zig/mach
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

gpu: replace &slice[0] with slice.ptr

Browse source 
This avoids UB if the slice is empty
This commit is contained in:
Silver 2022-04-07 00:21:07 +01:00 • committed by Stephen Gutekanst
parent f4c8a1908d
commit c62b5ba52f
4 changed files with 29 additions and 29 deletions
Download patch file Download diff file Expand all files Collapse all files

46
gpu/src/NativeInstance.zig
View file

@ -245,7 +245,7 @@ pub fn wrapAdapter(adapter: c.WGPUAdapter) Adapter {
.vtable = &adapter_vtable,
};
const features_len = c.wgpuAdapterEnumerateFeatures(adapter.?, @ptrCast(*c.WGPUFeatureName, &wrapped._features[0]));
const features_len = c.wgpuAdapterEnumerateFeatures(adapter.?, @ptrCast([*]c.WGPUFeatureName, &wrapped._features));
wrapped.features = wrapped._features[0..features_len];
return wrapped;
}
@ -278,7 +278,7 @@ const adapter_vtable = Adapter.VTable{
.nextInChain = null,
.label = if (descriptor.label) |l| l else null,
.requiredFeaturesCount = if (descriptor.required_features) |f| @intCast(u32, f.len) else 0,
.requiredFeatures = if (descriptor.required_features) |f| @ptrCast([*c]const c_uint, &f[0]) else null,
.requiredFeatures = if (descriptor.required_features) |f| @ptrCast([*]const c_uint, f.ptr) else null,
.requiredLimits = if (required_limits) |*l| l else null,
};
@ -320,7 +320,7 @@ fn wrapDevice(device: c.WGPUDevice) Device {
.vtable = &device_vtable,
};
const features_len = c.wgpuDeviceEnumerateFeatures(device.?, @ptrCast(*c.WGPUFeatureName, &wrapped._features[0]));
const features_len = c.wgpuDeviceEnumerateFeatures(device.?, @ptrCast([*]c.WGPUFeatureName, &wrapped._features));
wrapped.features = wrapped._features[0..features_len];
return wrapped;
}
@ -410,7 +410,7 @@ const device_vtable = Device.VTable{
.label = if (descriptor.label) |l| l else null,
.layout = @ptrCast(c.WGPUBindGroupLayout, descriptor.layout.ptr),
.entryCount = @intCast(u32, entries.len),
.entries = &entries[0],
.entries = entries.ptr,
};
return wrapBindGroup(c.wgpuDeviceCreateBindGroup(@ptrCast(c.WGPUDevice, ptr), &desc));
@ -451,7 +451,7 @@ const device_vtable = Device.VTable{
.nextInChain = null,
.label = if (descriptor.label) |l| l else null,
.entryCount = @intCast(u32, descriptor.entries.len),
.entries = @ptrCast(*const c.WGPUBindGroupLayoutEntry, &descriptor.entries[0]),
.entries = @ptrCast([*]const c.WGPUBindGroupLayoutEntry, descriptor.entries.ptr),
};
return wrapBindGroupLayout(c.wgpuDeviceCreateBindGroupLayout(@ptrCast(c.WGPUDevice, ptr), &desc));
}
@ -487,7 +487,7 @@ const device_vtable = Device.VTable{
.next = null,
.sType = c.WGPUSType_ShaderModuleSPIRVDescriptor,
},
.code = @ptrCast([*c]const u32, &spirv[0]),
.code = spirv.ptr,
.codeSize = @intCast(u32, spirv.len),
};
const desc = c.WGPUShaderModuleDescriptor{
@ -634,7 +634,7 @@ const device_vtable = Device.VTable{
.nextInChain = null,
.label = if (descriptor.label) |l| l else null,
.bindGroupLayoutCount = @intCast(u32, bind_group_layouts.len),
.bindGroupLayouts = &bind_group_layouts[0],
.bindGroupLayouts = bind_group_layouts.ptr,
};
return wrapPipelineLayout(c.wgpuDeviceCreatePipelineLayout(@ptrCast(c.WGPUDevice, ptr), &desc));
}
@ -646,7 +646,7 @@ const device_vtable = Device.VTable{
.label = if (descriptor.label) |l| l else null,
.type = @enumToInt(descriptor.type),
.count = descriptor.count,
.pipelineStatistics = @ptrCast(*const c.WGPUPipelineStatisticName, &descriptor.pipeline_statistics[0]),
.pipelineStatistics = @ptrCast([*]const c.WGPUPipelineStatisticName, descriptor.pipeline_statistics.ptr),
.pipelineStatisticsCount = @intCast(u32, descriptor.pipeline_statistics.len),
};
return wrapQuerySet(c.wgpuDeviceCreateQuerySet(@ptrCast(c.WGPUDevice, ptr), &desc));
@ -658,7 +658,7 @@ const device_vtable = Device.VTable{
.nextInChain = null,
.label = if (descriptor.label) |l| l else null,
.colorFormatsCount = @intCast(u32, descriptor.color_formats.len),
.colorFormats = @ptrCast(*const c.WGPUTextureFormat, &descriptor.color_formats[0]),
.colorFormats = @ptrCast([*]const c.WGPUTextureFormat, descriptor.color_formats.ptr),
.depthStencilFormat = @enumToInt(descriptor.depth_stencil_format),
.sampleCount = descriptor.sample_count,
.depthReadOnly = descriptor.depth_read_only,
@ -781,7 +781,7 @@ inline fn convertComputePipelineDescriptor(descriptor: *const ComputePipeline.De
.module = @ptrCast(c.WGPUShaderModule, descriptor.compute.module.ptr),
.entryPoint = descriptor.compute.entry_point,
.constantCount = if (descriptor.compute.constants) |v| @intCast(u32, v.len) else 0,
.constants = if (descriptor.compute.constants) |v| @ptrCast(*const c.WGPUConstantEntry, &v[0]) else null,
.constants = if (descriptor.compute.constants) |v| @ptrCast([*]const c.WGPUConstantEntry, v.ptr) else null,
},
};
}
@ -812,9 +812,9 @@ inline fn convertRenderPipelineDescriptor(
.module = @ptrCast(c.WGPUShaderModule, d.fragment.module.ptr),
.entryPoint = d.vertex.entry_point,
.constantCount = if (d.fragment.constants) |v| @intCast(u32, v.len) else 0,
.constants = if (d.fragment.constants) |v| @ptrCast(*const c.WGPUConstantEntry, &v[0]) else null,
.constants = if (d.fragment.constants) |v| @ptrCast([*]const c.WGPUConstantEntry, v.ptr) else null,
.targetCount = if (d.fragment.targets) |v| @intCast(u32, v.len) else 0,
.targets = if (d.fragment.targets) |v| @ptrCast(*const c.WGPUColorTargetState, &v[0]) else null,
.targets = if (d.fragment.targets) |v| @ptrCast([*]const c.WGPUColorTargetState, v.ptr) else null,
};
return c.WGPURenderPipelineDescriptor{
@ -826,9 +826,9 @@ inline fn convertRenderPipelineDescriptor(
.module = @ptrCast(c.WGPUShaderModule, d.vertex.module.ptr),
.entryPoint = d.vertex.entry_point,
.constantCount = if (d.vertex.constants) |v| @intCast(u32, v.len) else 0,
.constants = if (d.vertex.constants) |v| @ptrCast(*const c.WGPUConstantEntry, &v[0]) else null,
.constants = if (d.vertex.constants) |v| @ptrCast([*]const c.WGPUConstantEntry, v.ptr) else null,
.bufferCount = if (d.vertex.buffers) |v| @intCast(u32, v.len) else 0,
.buffers = if (d.vertex.buffers) |v| @ptrCast(*const c.WGPUVertexBufferLayout, &v[0]) else null,
.buffers = if (d.vertex.buffers) |v| @ptrCast([*]const c.WGPUVertexBufferLayout, v.ptr) else null,
},
.primitive = c.WGPUPrimitiveState{
.nextInChain = null,
@ -908,7 +908,7 @@ const queue_vtable = Queue.VTable{
c.wgpuQueueSubmit(
wgpu_queue,
@intCast(u32, commands.len),
@ptrCast(*c.WGPUCommandBuffer, &commands[0]),
@ptrCast([*]c.WGPUCommandBuffer, commands.ptr),
);
}
}).submit,
@ -1272,7 +1272,7 @@ const render_pass_encoder_vtable = RenderPassEncoder.VTable{
c.wgpuRenderPassEncoderExecuteBundles(
@ptrCast(c.WGPURenderPassEncoder, ptr),
@intCast(u32, c_bundles.len),
&c_bundles[0],
c_bundles.ptr,
);
}
}).executeBundles,
@ -1303,7 +1303,7 @@ const render_pass_encoder_vtable = RenderPassEncoder.VTable{
group_index,
@ptrCast(c.WGPUBindGroup, group.ptr),
@intCast(u32, dynamic_offsets.len),
&dynamic_offsets[0],
dynamic_offsets.ptr,
);
}
}).setBindGroup,
@ -1500,7 +1500,7 @@ const render_bundle_encoder_vtable = RenderBundleEncoder.VTable{
group_index,
@ptrCast(c.WGPUBindGroup, group.ptr),
@intCast(u32, dynamic_offsets.len),
&dynamic_offsets[0],
dynamic_offsets.ptr,
);
}
}).setBindGroup,
@ -1868,7 +1868,7 @@ const command_encoder_vtable = CommandEncoder.VTable{
.nextInChain = null,
.label = if (d.label) |l| l else null,
.timestampWriteCount = @intCast(u32, timestamp_writes.len),
.timestampWrites = @ptrCast(*const c.WGPUComputePassTimestampWrite, &timestamp_writes[0]),
.timestampWrites = @ptrCast([*]const c.WGPUComputePassTimestampWrite, timestamp_writes.ptr),
};
return wrapComputePassEncoder(c.wgpuCommandEncoderBeginComputePass(@ptrCast(c.WGPUCommandEncoder, ptr), &desc));
}
@ -1945,7 +1945,7 @@ const command_encoder_vtable = CommandEncoder.VTable{
.nextInChain = null,
.label = if (d.label) |l| l else null,
.colorAttachmentCount = @intCast(u32, color_attachments.len),
.colorAttachments = &color_attachments[0],
.colorAttachments = color_attachments.ptr,
.depthStencilAttachment = if (d.depth_stencil_attachment) |v| &c.WGPURenderPassDepthStencilAttachment{
.view = @ptrCast(c.WGPUTextureView, v.view.ptr),
.depthLoadOp = @enumToInt(v.depth_load_op),
@ -1961,7 +1961,7 @@ const command_encoder_vtable = CommandEncoder.VTable{
} else null,
.occlusionQuerySet = if (d.occlusion_query_set) |v| @ptrCast(c.WGPUQuerySet, v.ptr) else null,
.timestampWriteCount = if (timestamp_writes) |v| @intCast(u32, v.len) else 0,
.timestampWrites = if (timestamp_writes) |v| @ptrCast(*const c.WGPURenderPassTimestampWrite, &v[0]) else null,
.timestampWrites = if (timestamp_writes) |v| @ptrCast([*]const c.WGPURenderPassTimestampWrite, v.ptr) else null,
};
return wrapRenderPassEncoder(c.wgpuCommandEncoderBeginRenderPass(@ptrCast(c.WGPUCommandEncoder, ptr), &desc));
}
@ -2051,7 +2051,7 @@ const command_encoder_vtable = CommandEncoder.VTable{
}
}).pushDebugGroup,
.writeBuffer = (struct {
pub fn writeBuffer(ptr: *anyopaque, buffer: Buffer, buffer_offset: u64, data: *const u8, size: u64) void {
pub fn writeBuffer(ptr: *anyopaque, buffer: Buffer, buffer_offset: u64, data: [*]const u8, size: u64) void {
c.wgpuCommandEncoderWriteBuffer(
@ptrCast(c.WGPUCommandEncoder, ptr),
@ptrCast(c.WGPUBuffer, buffer.ptr),
@ -2162,7 +2162,7 @@ const compute_pass_encoder_vtable = ComputePassEncoder.VTable{
group_index,
@ptrCast(c.WGPUBindGroup, group.ptr),
@intCast(u32, dynamic_offsets.len),
&dynamic_offsets[0],
dynamic_offsets.ptr,
);
}
}).setBindGroup,