Skip to content

Commit

Permalink
simplify limits
Browse files Browse the repository at this point in the history
  • Loading branch information
@Vipitis
Vipitis committed Feb 22, 2025
1 parent 770f5a3 commit dc89134
Show file tree
Hide file tree
Showing 2 changed files with 134 additions and 59 deletions.
185 changes: 133 additions & 52 deletions wgpu/backends/wgpu_native/_api.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -250,6 +250,24 @@ def to_c_string(string: str):
return ffi.new("char []", string.encode())


def to_c_string_view(string: str):
if string == ffi.NULL:
# H: data: char *, length: int
struct = new_struct(
"WGPUStringView",
data=ffi.NULL,
length=0,
)
else:
# H: data: char *, length: int
struct = new_struct(
"WGPUStringView",
data=to_c_string(string),
length=len(string.encode(encoding="utf-8")),
)
return struct


def to_c_string_or_null(string: Optional[str]):
return ffi.NULL if string is None else ffi.new("char []", string.encode())

Expand All @@ -262,7 +280,7 @@ def to_c_label(label):
if not label:
return _empty_label
else:
return to_c_string(label)
return to_c_string_view(label)


def feature_flag_to_feature_names(flag):
Expand All @@ -288,30 +306,51 @@ def _get_limits(id: int, device: bool = False, adapter: bool = False):
"""Gets the limits for a device or an adapter"""
assert device + adapter == 1 # exactly one is set

# FIXME: unknown C struct WGPUSupportedLimitsExtras
c_supported_limits_extras = new_struct_p(
"WGPUSupportedLimitsExtras *",
# not used: chain
# not used: limits
)
c_supported_limits_extras.chain.sType = lib.WGPUSType_SupportedLimitsExtras
# FIXME: unknown C struct WGPUSupportedLimits
c_supported_limits = new_struct_p(
"WGPUSupportedLimits *",
nextInChain=ffi.cast("WGPUChainedStructOut *", c_supported_limits_extras),
# not used: limits
# H: nextInChain: WGPUChainedStructOut *, maxTextureDimension1D: int, maxTextureDimension2D: int, maxTextureDimension3D: int, maxTextureArrayLayers: int, maxBindGroups: int, maxBindGroupsPlusVertexBuffers: int, maxBindingsPerBindGroup: int, maxDynamicUniformBuffersPerPipelineLayout: int, maxDynamicStorageBuffersPerPipelineLayout: int, maxSampledTexturesPerShaderStage: int, maxSamplersPerShaderStage: int, maxStorageBuffersPerShaderStage: int, maxStorageTexturesPerShaderStage: int, maxUniformBuffersPerShaderStage: int, maxUniformBufferBindingSize: int, maxStorageBufferBindingSize: int, minUniformBufferOffsetAlignment: int, minStorageBufferOffsetAlignment: int, maxVertexBuffers: int, maxBufferSize: int, maxVertexAttributes: int, maxVertexBufferArrayStride: int, maxInterStageShaderVariables: int, maxColorAttachments: int, maxColorAttachmentBytesPerSample: int, maxComputeWorkgroupStorageSize: int, maxComputeInvocationsPerWorkgroup: int, maxComputeWorkgroupSizeX: int, maxComputeWorkgroupSizeY: int, maxComputeWorkgroupSizeZ: int, maxComputeWorkgroupsPerDimension: int
c_limits = new_struct_p(
"WGPULimits *",
# not used: nextInChain
# not used: maxTextureDimension1D
# not used: maxTextureDimension2D
# not used: maxTextureDimension3D
# not used: maxTextureArrayLayers
# not used: maxBindGroups
# not used: maxBindGroupsPlusVertexBuffers
# not used: maxBindingsPerBindGroup
# not used: maxDynamicUniformBuffersPerPipelineLayout
# not used: maxDynamicStorageBuffersPerPipelineLayout
# not used: maxSampledTexturesPerShaderStage
# not used: maxSamplersPerShaderStage
# not used: maxStorageBuffersPerShaderStage
# not used: maxStorageTexturesPerShaderStage
# not used: maxUniformBuffersPerShaderStage
# not used: maxUniformBufferBindingSize
# not used: maxStorageBufferBindingSize
# not used: minUniformBufferOffsetAlignment
# not used: minStorageBufferOffsetAlignment
# not used: maxVertexBuffers
# not used: maxBufferSize
# not used: maxVertexAttributes
# not used: maxVertexBufferArrayStride
# not used: maxInterStageShaderVariables
# not used: maxColorAttachments
# not used: maxColorAttachmentBytesPerSample
# not used: maxComputeWorkgroupStorageSize
# not used: maxComputeInvocationsPerWorkgroup
# not used: maxComputeWorkgroupSizeX
# not used: maxComputeWorkgroupSizeY
# not used: maxComputeWorkgroupSizeZ
# not used: maxComputeWorkgroupsPerDimension
)
if adapter:
# H: WGPUStatus f(WGPUAdapter adapter, WGPULimits * limits)
libf.wgpuAdapterGetLimits(id, c_supported_limits)
libf.wgpuAdapterGetLimits(id, c_limits)
else:
# H: WGPUStatus f(WGPUDevice device, WGPULimits * limits)
libf.wgpuDeviceGetLimits(id, c_supported_limits)
libf.wgpuDeviceGetLimits(id, c_limits)

key_value_pairs = [
(to_snake_case(name, "-"), getattr(c_limits, name))
for c_limits in (c_supported_limits.limits, c_supported_limits_extras.limits)
for name in dir(c_limits)
(to_snake_case(name, "-"), getattr(c_limits, name)) for name in dir(c_limits)
]
limits = dict(sorted(key_value_pairs))
return limits
Expand Down Expand Up @@ -982,21 +1021,42 @@ def _request_device(

# ----- Set limits

# FIXME: unknown C struct WGPURequiredLimitsExtras
c_required_limits_extras = new_struct_p(
"WGPURequiredLimitsExtras *",
# not used: chain
# not used: limits
)
c_required_limits_extras.chain.sType = lib.WGPUSType_RequiredLimitsExtras
# FIXME: unknown C struct WGPURequiredLimits
# H: nextInChain: WGPUChainedStructOut *, maxTextureDimension1D: int, maxTextureDimension2D: int, maxTextureDimension3D: int, maxTextureArrayLayers: int, maxBindGroups: int, maxBindGroupsPlusVertexBuffers: int, maxBindingsPerBindGroup: int, maxDynamicUniformBuffersPerPipelineLayout: int, maxDynamicStorageBuffersPerPipelineLayout: int, maxSampledTexturesPerShaderStage: int, maxSamplersPerShaderStage: int, maxStorageBuffersPerShaderStage: int, maxStorageTexturesPerShaderStage: int, maxUniformBuffersPerShaderStage: int, maxUniformBufferBindingSize: int, maxStorageBufferBindingSize: int, minUniformBufferOffsetAlignment: int, minStorageBufferOffsetAlignment: int, maxVertexBuffers: int, maxBufferSize: int, maxVertexAttributes: int, maxVertexBufferArrayStride: int, maxInterStageShaderVariables: int, maxColorAttachments: int, maxColorAttachmentBytesPerSample: int, maxComputeWorkgroupStorageSize: int, maxComputeInvocationsPerWorkgroup: int, maxComputeWorkgroupSizeX: int, maxComputeWorkgroupSizeY: int, maxComputeWorkgroupSizeZ: int, maxComputeWorkgroupsPerDimension: int
c_required_limits = new_struct_p(
"WGPURequiredLimits *",
nextInChain=ffi.cast("WGPUChainedStruct*", c_required_limits_extras),
# not used: limits
"WGPULimits *",
# not used: nextInChain
# not used: maxTextureDimension1D
# not used: maxTextureDimension2D
# not used: maxTextureDimension3D
# not used: maxTextureArrayLayers
# not used: maxBindGroups
# not used: maxBindGroupsPlusVertexBuffers
# not used: maxBindingsPerBindGroup
# not used: maxDynamicUniformBuffersPerPipelineLayout
# not used: maxDynamicStorageBuffersPerPipelineLayout
# not used: maxSampledTexturesPerShaderStage
# not used: maxSamplersPerShaderStage
# not used: maxStorageBuffersPerShaderStage
# not used: maxStorageTexturesPerShaderStage
# not used: maxUniformBuffersPerShaderStage
# not used: maxUniformBufferBindingSize
# not used: maxStorageBufferBindingSize
# not used: minUniformBufferOffsetAlignment
# not used: minStorageBufferOffsetAlignment
# not used: maxVertexBuffers
# not used: maxBufferSize
# not used: maxVertexAttributes
# not used: maxVertexBufferArrayStride
# not used: maxInterStageShaderVariables
# not used: maxColorAttachments
# not used: maxColorAttachmentBytesPerSample
# not used: maxComputeWorkgroupStorageSize
# not used: maxComputeInvocationsPerWorkgroup
# not used: maxComputeWorkgroupSizeX
# not used: maxComputeWorkgroupSizeY
# not used: maxComputeWorkgroupSizeZ
# not used: maxComputeWorkgroupsPerDimension
)
c_limits = c_required_limits.limits
c_limits_extras = c_required_limits_extras.limits

def canonicalize_limit_name(name):
if name in self._limits:
Expand All @@ -1021,15 +1081,14 @@ def canonicalize_limit_name(name):
# setting it to {}, but the loop below goes just a little bit faster.
required_limits = self._limits

for limit in (c_limits, c_limits_extras):
for key in dir(limit):
snake_key = to_snake_case(key, "-")
# Use the value in required_limits if it exists. Otherwise, the old value
try:
value = required_limits[snake_key]
except KeyError:
value = self._limits[snake_key]
setattr(limit, key, value)
for key in dir(c_required_limits):
snake_key = to_snake_case(key, "-")
# Use the value in required_limits if it exists. Otherwise, the old value
try:
value = required_limits[snake_key]
except KeyError:
value = self._limits[snake_key]
setattr(c_required_limits, key, value)

# ---- Set queue descriptor

Expand All @@ -1045,9 +1104,9 @@ def canonicalize_limit_name(name):

# ----- Compose device descriptor extras

c_trace_path = ffi.NULL
c_trace_path = to_c_string_view(ffi.NULL)
if trace_path: # no-cover
c_trace_path = to_c_string(trace_path)
c_trace_path = to_c_string_view(trace_path)

# H: chain: WGPUChainedStruct, tracePath: WGPUStringView
extras = new_struct_p(
Expand All @@ -1070,6 +1129,16 @@ def device_lost_callback(c_reason, c_message, userdata):
# We might want to do something similar, once we have async figured out.
error_handler.log_error(msg)

# H: nextInChain: WGPUChainedStruct *, mode: WGPUCallbackMode, callback: WGPUDeviceLostCallback, userdata1: void*, userdata2: void*
device_lost_callback_info = new_struct(
"WGPUDeviceLostCallbackInfo",
nextInChain=ffi.NULL,
mode=1,
callback=device_lost_callback,
# not used: userdata1
# not used: userdata2
)

# ----- Uncaptured error

# TODO: For some errors (seen for errors in wgsl, but not for some others) the error gets logged via the logger as well (duplicate). Probably an issue with wgpu-core.
Expand Down Expand Up @@ -1097,28 +1166,40 @@ def uncaptured_error_callback(c_type, c_message, userdata):
# H: nextInChain: WGPUChainedStruct *, label: WGPUStringView, requiredFeatureCount: int, requiredFeatures: WGPUFeatureName *, requiredLimits: WGPULimits *, defaultQueue: WGPUQueueDescriptor, deviceLostCallbackInfo: WGPUDeviceLostCallbackInfo, uncapturedErrorCallbackInfo: WGPUUncapturedErrorCallbackInfo
struct = new_struct_p(
"WGPUDeviceDescriptor *",
label=to_c_label(label),
nextInChain=ffi.cast("WGPUChainedStruct * ", extras),
label=to_c_label(label),
requiredFeatureCount=len(c_features),
requiredFeatures=new_array("WGPUFeatureName[]", c_features),
requiredLimits=c_required_limits,
defaultQueue=queue_struct,
# FIXME: unknown C struct field WGPUDeviceDescriptor.deviceLostCallback
deviceLostCallback=device_lost_callback,
deviceLostCallbackInfo=device_lost_callback_info,
uncapturedErrorCallbackInfo=uncaptured_error_callback_info,
# FIXME: unknown C struct field WGPUDeviceDescriptor.deviceLostUserdata
# not used: deviceLostUserdata
# not used: deviceLostCallbackInfo
)

@ffi.callback("void(WGPURequestDeviceStatus, WGPUDevice, char *, void *)")
@ffi.callback(
"void(WGPURequestDeviceStatus, WGPUDevice, WGPUStringView, void *)"
)
def callback(status, result, message, userdata):
if status != 0:
msg = "-" if message == ffi.NULL else ffi.string(message).decode()
if status != 1:
msg = (
"-"
if message.data == ffi.NULL
else ffi.string(message.data, message.length).decode()
)
awaitable.set_error(f"Request device failed ({status}): {msg}")
else:
awaitable.set_result(result)

# H: nextInChain: WGPUChainedStruct *, mode: WGPUCallbackMode, callback: WGPURequestDeviceCallback, userdata1: void*, userdata2: void*
callback_info = new_struct(
"WGPURequestDeviceCallbackInfo",
nextInChain=ffi.NULL,
mode=1,
callback=callback,
# not used: userdata1
# not used: userdata2
)

def finalizer(device_id):
limits = _get_limits(device_id, device=True)
features = _get_features(device_id, device=True)
Expand All @@ -1137,7 +1218,7 @@ def finalizer(device_id):
awaitable = WgpuAwaitable("request_device", callback, finalizer)

# H: WGPUFuture f(WGPUAdapter adapter, WGPUDeviceDescriptor const * descriptor, WGPURequestDeviceCallbackInfo callbackInfo)
libf.wgpuAdapterRequestDevice(self._internal, struct, callback, ffi.NULL)
libf.wgpuAdapterRequestDevice(self._internal, struct, callback_info)

return awaitable

Expand Down
8 changes: 1 addition & 7 deletions wgpu/resources/codegen_report.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -35,14 +35,8 @@
* Wrote 243 enum mappings and 47 struct-field mappings to wgpu_native/_mappings.py
* Validated 142 C function calls
* Not using 76 C functions
* ERROR: unknown C struct WGPUSupportedLimitsExtras
* ERROR: unknown C struct WGPUSupportedLimits
* ERROR: unknown C struct field WGPUSurfaceTexture.suboptimal
* ERROR: unknown C struct WGPURequiredLimitsExtras
* ERROR: unknown C struct WGPURequiredLimits
* ERROR: unknown C struct field WGPUUncapturedErrorCallbackInfo.userdata
* ERROR: unknown C struct field WGPUDeviceDescriptor.deviceLostCallback
* ERROR: unknown C struct field WGPUDeviceDescriptor.deviceLostUserdata
* ERROR: unknown C struct WGPUShaderModuleWGSLDescriptor
* ERROR: unknown C struct WGPUShaderModuleSPIRVDescriptor
* ERROR: unknown C struct field WGPUShaderModuleDescriptor.hintCount
Expand All @@ -55,4 +49,4 @@
* ERROR: unknown C struct WGPUImageCopyTexture
* ERROR: unknown C struct WGPUImageCopyTexture
* ERROR: unknown C struct WGPUTextureDataLayout
* Validated 83 C structs
* Validated 85 C structs

0 comments on commit dc89134

Please sign in to comment.