- Wine-Devel - list.winehq.org

[PATCH v5 1/3] winegstreamer: Duplicate source shutdown path into constructor with leak fixes.
by Zebediah Figura Sept. 15, 2021

Sept. 15, 2021

From: Derek Lesho <dlesho(a)codeweavers.com> Instead of having mixing together IMFMediaSource::Shutdown() and the constructors' failure paths, creating confusion about what should be released where, designate ::Shutdown/::Release to shutting down fully initialized objects without checks, and keep the partially-created object cleanup code in the constructor. Signed-off-by: Derek Lesho <dlesho(a)codeweavers.com> Signed-off-by: Zebediah Figura <zfigura(a)codeweavers.com> --- v5: Resent with a couple of spacing tweaks for consistency; also use UINT_MAX instead of UINT32_MAX to match the type. dlls/winegstreamer/gst_private.h | 1 + dlls/winegstreamer/media_source.c | 102 +++++++++++++++++++----------- 2 files changed, 66 insertions(+), 37 deletions(-) diff --git a/dlls/winegstreamer/gst_private.h b/dlls/winegstreamer/gst_private.h index c6c99b1dd55..c29fc4a2437 100644 --- a/dlls/winegstreamer/gst_private.h +++ b/dlls/winegstreamer/gst_private.h @@ -22,6 +22,7 @@ #define __GST_PRIVATE_INCLUDED__ #include <assert.h> +#include <limits.h> #include <stdarg.h> #include <stdbool.h> #include <stdint.h> diff --git a/dlls/winegstreamer/media_source.c b/dlls/winegstreamer/media_source.c index 01ab626254a..95bdb9b488e 100644 --- a/dlls/winegstreamer/media_source.c +++ b/dlls/winegstreamer/media_source.c @@ -733,6 +733,12 @@ static HRESULT new_media_stream(struct media_source *source, object->IMFMediaStream_iface.lpVtbl = &media_stream_vtbl; object->ref = 1; + if (FAILED(hr = MFCreateEventQueue(&object->event_queue))) + { + free(object); + return hr; + } + IMFMediaSource_AddRef(&source->IMFMediaSource_iface); object->parent_source = source; object->stream_id = stream_id; @@ -741,20 +747,11 @@ static HRESULT new_media_stream(struct media_source *source, object->eos = FALSE; object->wg_stream = wg_stream; - if (FAILED(hr = MFCreateEventQueue(&object->event_queue))) - goto fail; - TRACE("Created stream object %p.\n", object); *out_stream = object; return S_OK; - -fail: - WARN("Failed to construct media stream, hr %#x.\n", hr); - - IMFMediaStream_Release(&object->IMFMediaStream_iface); - return hr; } static HRESULT media_stream_init_desc(struct media_stream *stream) @@ -847,10 +844,16 @@ static HRESULT media_stream_init_desc(struct media_stream *stream) goto done; if (FAILED(hr = IMFStreamDescriptor_GetMediaTypeHandler(stream->descriptor, &type_handler))) + { + IMFStreamDescriptor_Release(stream->descriptor); goto done; + } if (FAILED(hr = IMFMediaTypeHandler_SetCurrentMediaType(type_handler, stream_types[0]))) + { + IMFStreamDescriptor_Release(stream->descriptor); goto done; + } done: if (type_handler) @@ -1213,19 +1216,13 @@ static HRESULT WINAPI media_source_Shutdown(IMFMediaSource *iface) unix_funcs->wg_parser_disconnect(source->wg_parser); - if (source->read_thread) - { - source->read_thread_shutdown = true; - WaitForSingleObject(source->read_thread, INFINITE); - CloseHandle(source->read_thread); - } + source->read_thread_shutdown = true; + WaitForSingleObject(source->read_thread, INFINITE); + CloseHandle(source->read_thread); - if (source->pres_desc) - IMFPresentationDescriptor_Release(source->pres_desc); - if (source->event_queue) - IMFMediaEventQueue_Shutdown(source->event_queue); - if (source->byte_stream) - IMFByteStream_Release(source->byte_stream); + IMFPresentationDescriptor_Release(source->pres_desc); + IMFMediaEventQueue_Shutdown(source->event_queue); + IMFByteStream_Release(source->byte_stream); for (i = 0; i < source->stream_count; i++) { @@ -1233,23 +1230,18 @@ static HRESULT WINAPI media_source_Shutdown(IMFMediaSource *iface) stream->state = STREAM_SHUTDOWN; - if (stream->event_queue) - IMFMediaEventQueue_Shutdown(stream->event_queue); - if (stream->descriptor) - IMFStreamDescriptor_Release(stream->descriptor); - if (stream->parent_source) - IMFMediaSource_Release(&stream->parent_source->IMFMediaSource_iface); + IMFMediaEventQueue_Shutdown(stream->event_queue); + IMFStreamDescriptor_Release(stream->descriptor); + IMFMediaSource_Release(&stream->parent_source->IMFMediaSource_iface); IMFMediaStream_Release(&stream->IMFMediaStream_iface); } unix_funcs->wg_parser_destroy(source->wg_parser); - if (source->stream_count) - free(source->streams); + free(source->streams); - if (source->async_commands_queue) - MFUnlockWorkQueue(source->async_commands_queue); + MFUnlockWorkQueue(source->async_commands_queue); return S_OK; } @@ -1274,6 +1266,7 @@ static const IMFMediaSourceVtbl IMFMediaSource_vtbl = static HRESULT media_source_constructor(IMFByteStream *bytestream, struct media_source **out_media_source) { IMFStreamDescriptor **descriptors = NULL; + unsigned int stream_count = UINT_MAX; struct media_source *object; UINT64 total_pres_time = 0; struct wg_parser *parser; @@ -1337,15 +1330,15 @@ static HRESULT media_source_constructor(IMFByteStream *bytestream, struct media_ * leak occurs with native. */ unix_funcs->wg_parser_set_unlimited_buffering(parser); - object->stream_count = unix_funcs->wg_parser_get_stream_count(parser); + stream_count = unix_funcs->wg_parser_get_stream_count(parser); - if (!(object->streams = calloc(object->stream_count, sizeof(*object->streams)))) + if (!(object->streams = calloc(stream_count, sizeof(*object->streams)))) { hr = E_OUTOFMEMORY; goto fail; } - for (i = 0; i < object->stream_count; ++i) + for (i = 0; i < stream_count; ++i) { if (FAILED(hr = new_media_stream(object, unix_funcs->wg_parser_get_stream(parser, i), i, &object->streams[i]))) goto fail; @@ -1353,9 +1346,13 @@ static HRESULT media_source_constructor(IMFByteStream *bytestream, struct media_ if (FAILED(hr = media_stream_init_desc(object->streams[i]))) { ERR("Failed to finish initialization of media stream %p, hr %x.\n", object->streams[i], hr); - IMFMediaStream_Release(&object->streams[i]->IMFMediaStream_iface); + IMFMediaSource_Release(&object->streams[i]->parent_source->IMFMediaSource_iface); + IMFMediaEventQueue_Release(object->streams[i]->event_queue); + free(object->streams[i]); goto fail; } + + object->stream_count++; } /* init presentation descriptor */ @@ -1392,8 +1389,39 @@ static HRESULT media_source_constructor(IMFByteStream *bytestream, struct media_ fail: WARN("Failed to construct MFMediaSource, hr %#x.\n", hr); - free(descriptors); - IMFMediaSource_Release(&object->IMFMediaSource_iface); + if (descriptors) + { + for (i = 0; i < object->stream_count; i++) + IMFStreamDescriptor_Release(descriptors[i]); + free(descriptors); + } + for (i = 0; i < object->stream_count; i++) + { + struct media_stream *stream = object->streams[i]; + + IMFMediaEventQueue_Release(stream->event_queue); + IMFStreamDescriptor_Release(stream->descriptor); + IMFMediaSource_Release(&stream->parent_source->IMFMediaSource_iface); + + free(stream); + } + free(object->streams); + if (stream_count != UINT_MAX) + unix_funcs->wg_parser_disconnect(object->wg_parser); + if (object->read_thread) + { + object->read_thread_shutdown = true; + WaitForSingleObject(object->read_thread, INFINITE); + CloseHandle(object->read_thread); + } + if (object->wg_parser) + unix_funcs->wg_parser_destroy(object->wg_parser); + if (object->async_commands_queue) + MFUnlockWorkQueue(object->async_commands_queue); + if (object->event_queue) + IMFMediaEventQueue_Release(object->event_queue); + IMFByteStream_Release(object->byte_stream); + free(object); return hr; } -- 2.33.0

2 4

[PATCH 1/4] winegstreamer: Wake up the read thread when wg_parser_connect fails.
by Derek Lesho Sept. 15, 2021

Sept. 15, 2021

This bug causes read_thread never to shut down when wg_parser_connect fails, which we rely on in the next commit. Signed-off-by: Derek Lesho <dlesho(a)codeweavers.com> --- dlls/winegstreamer/wg_parser.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/dlls/winegstreamer/wg_parser.c b/dlls/winegstreamer/wg_parser.c index cd12a23d0c8..2e409f5ae18 100644 --- a/dlls/winegstreamer/wg_parser.c +++ b/dlls/winegstreamer/wg_parser.c @@ -1518,7 +1518,11 @@ static HRESULT CDECL wg_parser_connect(struct wg_parser *parser, uint64_t file_s parser->next_pull_offset = 0; if (!parser->init_gst(parser)) + { + parser->sink_connected = false; + pthread_cond_signal(&parser->read_cond); return E_FAIL; + } pthread_mutex_lock(&parser->mutex); @@ -1559,6 +1563,8 @@ static HRESULT CDECL wg_parser_connect(struct wg_parser *parser, uint64_t file_s if (parser->error) { pthread_mutex_unlock(&parser->mutex); + parser->sink_connected = false; + pthread_cond_signal(&parser->read_cond); return E_FAIL; } if (gst_pad_query_duration(stream->their_src, GST_FORMAT_TIME, &duration)) -- 2.33.0

1 3

[PATCH] sxs: Don't assign a COM object to the 3rd param of QueryInterface
by Michael Stefaniuc Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Michael Stefaniuc <mstefani(a)winehq.org> --- dlls/sxs/cache.c | 10 ++++------ dlls/sxs/name.c | 10 ++++------ 2 files changed, 8 insertions(+), 12 deletions(-) diff --git a/dlls/sxs/cache.c b/dlls/sxs/cache.c index 0acf012032d..e4fdda8c860 100644 --- a/dlls/sxs/cache.c +++ b/dlls/sxs/cache.c @@ -57,19 +57,17 @@ static inline struct cache *impl_from_IAssemblyCache(IAssemblyCache *iface) static HRESULT WINAPI cache_QueryInterface( IAssemblyCache *iface, REFIID riid, - void **obj ) + void **ret_iface ) { - struct cache *cache = impl_from_IAssemblyCache(iface); - - TRACE("%p, %s, %p\n", cache, debugstr_guid(riid), obj); + TRACE("%p, %s, %p\n", iface, debugstr_guid(riid), ret_iface); - *obj = NULL; + *ret_iface = NULL; if (IsEqualIID(riid, &IID_IUnknown) || IsEqualIID(riid, &IID_IAssemblyCache)) { IAssemblyCache_AddRef( iface ); - *obj = cache; + *ret_iface = iface; return S_OK; } diff --git a/dlls/sxs/name.c b/dlls/sxs/name.c index 6b3ae5197ac..0fab45c5018 100644 --- a/dlls/sxs/name.c +++ b/dlls/sxs/name.c @@ -56,19 +56,17 @@ static inline struct name *impl_from_IAssemblyName( IAssemblyName *iface ) static HRESULT WINAPI name_QueryInterface( IAssemblyName *iface, REFIID riid, - void **obj ) + void **ret_iface ) { - struct name *name = impl_from_IAssemblyName( iface ); - - TRACE("%p, %s, %p\n", name, debugstr_guid(riid), obj); + TRACE("%p, %s, %p\n", iface, debugstr_guid(riid), ret_iface); - *obj = NULL; + *ret_iface = NULL; if (IsEqualIID( riid, &IID_IUnknown ) || IsEqualIID( riid, &IID_IAssemblyName )) { IAssemblyName_AddRef( iface ); - *obj = name; + *ret_iface = iface; return S_OK; } -- 2.31.1

2 2

[PATCH] mfplat/tests: Remove unnecessary check (Coverity).
by Nikolay Sivov Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> --- dlls/mfplat/tests/mfplat.c | 2 -- 1 file changed, 2 deletions(-) diff --git a/dlls/mfplat/tests/mfplat.c b/dlls/mfplat/tests/mfplat.c index 54af1ee151d..541459c715b 100644 --- a/dlls/mfplat/tests/mfplat.c +++ b/dlls/mfplat/tests/mfplat.c @@ -877,8 +877,6 @@ todo_wine for (i = 0; i < sample_count; ++i) { hr = IMFMediaStream_RequestSample(video_stream, NULL); - if (i == sample_count) - break; ok(hr == S_OK, "Failed to request sample %u, hr %#x.\n", i + 1, hr); if (hr != S_OK) break; -- 2.33.0

1 0

[PATCH v3 1/2] ntdll: Use posix_fallocate(2/3) instead of fallocate(2).
by Chip Davis Sept. 15, 2021

Sept. 15, 2021

From: Charles Davis <cdavis5x(a)gmail.com> The fallocate(2) call only exists on Linux. posix_fallocate() is more portable. Signed-off-by: Chip Davis <cdavis(a)codeweavers.com> --- v3: Fix typo. v2: Don't fall back to fallocate(2). --- configure.ac | 10 +--------- dlls/ntdll/unix/file.c | 9 +++++---- 2 files changed, 6 insertions(+), 13 deletions(-) diff --git a/configure.ac b/configure.ac index fec39b6f2ac..cc8a4b79aaf 100644 --- a/configure.ac +++ b/configure.ac @@ -2183,6 +2183,7 @@ AC_CHECK_FUNCS(\ poll \ port_create \ posix_fadvise \ + posix_fallocate \ prctl \ proc_pidinfo \ readlink \ @@ -2274,15 +2275,6 @@ then AC_DEFINE(HAVE_SCHED_SETAFFINITY, 1, [Define to 1 if you have the `sched_setaffinity' function.]) fi -AC_CACHE_CHECK([for fallocate],wine_cv_have_fallocate, - AC_LINK_IFELSE([AC_LANG_PROGRAM( -[[#define _GNU_SOURCE -#include <fcntl.h>]], [[fallocate(-1, 0, 0, 0);]])],[wine_cv_have_fallocate=yes],[wine_cv_have_fallocate=no])) -if test "$wine_cv_have_fallocate" = "yes" -then - AC_DEFINE(HAVE_FALLOCATE, 1, [Define to 1 if you have the `fallocate' function.]) -fi - dnl **** Check for types **** AC_C_INLINE diff --git a/dlls/ntdll/unix/file.c b/dlls/ntdll/unix/file.c index 046338b2a82..c92cbd1db4a 100644 --- a/dlls/ntdll/unix/file.c +++ b/dlls/ntdll/unix/file.c @@ -4574,11 +4574,12 @@ NTSTATUS WINAPI NtSetInformationFile( HANDLE handle, IO_STATUS_BLOCK *io, status = STATUS_INVALID_PARAMETER; else { -#ifdef HAVE_FALLOCATE - if (fallocate( fd, 0, 0, (off_t)info->ValidDataLength.QuadPart ) == -1) +#ifdef HAVE_POSIX_FALLOCATE + int err; + if ((err = posix_fallocate( fd, 0, (off_t)info->ValidDataLength.QuadPart )) != 0) { - if (errno == EOPNOTSUPP) WARN( "fallocate not supported on this filesystem\n" ); - else status = errno_to_status( errno ); + if (err == EOPNOTSUPP) WARN( "posix_fallocate not supported on this filesystem\n" ); + else status = errno_to_status( err ); } #else FIXME( "setting valid data length not supported\n" ); -- 2.33.0

3 6

[PATCH vkd3d v=2 1/2] tests: Test the shape of the sum of two numeric values.
by Giovanni Mascellani Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> --- Makefile.am | 2 + tests/hlsl-shape.shader_test | 355 +++++++++++++++++++++++++++++++++++ 2 files changed, 357 insertions(+) create mode 100644 tests/hlsl-shape.shader_test diff --git a/Makefile.am b/Makefile.am index 05d9f00c..5decb3ef 100644 --- a/Makefile.am +++ b/Makefile.am @@ -67,6 +67,7 @@ vkd3d_shader_tests = \ tests/hlsl-nested-arrays.shader_test \ tests/hlsl-return-implicit-conversion.shader_test \ tests/hlsl-return-void.shader_test \ + tests/hlsl-shape.shader_test \ tests/hlsl-static-initializer.shader_test \ tests/hlsl-storage-qualifiers.shader_test \ tests/hlsl-struct-assignment.shader_test \ @@ -278,6 +279,7 @@ XFAIL_TESTS = \ tests/hlsl-nested-arrays.shader_test \ tests/hlsl-return-implicit-conversion.shader_test \ tests/hlsl-return-void.shader_test \ + tests/hlsl-shape.shader_test \ tests/hlsl-static-initializer.shader_test \ tests/hlsl-storage-qualifiers.shader_test \ tests/hlsl-vector-indexing.shader_test \ diff --git a/tests/hlsl-shape.shader_test b/tests/hlsl-shape.shader_test new file mode 100644 index 00000000..b96f0fd2 --- /dev/null +++ b/tests/hlsl-shape.shader_test @@ -0,0 +1,355 @@ +[pixel shader] +float4 main() : sv_target +{ + float x = 1.0; + float y = 2.0; + return float4(x + y, 0.0, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (3.0, 0.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float1 x = float1(1.0); + float2 y = float2(1.0, 2.0); + return float4(x + y, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float1 x = float1(1.0); + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float2 x = float2(1.0, 2.0); + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return float4(x + y, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float x = 1.0; + float1 y = float1(1.0); + return float4(x + y, y + x, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 2.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float x = 1.0; + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float4 x = float4(1.0, 2.0, 3.0, 4.0); + float y = 1.0; + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float2x2 x = float2x2(1.0, 2.0, + 3.0, 4.0); + float2x3 y = float2x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 7.0, 9.0) + +[pixel shader] +float4 main() : sv_target +{ + float2x3 x = float2x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0); + float2x2 y = float2x2(1.0, 2.0, + 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 7.0, 9.0) + +[pixel shader] +float4 main() : sv_target +{ + float1x4 x = float1x4(1.0, 2.0, 3.0, 4.0); + float4x4 y = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x4 x = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + float1x4 y = float1x4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x1 x = float4x1(1.0, 2.0, 3.0, 4.0); + float4x4 y = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 7.0, 12.0, 17.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x4 x = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + float4x1 y = float4x1(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 7.0, 12.0, 17.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x4 x = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + float2x3 y = float2x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0); + return float4((x + y)[0], 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float2x3 x = float2x3(1.0, 2.0, 3.0, + 4.0, 5.0, 6.0); + float4x4 y = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + return float4((x + y)[1], 0.0); +} + +[test] +draw quad +probe all rgba (9.0, 11.0, 13.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float4 x = float4(1.0, 2.0, 3.0, 4.0); + float1x4 y = float1x4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float1x4 x = float1x4(1.0, 2.0, 3.0, 4.0); + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float4 x = float4(1.0, 2.0, 3.0, 4.0); + float4x1 y = float4x1(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x1 x = float4x1(1.0, 2.0, 3.0, 4.0); + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float4 x = float4(1.0, 2.0, 3.0, 4.0); + float2x2 y = float2x2(1.0, 2.0, + 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float2x2 x = float2x2(1.0, 2.0, + 3.0, 4.0); + float4 y = float4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 6.0, 8.0) + +[pixel shader] +float4 main() : sv_target +{ + float1 x = float1(1.0); + float1x4 y = float1x4(1.0, 2.0, 3.0, 4.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float1x4 x = float1x4(1.0, 2.0, 3.0, 4.0); + float1 y = float1(1.0); + return x + y; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float2 x = float2(1.0, 2.0); + float4x1 y = float4x1(1.0, 2.0, 3.0, 4.0); + return float4(x + y, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x1 x = float4x1(1.0, 2.0, 3.0, 4.0); + float2 y = float2(1.0, 2.0); + return float4(x + y, 0.0, 0.0); +} + +[test] +draw quad +probe all rgba (2.0, 4.0, 0.0, 0.0) + +[pixel shader] +float4 main() : sv_target +{ + float x = 1.0; + float4x4 y = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + return (x + y)[0]; +} + +[test] +draw quad +probe all rgba (2.0, 3.0, 4.0, 5.0) + +[pixel shader] +float4 main() : sv_target +{ + float4x4 x = float4x4(1.0, 2.0, 3.0, 4.0, + 5.0, 6.0, 7.0, 8.0, + 9.0, 10.0, 11.0, 12.0, + 13.0, 14.0, 15.0, 16.0); + float y = 1.0; + return (x + y)[1]; +} + +[test] +draw quad +probe all rgba (6.0, 7.0, 8.0, 9.0) -- 2.33.0

2 3

[PATCH vkd3d] vkd3d-shader/hlsl: Write dimension for constant buffer registers.
by Giovanni Mascellani Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> --- libs/vkd3d-shader/hlsl_sm4.c | 1 + 1 file changed, 1 insertion(+) diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index 160da47f..beeecf85 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -613,6 +613,7 @@ static void sm4_register_from_deref(struct hlsl_ctx *ctx, struct sm4_register *r if (var->is_uniform) { reg->type = VKD3D_SM4_RT_CONSTBUFFER; + reg->dim = VKD3D_SM4_DIMENSION_VEC4; reg->idx[0] = var->buffer->reg.id; reg->idx[1] = var->buffer_offset / 4; reg->idx_count = 2; -- 2.33.0

2 1

[PATCH vkd3d] vkd3d-shader/hlsl: Write constant buffer offsets in bytes.
by Giovanni Mascellani Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Giovanni Mascellani <gmascellani(a)codeweavers.com> --- libs/vkd3d-shader/hlsl_sm4.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/libs/vkd3d-shader/hlsl_sm4.c b/libs/vkd3d-shader/hlsl_sm4.c index 160da47f..8693f2f1 100644 --- a/libs/vkd3d-shader/hlsl_sm4.c +++ b/libs/vkd3d-shader/hlsl_sm4.c @@ -515,7 +515,7 @@ static void write_sm4_rdef(struct hlsl_ctx *ctx, struct dxbc_writer *dxbc) flags |= D3D_SVF_USED; put_u32(&buffer, 0); /* name */ - put_u32(&buffer, var->buffer_offset); + put_u32(&buffer, var->buffer_offset * sizeof(float)); put_u32(&buffer, var->data_type->reg_size * sizeof(float)); put_u32(&buffer, flags); put_u32(&buffer, 0); /* type */ -- 2.33.0

2 1

[PATCH] msvcrt: Use free() on msvcrt_wstrdupa() results (Coverity).
by Nikolay Sivov Sept. 15, 2021

Sept. 15, 2021

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> --- dlls/msvcrt/process.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/dlls/msvcrt/process.c b/dlls/msvcrt/process.c index cf1dcd3d279..a5f8f0bca3d 100644 --- a/dlls/msvcrt/process.c +++ b/dlls/msvcrt/process.c @@ -1165,14 +1165,14 @@ FILE* CDECL _popen(const char* command, const char* mode) if (!(cmdW = msvcrt_wstrdupa(command))) return NULL; if (!(modeW = msvcrt_wstrdupa(mode))) { - HeapFree(GetProcessHeap(), 0, cmdW); + free(cmdW); return NULL; } ret = _wpopen(cmdW, modeW); - HeapFree(GetProcessHeap(), 0, cmdW); - HeapFree(GetProcessHeap(), 0, modeW); + free(cmdW); + free(modeW); return ret; } @@ -1274,7 +1274,7 @@ int CDECL system(const char* cmd) if ((cmdW = msvcrt_wstrdupa(cmd))) { res = _wsystem(cmdW); - HeapFree(GetProcessHeap(), 0, cmdW); + free(cmdW); } return res; } -- 2.33.0

2 1

[PATCH v5] dsound: Commit next audio chunk between play cursor and write cursor to playing.
by Eduard Permyakov Sept. 15, 2021

Sept. 15, 2021

This region of the audio buffer is forbidden to be written to by the DirectSound specification. The documentation states: "The write cursor is the point after which it is safe to write data into the buffer. The block between the play cursor and the write cursor is already committed to be played, and cannot be changed safely." However, some applications still do this, which has lead to audio glitches only when using the Wine DirectSound implementation. Experiments showed that the native DirctSound implementation will still play the old audio the first time around when the buffer region gets overwritten. Use an approach of copying the next forbidden region into a "committed buffer" to add the same behavior to the Wine implementation. Out of performance considerations, only copy data to the committed buffer when we detect that an overwrite is possible (i.e. the current mixing region of the buffer gets locked). Signed-off-by: Eduard Permyakov <epermyakov(a)codeweavers.com> --- dlls/dsound/buffer.c | 48 +++++++++++++++++++++++++++++++++--- dlls/dsound/dsound_convert.c | 33 ++++++++++++------------- dlls/dsound/dsound_private.h | 11 ++++++--- dlls/dsound/mixer.c | 48 ++++++++++++++++++++++++++++-------- 4 files changed, 107 insertions(+), 33 deletions(-) diff --git a/dlls/dsound/buffer.c b/dlls/dsound/buffer.c index dc3b54906ce..63468732e1d 100644 --- a/dlls/dsound/buffer.c +++ b/dlls/dsound/buffer.c @@ -101,6 +101,24 @@ static int __cdecl notify_compar(const void *l, const void *r) return 1; } +static void commit_next_chunk(IDirectSoundBufferImpl *dsb) +{ + void *dstbuff = dsb->committedbuff, *srcbuff = dsb->buffer->memory; + DWORD srcoff = dsb->sec_mixpos, srcsize = dsb->buflen, cpysize = dsb->writelead; + + if(cpysize > srcsize - srcoff) { + DWORD overflow = cpysize - (srcsize - srcoff); + memcpy(dstbuff, (BYTE*)srcbuff + srcoff, srcsize - srcoff); + memcpy((BYTE*)dstbuff + (srcsize - srcoff), srcbuff, overflow); + }else{ + memcpy(dstbuff, (BYTE*)srcbuff + srcoff, cpysize); + } + + dsb->use_committed = TRUE; + dsb->committed_mixpos = 0; + TRACE("committing %u bytes from offset %u\n", dsb->writelead, dsb->sec_mixpos); +} + static HRESULT WINAPI IDirectSoundNotifyImpl_SetNotificationPositions(IDirectSoundNotify *iface, DWORD howmuch, const DSBPOSITIONNOTIFY *notify) { @@ -503,8 +521,10 @@ static HRESULT WINAPI IDirectSoundBufferImpl_Lock(IDirectSoundBuffer8 *iface, DW if (writecursor+writebytes <= This->buflen) { *(LPBYTE*)lplpaudioptr1 = This->buffer->memory+writecursor; - if (This->sec_mixpos >= writecursor && This->sec_mixpos < writecursor + writebytes && This->state == STATE_PLAYING) + if (This->sec_mixpos >= writecursor && This->sec_mixpos < writecursor + writebytes && This->state == STATE_PLAYING) { WARN("Overwriting mixing position, case 1\n"); + commit_next_chunk(This); + } *audiobytes1 = writebytes; if (lplpaudioptr2) *(LPBYTE*)lplpaudioptr2 = NULL; @@ -519,16 +539,20 @@ static HRESULT WINAPI IDirectSoundBufferImpl_Lock(IDirectSoundBuffer8 *iface, DW *(LPBYTE*)lplpaudioptr1 = This->buffer->memory+writecursor; *audiobytes1 = This->buflen-writecursor; This->buffer->lockedbytes += *audiobytes1; - if (This->sec_mixpos >= writecursor && This->sec_mixpos < writecursor + writebytes && This->state == STATE_PLAYING) + if (This->sec_mixpos >= writecursor && This->sec_mixpos < writecursor + writebytes && This->state == STATE_PLAYING) { WARN("Overwriting mixing position, case 2\n"); + commit_next_chunk(This); + } if (lplpaudioptr2) *(LPBYTE*)lplpaudioptr2 = This->buffer->memory; if (audiobytes2) { *audiobytes2 = writebytes-(This->buflen-writecursor); This->buffer->lockedbytes += *audiobytes2; } - if (audiobytes2 && This->sec_mixpos < remainder && This->state == STATE_PLAYING) + if (audiobytes2 && This->sec_mixpos < remainder && This->state == STATE_PLAYING) { WARN("Overwriting mixing position, case 3\n"); + commit_next_chunk(This); + } TRACE("Locked %p(%i bytes) and %p(%i bytes) writecursor=%d\n", *(LPBYTE*)lplpaudioptr1, *audiobytes1, lplpaudioptr2 ? *(LPBYTE*)lplpaudioptr2 : NULL, audiobytes2 ? *audiobytes2: 0, writecursor); } @@ -1081,6 +1105,12 @@ HRESULT secondarybuffer_create(DirectSoundDevice *device, const DSBUFFERDESC *ds /* calculate fragment size and write lead */ DSOUND_RecalcFormat(dsb); + dsb->committedbuff = HeapAlloc(GetProcessHeap(), 0, dsb->writelead); + if(!dsb->committedbuff) { + IDirectSoundBuffer8_Release(&dsb->IDirectSoundBuffer8_iface); + return DSERR_OUTOFMEMORY; + } + if (dsb->dsbd.dwFlags & DSBCAPS_CTRL3D) { dsb->ds3db_ds3db.dwSize = sizeof(DS3DBUFFER); dsb->ds3db_ds3db.vPosition.x = 0.0; @@ -1135,6 +1165,7 @@ void secondarybuffer_destroy(IDirectSoundBufferImpl *This) HeapFree(GetProcessHeap(), 0, This->notifies); HeapFree(GetProcessHeap(), 0, This->pwfx); + HeapFree(GetProcessHeap(), 0, This->committedbuff); if (This->filters) { int i; @@ -1157,6 +1188,7 @@ HRESULT IDirectSoundBufferImpl_Duplicate( { IDirectSoundBufferImpl *dsb; HRESULT hres = DS_OK; + VOID *committedbuff; TRACE("(%p,%p,%p)\n", device, ppdsb, pdsb); dsb = HeapAlloc(GetProcessHeap(),0,sizeof(*dsb)); @@ -1166,6 +1198,13 @@ HRESULT IDirectSoundBufferImpl_Duplicate( return DSERR_OUTOFMEMORY; } + committedbuff = HeapAlloc(GetProcessHeap(),0,pdsb->writelead); + if (committedbuff == NULL) { + HeapFree(GetProcessHeap(),0,dsb); + *ppdsb = NULL; + return DSERR_OUTOFMEMORY; + } + AcquireSRWLockShared(&pdsb->lock); CopyMemory(dsb, pdsb, sizeof(*dsb)); @@ -1175,6 +1214,7 @@ HRESULT IDirectSoundBufferImpl_Duplicate( ReleaseSRWLockShared(&pdsb->lock); if (dsb->pwfx == NULL) { + HeapFree(GetProcessHeap(),0,committedbuff); HeapFree(GetProcessHeap(),0,dsb); *ppdsb = NULL; return DSERR_OUTOFMEMORY; @@ -1192,6 +1232,7 @@ HRESULT IDirectSoundBufferImpl_Duplicate( dsb->notifies = NULL; dsb->nrofnotifies = 0; dsb->device = device; + dsb->committedbuff = committedbuff; DSOUND_RecalcFormat(dsb); InitializeSRWLock(&dsb->lock); @@ -1202,6 +1243,7 @@ HRESULT IDirectSoundBufferImpl_Duplicate( list_remove(&dsb->entry); dsb->buffer->ref--; HeapFree(GetProcessHeap(),0,dsb->pwfx); + HeapFree(GetProcessHeap(),0,dsb->committedbuff); HeapFree(GetProcessHeap(),0,dsb); dsb = NULL; }else diff --git a/dlls/dsound/dsound_convert.c b/dlls/dsound/dsound_convert.c index 4735178a9a8..67344565860 100644 --- a/dlls/dsound/dsound_convert.c +++ b/dlls/dsound/dsound_convert.c @@ -55,26 +55,25 @@ WINE_DEFAULT_DEBUG_CHANNEL(dsound); #define le32(x) (x) #endif -static float get8(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +static float get8(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { - const BYTE* buf = dsb->buffer->memory; - buf += pos + channel; + const BYTE *buf = base + channel; return (buf[0] - 0x80) / (float)0x80; } -static float get16(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +static float get16(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { - const BYTE* buf = dsb->buffer->memory; - const SHORT *sbuf = (const SHORT*)(buf + pos + 2 * channel); + const BYTE *buf = base + 2 * channel; + const SHORT *sbuf = (const SHORT*)(buf); SHORT sample = (SHORT)le16(*sbuf); return sample / (float)0x8000; } -static float get24(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +static float get24(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { LONG sample; - const BYTE* buf = dsb->buffer->memory; - buf += pos + 3 * channel; + const BYTE *buf = base + 3 * channel; + /* The next expression deliberately has an overflow for buf[2] >= 0x80, this is how negative values are made. */ @@ -82,32 +81,32 @@ static float get24(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) return sample / (float)0x80000000U; } -static float get32(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +static float get32(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { - const BYTE* buf = dsb->buffer->memory; - const LONG *sbuf = (const LONG*)(buf + pos + 4 * channel); + const BYTE *buf = base + 4 * channel; + const LONG *sbuf = (const LONG*)(buf); LONG sample = le32(*sbuf); return sample / (float)0x80000000U; } -static float getieee32(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +static float getieee32(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { - const BYTE* buf = dsb->buffer->memory; - const float *sbuf = (const float*)(buf + pos + 4 * channel); + const BYTE *buf = base + 4 * channel; + const float *sbuf = (const float*)(buf); /* The value will be clipped later, when put into some non-float buffer */ return *sbuf; } const bitsgetfunc getbpp[5] = {get8, get16, get24, get32, getieee32}; -float get_mono(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) +float get_mono(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) { DWORD channels = dsb->pwfx->nChannels; DWORD c; float val = 0; /* XXX: does Windows include LFE into the mix? */ for (c = 0; c < channels; c++) - val += dsb->get_aux(dsb, pos, c); + val += dsb->get_aux(dsb, base, c); val /= channels; return val; } diff --git a/dlls/dsound/dsound_private.h b/dlls/dsound/dsound_private.h index bdb9ebee544..9ec96504a50 100644 --- a/dlls/dsound/dsound_private.h +++ b/dlls/dsound/dsound_private.h @@ -43,7 +43,7 @@ typedef struct IDirectSoundBufferImpl IDirectSoundBufferImpl; typedef struct DirectSoundDevice DirectSoundDevice; /* dsound_convert.h */ -typedef float (*bitsgetfunc)(const IDirectSoundBufferImpl *, DWORD, DWORD); +typedef float (*bitsgetfunc)(const IDirectSoundBufferImpl *, BYTE *, DWORD); typedef void (*bitsputfunc)(const IDirectSoundBufferImpl *, DWORD, DWORD, float); extern const bitsgetfunc getbpp[5] DECLSPEC_HIDDEN; void putieee32(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel, float value) DECLSPEC_HIDDEN; @@ -153,7 +153,12 @@ struct IDirectSoundBufferImpl LONG64 freqAccNum; /* used for mixing */ DWORD sec_mixpos; - + /* Holds a copy of the next 'writelead' bytes, to be used for mixing. This makes it + * so that these bytes get played once even if this region of the buffer gets overwritten, + * which is more in-line with native DirectSound behavior. */ + BOOL use_committed; + LPVOID committedbuff; + DWORD committed_mixpos; /* IDirectSoundNotify fields */ LPDSBPOSITIONNOTIFY notifies; int nrofnotifies; @@ -171,7 +176,7 @@ struct IDirectSoundBufferImpl struct list entry; }; -float get_mono(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel) DECLSPEC_HIDDEN; +float get_mono(const IDirectSoundBufferImpl *dsb, BYTE *base, DWORD channel) DECLSPEC_HIDDEN; void put_mono2stereo(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel, float value) DECLSPEC_HIDDEN; void put_mono2quad(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel, float value) DECLSPEC_HIDDEN; void put_stereo2quad(const IDirectSoundBufferImpl *dsb, DWORD pos, DWORD channel, float value) DECLSPEC_HIDDEN; diff --git a/dlls/dsound/mixer.c b/dlls/dsound/mixer.c index 124d1e4fba1..e9d8a6e3cdd 100644 --- a/dlls/dsound/mixer.c +++ b/dlls/dsound/mixer.c @@ -276,22 +276,35 @@ void DSOUND_CheckEvent(const IDirectSoundBufferImpl *dsb, DWORD playpos, int len } static inline float get_current_sample(const IDirectSoundBufferImpl *dsb, - DWORD mixpos, DWORD channel) + BYTE *buffer, DWORD buflen, DWORD mixpos, DWORD channel) { - if (mixpos >= dsb->buflen && !(dsb->playflags & DSBPLAY_LOOPING)) + if (mixpos >= buflen && !(dsb->playflags & DSBPLAY_LOOPING)) return 0.0f; - return dsb->get(dsb, mixpos % dsb->buflen, channel); + return dsb->get(dsb, buffer + (mixpos % buflen), channel); } static UINT cp_fields_noresample(IDirectSoundBufferImpl *dsb, UINT count) { UINT istride = dsb->pwfx->nBlockAlign; UINT ostride = dsb->device->pwfx->nChannels * sizeof(float); + UINT committed_samples = 0; DWORD channel, i; - for (i = 0; i < count; i++) + + if(dsb->use_committed) { + committed_samples = (dsb->writelead - dsb->committed_mixpos) / istride; + committed_samples = committed_samples <= count ? committed_samples : count; + } + + for (i = 0; i < committed_samples; i++) + for (channel = 0; channel < dsb->mix_channels; channel++) + dsb->put(dsb, i * ostride, channel, get_current_sample(dsb, dsb->committedbuff, + dsb->writelead, dsb->committed_mixpos + i * istride, channel)); + + for (; i < count; i++) for (channel = 0; channel < dsb->mix_channels; channel++) - dsb->put(dsb, i * ostride, channel, get_current_sample(dsb, - dsb->sec_mixpos + i * istride, channel)); + dsb->put(dsb, i * ostride, channel, get_current_sample(dsb, dsb->buffer->memory, + dsb->buflen, dsb->sec_mixpos + i * istride, channel)); + return count; } @@ -300,6 +313,7 @@ static UINT cp_fields_resample(IDirectSoundBufferImpl *dsb, UINT count, LONG64 * UINT i, channel; UINT istride = dsb->pwfx->nBlockAlign; UINT ostride = dsb->device->pwfx->nChannels * sizeof(float); + UINT committed_samples = 0; LONG64 freqAcc_start = *freqAccNum; LONG64 freqAcc_end = freqAcc_start + count * dsb->freqAdjustNum; @@ -326,16 +340,24 @@ static UINT cp_fields_resample(IDirectSoundBufferImpl *dsb, UINT count, LONG64 * fir_copy = dsb->device->cp_buffer; intermediate = fir_copy + fir_cachesize; + if(dsb->use_committed) { + committed_samples = (dsb->writelead - dsb->committed_mixpos) / istride; + committed_samples = committed_samples <= required_input ? committed_samples : required_input; + } /* Important: this buffer MUST be non-interleaved * if you want -msse3 to have any effect. * This is good for CPU cache effects, too. */ itmp = intermediate; - for (channel = 0; channel < channels; channel++) - for (i = 0; i < required_input; i++) - *(itmp++) = get_current_sample(dsb, - dsb->sec_mixpos + i * istride, channel); + for (channel = 0; channel < channels; channel++) { + for (i = 0; i < committed_samples; i++) + *(itmp++) = get_current_sample(dsb, dsb->committedbuff, + dsb->writelead, dsb->committed_mixpos + i * istride, channel); + for (; i < required_input; i++) + *(itmp++) = get_current_sample(dsb, dsb->buffer->memory, + dsb->buflen, dsb->sec_mixpos + i * istride, channel); + } for(i = 0; i < count; ++i) { UINT int_fir_steps = (freqAcc_start + i * dsb->freqAdjustNum) * dsbfirstep / dsb->freqAdjustDen; @@ -389,6 +411,12 @@ static void cp_fields(IDirectSoundBufferImpl *dsb, UINT count, LONG64 *freqAccNu } dsb->sec_mixpos = ipos; + + if(dsb->use_committed) { + dsb->committed_mixpos += adv * dsb->pwfx->nBlockAlign; + if(dsb->committed_mixpos >= dsb->writelead) + dsb->use_committed = FALSE; + } } /** -- 2.25.1

2 6