wine-commits June 2022

wine-commits@winehq.org

2 participants
791 discussions

Rémi Bernon : winegstreamer: Support zero-copy in wg_transform_push_data.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: wine Branch: master Commit: c40ab0f9a62f02b80fe7604d698db4aa3a8b7713 URL: https://source.winehq.org/git/wine.git/?a=commit;h=c40ab0f9a62f02b80fe7604d… Author: Rémi Bernon <rbernon(a)codeweavers.com> Date: Mon Jun 6 17:52:12 2022 +0200 winegstreamer: Support zero-copy in wg_transform_push_data. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/mf/tests/mf.c | 1 - dlls/winegstreamer/gst_private.h | 9 +++- dlls/winegstreamer/h264_decoder.c | 17 +++++-- dlls/winegstreamer/mfplat.c | 87 +++++++++++++++++++++++++++++++++-- dlls/winegstreamer/quartz_transform.c | 9 ++++ dlls/winegstreamer/unixlib.h | 2 + dlls/winegstreamer/wg_transform.c | 28 +++++++++-- dlls/winegstreamer/wma_decoder.c | 15 ++++-- 8 files changed, 153 insertions(+), 15 deletions(-) Diff: https://source.winehq.org/git/wine.git/?a=commitdiff;h=c40ab0f9a62f02b80fe7…

1 0

Rémi Bernon : winegstreamer: Support zero-copy output using the allocator.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: wine Branch: master Commit: f17b80105681c2c025d61fe98175c4b83da2d5d2 URL: https://source.winehq.org/git/wine.git/?a=commit;h=f17b80105681c2c025d61fe9… Author: Rémi Bernon <rbernon(a)codeweavers.com> Date: Mon Jun 6 17:09:18 2022 +0200 winegstreamer: Support zero-copy output using the allocator. Through a custom allocator, by borrowing memory from the reading thread and mapping it instead of the allocated memory. We cannot use the buffer pool to share wrapped buffers, because some decoder will hold on the acquired buffers longer than they should and we cannot remove our memory from them as long as they keep a reference. Swapping the memory on map should be safe. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/winegstreamer/mfplat.c | 6 ++ dlls/winegstreamer/unix_private.h | 7 +- dlls/winegstreamer/unixlib.h | 1 + dlls/winegstreamer/wg_allocator.c | 135 ++++++++++++++++++++++++++++++++++++-- dlls/winegstreamer/wg_transform.c | 98 +++++++++++++++++++++++---- 5 files changed, 230 insertions(+), 17 deletions(-) Diff: https://source.winehq.org/git/wine.git/?a=commitdiff;h=f17b80105681c2c025d6…

1 0

Rémi Bernon : winegstreamer: Introduce a new custom memory allocator.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: wine Branch: master Commit: b38935f1cf701b019d1699197004c348e91963ad URL: https://source.winehq.org/git/wine.git/?a=commit;h=b38935f1cf701b019d169919… Author: Rémi Bernon <rbernon(a)codeweavers.com> Date: Fri May 13 13:17:13 2022 +0200 winegstreamer: Introduce a new custom memory allocator. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/winegstreamer/Makefile.in | 1 + dlls/winegstreamer/unix_private.h | 3 + dlls/winegstreamer/wg_allocator.c | 162 ++++++++++++++++++++++++++++++++++++++ dlls/winegstreamer/wg_transform.c | 12 ++- 4 files changed, 176 insertions(+), 2 deletions(-) diff --git a/dlls/winegstreamer/Makefile.in b/dlls/winegstreamer/Makefile.in index e4c2636d02d..50f4dc861d4 100644 --- a/dlls/winegstreamer/Makefile.in +++ b/dlls/winegstreamer/Makefile.in @@ -14,6 +14,7 @@ C_SRCS = \ mfplat.c \ quartz_parser.c \ quartz_transform.c \ + wg_allocator.c \ wg_format.c \ wg_parser.c \ wg_transform.c \ diff --git a/dlls/winegstreamer/unix_private.h b/dlls/winegstreamer/unix_private.h index 7bce8263aaf..16615ef0833 100644 --- a/dlls/winegstreamer/unix_private.h +++ b/dlls/winegstreamer/unix_private.h @@ -37,4 +37,7 @@ extern NTSTATUS wg_transform_destroy(void *args) DECLSPEC_HIDDEN; extern NTSTATUS wg_transform_push_data(void *args) DECLSPEC_HIDDEN; extern NTSTATUS wg_transform_read_data(void *args) DECLSPEC_HIDDEN; +extern GstAllocator *wg_allocator_create(void) DECLSPEC_HIDDEN; +extern void wg_allocator_destroy(GstAllocator *allocator) DECLSPEC_HIDDEN; + #endif /* __WINE_WINEGSTREAMER_UNIX_PRIVATE_H */ diff --git a/dlls/winegstreamer/wg_allocator.c b/dlls/winegstreamer/wg_allocator.c new file mode 100644 index 00000000000..90dada288ae --- /dev/null +++ b/dlls/winegstreamer/wg_allocator.c @@ -0,0 +1,162 @@ +/* + * GStreamer memory allocator + * + * Copyright 2022 Rémi Bernon for CodeWeavers + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA + */ + +#if 0 +#pragma makedep unix +#endif + +#include "config.h" + +#include <assert.h> +#include <stdarg.h> + +#include <gst/gst.h> +#include <gst/video/video.h> +#include <gst/audio/audio.h> + +#include "unix_private.h" + +GST_DEBUG_CATEGORY_EXTERN(wine); +#define GST_CAT_DEFAULT wine + +typedef struct +{ + GstMemory parent; + + GstMemory *unix_memory; + GstMapInfo unix_map_info; +} WgMemory; + +typedef struct +{ + GstAllocator parent; +} WgAllocator; + +typedef struct +{ + GstAllocatorClass parent_class; +} WgAllocatorClass; + +G_DEFINE_TYPE(WgAllocator, wg_allocator, GST_TYPE_ALLOCATOR); + +static gpointer wg_allocator_map(GstMemory *gst_memory, GstMapInfo *info, gsize maxsize) +{ + WgMemory *memory = (WgMemory *)gst_memory; + + if (gst_memory->parent) + return wg_allocator_map(gst_memory->parent, info, maxsize); + + GST_LOG("memory %p, info %p, maxsize %#zx", memory, info, maxsize); + + info->data = memory->unix_map_info.data; + + GST_INFO("Mapped memory %p to %p", memory, info->data); + return info->data; +} + +static void wg_allocator_unmap(GstMemory *gst_memory, GstMapInfo *info) +{ + WgMemory *memory = (WgMemory *)gst_memory; + + if (gst_memory->parent) + return wg_allocator_unmap(gst_memory->parent, info); + + GST_LOG("memory %p, info %p", memory, info); +} + +static void wg_allocator_init(WgAllocator *allocator) +{ + GST_LOG("allocator %p", allocator); + + allocator->parent.mem_type = "Wine"; + + allocator->parent.mem_map_full = wg_allocator_map; + allocator->parent.mem_unmap_full = wg_allocator_unmap; + + GST_OBJECT_FLAG_SET(allocator, GST_ALLOCATOR_FLAG_CUSTOM_ALLOC); +} + +static void wg_allocator_finalize(GObject *object) +{ + WgAllocator *allocator = (WgAllocator *)object; + + GST_LOG("allocator %p", allocator); + + G_OBJECT_CLASS(wg_allocator_parent_class)->finalize(object); +} + +static GstMemory *wg_allocator_alloc(GstAllocator *gst_allocator, gsize size, + GstAllocationParams *params) +{ + WgAllocator *allocator = (WgAllocator *)gst_allocator; + WgMemory *memory; + + GST_LOG("allocator %p, size %#zx, params %p", allocator, size, params); + + memory = g_slice_new0(WgMemory); + gst_memory_init(GST_MEMORY_CAST(memory), 0, GST_ALLOCATOR_CAST(allocator), + NULL, size, 0, 0, size); + memory->unix_memory = gst_allocator_alloc(NULL, size, params); + gst_memory_map(memory->unix_memory, &memory->unix_map_info, GST_MAP_WRITE); + + GST_INFO("Allocated memory %p, unix_memory %p, data %p", memory, memory->unix_memory, + memory->unix_map_info.data); + return (GstMemory *)memory; +} + +static void wg_allocator_free(GstAllocator *gst_allocator, GstMemory *gst_memory) +{ + WgAllocator *allocator = (WgAllocator *)gst_allocator; + WgMemory *memory = (WgMemory *)gst_memory; + + GST_LOG("allocator %p, memory %p", allocator, memory); + + gst_memory_unmap(memory->unix_memory, &memory->unix_map_info); + gst_memory_unref(memory->unix_memory); + g_slice_free(WgMemory, memory); +} + +static void wg_allocator_class_init(WgAllocatorClass *klass) +{ + GstAllocatorClass *parent_class = (GstAllocatorClass *)klass; + GObjectClass *root_class = (GObjectClass *)klass; + + GST_LOG("klass %p", klass); + + parent_class->alloc = wg_allocator_alloc; + parent_class->free = wg_allocator_free; + root_class->finalize = wg_allocator_finalize; +} + +GstAllocator *wg_allocator_create(void) +{ + return g_object_new(wg_allocator_get_type(), NULL); +} + +void wg_allocator_destroy(GstAllocator *gst_allocator) +{ + WgAllocator *allocator = (WgAllocator *)gst_allocator; + + GST_LOG("allocator %p", allocator); + + g_object_unref(allocator); + + GST_INFO("Destroyed buffer allocator %p", allocator); +} diff --git a/dlls/winegstreamer/wg_transform.c b/dlls/winegstreamer/wg_transform.c index 9a5c16afe74..16c73bcbcd6 100644 --- a/dlls/winegstreamer/wg_transform.c +++ b/dlls/winegstreamer/wg_transform.c @@ -47,6 +47,7 @@ GST_DEBUG_CATEGORY_EXTERN(wine); struct wg_transform { GstElement *container; + GstAllocator *allocator; GstPad *my_src, *my_sink; GstPad *their_sink, *their_src; GstSegment segment; @@ -153,6 +154,7 @@ static gboolean transform_sink_query_cb(GstPad *pad, GstObject *parent, GstQuery gst_buffer_pool_config_set_params(config, caps, info.size, 0, 0); + gst_buffer_pool_config_set_allocator(config, transform->allocator, NULL); if (!gst_buffer_pool_set_config(pool, config)) GST_ERROR("Failed to set pool %p config.", pool); } @@ -162,9 +164,10 @@ static gboolean transform_sink_query_cb(GstPad *pad, GstObject *parent, GstQuery GST_ERROR("Pool %p failed to activate.", pool); gst_query_add_allocation_pool(query, pool, info.size, 0, 0); + gst_query_add_allocation_param(query, transform->allocator, NULL); - GST_INFO("Proposing pool %p, buffer size %#zx, for query %p.", - pool, info.size, query); + GST_INFO("Proposing pool %p, buffer size %#zx, allocator %p, for query %p.", + pool, info.size, transform->allocator, query); g_object_unref(pool); return true; @@ -222,6 +225,7 @@ NTSTATUS wg_transform_destroy(void *args) while ((sample = gst_atomic_queue_pop(transform->output_queue))) gst_sample_unref(sample); + wg_allocator_destroy(transform->allocator); g_object_unref(transform->their_sink); g_object_unref(transform->their_src); g_object_unref(transform->container); @@ -326,6 +330,8 @@ NTSTATUS wg_transform_create(void *args) goto out; if (!(transform->output_queue = gst_atomic_queue_new(8))) goto out; + if (!(transform->allocator = wg_allocator_create())) + goto out; transform->input_max_length = 1; transform->output_plane_align = 0; @@ -482,6 +488,8 @@ out: gst_object_unref(transform->my_src); if (src_caps) gst_caps_unref(src_caps); + if (transform->allocator) + wg_allocator_destroy(transform->allocator); if (transform->output_queue) gst_atomic_queue_unref(transform->output_queue); if (transform->input)

1 0

Rémi Bernon : winegstreamer: Rename mf_(create|destroy)_wg_sample helpers.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: wine Branch: master Commit: 4f242f09c006e96f8518092b746648b2fdbd43f6 URL: https://source.winehq.org/git/wine.git/?a=commit;h=4f242f09c006e96f8518092b… Author: Rémi Bernon <rbernon(a)codeweavers.com> Date: Tue Jun 7 10:19:17 2022 +0200 winegstreamer: Rename mf_(create|destroy)_wg_sample helpers. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/winegstreamer/gst_private.h | 4 ++-- dlls/winegstreamer/h264_decoder.c | 10 +++++----- dlls/winegstreamer/mfplat.c | 4 ++-- dlls/winegstreamer/wma_decoder.c | 12 ++++++------ 4 files changed, 15 insertions(+), 15 deletions(-) diff --git a/dlls/winegstreamer/gst_private.h b/dlls/winegstreamer/gst_private.h index ca7396ad5cd..ab89942847a 100644 --- a/dlls/winegstreamer/gst_private.h +++ b/dlls/winegstreamer/gst_private.h @@ -120,8 +120,8 @@ extern HRESULT mfplat_DllRegisterServer(void); IMFMediaType *mf_media_type_from_wg_format(const struct wg_format *format); void mf_media_type_to_wg_format(IMFMediaType *type, struct wg_format *format); -HRESULT mf_create_wg_sample(IMFSample *sample, struct wg_sample **out); -void mf_destroy_wg_sample(struct wg_sample *wg_sample); +HRESULT wg_sample_create_mf(IMFSample *sample, struct wg_sample **out); +void wg_sample_release(struct wg_sample *wg_sample); HRESULT wg_transform_push_mf(struct wg_transform *transform, struct wg_sample *sample); HRESULT wg_transform_read_mf(struct wg_transform *transform, struct wg_sample *sample, diff --git a/dlls/winegstreamer/h264_decoder.c b/dlls/winegstreamer/h264_decoder.c index 0ee9084e12e..8d3de4355bb 100644 --- a/dlls/winegstreamer/h264_decoder.c +++ b/dlls/winegstreamer/h264_decoder.c @@ -540,11 +540,11 @@ static HRESULT WINAPI transform_ProcessInput(IMFTransform *iface, DWORD id, IMFS if (!decoder->wg_transform) return MF_E_TRANSFORM_TYPE_NOT_SET; - if (FAILED(hr = mf_create_wg_sample(sample, &wg_sample))) + if (FAILED(hr = wg_sample_create_mf(sample, &wg_sample))) return hr; hr = wg_transform_push_mf(decoder->wg_transform, wg_sample); - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); return hr; } @@ -573,17 +573,17 @@ static HRESULT WINAPI transform_ProcessOutput(IMFTransform *iface, DWORD flags, samples[0].dwStatus = 0; if (!samples[0].pSample) return E_INVALIDARG; - if (FAILED(hr = mf_create_wg_sample(samples[0].pSample, &wg_sample))) + if (FAILED(hr = wg_sample_create_mf(samples[0].pSample, &wg_sample))) return hr; if (wg_sample->max_size < info.cbSize) { - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); return MF_E_BUFFERTOOSMALL; } hr = wg_transform_read_mf(decoder->wg_transform, wg_sample, &wg_format); - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); if (hr == MF_E_TRANSFORM_STREAM_CHANGE) { diff --git a/dlls/winegstreamer/mfplat.c b/dlls/winegstreamer/mfplat.c index 026237bdf5d..28a3fc20ead 100644 --- a/dlls/winegstreamer/mfplat.c +++ b/dlls/winegstreamer/mfplat.c @@ -965,7 +965,7 @@ struct mf_sample struct wg_sample wg_sample; }; -HRESULT mf_create_wg_sample(IMFSample *sample, struct wg_sample **out) +HRESULT wg_sample_create_mf(IMFSample *sample, struct wg_sample **out) { DWORD current_length, max_length; struct mf_sample *mf_sample; @@ -995,7 +995,7 @@ out: return hr; } -void mf_destroy_wg_sample(struct wg_sample *wg_sample) +void wg_sample_release(struct wg_sample *wg_sample) { struct mf_sample *mf_sample = CONTAINING_RECORD(wg_sample, struct mf_sample, wg_sample); diff --git a/dlls/winegstreamer/wma_decoder.c b/dlls/winegstreamer/wma_decoder.c index fdb68328c3a..81285f5ad44 100644 --- a/dlls/winegstreamer/wma_decoder.c +++ b/dlls/winegstreamer/wma_decoder.c @@ -534,18 +534,18 @@ static HRESULT WINAPI transform_ProcessInput(IMFTransform *iface, DWORD id, IMFS if (FAILED(hr = IMFTransform_GetInputStreamInfo(iface, 0, &info))) return hr; - if (FAILED(hr = mf_create_wg_sample(sample, &wg_sample))) + if (FAILED(hr = wg_sample_create_mf(sample, &wg_sample))) return hr; /* WMA transform uses fixed size input samples and ignores samples with invalid sizes */ if (wg_sample->size % info.cbSize) { - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); return S_OK; } hr = wg_transform_push_mf(decoder->wg_transform, wg_sample); - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); return hr; } @@ -576,13 +576,13 @@ static HRESULT WINAPI transform_ProcessOutput(IMFTransform *iface, DWORD flags, return MF_E_TRANSFORM_NEED_MORE_INPUT; } - if (FAILED(hr = mf_create_wg_sample(samples[0].pSample, &wg_sample))) + if (FAILED(hr = wg_sample_create_mf(samples[0].pSample, &wg_sample))) return hr; wg_sample->size = 0; if (wg_sample->max_size < info.cbSize) { - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); return MF_E_BUFFERTOOSMALL; } @@ -592,7 +592,7 @@ static HRESULT WINAPI transform_ProcessOutput(IMFTransform *iface, DWORD flags, samples[0].dwStatus |= MFT_OUTPUT_DATA_BUFFER_INCOMPLETE; } - mf_destroy_wg_sample(wg_sample); + wg_sample_release(wg_sample); if (hr == MF_E_TRANSFORM_STREAM_CHANGE) {

1 0

Rémi Bernon : winegstreamer: Introduce new wg_transform_(push|read)_mf helpers.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: wine Branch: master Commit: 240aff382def741f306783227d4d46d67e931a36 URL: https://source.winehq.org/git/wine.git/?a=commit;h=240aff382def741f30678322… Author: Rémi Bernon <rbernon(a)codeweavers.com> Date: Mon Jun 6 17:06:32 2022 +0200 winegstreamer: Introduce new wg_transform_(push|read)_mf helpers. To read MF sample properties before pushing, and update them after sucessfully reading a sample. Signed-off-by: Rémi Bernon <rbernon(a)codeweavers.com> --- dlls/winegstreamer/gst_private.h | 4 +++ dlls/winegstreamer/h264_decoder.c | 6 ++--- dlls/winegstreamer/mfplat.c | 55 ++++++++++++++++++++++++++------------- dlls/winegstreamer/wma_decoder.c | 4 +-- 4 files changed, 45 insertions(+), 24 deletions(-) diff --git a/dlls/winegstreamer/gst_private.h b/dlls/winegstreamer/gst_private.h index 159143d7e54..ca7396ad5cd 100644 --- a/dlls/winegstreamer/gst_private.h +++ b/dlls/winegstreamer/gst_private.h @@ -123,6 +123,10 @@ void mf_media_type_to_wg_format(IMFMediaType *type, struct wg_format *format); HRESULT mf_create_wg_sample(IMFSample *sample, struct wg_sample **out); void mf_destroy_wg_sample(struct wg_sample *wg_sample); +HRESULT wg_transform_push_mf(struct wg_transform *transform, struct wg_sample *sample); +HRESULT wg_transform_read_mf(struct wg_transform *transform, struct wg_sample *sample, + struct wg_format *format); + HRESULT winegstreamer_stream_handler_create(REFIID riid, void **obj); HRESULT h264_decoder_create(REFIID riid, void **ret); diff --git a/dlls/winegstreamer/h264_decoder.c b/dlls/winegstreamer/h264_decoder.c index 19a36a9a77a..0ee9084e12e 100644 --- a/dlls/winegstreamer/h264_decoder.c +++ b/dlls/winegstreamer/h264_decoder.c @@ -543,8 +543,7 @@ static HRESULT WINAPI transform_ProcessInput(IMFTransform *iface, DWORD id, IMFS if (FAILED(hr = mf_create_wg_sample(sample, &wg_sample))) return hr; - hr = wg_transform_push_data(decoder->wg_transform, wg_sample); - + hr = wg_transform_push_mf(decoder->wg_transform, wg_sample); mf_destroy_wg_sample(wg_sample); return hr; } @@ -583,8 +582,7 @@ static HRESULT WINAPI transform_ProcessOutput(IMFTransform *iface, DWORD flags, return MF_E_BUFFERTOOSMALL; } - hr = wg_transform_read_data(decoder->wg_transform, wg_sample, - &wg_format); + hr = wg_transform_read_mf(decoder->wg_transform, wg_sample, &wg_format); mf_destroy_wg_sample(wg_sample); if (hr == MF_E_TRANSFORM_STREAM_CHANGE) diff --git a/dlls/winegstreamer/mfplat.c b/dlls/winegstreamer/mfplat.c index 0226e7a2e45..026237bdf5d 100644 --- a/dlls/winegstreamer/mfplat.c +++ b/dlls/winegstreamer/mfplat.c @@ -969,8 +969,6 @@ HRESULT mf_create_wg_sample(IMFSample *sample, struct wg_sample **out) { DWORD current_length, max_length; struct mf_sample *mf_sample; - LONGLONG time, duration; - UINT32 value; BYTE *buffer; HRESULT hr; @@ -981,19 +979,6 @@ HRESULT mf_create_wg_sample(IMFSample *sample, struct wg_sample **out) if (FAILED(hr = IMFMediaBuffer_Lock(mf_sample->media_buffer, &buffer, &max_length, &current_length))) goto out; - if (SUCCEEDED(IMFSample_GetSampleTime(sample, &time))) - { - mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_HAS_PTS; - mf_sample->wg_sample.pts = time; - } - if (SUCCEEDED(IMFSample_GetSampleDuration(sample, &duration))) - { - mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_HAS_DURATION; - mf_sample->wg_sample.duration = duration; - } - if (SUCCEEDED(IMFSample_GetUINT32(sample, &MFSampleExtension_CleanPoint, &value)) && value) - mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_SYNC_POINT; - IMFSample_AddRef((mf_sample->sample = sample)); mf_sample->wg_sample.data = buffer; mf_sample->wg_sample.size = current_length; @@ -1015,8 +1000,43 @@ void mf_destroy_wg_sample(struct wg_sample *wg_sample) struct mf_sample *mf_sample = CONTAINING_RECORD(wg_sample, struct mf_sample, wg_sample); IMFMediaBuffer_Unlock(mf_sample->media_buffer); - IMFMediaBuffer_SetCurrentLength(mf_sample->media_buffer, wg_sample->size); IMFMediaBuffer_Release(mf_sample->media_buffer); + IMFSample_Release(mf_sample->sample); + free(mf_sample); +} + +HRESULT wg_transform_push_mf(struct wg_transform *transform, struct wg_sample *sample) +{ + struct mf_sample *mf_sample = CONTAINING_RECORD(sample, struct mf_sample, wg_sample); + LONGLONG time, duration; + UINT32 value; + + if (SUCCEEDED(IMFSample_GetSampleTime(mf_sample->sample, &time))) + { + mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_HAS_PTS; + mf_sample->wg_sample.pts = time; + } + if (SUCCEEDED(IMFSample_GetSampleDuration(mf_sample->sample, &duration))) + { + mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_HAS_DURATION; + mf_sample->wg_sample.duration = duration; + } + if (SUCCEEDED(IMFSample_GetUINT32(mf_sample->sample, &MFSampleExtension_CleanPoint, &value)) && value) + mf_sample->wg_sample.flags |= WG_SAMPLE_FLAG_SYNC_POINT; + + return wg_transform_push_data(transform, sample); +} + +HRESULT wg_transform_read_mf(struct wg_transform *transform, struct wg_sample *wg_sample, + struct wg_format *format) +{ + struct mf_sample *mf_sample = CONTAINING_RECORD(wg_sample, struct mf_sample, wg_sample); + HRESULT hr; + + if (FAILED(hr = wg_transform_read_data(transform, wg_sample, format))) + return hr; + + IMFMediaBuffer_SetCurrentLength(mf_sample->media_buffer, wg_sample->size); if (wg_sample->flags & WG_SAMPLE_FLAG_HAS_PTS) IMFSample_SetSampleTime(mf_sample->sample, wg_sample->pts); @@ -1025,6 +1045,5 @@ void mf_destroy_wg_sample(struct wg_sample *wg_sample) if (wg_sample->flags & WG_SAMPLE_FLAG_SYNC_POINT) IMFSample_SetUINT32(mf_sample->sample, &MFSampleExtension_CleanPoint, 1); - IMFSample_Release(mf_sample->sample); - free(mf_sample); + return S_OK; } diff --git a/dlls/winegstreamer/wma_decoder.c b/dlls/winegstreamer/wma_decoder.c index 106d32adce9..fdb68328c3a 100644 --- a/dlls/winegstreamer/wma_decoder.c +++ b/dlls/winegstreamer/wma_decoder.c @@ -544,7 +544,7 @@ static HRESULT WINAPI transform_ProcessInput(IMFTransform *iface, DWORD id, IMFS return S_OK; } - hr = wg_transform_push_data(decoder->wg_transform, wg_sample); + hr = wg_transform_push_mf(decoder->wg_transform, wg_sample); mf_destroy_wg_sample(wg_sample); return hr; } @@ -586,7 +586,7 @@ static HRESULT WINAPI transform_ProcessOutput(IMFTransform *iface, DWORD flags, return MF_E_BUFFERTOOSMALL; } - if (SUCCEEDED(hr = wg_transform_read_data(decoder->wg_transform, wg_sample, NULL))) + if (SUCCEEDED(hr = wg_transform_read_mf(decoder->wg_transform, wg_sample, NULL))) { if (wg_sample->flags & WG_SAMPLE_FLAG_INCOMPLETE) samples[0].dwStatus |= MFT_OUTPUT_DATA_BUFFER_INCOMPLETE;

1 0

Giovanni Mascellani : vkd3d-shader/hlsl: Parse the mul() intrinsic.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: vkd3d Branch: master Commit: 10917a7f479b02a1313bc2e94ee904ea15ad3106 URL: https://source.winehq.org/git/vkd3d.git/?a=commit;h=10917a7f479b02a1313bc2e… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Wed Jun 8 18:18:43 2022 -0500 vkd3d-shader/hlsl: Parse the mul() intrinsic. Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> --- libs/vkd3d-shader/hlsl.y | 118 ++++++++++++++++++++++++++++++++ tests/hlsl-majority-pragma.shader_test | 2 +- tests/hlsl-majority-typedef.shader_test | 2 +- tests/hlsl-mul.shader_test | 36 +++++----- 4 files changed, 138 insertions(+), 20 deletions(-) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 2aee5153..df5fda47 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -2046,6 +2046,123 @@ static bool intrinsic_min(struct hlsl_ctx *ctx, return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MIN, params->args[0], params->args[1], loc); } +static bool intrinsic_mul(struct hlsl_ctx *ctx, + const struct parse_initializer *params, const struct vkd3d_shader_location *loc) +{ + struct hlsl_ir_node *arg1 = params->args[0], *arg2 = params->args[1], *cast1, *cast2; + enum hlsl_base_type base = expr_common_base_type(arg1->data_type->base_type, arg2->data_type->base_type); + struct hlsl_type *cast_type1 = arg1->data_type, *cast_type2 = arg2->data_type, *matrix_type, *ret_type; + unsigned int i, j, k, vect_count = 0; + struct vkd3d_string_buffer *name; + static unsigned int counter = 0; + struct hlsl_ir_load *load; + struct hlsl_ir_var *var; + + if (arg1->data_type->type == HLSL_CLASS_SCALAR || arg2->data_type->type == HLSL_CLASS_SCALAR) + return !!add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, arg1, arg2, loc); + + if (arg1->data_type->type == HLSL_CLASS_VECTOR) + { + vect_count++; + cast_type1 = hlsl_get_matrix_type(ctx, base, arg1->data_type->dimx, 1); + } + if (arg2->data_type->type == HLSL_CLASS_VECTOR) + { + vect_count++; + cast_type2 = hlsl_get_matrix_type(ctx, base, 1, arg2->data_type->dimx); + } + + matrix_type = hlsl_get_matrix_type(ctx, base, cast_type2->dimx, cast_type1->dimy); + + if (vect_count == 0) + { + ret_type = matrix_type; + } + else if (vect_count == 1) + { + assert(matrix_type->dimx == 1 || matrix_type->dimy == 1); + ret_type = hlsl_get_vector_type(ctx, base, matrix_type->dimx * matrix_type->dimy); + } + else + { + assert(matrix_type->dimx == 1 && matrix_type->dimy == 1); + ret_type = hlsl_get_scalar_type(ctx, base); + } + + if (!(cast1 = add_implicit_conversion(ctx, params->instrs, arg1, cast_type1, loc))) + return false; + + if (!(cast2 = add_implicit_conversion(ctx, params->instrs, arg2, cast_type2, loc))) + return false; + + name = vkd3d_string_buffer_get(&ctx->string_buffers); + vkd3d_string_buffer_printf(name, "<mul-%u>", counter++); + var = hlsl_new_synthetic_var(ctx, name->buffer, matrix_type, *loc); + vkd3d_string_buffer_release(&ctx->string_buffers, name); + if (!var) + return false; + + for (i = 0; i < matrix_type->dimx; ++i) + for (j = 0; j < matrix_type->dimy; ++j) + { + struct hlsl_ir_node *node = NULL; + struct hlsl_type *scalar_type; + struct hlsl_ir_store *store; + struct hlsl_ir_constant *c; + unsigned int offset; + + for (k = 0; k < cast_type1->dimx && k < cast_type2->dimy; ++k) + { + struct hlsl_ir_load *value1, *value2; + struct hlsl_ir_node *mul; + + offset = hlsl_compute_component_offset(ctx, cast_type1, j * cast_type1->dimx + k, &scalar_type); + if (!(c = hlsl_new_uint_constant(ctx, offset, loc))) + return false; + list_add_tail(params->instrs, &c->node.entry); + + if (!(value1 = add_load(ctx, params->instrs, cast1, &c->node, scalar_type, *loc))) + return false; + + offset = hlsl_compute_component_offset(ctx, cast_type2, k * cast_type2->dimx + i, &scalar_type); + if (!(c = hlsl_new_uint_constant(ctx, offset, loc))) + return false; + list_add_tail(params->instrs, &c->node.entry); + + if (!(value2 = add_load(ctx, params->instrs, cast2, &c->node, scalar_type, *loc))) + return false; + + if (!(mul = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_MUL, &value1->node, &value2->node, loc))) + return false; + + if (node) + { + if (!(node = add_binary_arithmetic_expr(ctx, params->instrs, HLSL_OP2_ADD, node, mul, loc))) + return false; + } + else + { + node = mul; + } + } + + offset = hlsl_compute_component_offset(ctx, matrix_type, j * matrix_type->dimx + i, &scalar_type); + if (!(c = hlsl_new_uint_constant(ctx, offset, loc))) + return false; + list_add_tail(params->instrs, &c->node.entry); + + if (!(store = hlsl_new_store(ctx, var, &c->node, node, 0, *loc))) + return false; + list_add_tail(params->instrs, &store->node.entry); + } + + if (!(load = hlsl_new_load(ctx, var, NULL, matrix_type, *loc))) + return false; + list_add_tail(params->instrs, &load->node.entry); + + return !!add_implicit_conversion(ctx, params->instrs, &load->node, ret_type, loc); +} + static bool intrinsic_pow(struct hlsl_ctx *ctx, const struct parse_initializer *params, const struct vkd3d_shader_location *loc) { @@ -2106,6 +2223,7 @@ intrinsic_functions[] = {"floor", 1, true, intrinsic_floor}, {"max", 2, true, intrinsic_max}, {"min", 2, true, intrinsic_min}, + {"mul", 2, true, intrinsic_mul}, {"pow", 2, true, intrinsic_pow}, {"round", 1, true, intrinsic_round}, {"saturate", 1, true, intrinsic_saturate}, diff --git a/tests/hlsl-majority-pragma.shader_test b/tests/hlsl-majority-pragma.shader_test index a9f917ff..e7fc75cd 100644 --- a/tests/hlsl-majority-pragma.shader_test +++ b/tests/hlsl-majority-pragma.shader_test @@ -17,5 +17,5 @@ uniform 0 float4 0.1 0.2 0.0 0.0 uniform 4 float4 0.3 0.4 0.0 0.0 uniform 8 float4 0.1 0.3 0.0 0.0 uniform 12 float4 0.2 0.4 0.0 0.0 -todo draw quad +draw quad probe all rgba (0.17, 0.39, 0.17, 0.39) 1 diff --git a/tests/hlsl-majority-typedef.shader_test b/tests/hlsl-majority-typedef.shader_test index 192c96db..1460e9a0 100644 --- a/tests/hlsl-majority-typedef.shader_test +++ b/tests/hlsl-majority-typedef.shader_test @@ -18,5 +18,5 @@ uniform 0 float4 0.1 0.2 0.0 0.0 uniform 4 float4 0.3 0.4 0.0 0.0 uniform 8 float4 0.1 0.3 0.0 0.0 uniform 12 float4 0.2 0.4 0.0 0.0 -todo draw quad +draw quad probe all rgba (0.17, 0.39, 0.17, 0.39) 1 diff --git a/tests/hlsl-mul.shader_test b/tests/hlsl-mul.shader_test index 1d137e69..7b453187 100644 --- a/tests/hlsl-mul.shader_test +++ b/tests/hlsl-mul.shader_test @@ -12,7 +12,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (30.0, 70.0, 110.0, 150.0) [pixel shader] @@ -28,7 +28,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (90.0, 100.0, 110.0, 120.0) [pixel shader] @@ -44,7 +44,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (14.0, 38.0, 62.0, 86.0) [pixel shader] @@ -60,7 +60,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (38.0, 44.0, 50.0, 56.0) [pixel shader] @@ -75,7 +75,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (14.0, 32.0, 50.0, 0.0) [pixel shader] @@ -90,7 +90,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (30.0, 36.0, 42.0, 0.0) [pixel shader] @@ -106,7 +106,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (50.0, 60.0, 70.0, 80.0) [pixel shader] @@ -122,7 +122,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (50.0, 60.0, 70.0, 80.0) [pixel shader] @@ -138,7 +138,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (10.0, 20.0, 30.0, 40.0) [pixel shader] @@ -154,7 +154,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (10.0, 50.0, 90.0, 130.0) [pixel shader] @@ -170,7 +170,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (10.0, 20.0, 30.0, 40.0) [pixel shader] @@ -186,7 +186,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (10.0, 50.0, 90.0, 130.0) [pixel shader] @@ -202,7 +202,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (90.0, 100.0, 110.0, 120.0) [pixel shader] @@ -218,7 +218,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (5.0, 10.0, 15.0, 20.0) [pixel shader] @@ -234,7 +234,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -250,7 +250,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (30.0, 70.0, 110.0, 150.0) [pixel shader] @@ -268,7 +268,7 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (83.0, 98.0, 113.0, 128.0) [pixel shader] @@ -286,5 +286,5 @@ float4 main(float4 pos : sv_position) : sv_target } [test] -todo draw quad +draw quad probe all rgba (78.0, 96.0, 114.0, 0.0)

1 0

Giovanni Mascellani : vkd3d-shader/hlsl: Split matrix operations.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: vkd3d Branch: master Commit: 62fa53f9b65486c90ca434c51a608e77ae154835 URL: https://source.winehq.org/git/vkd3d.git/?a=commit;h=62fa53f9b65486c90ca434c… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Wed Jun 8 18:18:42 2022 -0500 vkd3d-shader/hlsl: Split matrix operations. Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> --- libs/vkd3d-shader/hlsl.y | 74 ++++++++++++++++++++++++++++++++++++++++++++ tests/hlsl-shape.shader_test | 30 +++++++++--------- 2 files changed, 89 insertions(+), 15 deletions(-) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index 687569e8..2aee5153 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -1201,6 +1201,22 @@ static bool expr_common_shape(struct hlsl_ctx *ctx, struct hlsl_type *t1, struct return true; } +static unsigned int minor_size(const struct hlsl_type *type) +{ + if (type->modifiers & HLSL_MODIFIER_ROW_MAJOR) + return type->dimx; + else + return type->dimy; +} + +static unsigned int major_size(const struct hlsl_type *type) +{ + if (type->modifiers & HLSL_MODIFIER_ROW_MAJOR) + return type->dimy; + else + return type->dimx; +} + static struct hlsl_ir_node *add_expr(struct hlsl_ctx *ctx, struct list *instrs, enum hlsl_ir_expr_op op, struct hlsl_ir_node *operands[HLSL_MAX_OPERANDS], struct hlsl_type *type, const struct vkd3d_shader_location *loc) @@ -1208,6 +1224,64 @@ static struct hlsl_ir_node *add_expr(struct hlsl_ctx *ctx, struct list *instrs, struct hlsl_ir_expr *expr; unsigned int i; + if (type->type == HLSL_CLASS_MATRIX) + { + struct vkd3d_string_buffer *name; + static unsigned int counter = 0; + struct hlsl_type *vector_type; + struct hlsl_ir_load *load; + struct hlsl_ir_var *var; + + vector_type = hlsl_get_vector_type(ctx, type->base_type, minor_size(type)); + + name = vkd3d_string_buffer_get(&ctx->string_buffers); + vkd3d_string_buffer_printf(name, "<split_op-%u>", counter++); + var = hlsl_new_synthetic_var(ctx, name->buffer, type, *loc); + vkd3d_string_buffer_release(&ctx->string_buffers, name); + if (!var) + return NULL; + + for (i = 0; i < major_size(type); i++) + { + struct hlsl_ir_node *value, *vector_operands[HLSL_MAX_OPERANDS] = { NULL }; + struct hlsl_ir_store *store; + struct hlsl_ir_constant *c; + unsigned int j; + + if (!(c = hlsl_new_uint_constant(ctx, 4 * i, loc))) + return NULL; + list_add_tail(instrs, &c->node.entry); + + for (j = 0; j < HLSL_MAX_OPERANDS; j++) + { + if (operands[j]) + { + struct hlsl_type *vector_arg_type; + struct hlsl_ir_load *load; + + vector_arg_type = hlsl_get_vector_type(ctx, operands[j]->data_type->base_type, minor_size(type)); + + if (!(load = add_load(ctx, instrs, operands[j], &c->node, vector_arg_type, *loc))) + return NULL; + vector_operands[j] = &load->node; + } + } + + if (!(value = add_expr(ctx, instrs, op, vector_operands, vector_type, loc))) + return NULL; + + if (!(store = hlsl_new_store(ctx, var, &c->node, value, 0, *loc))) + return NULL; + list_add_tail(instrs, &store->node.entry); + } + + if (!(load = hlsl_new_load(ctx, var, NULL, type, *loc))) + return NULL; + list_add_tail(instrs, &load->node.entry); + + return &load->node; + } + if (!(expr = hlsl_alloc(ctx, sizeof(*expr)))) return NULL; init_node(&expr->node, HLSL_IR_EXPR, type, *loc); diff --git a/tests/hlsl-shape.shader_test b/tests/hlsl-shape.shader_test index 65cc322c..b96f0fd2 100644 --- a/tests/hlsl-shape.shader_test +++ b/tests/hlsl-shape.shader_test @@ -93,7 +93,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 7.0, 9.0) [pixel shader] @@ -107,7 +107,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 7.0, 9.0) [pixel shader] @@ -122,7 +122,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -137,7 +137,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -152,7 +152,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 7.0, 12.0, 17.0) [pixel shader] @@ -167,7 +167,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 7.0, 12.0, 17.0) [pixel shader] @@ -183,7 +183,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 0.0) [pixel shader] @@ -199,7 +199,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (9.0, 11.0, 13.0, 0.0) [pixel shader] @@ -223,7 +223,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -247,7 +247,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -273,7 +273,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -285,7 +285,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 3.0, 4.0, 5.0) [pixel shader] @@ -297,7 +297,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 3.0, 4.0, 5.0) [pixel shader] @@ -336,7 +336,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 3.0, 4.0, 5.0) [pixel shader] @@ -351,5 +351,5 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (6.0, 7.0, 8.0, 9.0)

1 0

Giovanni Mascellani : vkd3d-shader/hlsl: Lower matrix casts.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: vkd3d Branch: master Commit: 3a755cb0e2a48b2a2fd82e195dfe4ba079c7ff92 URL: https://source.winehq.org/git/vkd3d.git/?a=commit;h=3a755cb0e2a48b2a2fd82e1… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Wed Jun 8 18:18:41 2022 -0500 vkd3d-shader/hlsl: Lower matrix casts. Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> --- libs/vkd3d-shader/hlsl.y | 93 ++++++++++++++++++++++- tests/hlsl-duplicate-modifiers.shader_test | 2 +- tests/hlsl-initializer-matrix.shader_test | 2 +- tests/hlsl-return-implicit-conversion.shader_test | 8 +- tests/hlsl-shape.shader_test | 10 +-- tests/matrix-semantics.shader_test | 12 +-- 6 files changed, 106 insertions(+), 21 deletions(-) diff --git a/libs/vkd3d-shader/hlsl.y b/libs/vkd3d-shader/hlsl.y index cf0f51f2..687569e8 100644 --- a/libs/vkd3d-shader/hlsl.y +++ b/libs/vkd3d-shader/hlsl.y @@ -266,6 +266,9 @@ static bool implicit_compatible_data_types(struct hlsl_type *t1, struct hlsl_typ return false; } +static struct hlsl_ir_load *add_load(struct hlsl_ctx *ctx, struct list *instrs, struct hlsl_ir_node *var_node, + struct hlsl_ir_node *offset, struct hlsl_type *data_type, const struct vkd3d_shader_location loc); + static struct hlsl_ir_node *add_cast(struct hlsl_ctx *ctx, struct list *instrs, struct hlsl_ir_node *node, struct hlsl_type *dst_type, const struct vkd3d_shader_location *loc) { @@ -275,10 +278,92 @@ static struct hlsl_ir_node *add_cast(struct hlsl_ctx *ctx, struct list *instrs, if (hlsl_types_are_equal(src_type, dst_type)) return node; - if (!(cast = hlsl_new_cast(ctx, node, dst_type, loc))) - return NULL; - list_add_tail(instrs, &cast->node.entry); - return &cast->node; + if ((src_type->type == HLSL_CLASS_MATRIX || dst_type->type == HLSL_CLASS_MATRIX) + && src_type->type <= HLSL_CLASS_LAST_NUMERIC && dst_type->type <= HLSL_CLASS_LAST_NUMERIC) + { + struct vkd3d_string_buffer *name; + static unsigned int counter = 0; + struct hlsl_ir_load *load; + struct hlsl_ir_var *var; + unsigned int dst_idx; + bool broadcast; + + broadcast = src_type->dimx == 1 && src_type->dimy == 1; + assert(dst_type->dimx * dst_type->dimy <= src_type->dimx * src_type->dimy || broadcast); + if (src_type->type == HLSL_CLASS_MATRIX && dst_type->type == HLSL_CLASS_MATRIX && !broadcast) + { + assert(dst_type->dimx <= src_type->dimx); + assert(dst_type->dimy <= src_type->dimy); + } + + name = vkd3d_string_buffer_get(&ctx->string_buffers); + vkd3d_string_buffer_printf(name, "<cast-%u>", counter++); + var = hlsl_new_synthetic_var(ctx, name->buffer, dst_type, *loc); + vkd3d_string_buffer_release(&ctx->string_buffers, name); + if (!var) + return NULL; + + for (dst_idx = 0; dst_idx < dst_type->dimx * dst_type->dimy; ++dst_idx) + { + struct hlsl_type *src_scalar_type, *dst_scalar_type; + unsigned int src_idx, src_offset, dst_offset; + struct hlsl_ir_store *store; + struct hlsl_ir_constant *c; + + if (broadcast) + { + src_idx = 0; + } + else + { + if (src_type->type == HLSL_CLASS_MATRIX && dst_type->type == HLSL_CLASS_MATRIX) + { + unsigned int x = dst_idx % dst_type->dimx, y = dst_idx / dst_type->dimx; + + src_idx = y * src_type->dimx + x; + } + else + { + src_idx = dst_idx; + } + } + + dst_offset = hlsl_compute_component_offset(ctx, dst_type, dst_idx, &dst_scalar_type); + src_offset = hlsl_compute_component_offset(ctx, src_type, src_idx, &src_scalar_type); + + if (!(c = hlsl_new_uint_constant(ctx, src_offset, loc))) + return NULL; + list_add_tail(instrs, &c->node.entry); + + if (!(load = add_load(ctx, instrs, node, &c->node, src_scalar_type, *loc))) + return NULL; + + if (!(cast = hlsl_new_cast(ctx, &load->node, dst_scalar_type, loc))) + return NULL; + list_add_tail(instrs, &cast->node.entry); + + if (!(c = hlsl_new_uint_constant(ctx, dst_offset, loc))) + return NULL; + list_add_tail(instrs, &c->node.entry); + + if (!(store = hlsl_new_store(ctx, var, &c->node, &cast->node, 0, *loc))) + return NULL; + list_add_tail(instrs, &store->node.entry); + } + + if (!(load = hlsl_new_load(ctx, var, NULL, dst_type, *loc))) + return NULL; + list_add_tail(instrs, &load->node.entry); + + return &load->node; + } + else + { + if (!(cast = hlsl_new_cast(ctx, node, dst_type, loc))) + return NULL; + list_add_tail(instrs, &cast->node.entry); + return &cast->node; + } } static struct hlsl_ir_node *add_implicit_conversion(struct hlsl_ctx *ctx, struct list *instrs, diff --git a/tests/hlsl-duplicate-modifiers.shader_test b/tests/hlsl-duplicate-modifiers.shader_test index fcae12da..6491701a 100644 --- a/tests/hlsl-duplicate-modifiers.shader_test +++ b/tests/hlsl-duplicate-modifiers.shader_test @@ -7,5 +7,5 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (0.1, 0.2, 0.3, 0.4) diff --git a/tests/hlsl-initializer-matrix.shader_test b/tests/hlsl-initializer-matrix.shader_test index ea9de9c0..7e12b0a0 100644 --- a/tests/hlsl-initializer-matrix.shader_test +++ b/tests/hlsl-initializer-matrix.shader_test @@ -55,7 +55,7 @@ float4 main() : SV_TARGET } [test] -todo draw quad +draw quad probe all rgba (21, 22, 31, 32) diff --git a/tests/hlsl-return-implicit-conversion.shader_test b/tests/hlsl-return-implicit-conversion.shader_test index bf99d9cb..545340eb 100644 --- a/tests/hlsl-return-implicit-conversion.shader_test +++ b/tests/hlsl-return-implicit-conversion.shader_test @@ -5,7 +5,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (0.4, 0.3, 0.2, 0.1) [pixel shader] @@ -15,7 +15,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (0.4, 0.3, 0.2, 0.1) [pixel shader] @@ -25,7 +25,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (0.4, 0.3, 0.2, 0.1) [pixel shader] @@ -35,7 +35,7 @@ float4x1 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (0.4, 0.3, 0.2, 0.1) [pixel shader] diff --git a/tests/hlsl-shape.shader_test b/tests/hlsl-shape.shader_test index 57d59534..65cc322c 100644 --- a/tests/hlsl-shape.shader_test +++ b/tests/hlsl-shape.shader_test @@ -211,7 +211,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -235,7 +235,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -260,7 +260,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 6.0, 8.0) [pixel shader] @@ -309,7 +309,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 0.0, 0.0) [pixel shader] @@ -321,7 +321,7 @@ float4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (2.0, 4.0, 0.0, 0.0) [pixel shader] diff --git a/tests/matrix-semantics.shader_test b/tests/matrix-semantics.shader_test index 1be50406..43f467ec 100644 --- a/tests/matrix-semantics.shader_test +++ b/tests/matrix-semantics.shader_test @@ -5,7 +5,7 @@ float4x1 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (1.0, 2.0, 3.0, 4.0) [pixel shader] @@ -15,7 +15,7 @@ row_major float1x4 main() : sv_target } [test] -todo draw quad +draw quad probe all rgba (1.0, 2.0, 3.0, 4.0) [require] @@ -44,7 +44,7 @@ row_major float4x1 main() : sv_target } [test] -todo draw quad +draw quad probe render target 0 all r (1.0) probe render target 1 all r (2.0) probe render target 2 all r (3.0) @@ -57,7 +57,7 @@ float1x4 main() : sv_target } [test] -todo draw quad +draw quad probe render target 0 all r (1.0) probe render target 1 all r (2.0) probe render target 2 all r (3.0) @@ -78,7 +78,7 @@ void main(out float1x2 x : sv_target0, out float1x2 y : sv_target2) } [test] -todo draw quad +draw quad probe render target 0 all r (1.0) probe render target 1 all r (2.0) probe render target 2 all r (5.0) @@ -96,7 +96,7 @@ void main(out row_major float1x4 x : sv_target0, out float1x2 y : sv_target1) } [test] -todo draw quad +draw quad probe render target 0 all rgba (1.0, 2.0, 3.0, 4.0) probe render target 1 all r (5.0) probe render target 2 all r (6.0)

1 0

Giovanni Mascellani : tests: Test how matrix types interact with semantics.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: vkd3d Branch: master Commit: 41df40b4ed791f14d6157ab62251363077faa029 URL: https://source.winehq.org/git/vkd3d.git/?a=commit;h=41df40b4ed791f14d6157ab… Author: Giovanni Mascellani <gmascellani(a)codeweavers.com> Date: Wed Jun 8 18:18:40 2022 -0500 tests: Test how matrix types interact with semantics. Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> --- Makefile.am | 1 + tests/matrix-semantics.shader_test | 102 +++++++++++++++++++++++++++++++++++++ 2 files changed, 103 insertions(+) diff --git a/Makefile.am b/Makefile.am index 6e827175..bdbf8e40 100644 --- a/Makefile.am +++ b/Makefile.am @@ -108,6 +108,7 @@ vkd3d_shader_tests = \ tests/hlsl-vector-indexing-uniform.shader_test \ tests/logic-operations.shader_test \ tests/math.shader_test \ + tests/matrix-semantics.shader_test \ tests/multiple-rt.shader_test \ tests/nointerpolation.shader_test \ tests/pow.shader_test \ diff --git a/tests/matrix-semantics.shader_test b/tests/matrix-semantics.shader_test new file mode 100644 index 00000000..1be50406 --- /dev/null +++ b/tests/matrix-semantics.shader_test @@ -0,0 +1,102 @@ +[pixel shader] +float4x1 main() : sv_target +{ + return float4(1.0, 2.0, 3.0, 4.0); +} + +[test] +todo draw quad +probe all rgba (1.0, 2.0, 3.0, 4.0) + +[pixel shader] +row_major float1x4 main() : sv_target +{ + return float4(1.0, 2.0, 3.0, 4.0); +} + +[test] +todo draw quad +probe all rgba (1.0, 2.0, 3.0, 4.0) + +[require] +shader model >= 4.0 + +[render target 0] +format r32 float +size (640, 480) + +[render target 1] +format r32 float +size (640, 480) + +[render target 2] +format r32 float +size (640, 480) + +[render target 3] +format r32 float +size (640, 480) + +[pixel shader] +row_major float4x1 main() : sv_target +{ + return float4(1.0, 2.0, 3.0, 4.0); +} + +[test] +todo draw quad +probe render target 0 all r (1.0) +probe render target 1 all r (2.0) +probe render target 2 all r (3.0) +probe render target 3 all r (4.0) + +[pixel shader] +float1x4 main() : sv_target +{ + return float4(1.0, 2.0, 3.0, 4.0); +} + +[test] +todo draw quad +probe render target 0 all r (1.0) +probe render target 1 all r (2.0) +probe render target 2 all r (3.0) +probe render target 3 all r (4.0) + +[pixel shader fail todo] +void main(out float1x2 x : sv_target0, out float1x2 y : sv_target1) +{ + x = float2(1.0, 2.0); + y = float2(5.0, 6.0); +} + +[pixel shader] +void main(out float1x2 x : sv_target0, out float1x2 y : sv_target2) +{ + x = float2(1.0, 2.0); + y = float2(5.0, 6.0); +} + +[test] +todo draw quad +probe render target 0 all r (1.0) +probe render target 1 all r (2.0) +probe render target 2 all r (5.0) +probe render target 3 all r (6.0) + +[render target 0] +format r32g32b32a32 float +size (640, 480) + +[pixel shader] +void main(out row_major float1x4 x : sv_target0, out float1x2 y : sv_target1) +{ + x = float4(1.0, 2.0, 3.0, 4.0); + y = float2(5.0, 6.0); +} + +[test] +todo draw quad +probe render target 0 all rgba (1.0, 2.0, 3.0, 4.0) +probe render target 1 all r (5.0) +probe render target 2 all r (6.0)

1 0

Zebediah Figura : tests: Add a test for multiple render targets.
by Alexandre Julliard 09 Jun '22

09 Jun '22

Module: vkd3d Branch: master Commit: f821a1b29a1f07c4e82e7fbcbb40645b18f6ed31 URL: https://source.winehq.org/git/vkd3d.git/?a=commit;h=f821a1b29a1f07c4e82e7fb… Author: Zebediah Figura <zfigura(a)codeweavers.com> Date: Wed Jun 8 18:18:39 2022 -0500 tests: Add a test for multiple render targets. Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> Signed-off-by: Henri Verbeet <hverbeet(a)codeweavers.com> Signed-off-by: Alexandre Julliard <julliard(a)winehq.org> --- Makefile.am | 1 + tests/multiple-rt.shader_test | 19 +++++++++++++++++++ tests/shader_runner.c | 26 ++++++++++++++++++++++++-- 3 files changed, 44 insertions(+), 2 deletions(-) diff --git a/Makefile.am b/Makefile.am index 400dc8aa..6e827175 100644 --- a/Makefile.am +++ b/Makefile.am @@ -108,6 +108,7 @@ vkd3d_shader_tests = \ tests/hlsl-vector-indexing-uniform.shader_test \ tests/logic-operations.shader_test \ tests/math.shader_test \ + tests/multiple-rt.shader_test \ tests/nointerpolation.shader_test \ tests/pow.shader_test \ tests/preproc-if.shader_test \ diff --git a/tests/multiple-rt.shader_test b/tests/multiple-rt.shader_test new file mode 100644 index 00000000..59e4acbb --- /dev/null +++ b/tests/multiple-rt.shader_test @@ -0,0 +1,19 @@ +[render target 0] +format r32g32b32a32 float +size (640, 480) + +[render target 1] +format r32g32b32a32 float +size (640, 480) + +[pixel shader] +void main(out float4 o0 : sv_target0, out float4 o1 : sv_target1) +{ + o0 = float4(0.1, 0.2, 0.3, 0.4); + o1 = float4(0.5, 0.6, 0.7, 0.8); +} + +[test] +draw quad +probe render target 0 all rgba (0.1, 0.2, 0.3, 0.4) +probe render target 1 all rgba (0.5, 0.6, 0.7, 0.8) diff --git a/tests/shader_runner.c b/tests/shader_runner.c index e2dc2dda..25986b31 100644 --- a/tests/shader_runner.c +++ b/tests/shader_runner.c @@ -468,7 +468,7 @@ static void parse_test_directive(struct shader_runner *runner, const char *line) } else if (match_string(line, "probe", &line)) { - unsigned int left, top, right, bottom, ulps; + unsigned int left, top, right, bottom, ulps, slot; struct resource_readback *rb; struct resource *resource; int ret, len; @@ -479,7 +479,7 @@ static void parse_test_directive(struct shader_runner *runner, const char *line) if (match_string(line, "uav", &line)) { - unsigned int slot = strtoul(line, &rest, 10); + slot = strtoul(line, &rest, 10); if (rest == line) fatal_error("Malformed UAV index '%s'.\n", line); @@ -487,6 +487,16 @@ static void parse_test_directive(struct shader_runner *runner, const char *line) resource = get_resource(runner, RESOURCE_TYPE_UAV, slot); } + else if (match_string(line, "render target", &line)) + { + slot = strtoul(line, &rest, 10); + + if (rest == line) + fatal_error("Malformed render target index '%s'.\n", line); + line = rest; + + resource = get_resource(runner, RESOURCE_TYPE_RENDER_TARGET, slot); + } else { resource = get_resource(runner, RESOURCE_TYPE_RENDER_TARGET, 0); @@ -836,6 +846,18 @@ void run_shader_tests(struct shader_runner *runner, int argc, char **argv, const current_sampler->v_address = D3D12_TEXTURE_ADDRESS_MODE_CLAMP; current_sampler->w_address = D3D12_TEXTURE_ADDRESS_MODE_CLAMP; } + else if (sscanf(line, "[render target %u]\n", &index)) + { + state = STATE_RESOURCE; + + memset(&current_resource, 0, sizeof(current_resource)); + + current_resource.slot = index; + current_resource.type = RESOURCE_TYPE_RENDER_TARGET; + current_resource.format = DXGI_FORMAT_R32G32B32A32_FLOAT; + current_resource.data_type = TEXTURE_DATA_FLOAT; + current_resource.texel_size = 16; + } else if (sscanf(line, "[texture %u]\n", &index)) { state = STATE_RESOURCE;

1 0

← Newer
1
...
52
53
54
55
56
57
58
...
80
Older →

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

wine-commits June 2022