May 2022 - Wine-devel - list.winehq.org

Envisioned Gitlab workflow
by Alexandre Julliard 11 May '22

11 May '22

Folks, Here's a description of the envisioned workflow for using Gitlab as the primary platform for Wine development, should we decide to adopt it. This will probably evolve as we experiment and find out what works or doesn't work for us. All feedback/suggestions are welcome. 1. All patch series are submitted as Merge Requests, so that the proper information can be captured. The usual rules apply (small patches, small number of patches per MR, sign-off by submitter). There are various methods available for creating merge requests: - using the web interface - directly when pushing: git push -o merge_request.create cf. https://docs.gitlab.com/ee/user/project/push_options.html - from the command line, eg. 'glab mr create' cf. https://glab.readthedocs.io/en/latest/mr/create.html - by email cf. https://docs.gitlab.com/ee/user/project/merge_requests/creating_merge_reque… 2. The merge request will run through the CI pipeline, ideally including a testbot run (not implemented yet). 3. The patches of the merge request will be forwarded to wine-devel by the mail gateway, for people who prefer to use their existing email tools to manage patches. 4. Reviewers will be automatically assigned to the merge request based on the MAINTAINERS file (not implemented yet). Reviewers can also be assigned manually. People assigned as reviewers should receive a notification from gitlab. 5. When someone starts reviewing a patch, they can optionally assign the patch to themselves, so that the submitter is aware that someone is working on their patch. 6. Feedback on patches can be sent by replying to the wine-devel email, or by adding a comment on gitlab. The mail gateway will bridge things so that all comments are reflected both on gitlab and on wine-devel. Note: if you subscribe to gitlab notifications (recommended) currently you'll get many duplicate emails. There are mail headers that can be used for filtering, but we may want to handle this differently, suggestions are welcome. 7. If there are comments, the submitter can address them and update the merge request by pushing to the MR branch. The mail gateway will send the new version of the patches to wine-devel (except for trivial changes) for further review. 8. Once a reviewer approves a merge request, they should mark it approved in gitlab. The idea is that this will replace the existing "send a signoff by email" mechanism. Note that an approval applies to the whole merge request, not to individual patches. This should be one more reason to keep patch series small. Approvals can be sent by: - using the approve button on the web interface - replying to a gitlab notification with "/approve" (note: replying to a wine-devel mail won't work, it needs the magic gitlab headers) cf. https://docs.gitlab.com/ee/user/project/quick_actions.html - from the command line, eg. 'glab mr approve' cf. https://glab.readthedocs.io/en/latest/mr/approve.html 9. Once a merge request has been approved by all reviewers, I'll merge it into my tree, do a full testbot run to make sure it doesn't break anything, then push the results and mark the merge request as 'merged'. 10. If a merge request is no longer relevant, the submitter should close it. We may want to add an auto-close process for stale merge requests at some point. -- Alexandre Julliard julliard(a)winehq.org

6 9

[PATCH 1/3] include: Add MIL/DWM HRESULT codes.
by Jactry Zeng 11 May '22

11 May '22

Signed-off-by: Jactry Zeng <jzeng(a)codeweavers.com> --- include/winerror.h | 41 +++++++++++++++++++++++++++++++++++++++++ 1 file changed, 41 insertions(+) diff --git a/include/winerror.h b/include/winerror.h index d51d5648cdb..fb44eaacbed 100644 --- a/include/winerror.h +++ b/include/winerror.h @@ -3884,6 +3884,47 @@ static inline HRESULT HRESULT_FROM_WIN32(unsigned int x) #define WINCODEC_ERR_WIN32ERROR _HRESULT_TYPEDEF_(0x88982f94) #define WINCODEC_ERR_INVALIDPROGRESSIVELEVEL _HRESULT_TYPEDEF_(0x88982f95) +#define MILERR_OBJECTBUSY _HRESULT_TYPEDEF_(0x88980001) +#define MILERR_INSUFFICIENTBUFFER _HRESULT_TYPEDEF_(0x88980002) +#define MILERR_WIN32ERROR _HRESULT_TYPEDEF_(0x88980003) +#define MILERR_SCANNER_FAILED _HRESULT_TYPEDEF_(0x88980004) +#define MILERR_SCREENACCESSDENIED _HRESULT_TYPEDEF_(0x88980005) +#define MILERR_DISPLAYSTATEINVALID _HRESULT_TYPEDEF_(0x88980006) +#define MILERR_NONINVERTIBLEMATRIX _HRESULT_TYPEDEF_(0x88980007) +#define MILERR_ZEROVECTOR _HRESULT_TYPEDEF_(0x88980008) +#define MILERR_TERMINATED _HRESULT_TYPEDEF_(0x88980009) +#define MILERR_BADNUMBER _HRESULT_TYPEDEF_(0x8898000a) +#define MILERR_INTERNALERROR _HRESULT_TYPEDEF_(0x88980080) +#define MILERR_DISPLAYFORMATNOTSUPPORTED _HRESULT_TYPEDEF_(0x88980084) +#define MILERR_INVALIDCALL _HRESULT_TYPEDEF_(0x88980085) +#define MILERR_ALREADYLOCKED _HRESULT_TYPEDEF_(0x88980086) +#define MILERR_NOTLOCKED _HRESULT_TYPEDEF_(0x88980087) +#define MILERR_DEVICECANNOTRENDERTEXT _HRESULT_TYPEDEF_(0x88980088) +#define MILERR_GLYPHBITMAPMISSED _HRESULT_TYPEDEF_(0x88980089) +#define MILERR_MALFORMEDGLYPHCACHE _HRESULT_TYPEDEF_(0x8898008a) +#define MILERR_GENERIC_IGNORE _HRESULT_TYPEDEF_(0x8898008b) +#define MILERR_MALFORMED_GUIDELINE_DATA _HRESULT_TYPEDEF_(0x8898008c) +#define MILERR_NO_HARDWARE_DEVICE _HRESULT_TYPEDEF_(0x8898008d) +#define MILERR_NEED_RECREATE_AND_PRESENT _HRESULT_TYPEDEF_(0x8898008e) +#define MILERR_ALREADY_INITIALIZED _HRESULT_TYPEDEF_(0x8898008f) +#define MILERR_MISMATCHED_SIZE _HRESULT_TYPEDEF_(0x88980090) +#define MILERR_NO_REDIRECTION_SURFACE_AVAILABLE _HRESULT_TYPEDEF_(0x88980091) +#define MILERR_REMOTING_NOT_SUPPORTED _HRESULT_TYPEDEF_(0x88980092) +#define MILERR_QUEUED_PRESENT_NOT_SUPPORTED _HRESULT_TYPEDEF_(0x88980093) +#define MILERR_NOT_QUEUING_PRESENTS _HRESULT_TYPEDEF_(0x88980094) +#define MILERR_NO_REDIRECTION_SURFACE_RETRY_LATER _HRESULT_TYPEDEF_(0x88980095) +#define MILERR_TOOMANYSHADERELEMNTS _HRESULT_TYPEDEF_(0x88980096) +#define MILERR_MROW_READLOCK_FAILED _HRESULT_TYPEDEF_(0x88980097) +#define MILERR_MROW_UPDATE_FAILED _HRESULT_TYPEDEF_(0x88980098) +#define MILERR_SHADER_COMPILE_FAILED _HRESULT_TYPEDEF_(0x88980099) +#define MILERR_MAX_TEXTURE_SIZE_EXCEEDED _HRESULT_TYPEDEF_(0x8898009a) +#define MILERR_QPC_TIME_WENT_BACKWARD _HRESULT_TYPEDEF_(0x8898009b) +#define MILERR_DXGI_ENUMERATION_OUT_OF_SYNC _HRESULT_TYPEDEF_(0x8898009d) +#define MILERR_ADAPTER_NOT_FOUND _HRESULT_TYPEDEF_(0x8898009e) +#define MILERR_COLORSPACE_NOT_SUPPORTED _HRESULT_TYPEDEF_(0x8898009f) +#define MILERR_PREFILTER_NOT_SUPPORTED _HRESULT_TYPEDEF_(0x889800a0) +#define MILERR_DISPLAYID_ACCESS_DENIED _HRESULT_TYPEDEF_(0x889800a1) + #define DWRITE_E_FILEFORMAT _HRESULT_TYPEDEF_(0x88985000) #define DWRITE_E_UNEXPECTED _HRESULT_TYPEDEF_(0x88985001) #define DWRITE_E_NOFONT _HRESULT_TYPEDEF_(0x88985002) -- 2.35.1

1 2

[PATCH 0/1] MR52: include: Add TF_TMF/TMAW defines
by Alistair Leslie-Hughes (＠alesliehughes) 10 May '22

10 May '22

Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> -- https://gitlab.winehq.org/wine/wine/-/merge_requests/52

2 1

[PATCH 0/1] MR46: d3dx10/tests: Fix d3dx10_34 test failures on all Windows versions.
by Sven Baars (＠sbaars) 10 May '22

10 May '22

Should fix https://bugs.winehq.org/show_bug.cgi?id=51827 I tested this on a Windows 10 VM. With last=4 it failed about 50% of the time, with last=3 it passed 100 tests without any failures. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/46

3 4

[PATCH v2 0/5] MR49: MPEG audio decoder
by Anton Baskanov (＠baskanov) 10 May '22

10 May '22

-- v2: winegstreamer: Set sample timestamps in MPEG audio decoder. quartz/tests: Add tests for MPEG audio decoder sample timestamps. winegstreamer: Register MPEG audio decoder. winegstreamer: Add IMpegAudioDecoder stub to MPEG audio decoder. include: Add IMpegAudioDecoder definition. https://gitlab.winehq.org/wine/wine/-/merge_requests/49

3 7

[PATCH] wineusb.sys: Remove erroneous call to LeaveCriticalSection() in add_usb_device.
by Jinoh Kang 10 May '22

10 May '22

Signed-off-by: Jinoh Kang <jinoh.kang.kr(a)gmail.com> --- dlls/wineusb.sys/wineusb.c | 1 - 1 file changed, 1 deletion(-) diff --git a/dlls/wineusb.sys/wineusb.c b/dlls/wineusb.sys/wineusb.c index cca0257bb02..c41ea580137 100644 --- a/dlls/wineusb.sys/wineusb.c +++ b/dlls/wineusb.sys/wineusb.c @@ -114,7 +114,6 @@ static void add_usb_device(libusb_device *libusb_device) FILE_DEVICE_USB, 0, FALSE, &device_obj))) { ERR("Failed to create device, status %#x.\n", status); - LeaveCriticalSection(&wineusb_cs); libusb_close(handle); return; } -- 2.34.1

2 1

[PATCH 1/2] evr/sample: Avoid a race condition when releasing a tracked sample.
by Nikolay Sivov 10 May '22

10 May '22

Same as b1807751219e256867eaac5cc4d8b97a5482978f. Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> --- dlls/evr/sample.c | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/dlls/evr/sample.c b/dlls/evr/sample.c index 1d29b6cdc2c..738f803fe66 100644 --- a/dlls/evr/sample.c +++ b/dlls/evr/sample.c @@ -920,19 +920,21 @@ static ULONG WINAPI video_sample_AddRef(IMFSample *iface) static ULONG WINAPI video_sample_Release(IMFSample *iface) { struct video_sample *sample = impl_from_IMFSample(iface); - ULONG refcount; + ULONG refcount = InterlockedDecrement(&sample->refcount); + IMFAsyncResult *tracked_result = NULL; IMFSample_LockStore(sample->sample); - if (sample->tracked_result && sample->tracked_refcount == (sample->refcount - 1)) + if (sample->tracked_result && sample->tracked_refcount == refcount) { - video_sample_tracking_thread_invoke(sample->tracked_result); - IMFAsyncResult_Release(sample->tracked_result); + tracked_result = sample->tracked_result; + video_sample_tracking_thread_invoke(tracked_result); sample->tracked_result = NULL; sample->tracked_refcount = 0; } IMFSample_UnlockStore(sample->sample); - refcount = InterlockedDecrement(&sample->refcount); + if (tracked_result) + IMFAsyncResult_Release(tracked_result); TRACE("%p, refcount %lu.\n", iface, refcount); -- 2.35.1

1 1

[PATCH] mfplat/sample: Fix use-after-free of a critical section.
by Nikolay Sivov 10 May '22

10 May '22

Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> --- dlls/mfplat/sample.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) diff --git a/dlls/mfplat/sample.c b/dlls/mfplat/sample.c index ed39b5f7297..687ada1a477 100644 --- a/dlls/mfplat/sample.c +++ b/dlls/mfplat/sample.c @@ -183,22 +183,25 @@ static ULONG WINAPI sample_tracked_Release(IMFSample *iface) { struct sample *sample = impl_from_IMFSample(iface); ULONG refcount = InterlockedDecrement(&sample->attributes.ref); + IRtwqAsyncResult *tracked_result = NULL; HRESULT hr; EnterCriticalSection(&sample->attributes.cs); if (sample->tracked_result && sample->tracked_refcount == refcount) { - IRtwqAsyncResult *tracked_result = sample->tracked_result; + tracked_result = sample->tracked_result; sample->tracked_result = NULL; sample->tracked_refcount = 0; /* Call could fail if queue system is not initialized, it's not critical. */ if (FAILED(hr = RtwqInvokeCallback(tracked_result))) WARN("Failed to invoke tracking callback, hr %#lx.\n", hr); - IRtwqAsyncResult_Release(tracked_result); } LeaveCriticalSection(&sample->attributes.cs); + if (tracked_result) + IRtwqAsyncResult_Release(tracked_result); + TRACE("%p, refcount %lu.\n", iface, refcount); if (!refcount) -- 2.35.1

1 0

[PATCH] mf/session: Do not request more samples when SA sample becomes available.
by Nikolay Sivov 10 May '22

10 May '22

The issue here is that upstream nodes get their request counters increased while sink hasn't requested any. This causes desync between sink request counters and upstream node counters, potentially causing sample drops and sink waiting forever for samples that will never come. Case like that is easy to reproduce with mfplay-based player, where occasionally end-of-presentation condition does not trigger, but new samples are coming in. Signed-off-by: Nikolay Sivov <nsivov(a)codeweavers.com> --- dlls/mf/session.c | 71 +++++++++++++++++++++++++++++++++++++++++------ 1 file changed, 62 insertions(+), 9 deletions(-) diff --git a/dlls/mf/session.c b/dlls/mf/session.c index 85b63f767fe..a78e7cfc473 100644 --- a/dlls/mf/session.c +++ b/dlls/mf/session.c @@ -2369,6 +2369,8 @@ static HRESULT WINAPI session_commands_callback_GetParameters(IMFAsyncCallback * return E_NOTIMPL; } +static void session_deliver_pending_samples(struct media_session *session, IMFTopologyNode *node); + static HRESULT WINAPI session_commands_callback_Invoke(IMFAsyncCallback *iface, IMFAsyncResult *result) { struct session_op *op = impl_op_from_IUnknown(IMFAsyncResult_GetStateNoAddRef(result)); @@ -2411,7 +2413,7 @@ static HRESULT WINAPI session_commands_callback_Invoke(IMFAsyncCallback *iface, { if (SUCCEEDED(IMFTopologyNode_GetInput(topo_node->node, 0, &upstream_node, &upstream_output))) { - session_request_sample_from_node(session, upstream_node, upstream_output); + session_deliver_pending_samples(session, upstream_node); IMFTopologyNode_Release(upstream_node); } } @@ -2952,19 +2954,19 @@ static void session_deliver_sample_to_node(struct media_session *session, IMFTop switch (node_type) { case MF_TOPOLOGY_OUTPUT_NODE: - if (sample) + if (topo_node->u.sink.requests) { - if (topo_node->u.sink.requests) + if (sample) { if (FAILED(hr = IMFStreamSink_ProcessSample(topo_node->object.sink_stream, sample))) WARN("Stream sink failed to process sample, hr %#lx.\n", hr); - topo_node->u.sink.requests--; } - } - else if (FAILED(hr = IMFStreamSink_PlaceMarker(topo_node->object.sink_stream, MFSTREAMSINK_MARKER_ENDOFSEGMENT, - NULL, NULL))) - { - WARN("Failed to place sink marker, hr %#lx.\n", hr); + else if (FAILED(hr = IMFStreamSink_PlaceMarker(topo_node->object.sink_stream, MFSTREAMSINK_MARKER_ENDOFSEGMENT, + NULL, NULL))) + { + WARN("Failed to place sink marker, hr %#lx.\n", hr); + } + topo_node->u.sink.requests--; } break; case MF_TOPOLOGY_TRANSFORM_NODE: @@ -3048,6 +3050,57 @@ static void session_deliver_sample_to_node(struct media_session *session, IMFTop } } +static void session_deliver_pending_samples(struct media_session *session, IMFTopologyNode *node) +{ + struct sample *sample_entry, *sample_entry2; + IMFTopologyNode *downstream_node; + struct topo_node *topo_node; + MF_TOPOLOGY_TYPE node_type; + DWORD downstream_input; + TOPOID node_id; + unsigned int i; + + IMFTopologyNode_GetNodeType(node, &node_type); + IMFTopologyNode_GetTopoNodeID(node, &node_id); + + topo_node = session_get_node_by_id(session, node_id); + + switch (node_type) + { + case MF_TOPOLOGY_TRANSFORM_NODE: + + transform_node_pull_samples(session, topo_node); + + /* Push down all available output. */ + for (i = 0; i < topo_node->u.transform.output_count; ++i) + { + if (FAILED(IMFTopologyNode_GetOutput(node, i, &downstream_node, &downstream_input))) + { + WARN("Failed to get connected node for output %u.\n", i); + continue; + } + + LIST_FOR_EACH_ENTRY_SAFE(sample_entry, sample_entry2, &topo_node->u.transform.outputs[i].samples, + struct sample, entry) + { + if (!topo_node->u.transform.outputs[i].requests) + break; + + session_deliver_sample_to_node(session, downstream_node, downstream_input, sample_entry->sample); + topo_node->u.transform.outputs[i].requests--; + + transform_release_sample(sample_entry); + } + + IMFTopologyNode_Release(downstream_node); + } + break; + default: + FIXME("Unexpected node type %u.\n", node_type); + } +} + + static HRESULT session_request_sample_from_node(struct media_session *session, IMFTopologyNode *node, DWORD output) { IMFTopologyNode *downstream_node, *upstream_node; -- 2.35.1

1 0

[PATCH 0/5] MR51: Fix handling of dbghelp's basic type
by eric pouech (＠epo) 10 May '22

10 May '22

Wine's dbghelp's implementation differ from native regarding basic types' handling (the basic types like char, int, long...) - Wine's dbghelp exposes a name for such a type, while native doesn't - parsing of basic type from dwarf & pdb don't always exactly return the correct information. Two examples (among others) of discrepancies: - in C, under Windows assumption, char, signed char and unsigned char are mapped to 2 types (as char = signed char); while in C++, they are mapped to three different types (char and signed char are two different types). Wine's dbhelp doesn't report correctly the C++ char types, and the char/signed char is broken too. - MS compiler shamelessly maps WCHAR to btWChar (basic type), while it's a typedef to unsigned short (which is mapped to btInt of size 2) This series's result is that Wine dbghelp's now returns the same basic types as native. (Tested with most of the C & C++ basic types) This requires: - to implement the re-generation of the basic types' name in WineDbg. This is done by adding data model tables (ILP32, LP64 and LLP64), and picking up the right one depending on current executable. Option is also added in WineDbg to force a(nother) data model. This is only used when printing types, not when reading/writing integer values as this (mostly) relies on the size of the integer. - fixing the pdb and dwarf backends in dbghelp to retarget the correct base types (for C and C++ basic types) - removing the name field from dbghelp's struct symt_basic allowed some simplification inside dbghelp: + all basic types are now only allocated once across every compilation unit and module) (small speed and memory consumption gain) + dwarf's backend used a basic type cache which is also removed Note: when using a basic type in WineDbg (like casting an integral value to another integral value), this serie doesn't change the data model used for getting the type's size. The data model used is currently ILP32 on 32 bit, and always LP64 on 64 bit. Yet to come serie shall use the current data model instead. -- https://gitlab.winehq.org/wine/wine/-/merge_requests/51

2 5