Commit 67d7dd59 authored by José Casanova Crespo's avatar José Casanova Crespo Committed by Jason Ekstrand

isl/i965/fs: SSBO/UBO buffers need size padding if not multiple of 32-bit

The surfaces that backup the GPU buffers have a boundary check that
considers that access to partial dwords are considered out-of-bounds.
For example, buffers with 1,3 16-bit elements has size 2 or 6 and the
last two bytes would always be read as 0 or its writting ignored.

The introduction of 16-bit types implies that we need to align the size
to 4-bytew multiples so that partial dwords could be read/written.
Adding an inconditional +2 size to buffers not being multiple of 2
solves this issue for the general cases of UBO or SSBO.

But, when unsized arrays of 16-bit elements are used it is not possible
to know if the size was padded or not. To solve this issue the
implementation calculates the needed size of the buffer surfaces,
as suggested by Jason:

surface_size = isl_align(buffer_size, 4) +
               (isl_align(buffer_size, 4) - buffer_size)

So when we calculate backwards the buffer_size in the backend we
update the resinfo return value with:

buffer_size = (surface_size & ~3) - (surface_size & 3)

It is also exposed this buffer requirements when robust buffer access
is enabled so these buffer sizes recommend being multiple of 4.

v2: (Jason Ekstrand)
    Move padding logic fron anv to isl_surface_state.
    Move calculus of original size from spirv to driver backend.
v3: (Jason Ekstrand)
    Rename some variables and use a similar expresion when calculating.
    padding than when obtaining the original buffer size.
    Avoid use of unnecesary component call at brw_fs_nir.
v4: (Jason Ekstrand)
    Complete comment with buffer size calculus explanation in brw_fs_nir.
Reviewed-by: Jason Ekstrand's avatarJason Ekstrand <>
parent 4c232dc7
......@@ -4290,7 +4290,36 @@ fs_visitor::nir_emit_intrinsic(const fs_builder &bld, nir_intrinsic_instr *instr
inst->mlen = 1;
inst->size_written = 4 * REG_SIZE;
bld.MOV(retype(dest, ret_payload.type), component(ret_payload, 0));
/* SKL PRM, vol07, 3D Media GPGPU Engine, Bounds Checking and Faulting:
* "Out-of-bounds checking is always performed at a DWord granularity. If
* any part of the DWord is out-of-bounds then the whole DWord is
* considered out-of-bounds."
* This implies that types with size smaller than 4-bytes need to be
* padded if they don't complete the last dword of the buffer. But as we
* need to maintain the original size we need to reverse the padding
* calculation to return the correct size to know the number of elements
* of an unsized array. As we stored in the last two bits of the surface
* size the needed padding for the buffer, we calculate here the
* original buffer_size reversing the surface_size calculation:
* surface_size = isl_align(buffer_size, 4) +
* (isl_align(buffer_size) - buffer_size)
* buffer_size = surface_size & ~3 - surface_size & 3
fs_reg size_aligned4 = ubld.vgrf(BRW_REGISTER_TYPE_UD);
fs_reg size_padding = ubld.vgrf(BRW_REGISTER_TYPE_UD);
fs_reg buffer_size = ubld.vgrf(BRW_REGISTER_TYPE_UD);
ubld.AND(size_padding, ret_payload, brw_imm_ud(3));
ubld.AND(size_aligned4, ret_payload, brw_imm_ud(~3));
ubld.ADD(buffer_size, size_aligned4, negate(size_padding));
bld.MOV(retype(dest, ret_payload.type), component(buffer_size, 0));
brw_mark_surface_used(prog_data, index);
......@@ -673,7 +673,27 @@ void
isl_genX(buffer_fill_state_s)(void *state,
const struct isl_buffer_fill_state_info *restrict info)
uint32_t num_elements = info->size / info->stride;
uint64_t buffer_size = info->size;
/* Uniform and Storage buffers need to have surface size not less that the
* aligned 32-bit size of the buffer. To calculate the array lenght on
* unsized arrays in StorageBuffer the last 2 bits store the padding size
* added to the surface, so we can calculate latter the original buffer
* size to know the number of elements.
* surface_size = isl_align(buffer_size, 4) +
* (isl_align(buffer_size) - buffer_size)
* buffer_size = (surface_size & ~3) - (surface_size & 3)
if (info->format == ISL_FORMAT_RAW ||
info->stride < isl_format_get_layout(info->format)->bpb / 8) {
assert(info->stride == 1);
uint64_t aligned_size = isl_align(buffer_size, 4);
buffer_size = aligned_size + (aligned_size - buffer_size);
uint32_t num_elements = buffer_size / info->stride;
if (GEN_GEN >= 7) {
/* From the IVB PRM, SURFACE_STATE::Height,
......@@ -2155,6 +2155,17 @@ void anv_GetBufferMemoryRequirements(
pMemoryRequirements->size = buffer->size;
pMemoryRequirements->alignment = alignment;
/* Storage and Uniform buffers should have their size aligned to
* 32-bits to avoid boundary checks when last DWord is not complete.
* This would ensure that not internal padding would be needed for
* 16-bit types.
if (device->robust_buffer_access &&
pMemoryRequirements->size = align_u64(buffer->size, 4);
pMemoryRequirements->memoryTypeBits = memory_types;
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment