From d76c204d0564701b4b8b6a2bdda50e2939683e66 Mon Sep 17 00:00:00 2001 From: Ian Romanick Date: Mon, 13 Nov 2017 11:17:41 -0800 Subject: [PATCH] util: Move util_is_power_of_two to bitscan.h and rename to util_is_power_of_two_or_zero MIME-Version: 1.0 Content-Type: text/plain; charset=utf8 Content-Transfer-Encoding: 8bit The new name make the zero-input behavior more obvious. The next patch adds a new function with different zero-input behavior. Signed-off-by: Ian Romanick Suggested-by: Matt Turner Reviewed-by: Alejandro Piñeiro --- src/amd/common/ac_gpu_info.c | 4 ++-- src/amd/common/ac_surface.c | 2 +- src/amd/vulkan/radv_formats.c | 4 ++-- src/broadcom/compiler/nir_to_vir.c | 4 ++-- src/gallium/auxiliary/gallivm/lp_bld_arit.c | 2 +- src/gallium/auxiliary/gallivm/lp_bld_debug.cpp | 2 +- src/gallium/auxiliary/gallivm/lp_bld_format_aos.c | 4 ++-- src/gallium/auxiliary/gallivm/lp_bld_gather.c | 8 ++++---- src/gallium/auxiliary/gallivm/lp_bld_pack.c | 2 +- src/gallium/auxiliary/gallivm/lp_bld_sample.c | 6 +++--- src/gallium/auxiliary/util/u_math.h | 10 +--------- src/gallium/auxiliary/util/u_ringbuffer.c | 2 +- src/gallium/drivers/etnaviv/etnaviv_blt.c | 2 +- src/gallium/drivers/etnaviv/etnaviv_texture_state.c | 3 ++- src/gallium/drivers/freedreno/freedreno_query_hw.c | 2 +- src/gallium/drivers/i915/i915_state_sampler.c | 3 ++- src/gallium/drivers/llvmpipe/lp_state_fs.c | 2 +- src/gallium/drivers/llvmpipe/lp_texture.c | 2 +- src/gallium/drivers/nouveau/codegen/nv50_ir.cpp | 2 +- .../drivers/nouveau/codegen/nv50_ir_peephole.cpp | 4 ++-- src/gallium/drivers/nouveau/nv30/nv30_miptree.c | 6 +++--- .../drivers/nouveau/nvc0/nvc0_state_validate.c | 2 +- src/gallium/drivers/r300/r300_texture_desc.c | 6 +++--- src/gallium/drivers/r600/r600_texture.c | 2 +- src/gallium/drivers/radeon/r600_texture.c | 2 +- src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c | 2 +- src/gallium/drivers/softpipe/sp_texture.c | 12 ++++++------ src/gallium/drivers/swr/swr_screen.cpp | 4 ++-- src/gallium/drivers/vc4/vc4_program.c | 4 ++-- src/intel/compiler/brw_fs.cpp | 4 ++-- src/intel/vulkan/anv_allocator.c | 6 +++--- src/intel/vulkan/anv_formats.c | 4 ++-- src/intel/vulkan/anv_nir_lower_multiview.c | 2 +- src/mesa/state_tracker/st_cb_readpixels.c | 4 ++-- src/mesa/state_tracker/st_cb_texture.c | 6 +++--- src/util/bitscan.h | 12 ++++++++++++ src/util/u_vector.c | 4 ++-- 37 files changed, 79 insertions(+), 73 deletions(-) diff --git a/src/amd/common/ac_gpu_info.c b/src/amd/common/ac_gpu_info.c index 73fc36203c1..f93ef4cbade 100644 --- a/src/amd/common/ac_gpu_info.c +++ b/src/amd/common/ac_gpu_info.c @@ -333,8 +333,8 @@ bool ac_query_gpu_info(int fd, amdgpu_device_handle dev, } info->has_virtual_memory = true; - assert(util_is_power_of_two(dma.available_rings + 1)); - assert(util_is_power_of_two(compute.available_rings + 1)); + assert(util_is_power_of_two_or_zero(dma.available_rings + 1)); + assert(util_is_power_of_two_or_zero(compute.available_rings + 1)); info->num_sdma_rings = util_bitcount(dma.available_rings); info->num_compute_rings = util_bitcount(compute.available_rings); diff --git a/src/amd/common/ac_surface.c b/src/amd/common/ac_surface.c index 81882576baf..1e1641daeec 100644 --- a/src/amd/common/ac_surface.c +++ b/src/amd/common/ac_surface.c @@ -275,7 +275,7 @@ static int gfx6_compute_level(ADDR_HANDLE addrlib, AddrSurfInfoIn->bpp) { unsigned alignment = 256 / (AddrSurfInfoIn->bpp / 8); - assert(util_is_power_of_two(AddrSurfInfoIn->bpp)); + assert(util_is_power_of_two_or_zero(AddrSurfInfoIn->bpp)); AddrSurfInfoIn->width = align(AddrSurfInfoIn->width, alignment); } diff --git a/src/amd/vulkan/radv_formats.c b/src/amd/vulkan/radv_formats.c index efb1d78790a..1bafe09e77f 100644 --- a/src/amd/vulkan/radv_formats.c +++ b/src/amd/vulkan/radv_formats.c @@ -638,13 +638,13 @@ radv_physical_device_get_format_properties(struct radv_physical_device *physical tiled |= VK_FORMAT_FEATURE_COLOR_ATTACHMENT_BLEND_BIT; } } - if (tiled && util_is_power_of_two(vk_format_get_blocksize(format)) && !scaled) { + if (tiled && util_is_power_of_two_or_zero(vk_format_get_blocksize(format)) && !scaled) { tiled |= VK_FORMAT_FEATURE_TRANSFER_SRC_BIT_KHR | VK_FORMAT_FEATURE_TRANSFER_DST_BIT_KHR; } } - if (linear && util_is_power_of_two(vk_format_get_blocksize(format)) && !scaled) { + if (linear && util_is_power_of_two_or_zero(vk_format_get_blocksize(format)) && !scaled) { linear |= VK_FORMAT_FEATURE_TRANSFER_SRC_BIT_KHR | VK_FORMAT_FEATURE_TRANSFER_DST_BIT_KHR; } diff --git a/src/broadcom/compiler/nir_to_vir.c b/src/broadcom/compiler/nir_to_vir.c index 51710001784..d6c2d1902b6 100644 --- a/src/broadcom/compiler/nir_to_vir.c +++ b/src/broadcom/compiler/nir_to_vir.c @@ -233,7 +233,7 @@ static struct qreg ntq_get_alu_src(struct v3d_compile *c, nir_alu_instr *instr, unsigned src) { - assert(util_is_power_of_two(instr->dest.write_mask)); + assert(util_is_power_of_two_or_zero(instr->dest.write_mask)); unsigned chan = ffs(instr->dest.write_mask) - 1; struct qreg r = ntq_get_src(c, instr->src[src].src, instr->src[src].swizzle[chan]); @@ -862,7 +862,7 @@ ntq_emit_alu(struct v3d_compile *c, nir_alu_instr *instr) /* We have a scalar result, so the instruction should only have a * single channel written to. */ - assert(util_is_power_of_two(instr->dest.write_mask)); + assert(util_is_power_of_two_or_zero(instr->dest.write_mask)); ntq_store_dest(c, &instr->dest.dest, ffs(instr->dest.write_mask) - 1, result); } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_arit.c b/src/gallium/auxiliary/gallivm/lp_bld_arit.c index 321c6e4edf0..e922474ef61 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_arit.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_arit.c @@ -1307,7 +1307,7 @@ lp_build_mul_imm(struct lp_build_context *bld, if(b == 2 && bld->type.floating) return lp_build_add(bld, a, a); - if(util_is_power_of_two(b)) { + if(util_is_power_of_two_or_zero(b)) { unsigned shift = ffs(b) - 1; if(bld->type.floating) { diff --git a/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp b/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp index f311fe7f690..23ada3d0433 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp +++ b/src/gallium/auxiliary/gallivm/lp_bld_debug.cpp @@ -60,7 +60,7 @@ extern "C" boolean lp_check_alignment(const void *ptr, unsigned alignment) { - assert(util_is_power_of_two(alignment)); + assert(util_is_power_of_two_or_zero(alignment)); return ((uintptr_t)ptr & (alignment - 1)) == 0; } diff --git a/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c b/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c index 36dedba34f0..b52acca1b3e 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_format_aos.c @@ -496,7 +496,7 @@ lp_build_fetch_rgba_aos(struct gallivm_state *gallivm, if (format_matches_type(format_desc, type) && format_desc->block.bits <= type.width * 4 && /* XXX this shouldn't be needed */ - util_is_power_of_two(format_desc->block.bits)) { + util_is_power_of_two_or_zero(format_desc->block.bits)) { LLVMValueRef packed; LLVMTypeRef dst_vec_type = lp_build_vec_type(gallivm, type); struct lp_type fetch_type; @@ -609,7 +609,7 @@ lp_build_fetch_rgba_aos(struct gallivm_state *gallivm, format_desc->block.width == 1 && format_desc->block.height == 1 && /* XXX this shouldn't be needed */ - util_is_power_of_two(format_desc->block.bits) && + util_is_power_of_two_or_zero(format_desc->block.bits) && format_desc->block.bits <= 32 && format_desc->is_bitmask && !format_desc->is_mixed && diff --git a/src/gallium/auxiliary/gallivm/lp_bld_gather.c b/src/gallium/auxiliary/gallivm/lp_bld_gather.c index 7d11dcd3b64..8cabe9ef01e 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_gather.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_gather.c @@ -118,7 +118,7 @@ lp_build_gather_elem(struct gallivm_state *gallivm, */ if (!aligned) { LLVMSetAlignment(res, 1); - } else if (!util_is_power_of_two(src_width)) { + } else if (!util_is_power_of_two_or_zero(src_width)) { /* * Full alignment is impossible, assume the caller really meant * the individual elements were aligned (e.g. 3x32bit format). @@ -130,7 +130,7 @@ lp_build_gather_elem(struct gallivm_state *gallivm, * this should cover all the 3-channel formats. */ if (((src_width / 24) * 24 == src_width) && - util_is_power_of_two(src_width / 24)) { + util_is_power_of_two_or_zero(src_width / 24)) { LLVMSetAlignment(res, src_width / 24); } else { LLVMSetAlignment(res, 1); @@ -199,7 +199,7 @@ lp_build_gather_elem_vec(struct gallivm_state *gallivm, */ if (!aligned) { LLVMSetAlignment(res, 1); - } else if (!util_is_power_of_two(src_width)) { + } else if (!util_is_power_of_two_or_zero(src_width)) { /* * Full alignment is impossible, assume the caller really meant * the individual elements were aligned (e.g. 3x32bit format). @@ -211,7 +211,7 @@ lp_build_gather_elem_vec(struct gallivm_state *gallivm, * this should cover all the 3-channel formats. */ if (((src_width / 24) * 24 == src_width) && - util_is_power_of_two(src_width / 24)) { + util_is_power_of_two_or_zero(src_width / 24)) { LLVMSetAlignment(res, src_width / 24); } else { LLVMSetAlignment(res, 1); diff --git a/src/gallium/auxiliary/gallivm/lp_bld_pack.c b/src/gallium/auxiliary/gallivm/lp_bld_pack.c index 7879826422d..b8b53a7d6eb 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_pack.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_pack.c @@ -219,7 +219,7 @@ lp_build_concat(struct gallivm_state *gallivm, LLVMValueRef shuffles[LP_MAX_VECTOR_LENGTH]; assert(src_type.length * num_vectors <= ARRAY_SIZE(shuffles)); - assert(util_is_power_of_two(num_vectors)); + assert(util_is_power_of_two_or_zero(num_vectors)); new_length = src_type.length; diff --git a/src/gallium/auxiliary/gallivm/lp_bld_sample.c b/src/gallium/auxiliary/gallivm/lp_bld_sample.c index ab9d051c911..81cb5060711 100644 --- a/src/gallium/auxiliary/gallivm/lp_bld_sample.c +++ b/src/gallium/auxiliary/gallivm/lp_bld_sample.c @@ -114,9 +114,9 @@ lp_sampler_static_texture_state(struct lp_static_texture_state *state, state->swizzle_a = view->swizzle_a; state->target = view->target; - state->pot_width = util_is_power_of_two(texture->width0); - state->pot_height = util_is_power_of_two(texture->height0); - state->pot_depth = util_is_power_of_two(texture->depth0); + state->pot_width = util_is_power_of_two_or_zero(texture->width0); + state->pot_height = util_is_power_of_two_or_zero(texture->height0); + state->pot_depth = util_is_power_of_two_or_zero(texture->depth0); state->level_zero_only = !view->u.tex.last_level; /* diff --git a/src/gallium/auxiliary/util/u_math.h b/src/gallium/auxiliary/util/u_math.h index a441b5457b2..46d02978fd6 100644 --- a/src/gallium/auxiliary/util/u_math.h +++ b/src/gallium/auxiliary/util/u_math.h @@ -179,14 +179,6 @@ util_fast_pow(float x, float y) return util_fast_exp2(util_fast_log2(x) * y); } -/* Note that this counts zero as a power of two. - */ -static inline boolean -util_is_power_of_two( unsigned v ) -{ - return (v & (v-1)) == 0; -} - /** * Floor(x), returned as int. @@ -459,7 +451,7 @@ util_next_power_of_two(unsigned x) if (x <= 1) return 1; - if (util_is_power_of_two(x)) + if (util_is_power_of_two_or_zero(x)) return x; val--; diff --git a/src/gallium/auxiliary/util/u_ringbuffer.c b/src/gallium/auxiliary/util/u_ringbuffer.c index 4d6166833e4..f6bb910671e 100644 --- a/src/gallium/auxiliary/util/u_ringbuffer.c +++ b/src/gallium/auxiliary/util/u_ringbuffer.c @@ -27,7 +27,7 @@ struct util_ringbuffer *util_ringbuffer_create( unsigned dwords ) if (!ring) return NULL; - assert(util_is_power_of_two(dwords)); + assert(util_is_power_of_two_or_zero(dwords)); ring->buf = MALLOC( dwords * sizeof(unsigned) ); if (ring->buf == NULL) diff --git a/src/gallium/drivers/etnaviv/etnaviv_blt.c b/src/gallium/drivers/etnaviv/etnaviv_blt.c index 5d783a4ad9c..c30c11ab61e 100644 --- a/src/gallium/drivers/etnaviv/etnaviv_blt.c +++ b/src/gallium/drivers/etnaviv/etnaviv_blt.c @@ -178,7 +178,7 @@ emit_blt_copyimage(struct etna_cmd_stream *stream, const struct blt_imgcopy_op * static void emit_blt_inplace(struct etna_cmd_stream *stream, const struct blt_inplace_op *op) { - assert(op->bpp > 0 && util_is_power_of_two(op->bpp)); + assert(op->bpp > 0 && util_is_power_of_two_or_zero(op->bpp)); etna_cmd_stream_reserve(stream, 64*2); /* Never allow BLT sequences to be broken up */ etna_set_state(stream, VIVS_BLT_ENABLE, 0x00000001); etna_set_state(stream, VIVS_BLT_CONFIG, diff --git a/src/gallium/drivers/etnaviv/etnaviv_texture_state.c b/src/gallium/drivers/etnaviv/etnaviv_texture_state.c index faa073a71af..7100865f925 100644 --- a/src/gallium/drivers/etnaviv/etnaviv_texture_state.c +++ b/src/gallium/drivers/etnaviv/etnaviv_texture_state.c @@ -158,7 +158,8 @@ etna_create_sampler_view_state(struct pipe_context *pctx, struct pipe_resource * /* Workaround for npot textures -- it appears that only CLAMP_TO_EDGE is * supported when the appropriate capability is not set. */ if (!ctx->specs.npot_tex_any_wrap && - (!util_is_power_of_two(res->base.width0) || !util_is_power_of_two(res->base.height0))) { + (!util_is_power_of_two_or_zero(res->base.width0) || + !util_is_power_of_two_or_zero(res->base.height0))) { sv->TE_SAMPLER_CONFIG0_MASK = ~(VIVS_TE_SAMPLER_CONFIG0_UWRAP__MASK | VIVS_TE_SAMPLER_CONFIG0_VWRAP__MASK); sv->TE_SAMPLER_CONFIG0 |= diff --git a/src/gallium/drivers/freedreno/freedreno_query_hw.c b/src/gallium/drivers/freedreno/freedreno_query_hw.c index 8b25e9cbcca..86a46faa730 100644 --- a/src/gallium/drivers/freedreno/freedreno_query_hw.c +++ b/src/gallium/drivers/freedreno/freedreno_query_hw.c @@ -301,7 +301,7 @@ fd_hw_sample_init(struct fd_batch *batch, uint32_t size) struct fd_hw_sample *samp = slab_alloc_st(&batch->ctx->sample_pool); pipe_reference_init(&samp->reference, 1); samp->size = size; - debug_assert(util_is_power_of_two(size)); + debug_assert(util_is_power_of_two_or_zero(size)); batch->next_sample_offset = align(batch->next_sample_offset, size); samp->offset = batch->next_sample_offset; /* NOTE: slab_alloc_st() does not zero out the buffer: */ diff --git a/src/gallium/drivers/i915/i915_state_sampler.c b/src/gallium/drivers/i915/i915_state_sampler.c index 84ed1514630..66a5778717a 100644 --- a/src/gallium/drivers/i915/i915_state_sampler.c +++ b/src/gallium/drivers/i915/i915_state_sampler.c @@ -307,7 +307,8 @@ static void update_map(struct i915_context *i915, int first_level = view->u.tex.first_level; const uint num_levels = pt->last_level - first_level; unsigned max_lod = num_levels * 4; - bool is_npot = (!util_is_power_of_two(pt->width0) || !util_is_power_of_two(pt->height0)); + bool is_npot = (!util_is_power_of_two_or_zero(pt->width0) || + !util_is_power_of_two_or_zero(pt->height0)); uint format, pitch; /* diff --git a/src/gallium/drivers/llvmpipe/lp_state_fs.c b/src/gallium/drivers/llvmpipe/lp_state_fs.c index 66645b07ac7..74b8d4dd96e 100644 --- a/src/gallium/drivers/llvmpipe/lp_state_fs.c +++ b/src/gallium/drivers/llvmpipe/lp_state_fs.c @@ -2240,7 +2240,7 @@ generate_unswizzled_blend(struct gallivm_state *gallivm, if (dst_count > src_count) { if ((dst_type.width == 8 || dst_type.width == 16) && - util_is_power_of_two(dst_type.length) && + util_is_power_of_two_or_zero(dst_type.length) && dst_type.length * dst_type.width < 128) { /* * Never try to load values as 4xi8 which we will then diff --git a/src/gallium/drivers/llvmpipe/lp_texture.c b/src/gallium/drivers/llvmpipe/lp_texture.c index 162c74ad7dd..89852cc95c3 100644 --- a/src/gallium/drivers/llvmpipe/lp_texture.c +++ b/src/gallium/drivers/llvmpipe/lp_texture.c @@ -670,7 +670,7 @@ llvmpipe_get_format_alignment( enum pipe_format format ) bytes = size / 8; - if (!util_is_power_of_two(bytes)) { + if (!util_is_power_of_two_or_zero(bytes)) { bytes /= desc->nr_channels; } diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir.cpp index 6f12df70a11..c987da99085 100644 --- a/src/gallium/drivers/nouveau/codegen/nv50_ir.cpp +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir.cpp @@ -423,7 +423,7 @@ ImmediateValue::isNegative() const bool ImmediateValue::isPow2() const { - return util_is_power_of_two(reg.data.u32); + return util_is_power_of_two_or_zero(reg.data.u32); } void diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp index 48cf74950df..39177bd044b 100644 --- a/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp +++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_peephole.cpp @@ -1305,7 +1305,7 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) src->op == OP_SHR && src->src(1).getImmediate(imm1) && i->src(t).mod == Modifier(0) && - util_is_power_of_two(imm0.reg.data.u32 + 1)) { + util_is_power_of_two_or_zero(imm0.reg.data.u32 + 1)) { // low byte = offset, high byte = width uint32_t ext = (util_last_bit(imm0.reg.data.u32) << 8) | imm1.reg.data.u32; i->op = OP_EXTBF; @@ -1314,7 +1314,7 @@ ConstantFolding::opnd(Instruction *i, ImmediateValue &imm0, int s) } else if (src->op == OP_SHL && src->src(1).getImmediate(imm1) && i->src(t).mod == Modifier(0) && - util_is_power_of_two(~imm0.reg.data.u32 + 1) && + util_is_power_of_two_or_zero(~imm0.reg.data.u32 + 1) && util_last_bit(~imm0.reg.data.u32) <= imm1.reg.data.u32) { i->op = OP_MOV; i->setSrc(s, NULL); diff --git a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c index 165b8f29b4b..4f991776323 100644 --- a/src/gallium/drivers/nouveau/nv30/nv30_miptree.c +++ b/src/gallium/drivers/nouveau/nv30/nv30_miptree.c @@ -401,9 +401,9 @@ nv30_miptree_create(struct pipe_screen *pscreen, if ((pt->target == PIPE_TEXTURE_RECT) || (pt->bind & PIPE_BIND_SCANOUT) || - !util_is_power_of_two(pt->width0) || - !util_is_power_of_two(pt->height0) || - !util_is_power_of_two(pt->depth0) || + !util_is_power_of_two_or_zero(pt->width0) || + !util_is_power_of_two_or_zero(pt->height0) || + !util_is_power_of_two_or_zero(pt->depth0) || util_format_is_compressed(pt->format) || util_format_is_float(pt->format) || mt->ms_mode) { mt->uniform_pitch = util_format_get_nblocksx(pt->format, w) * blocksz; diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c b/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c index 37a67619588..8e2192d3de2 100644 --- a/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c +++ b/src/gallium/drivers/nouveau/nvc0/nvc0_state_validate.c @@ -183,7 +183,7 @@ nvc0_validate_fb(struct nvc0_context *nvc0) } if (nr_cbufs == 0 && !fb->zsbuf) { - assert(util_is_power_of_two(fb->samples)); + assert(util_is_power_of_two_or_zero(fb->samples)); assert(fb->samples <= 8); nvc0_fb_set_null_rt(push, 0, fb->layers); diff --git a/src/gallium/drivers/r300/r300_texture_desc.c b/src/gallium/drivers/r300/r300_texture_desc.c index 2442d726cd1..37e7b5fc4e2 100644 --- a/src/gallium/drivers/r300/r300_texture_desc.c +++ b/src/gallium/drivers/r300/r300_texture_desc.c @@ -270,15 +270,15 @@ static void r300_setup_miptree(struct r300_screen *screen, static void r300_setup_flags(struct r300_resource *tex) { tex->tex.uses_stride_addressing = - !util_is_power_of_two(tex->b.b.width0) || + !util_is_power_of_two_or_zero(tex->b.b.width0) || (tex->tex.stride_in_bytes_override && r300_stride_to_width(tex->b.b.format, tex->tex.stride_in_bytes_override) != tex->b.b.width0); tex->tex.is_npot = tex->tex.uses_stride_addressing || - !util_is_power_of_two(tex->b.b.height0) || - !util_is_power_of_two(tex->b.b.depth0); + !util_is_power_of_two_or_zero(tex->b.b.height0) || + !util_is_power_of_two_or_zero(tex->b.b.depth0); } static void r300_setup_cbzb_flags(struct r300_screen *rscreen, diff --git a/src/gallium/drivers/r600/r600_texture.c b/src/gallium/drivers/r600/r600_texture.c index 806bc278b08..c39c00c2e3e 100644 --- a/src/gallium/drivers/r600/r600_texture.c +++ b/src/gallium/drivers/r600/r600_texture.c @@ -217,7 +217,7 @@ static int r600_init_surface(struct r600_common_screen *rscreen, bpe = 4; /* stencil is allocated separately on evergreen */ } else { bpe = util_format_get_blocksize(ptex->format); - assert(util_is_power_of_two(bpe)); + assert(util_is_power_of_two_or_zero(bpe)); } if (!is_flushed_depth && is_depth) { diff --git a/src/gallium/drivers/radeon/r600_texture.c b/src/gallium/drivers/radeon/r600_texture.c index eb6706c9dcd..50c5dbe3e1a 100644 --- a/src/gallium/drivers/radeon/r600_texture.c +++ b/src/gallium/drivers/radeon/r600_texture.c @@ -240,7 +240,7 @@ static int r600_init_surface(struct si_screen *sscreen, bpe = 4; /* stencil is allocated separately on evergreen */ } else { bpe = util_format_get_blocksize(ptex->format); - assert(util_is_power_of_two(bpe)); + assert(util_is_power_of_two_or_zero(bpe)); } if (!is_flushed_depth && is_depth) { diff --git a/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c b/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c index 90cc2e0d981..79fdebe8388 100644 --- a/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c +++ b/src/gallium/drivers/radeonsi/si_shader_tgsi_setup.c @@ -192,7 +192,7 @@ LLVMValueRef si_llvm_bound_index(struct si_shader_context *ctx, LLVMValueRef c_max = LLVMConstInt(ctx->i32, num - 1, 0); LLVMValueRef cc; - if (util_is_power_of_two(num)) { + if (util_is_power_of_two_or_zero(num)) { index = LLVMBuildAnd(builder, index, c_max, ""); } else { /* In theory, this MAX pattern should result in code that is diff --git a/src/gallium/drivers/softpipe/sp_texture.c b/src/gallium/drivers/softpipe/sp_texture.c index ea5e2c64b84..c49bfcaba55 100644 --- a/src/gallium/drivers/softpipe/sp_texture.c +++ b/src/gallium/drivers/softpipe/sp_texture.c @@ -166,9 +166,9 @@ softpipe_resource_create_front(struct pipe_screen *screen, pipe_reference_init(&spr->base.reference, 1); spr->base.screen = screen; - spr->pot = (util_is_power_of_two(templat->width0) && - util_is_power_of_two(templat->height0) && - util_is_power_of_two(templat->depth0)); + spr->pot = (util_is_power_of_two_or_zero(templat->width0) && + util_is_power_of_two_or_zero(templat->height0) && + util_is_power_of_two_or_zero(templat->depth0)); if (spr->base.bind & (PIPE_BIND_DISPLAY_TARGET | PIPE_BIND_SCANOUT | @@ -231,9 +231,9 @@ softpipe_resource_from_handle(struct pipe_screen *screen, pipe_reference_init(&spr->base.reference, 1); spr->base.screen = screen; - spr->pot = (util_is_power_of_two(templat->width0) && - util_is_power_of_two(templat->height0) && - util_is_power_of_two(templat->depth0)); + spr->pot = (util_is_power_of_two_or_zero(templat->width0) && + util_is_power_of_two_or_zero(templat->height0) && + util_is_power_of_two_or_zero(templat->depth0)); spr->dt = winsys->displaytarget_from_handle(winsys, templat, diff --git a/src/gallium/drivers/swr/swr_screen.cpp b/src/gallium/drivers/swr/swr_screen.cpp index dd2d003d15d..880a177c399 100644 --- a/src/gallium/drivers/swr/swr_screen.cpp +++ b/src/gallium/drivers/swr/swr_screen.cpp @@ -105,7 +105,7 @@ swr_is_format_supported(struct pipe_screen *_screen, return FALSE; if ((sample_count > screen->msaa_max_count) - || !util_is_power_of_two(sample_count)) + || !util_is_power_of_two_or_zero(sample_count)) return FALSE; if (bind & PIPE_BIND_DISPLAY_TARGET) { @@ -1103,7 +1103,7 @@ swr_validate_env_options(struct swr_screen *screen) int msaa_max_count = debug_get_num_option("SWR_MSAA_MAX_COUNT", 1); if (msaa_max_count != 1) { if ((msaa_max_count < 1) || (msaa_max_count > SWR_MAX_NUM_MULTISAMPLES) - || !util_is_power_of_two(msaa_max_count)) { + || !util_is_power_of_two_or_zero(msaa_max_count)) { fprintf(stderr, "SWR_MSAA_MAX_COUNT invalid: %d\n", msaa_max_count); fprintf(stderr, "must be power of 2 between 1 and %d" \ " (or 1 to disable msaa)\n", diff --git a/src/gallium/drivers/vc4/vc4_program.c b/src/gallium/drivers/vc4/vc4_program.c index 2ec6aa471d4..367613130bb 100644 --- a/src/gallium/drivers/vc4/vc4_program.c +++ b/src/gallium/drivers/vc4/vc4_program.c @@ -313,7 +313,7 @@ static struct qreg ntq_get_alu_src(struct vc4_compile *c, nir_alu_instr *instr, unsigned src) { - assert(util_is_power_of_two(instr->dest.write_mask)); + assert(util_is_power_of_two_or_zero(instr->dest.write_mask)); unsigned chan = ffs(instr->dest.write_mask) - 1; struct qreg r = ntq_get_src(c, instr->src[src].src, instr->src[src].swizzle[chan]); @@ -1363,7 +1363,7 @@ ntq_emit_alu(struct vc4_compile *c, nir_alu_instr *instr) /* We have a scalar result, so the instruction should only have a * single channel written to. */ - assert(util_is_power_of_two(instr->dest.write_mask)); + assert(util_is_power_of_two_or_zero(instr->dest.write_mask)); ntq_store_dest(c, &instr->dest.dest, ffs(instr->dest.write_mask) - 1, result); } diff --git a/src/intel/compiler/brw_fs.cpp b/src/intel/compiler/brw_fs.cpp index 3d454c3db14..fdcc909a3d5 100644 --- a/src/intel/compiler/brw_fs.cpp +++ b/src/intel/compiler/brw_fs.cpp @@ -1976,7 +1976,7 @@ struct cplx_align { static void cplx_align_assert_sane(struct cplx_align a) { - assert(a.mul > 0 && util_is_power_of_two(a.mul)); + assert(a.mul > 0 && util_is_power_of_two_or_zero(a.mul)); assert(a.offset < a.mul); } @@ -2028,7 +2028,7 @@ static void mark_uniform_slots_read(struct uniform_slot_info *slots, unsigned num_slots, unsigned alignment) { - assert(alignment > 0 && util_is_power_of_two(alignment)); + assert(alignment > 0 && util_is_power_of_two_or_zero(alignment)); assert(alignment <= CPLX_ALIGN_MAX_MUL); /* We can't align a slot to anything less than the slot size */ diff --git a/src/intel/vulkan/anv_allocator.c b/src/intel/vulkan/anv_allocator.c index a27af4eccc4..f884ac3b827 100644 --- a/src/intel/vulkan/anv_allocator.c +++ b/src/intel/vulkan/anv_allocator.c @@ -619,7 +619,7 @@ anv_state_pool_init(struct anv_state_pool *pool, if (result != VK_SUCCESS) return result; - assert(util_is_power_of_two(block_size)); + assert(util_is_power_of_two_or_zero(block_size)); pool->block_size = block_size; pool->back_alloc_free_list = ANV_FREE_LIST_EMPTY; for (unsigned i = 0; i < ANV_STATE_BUCKETS; i++) { @@ -814,7 +814,7 @@ done: static void anv_state_pool_free_no_vg(struct anv_state_pool *pool, struct anv_state state) { - assert(util_is_power_of_two(state.alloc_size)); + assert(util_is_power_of_two_or_zero(state.alloc_size)); unsigned bucket = anv_state_pool_get_bucket(state.alloc_size); if (state.offset < 0) { @@ -1041,7 +1041,7 @@ anv_bo_pool_free(struct anv_bo_pool *pool, const struct anv_bo *bo_in) struct bo_pool_bo_link *link = bo.map; VG_NOACCESS_WRITE(&link->bo, bo); - assert(util_is_power_of_two(bo.size)); + assert(util_is_power_of_two_or_zero(bo.size)); const unsigned size_log2 = ilog2_round_up(bo.size); const unsigned bucket = size_log2 - 12; assert(bucket < ARRAY_SIZE(pool->free_list)); diff --git a/src/intel/vulkan/anv_formats.c b/src/intel/vulkan/anv_formats.c index 8d1ea402f0f..085706402c1 100644 --- a/src/intel/vulkan/anv_formats.c +++ b/src/intel/vulkan/anv_formats.c @@ -437,7 +437,7 @@ anv_get_format_plane(const struct gen_device_info *devinfo, VkFormat vk_format, isl_format_get_layout(plane_format.isl_format); if (tiling == VK_IMAGE_TILING_OPTIMAL && - !util_is_power_of_two(isl_layout->bpb)) { + !util_is_power_of_two_or_zero(isl_layout->bpb)) { /* Tiled formats *must* be power-of-two because we need up upload * them with the render pipeline. For 3-channel formats, we fix * this by switching them over to RGBX or RGBA formats under the @@ -563,7 +563,7 @@ get_image_format_features(const struct gen_device_info *devinfo, */ if (vk_tiling == VK_IMAGE_TILING_OPTIMAL && base_isl_format != ISL_FORMAT_UNSUPPORTED && - !util_is_power_of_two(isl_format_layouts[base_isl_format].bpb) && + !util_is_power_of_two_or_zero(isl_format_layouts[base_isl_format].bpb) && isl_format_rgb_to_rgbx(base_isl_format) == ISL_FORMAT_UNSUPPORTED) { flags &= ~VK_FORMAT_FEATURE_COLOR_ATTACHMENT_BIT; flags &= ~VK_FORMAT_FEATURE_BLIT_DST_BIT; diff --git a/src/intel/vulkan/anv_nir_lower_multiview.c b/src/intel/vulkan/anv_nir_lower_multiview.c index 365a70d7579..6822595728c 100644 --- a/src/intel/vulkan/anv_nir_lower_multiview.c +++ b/src/intel/vulkan/anv_nir_lower_multiview.c @@ -87,7 +87,7 @@ build_view_index(struct lower_multiview_state *state) nir_umod(b, nir_load_instance_id(b), nir_imm_int(b, _mesa_bitcount(state->view_mask))); - if (util_is_power_of_two(state->view_mask + 1)) { + if (util_is_power_of_two_or_zero(state->view_mask + 1)) { /* If we have a full view mask, then compacted is what we want */ state->view_index = compacted; } else { diff --git a/src/mesa/state_tracker/st_cb_readpixels.c b/src/mesa/state_tracker/st_cb_readpixels.c index 84dd2d548e3..29181216071 100644 --- a/src/mesa/state_tracker/st_cb_readpixels.c +++ b/src/mesa/state_tracker/st_cb_readpixels.c @@ -270,8 +270,8 @@ blit_to_staging(struct st_context *st, struct st_renderbuffer *strb, /* We are creating a texture of the size of the region being read back. * Need to check for NPOT texture support. */ if (!screen->get_param(screen, PIPE_CAP_NPOT_TEXTURES) && - (!util_is_power_of_two(width) || - !util_is_power_of_two(height))) + (!util_is_power_of_two_or_zero(width) || + !util_is_power_of_two_or_zero(height))) return NULL; /* create the destination texture */ diff --git a/src/mesa/state_tracker/st_cb_texture.c b/src/mesa/state_tracker/st_cb_texture.c index 3a793a72654..628e4af1e67 100644 --- a/src/mesa/state_tracker/st_cb_texture.c +++ b/src/mesa/state_tracker/st_cb_texture.c @@ -1527,9 +1527,9 @@ st_TexSubImage(struct gl_context *ctx, GLuint dims, /* Check for NPOT texture support. */ if (!screen->get_param(screen, PIPE_CAP_NPOT_TEXTURES) && - (!util_is_power_of_two(src_templ.width0) || - !util_is_power_of_two(src_templ.height0) || - !util_is_power_of_two(src_templ.depth0))) { + (!util_is_power_of_two_or_zero(src_templ.width0) || + !util_is_power_of_two_or_zero(src_templ.height0) || + !util_is_power_of_two_or_zero(src_templ.depth0))) { goto fallback; } diff --git a/src/util/bitscan.h b/src/util/bitscan.h index 611e8120596..2d4e46ec0f1 100644 --- a/src/util/bitscan.h +++ b/src/util/bitscan.h @@ -31,6 +31,7 @@ #include #include +#include #include #if defined(_MSC_VER) @@ -107,6 +108,17 @@ u_bit_scan64(uint64_t *mask) return i; } +/* Determine if an unsigned value is a power of two. + * + * \note + * Zero is treated as a power of two. + */ +static inline bool +util_is_power_of_two_or_zero(unsigned v) +{ + return (v & (v - 1)) == 0; +} + /* For looping over a bitmask when you want to loop over consecutive bits * manually, for example: * diff --git a/src/util/u_vector.c b/src/util/u_vector.c index 0de492ccf9a..bec6e5bbc30 100644 --- a/src/util/u_vector.c +++ b/src/util/u_vector.c @@ -37,8 +37,8 @@ int u_vector_init(struct u_vector *vector, uint32_t element_size, uint32_t size) { - assert(util_is_power_of_two(size)); - assert(element_size < size && util_is_power_of_two(element_size)); + assert(util_is_power_of_two_or_zero(size)); + assert(element_size < size && util_is_power_of_two_or_zero(element_size)); vector->head = 0; vector->tail = 0; -- 2.30.2