing drm_intel_bo_emit_reloc function should set >>> the >>> use_48b_address_range flag beforehand, in order to use full ppgtt >>> range. >>> >>> v2: Make set/clear functions nops on pre-gen8 platforms, and use >>> them >>> internally in emit_reloc functions (Ben) >>> s/48BADDRESS/48B_ADDRESS/ (Dave) >>> v3: Keep set/clear functions internal, no-one needs to use them >>> directly. >>> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type >>> before enabling set/clear function, print full offsets in >>> debug >>> statements, using port of lower_32_bits and upper_32_bits >>> from linux >>> kernel (Michał) >>> >>> References: >>> >>> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >>> Cc: Ben Widawsky ben@bwidawsk.net >>> Cc: Michał Winiarski michal.winiarski@intel.com >> >> >> >> >> +Kristian >> >> LGTM. >> Acked-by: Michał Winiarski michal.winiarski@intel.com >> >>> Signed-off-by: Michel Thierry michel.thierry@intel.com > > > > > > Hi Kristian, > > More comments on this? > I've resent the mesa patch with the last changes > > > (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html). > > > Michał has agreed with the interface and will use it for his work. > If mesa doesn't plan to use this for now, it's ok. The kernel changes > have > been done to prevent any regressions when the support 48-bit flag is > not > set.
I didn't get any replies to my last comments on this:
http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html
Basically, I tried explicitly placing buffers above 8G and didn't see the HW problem described (ie it all worked fine). I still think there's some confusion as to what the W/A is about.
Here's my take: the W/A is a SW-only workaround to handle the cases where a driver uses heapless and 48-bit ppgtt. The problem is that the heaps are limited to 4G but with 48bit ppgtt a buffer can be placed anywhere it the 48 bit address space. If that happens it's consideredd out-of-bounds for the heap and access fails. To prevent this we need to make sure the bos we address in a heapless fashion are placed below 4g.
For mesa, we only configure general state base as heap-less, which affects scratch bos. What this boils down to is that we need the 4G restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS and 3DSTATE_PS (and tesselation stage eventually). Look for the OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here: https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Ok, so we do still need the 4G limit bit (i.e. I don't have to revert the kernel side for lack of open-source userspace), just not at the place we originally thought. I guess lesson learned is to actually test stuff first before I pull in the kernel side. Michel, did you not see the spilling corruptions when testing the mesa patches? Kristian, does this also blow up with just a piglit run?
Thanks, Daniel
On 10/14/2015 8:19 AM, Daniel Vetter wrote:
On Tue, Oct 13, 2015 at 02:51:36PM -0700, Kristian Høgsberg wrote:
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote:
On 10/5/2015 7:06 PM, Kristian Høgsberg wrote: > > > On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry > michel.thierry@intel.com wrote: >> >> >> On 9/14/2015 2:54 PM, Michał Winiarski wrote: >>> >>> >>> >>> On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >>>> >>>> >>>> >>>> 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 (GSH) or Instruction State Heap (ISH) must be in >>>> a >>>> 32-bit range, because the General State Offset and Instruction State >>>> Offset >>>> are limited to 32-bits. >>>> >>>> The i915 driver has been modified to provide a flag to set when the >>>> 4GB >>>> limit is not necessary in a given bo >>>> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >>>> 48-bit range will only be used when explicitly requested. >>>> >>>> Callers to the existing drm_intel_bo_emit_reloc function should set >>>> the >>>> use_48b_address_range flag beforehand, in order to use full ppgtt >>>> range. >>>> >>>> v2: Make set/clear functions nops on pre-gen8 platforms, and use >>>> them >>>> internally in emit_reloc functions (Ben) >>>> s/48BADDRESS/48B_ADDRESS/ (Dave) >>>> v3: Keep set/clear functions internal, no-one needs to use them >>>> directly. >>>> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt type >>>> before enabling set/clear function, print full offsets in >>>> debug >>>> statements, using port of lower_32_bits and upper_32_bits >>>> from linux >>>> kernel (Michał) >>>> >>>> References: >>>> >>>> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >>>> Cc: Ben Widawsky ben@bwidawsk.net >>>> Cc: Michał Winiarski michal.winiarski@intel.com >>> >>> >>> >>> >>> +Kristian >>> >>> LGTM. >>> Acked-by: Michał Winiarski michal.winiarski@intel.com >>> >>>> Signed-off-by: Michel Thierry michel.thierry@intel.com >> >> >> >> >> >> Hi Kristian, >> >> More comments on this? >> I've resent the mesa patch with the last changes >> >> >> (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html). >> >> >> Michał has agreed with the interface and will use it for his work. >> If mesa doesn't plan to use this for now, it's ok. The kernel changes >> have >> been done to prevent any regressions when the support 48-bit flag is >> not >> set. > > > > I didn't get any replies to my last comments on this: > > http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html > > Basically, I tried explicitly placing buffers above 8G and didn't see > the HW problem described (ie it all worked fine). I still think > there's some confusion as to what the W/A is about. > > Here's my take: the W/A is a SW-only workaround to handle the cases > where a driver uses heapless and 48-bit ppgtt. The problem is that the > heaps are limited to 4G but with 48bit ppgtt a buffer can be placed > anywhere it the 48 bit address space. If that happens it's consideredd > out-of-bounds for the heap and access fails. To prevent this we need > to make sure the bos we address in a heapless fashion are placed below > 4g. > > For mesa, we only configure general state base as heap-less, which > affects scratch bos. What this boils down to is that we need the 4G > restriction only for the scratch bos set up on 3DSTATE_VS, 3DSTATE_GS > and 3DSTATE_PS (and tesselation stage eventually). Look for the > OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, gen8_gs_state.c > and gen8_ps_state.c
I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it isn't exclusive to the scratch bos (the error state points to that instruction).
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here: https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Ok, so we do still need the 4G limit bit (i.e. I don't have to revert the kernel side for lack of open-source userspace), just not at the place we originally thought. I guess lesson learned is to actually test stuff first before I pull in the kernel side. Michel, did you not see the spilling corruptions when testing the mesa patches? Kristian, does this also blow up with just a piglit run?
There's must be something else we have different in our systems.
Kristian doesn't get the hang, and if I back out the 4G restriction in the STATE_BASE_ADDRESS, I don't see any corruption while running INTEL_DEBUG=spill_fs glxgears.
On Wed, Oct 14, 2015 at 5:11 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/14/2015 8:19 AM, Daniel Vetter wrote:
On Tue, Oct 13, 2015 at 02:51:36PM -0700, Kristian Høgsberg wrote:
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote:
On 10/6/2015 2:12 PM, Michel Thierry wrote: > > > > On 10/5/2015 7:06 PM, Kristian Høgsberg wrote: >> >> >> >> On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry >> michel.thierry@intel.com wrote: >>> >>> >>> >>> On 9/14/2015 2:54 PM, Michał Winiarski wrote: >>>> >>>> >>>> >>>> >>>> On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >>>>> >>>>> >>>>> >>>>> >>>>> 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 (GSH) or Instruction State Heap (ISH) must be >>>>> in >>>>> a >>>>> 32-bit range, because the General State Offset and Instruction >>>>> State >>>>> Offset >>>>> are limited to 32-bits. >>>>> >>>>> The i915 driver has been modified to provide a flag to set when >>>>> the >>>>> 4GB >>>>> limit is not necessary in a given bo >>>>> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >>>>> 48-bit range will only be used when explicitly requested. >>>>> >>>>> Callers to the existing drm_intel_bo_emit_reloc function should >>>>> set >>>>> the >>>>> use_48b_address_range flag beforehand, in order to use full ppgtt >>>>> range. >>>>> >>>>> v2: Make set/clear functions nops on pre-gen8 platforms, and use >>>>> them >>>>> internally in emit_reloc functions (Ben) >>>>> s/48BADDRESS/48B_ADDRESS/ (Dave) >>>>> v3: Keep set/clear functions internal, no-one needs to use them >>>>> directly. >>>>> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt >>>>> type >>>>> before enabling set/clear function, print full offsets in >>>>> debug >>>>> statements, using port of lower_32_bits and upper_32_bits >>>>> from linux >>>>> kernel (Michał) >>>>> >>>>> References: >>>>> >>>>> >>>>> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >>>>> Cc: Ben Widawsky ben@bwidawsk.net >>>>> Cc: Michał Winiarski michal.winiarski@intel.com >>>> >>>> >>>> >>>> >>>> >>>> +Kristian >>>> >>>> LGTM. >>>> Acked-by: Michał Winiarski michal.winiarski@intel.com >>>> >>>>> Signed-off-by: Michel Thierry michel.thierry@intel.com >>> >>> >>> >>> >>> >>> >>> Hi Kristian, >>> >>> More comments on this? >>> I've resent the mesa patch with the last changes >>> >>> >>> >>> (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html). >>> >>> >>> Michał has agreed with the interface and will use it for his work. >>> If mesa doesn't plan to use this for now, it's ok. The kernel >>> changes >>> have >>> been done to prevent any regressions when the support 48-bit flag >>> is >>> not >>> set. >> >> >> >> >> I didn't get any replies to my last comments on this: >> >> >> http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html >> >> Basically, I tried explicitly placing buffers above 8G and didn't >> see >> the HW problem described (ie it all worked fine). I still think >> there's some confusion as to what the W/A is about. >> >> Here's my take: the W/A is a SW-only workaround to handle the cases >> where a driver uses heapless and 48-bit ppgtt. The problem is that >> the >> heaps are limited to 4G but with 48bit ppgtt a buffer can be placed >> anywhere it the 48 bit address space. If that happens it's >> consideredd >> out-of-bounds for the heap and access fails. To prevent this we need >> to make sure the bos we address in a heapless fashion are placed >> below >> 4g. >> >> For mesa, we only configure general state base as heap-less, which >> affects scratch bos. What this boils down to is that we need the 4G >> restriction only for the scratch bos set up on 3DSTATE_VS, >> 3DSTATE_GS >> and 3DSTATE_PS (and tesselation stage eventually). Look for the >> OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, >> gen8_gs_state.c >> and gen8_ps_state.c > > > > > I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it > isn't exclusive to the scratch bos (the error state points to that > instruction). > >
Hi,
Anymore inputs about this patch? AFAIK, if changes are required based on further comments from Kristian, these will be in the mesa patch[1], not libdrm. Also, Michał will use this flag in another project.
The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here:
https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Ok, so we do still need the 4G limit bit (i.e. I don't have to revert the kernel side for lack of open-source userspace), just not at the place we originally thought. I guess lesson learned is to actually test stuff first before I pull in the kernel side. Michel, did you not see the spilling corruptions when testing the mesa patches? Kristian, does this also blow up with just a piglit run?
There's must be something else we have different in our systems.
Kristian doesn't get the hang, and if I back out the 4G restriction in the STATE_BASE_ADDRESS, I don't see any corruption while running INTEL_DEBUG=spill_fs glxgears.
I don't want to block this, in fact, I very much want to see this move forward. However, I can't ack patches that 1) fix a gpu hang I can't reproduce and I don't agree should be a problem (the STATE_BASE_ADDRESS probllem) and 2) don't fix the issue that I see and reproduced on my BDW (the spilling problem).
I'll try a full piglit run tomorrow with both my mesa patch and Michels and see what happens. However, this isn't the kind of problem that occurs once in a blue moon or for some corner case piglit test. If Michels understanding of the hw and the workaround is correct, nothing should work without his patch.
To move forward on this, unless something new comes up when running piglit, I can take ownership of making mesa work one way or the other. The kernel and libdrm work is fine and most of Michels mesa patch is good. If mesa breaks down the road because we need to limit the STATE_BASE_ADDRESS relocs, I'll own the problem then.
Kristian
On 11/18/2015 10:53 PM, Kristian Høgsberg wrote:
On Wed, Oct 14, 2015 at 5:11 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/14/2015 8:19 AM, Daniel Vetter wrote:
On Tue, Oct 13, 2015 at 02:51:36PM -0700, Kristian Høgsberg wrote:
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote:
On 13 October 2015 at 13:16, Michel Thierry michel.thierry@intel.com wrote: > > > On 10/6/2015 2:12 PM, Michel Thierry wrote: >> >> >> >> On 10/5/2015 7:06 PM, Kristian Høgsberg wrote: >>> >>> >>> >>> On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry >>> michel.thierry@intel.com wrote: >>>> >>>> >>>> >>>> On 9/14/2015 2:54 PM, Michał Winiarski wrote: >>>>> >>>>> >>>>> >>>>> >>>>> On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> 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 (GSH) or Instruction State Heap (ISH) must be >>>>>> in >>>>>> a >>>>>> 32-bit range, because the General State Offset and Instruction >>>>>> State >>>>>> Offset >>>>>> are limited to 32-bits. >>>>>> >>>>>> The i915 driver has been modified to provide a flag to set when >>>>>> the >>>>>> 4GB >>>>>> limit is not necessary in a given bo >>>>>> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >>>>>> 48-bit range will only be used when explicitly requested. >>>>>> >>>>>> Callers to the existing drm_intel_bo_emit_reloc function should >>>>>> set >>>>>> the >>>>>> use_48b_address_range flag beforehand, in order to use full ppgtt >>>>>> range. >>>>>> >>>>>> v2: Make set/clear functions nops on pre-gen8 platforms, and use >>>>>> them >>>>>> internally in emit_reloc functions (Ben) >>>>>> s/48BADDRESS/48B_ADDRESS/ (Dave) >>>>>> v3: Keep set/clear functions internal, no-one needs to use them >>>>>> directly. >>>>>> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt >>>>>> type >>>>>> before enabling set/clear function, print full offsets in >>>>>> debug >>>>>> statements, using port of lower_32_bits and upper_32_bits >>>>>> from linux >>>>>> kernel (Michał) >>>>>> >>>>>> References: >>>>>> >>>>>> >>>>>> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >>>>>> Cc: Ben Widawsky ben@bwidawsk.net >>>>>> Cc: Michał Winiarski michal.winiarski@intel.com >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> +Kristian >>>>> >>>>> LGTM. >>>>> Acked-by: Michał Winiarski michal.winiarski@intel.com >>>>> >>>>>> Signed-off-by: Michel Thierry michel.thierry@intel.com >>>> >>>> >>>> >>>> >>>> >>>> >>>> Hi Kristian, >>>> >>>> More comments on this? >>>> I've resent the mesa patch with the last changes >>>> >>>> >>>> >>>> (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html). >>>> >>>> >>>> Michał has agreed with the interface and will use it for his work. >>>> If mesa doesn't plan to use this for now, it's ok. The kernel >>>> changes >>>> have >>>> been done to prevent any regressions when the support 48-bit flag >>>> is >>>> not >>>> set. >>> >>> >>> >>> >>> I didn't get any replies to my last comments on this: >>> >>> >>> http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html >>> >>> Basically, I tried explicitly placing buffers above 8G and didn't >>> see >>> the HW problem described (ie it all worked fine). I still think >>> there's some confusion as to what the W/A is about. >>> >>> Here's my take: the W/A is a SW-only workaround to handle the cases >>> where a driver uses heapless and 48-bit ppgtt. The problem is that >>> the >>> heaps are limited to 4G but with 48bit ppgtt a buffer can be placed >>> anywhere it the 48 bit address space. If that happens it's >>> consideredd >>> out-of-bounds for the heap and access fails. To prevent this we need >>> to make sure the bos we address in a heapless fashion are placed >>> below >>> 4g. >>> >>> For mesa, we only configure general state base as heap-less, which >>> affects scratch bos. What this boils down to is that we need the 4G >>> restriction only for the scratch bos set up on 3DSTATE_VS, >>> 3DSTATE_GS >>> and 3DSTATE_PS (and tesselation stage eventually). Look for the >>> OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, >>> gen8_gs_state.c >>> and gen8_ps_state.c >> >> >> >> >> I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe it >> isn't exclusive to the scratch bos (the error state points to that >> instruction). >> >> > > Hi, > > Anymore inputs about this patch? > AFAIK, if changes are required based on further comments from > Kristian, > these will be in the mesa patch[1], not libdrm. Also, Michał will use > this > flag in another project. > The comment seems quite brief and I'm not sure it fully addresses Kristian's concern. I'd suspect that providing reference to the HW documentation (confirming your assumption) might be beneficial.
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here:
https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Ok, so we do still need the 4G limit bit (i.e. I don't have to revert the kernel side for lack of open-source userspace), just not at the place we originally thought. I guess lesson learned is to actually test stuff first before I pull in the kernel side. Michel, did you not see the spilling corruptions when testing the mesa patches? Kristian, does this also blow up with just a piglit run?
There's must be something else we have different in our systems.
Kristian doesn't get the hang, and if I back out the 4G restriction in the STATE_BASE_ADDRESS, I don't see any corruption while running INTEL_DEBUG=spill_fs glxgears.
I don't want to block this, in fact, I very much want to see this move forward. However, I can't ack patches that 1) fix a gpu hang I can't reproduce and I don't agree should be a problem (the STATE_BASE_ADDRESS probllem) and 2) don't fix the issue that I see and reproduced on my BDW (the spilling problem).
I'll try a full piglit run tomorrow with both my mesa patch and Michels and see what happens. However, this isn't the kind of problem that occurs once in a blue moon or for some corner case piglit test. If Michels understanding of the hw and the workaround is correct, nothing should work without his patch.
To move forward on this, unless something new comes up when running piglit, I can take ownership of making mesa work one way or the other. The kernel and libdrm work is fine and most of Michels mesa patch is good. If mesa breaks down the road because we need to limit the STATE_BASE_ADDRESS relocs, I'll own the problem then.
Kristian
Thanks Kristian,
Let me know if you need something from i915 or libdrm side.
On Fri, Dec 4, 2015 at 6:24 AM, Michel Thierry michel.thierry@intel.com wrote:
On 11/18/2015 10:53 PM, Kristian Høgsberg wrote:
On Wed, Oct 14, 2015 at 5:11 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/14/2015 8:19 AM, Daniel Vetter wrote:
On Tue, Oct 13, 2015 at 02:51:36PM -0700, Kristian Høgsberg wrote:
On Tue, Oct 13, 2015 at 7:55 AM, Michel Thierry michel.thierry@intel.com wrote:
On 10/13/2015 3:13 PM, Emil Velikov wrote: > > > > On 13 October 2015 at 13:16, Michel Thierry > michel.thierry@intel.com > wrote: >> >> >> >> On 10/6/2015 2:12 PM, Michel Thierry wrote: >>> >>> >>> >>> >>> On 10/5/2015 7:06 PM, Kristian Høgsberg wrote: >>>> >>>> >>>> >>>> >>>> On Mon, Oct 5, 2015 at 7:03 AM, Michel Thierry >>>> michel.thierry@intel.com wrote: >>>>> >>>>> >>>>> >>>>> >>>>> On 9/14/2015 2:54 PM, Michał Winiarski wrote: >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> On Thu, Sep 03, 2015 at 03:23:58PM +0100, Michel Thierry wrote: >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> 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 (GSH) or Instruction State Heap (ISH) must >>>>>>> be >>>>>>> in >>>>>>> a >>>>>>> 32-bit range, because the General State Offset and Instruction >>>>>>> State >>>>>>> Offset >>>>>>> are limited to 32-bits. >>>>>>> >>>>>>> The i915 driver has been modified to provide a flag to set when >>>>>>> the >>>>>>> 4GB >>>>>>> limit is not necessary in a given bo >>>>>>> (EXEC_OBJECT_SUPPORTS_48B_ADDRESS). >>>>>>> 48-bit range will only be used when explicitly requested. >>>>>>> >>>>>>> Callers to the existing drm_intel_bo_emit_reloc function should >>>>>>> set >>>>>>> the >>>>>>> use_48b_address_range flag beforehand, in order to use full >>>>>>> ppgtt >>>>>>> range. >>>>>>> >>>>>>> v2: Make set/clear functions nops on pre-gen8 platforms, and >>>>>>> use >>>>>>> them >>>>>>> internally in emit_reloc functions (Ben) >>>>>>> s/48BADDRESS/48B_ADDRESS/ (Dave) >>>>>>> v3: Keep set/clear functions internal, no-one needs to use them >>>>>>> directly. >>>>>>> v4: Don't set 48bit-support flag in emit reloc, check for ppgtt >>>>>>> type >>>>>>> before enabling set/clear function, print full offsets >>>>>>> in >>>>>>> debug >>>>>>> statements, using port of lower_32_bits and >>>>>>> upper_32_bits >>>>>>> from linux >>>>>>> kernel (Michał) >>>>>>> >>>>>>> References: >>>>>>> >>>>>>> >>>>>>> >>>>>>> http://lists.freedesktop.org/archives/intel-gfx/2015-July/072612.html >>>>>>> Cc: Ben Widawsky ben@bwidawsk.net >>>>>>> Cc: Michał Winiarski michal.winiarski@intel.com >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> >>>>>> +Kristian >>>>>> >>>>>> LGTM. >>>>>> Acked-by: Michał Winiarski michal.winiarski@intel.com >>>>>> >>>>>>> Signed-off-by: Michel Thierry michel.thierry@intel.com >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> Hi Kristian, >>>>> >>>>> More comments on this? >>>>> I've resent the mesa patch with the last changes >>>>> >>>>> >>>>> >>>>> >>>>> (http://lists.freedesktop.org/archives/dri-devel/2015-October/091752.html). >>>>> >>>>> >>>>> Michał has agreed with the interface and will use it for his >>>>> work. >>>>> If mesa doesn't plan to use this for now, it's ok. The kernel >>>>> changes >>>>> have >>>>> been done to prevent any regressions when the support 48-bit flag >>>>> is >>>>> not >>>>> set. >>>> >>>> >>>> >>>> >>>> >>>> I didn't get any replies to my last comments on this: >>>> >>>> >>>> >>>> http://lists.freedesktop.org/archives/mesa-dev/2015-August/091398.html >>>> >>>> Basically, I tried explicitly placing buffers above 8G and didn't >>>> see >>>> the HW problem described (ie it all worked fine). I still think >>>> there's some confusion as to what the W/A is about. >>>> >>>> Here's my take: the W/A is a SW-only workaround to handle the >>>> cases >>>> where a driver uses heapless and 48-bit ppgtt. The problem is that >>>> the >>>> heaps are limited to 4G but with 48bit ppgtt a buffer can be >>>> placed >>>> anywhere it the 48 bit address space. If that happens it's >>>> consideredd >>>> out-of-bounds for the heap and access fails. To prevent this we >>>> need >>>> to make sure the bos we address in a heapless fashion are placed >>>> below >>>> 4g. >>>> >>>> For mesa, we only configure general state base as heap-less, which >>>> affects scratch bos. What this boils down to is that we need the >>>> 4G >>>> restriction only for the scratch bos set up on 3DSTATE_VS, >>>> 3DSTATE_GS >>>> and 3DSTATE_PS (and tesselation stage eventually). Look for the >>>> OUT_RELOC64 for stage->scratch_bo in gen8_vs_state.c, >>>> gen8_gs_state.c >>>> and gen8_ps_state.c >>> >>> >>> >>> >>> >>> I think it also affects _3DSTATE_VIEWPORT_STATE_POINTERS_CC, maybe >>> it >>> isn't exclusive to the scratch bos (the error state points to that >>> instruction). >>> >>> >> >> Hi, >> >> Anymore inputs about this patch? >> AFAIK, if changes are required based on further comments from >> Kristian, >> these will be in the mesa patch[1], not libdrm. Also, Michał will >> use >> this >> flag in another project. >> > The comment seems quite brief and I'm not sure it fully addresses > Kristian's concern. I'd suspect that providing reference to the HW > documentation (confirming your assumption) might be beneficial. >
Sure, attached is the hang I get if I don't set the restriction in gen8_misc_state.c and try to use the full 48-bit range (i915_error_state_nowa.txt). This is just running gfxbench t-rex.
I see the same hang signature when it is only applied to the scratch bos (in gen8_vs_state.c, gen8_gs_state.c and gen8_ps_state.c - i915_error_state_scratchbo.txt).
3DSTATE_VIEWPORT_STATE_POINTERS_CC (0x78230000) is defined here:
https://01.org/sites/default/files/documentation/intel-gfx-prm-osrc-bdw-vol0... (page 256)
I applied your mesa and libdrm patches and then backed out the 4G restriction in the STATE_BASE_ADDRESS relocations. I was not able to reproduce any hang with trex or glxgears. I confirmed (using INTEL_DEBUG=bat) that gem actually placed BOs above the 4G mark and got this:
0: 8 @ 0xfffffff90000 (miptree) 1: 9 @ 0xfffffff78000 (hiz) 2: 4 @ 0xfffffff77000 (pipe_control workaround) 3: 5 @ 0xfffffff76000 (program cache) 4: 12 @ 0x000000000000 (miptree) 5: 7 @ 0x000000001000 (image) 6: 10 @ 0xfffffff5a000 (miptree) 7: 13 @ 0xfffffff59000 (bufferobj) 8: 11 @ 0xfffffff51000 (bufferobj) 9: 14 @ 0xfffffff50000 (bufferobj) 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000020 -> 8 (miptree)@0x0000ffff fff90000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000040 -> 9 (hiz)@0x0000ffff fff78000 + 0x00000000 10: 15 @ 0xfffffff40000 (batchbuffer)@0x00000000 00000098 -> 4 (pipe_control workaround)@0x0000ffff fff77000 + 0x00000000 ...
In particular, the mesa batchbuffer is placed at 0xfffffff40000, and all the STATE_BASE_ADDRESS heaps are in that bo. In other words, all heap bases are set to 0xfffffff40000.
The offset in 3DSTATE_VIEWPORT_STATE_POINTERS_CC is indeed only 32 bits, but that only affects how far from the dynamic state base address we can place that. In mesas case, it's always inside the batch buffer, which is at most 32k, so that's never a problem. If a driver uses dynamic state in a heapless fashion, then you need to be careful to place the CC viewport state below 4g.
What I was able to confirm is that scratch buffer I/O (which we use for spilling) does break with 48 bit ppgtt. If you run glxgears with INTEL_DEBUG=spill_fs set, you can see lots of rendering artifacts as the 4G limit on the general state heap caps attempts to spill and fill from the vertex shader. Using OUT_RELOC64_INSIDE_4G() when setting up the scratch BOs in 3DSTATE_VS and 3DSTATE_PS fixes the problem.
Ok, so we do still need the 4G limit bit (i.e. I don't have to revert the kernel side for lack of open-source userspace), just not at the place we originally thought. I guess lesson learned is to actually test stuff first before I pull in the kernel side. Michel, did you not see the spilling corruptions when testing the mesa patches? Kristian, does this also blow up with just a piglit run?
There's must be something else we have different in our systems.
Kristian doesn't get the hang, and if I back out the 4G restriction in the STATE_BASE_ADDRESS, I don't see any corruption while running INTEL_DEBUG=spill_fs glxgears.
I don't want to block this, in fact, I very much want to see this move forward. However, I can't ack patches that 1) fix a gpu hang I can't reproduce and I don't agree should be a problem (the STATE_BASE_ADDRESS probllem) and 2) don't fix the issue that I see and reproduced on my BDW (the spilling problem).
I'll try a full piglit run tomorrow with both my mesa patch and Michels and see what happens. However, this isn't the kind of problem that occurs once in a blue moon or for some corner case piglit test. If Michels understanding of the hw and the workaround is correct, nothing should work without his patch.
To move forward on this, unless something new comes up when running piglit, I can take ownership of making mesa work one way or the other. The kernel and libdrm work is fine and most of Michels mesa patch is good. If mesa breaks down the road because we need to limit the STATE_BASE_ADDRESS relocs, I'll own the problem then.
Kristian
Thanks Kristian,
Let me know if you need something from i915 or libdrm side.
I've reviewed and pushed these two patches. I'll send out the mesa patch for review in a bit and once that lands we can do a libdrm release.
Kristian
Signed-off-by: Michel Thierry michel.thierry@intel.com --- intel/intel-symbol-check | 1 + 1 file changed, 1 insertion(+)
diff --git a/intel/intel-symbol-check b/intel/intel-symbol-check index c555e6d..64ec4ed 100755 --- a/intel/intel-symbol-check +++ b/intel/intel-symbol-check @@ -39,6 +39,7 @@ drm_intel_bo_subdata drm_intel_bo_unmap drm_intel_bo_unpin drm_intel_bo_unreference +drm_intel_bo_use_48b_address_range drm_intel_bo_wait_rendering drm_intel_bufmgr_check_aperture_space drm_intel_bufmgr_destroy
HI all,
On 3 September 2015 at 15:23, Michel Thierry michel.thierry@intel.com wrote:
This new version picks the suggestion from Michał Winiarski related to v3 [1].
48-bit virtual address range will be enabled in i915 soon, but some objects must be referenced by 32-bit offsets. These patches use a new kernel flag to specify if this restriction applies or not.
I'm sending these patches to comply with the i915 merge process. Once the kernel patch is merged, I'll make a new libdrm release and address the mesa build dependency.
[1] http://lists.freedesktop.org/archives/dri-devel/2015-August/087837.html
Michel Thierry (2): intel: 48b ppgtt support (EXEC_OBJECT_SUPPORTS_48B_ADDRESS flag) intel: add drm_intel_bo_use_48b_address_range to symbol-check test
Can we get some eyes on these and the mesa patches ?
Michel, Afaics Kristian had some concerns wrt the mesa patches. Did you address these and forgot to keep mesa-dev in the CC list ?
-Emil
P.S. Michel, lets keep patch discussion on the ML rather than bugzilla ;-)
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Emil Velikov -
Kristian Høgsberg -
Michał Winiarski -
Michel Thierry