The video decoder thread is the pilot user.
For now this is an Android-only feature, since that's the only
platform we can print stack traces on.
Bug: webrtc:9987
Change-Id: Ie638c619673b5f159d91a32683fd787baf46479a
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/126222
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27127}
This cl deprecates the FrameType enum, and adds aliases AudioFrameType
and VideoFrameType.
After downstream usage is updated, the enums will be separated
and be moved out of common_types.h.
Bug: webrtc:6883
Change-Id: I2aaf660169da45f22574b4cbb16aea8522cc07a6
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/123184
Commit-Queue: Niels Moller <nisse@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#27011}
Ignore rtc_link_task_queue_impl flag,
instead use build_with_chromium for custom chromium implementation injection
This changes TaskQueue implementation used in webrtc fuzzers in chromium:
from own webrtc implementation to chromium's.
Bug: webrtc:10191
Change-Id: I63be28b680ae8ea8ee1dbf0c699263c392ce29d3
Reviewed-on: https://webrtc-review.googlesource.com/c/src/+/125196
Commit-Queue: Danil Chapovalov <danilchap@webrtc.org>
Reviewed-by: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26977}
Previously, we have created a Legacy ADM when no ADM is supplied.
With this change we will start creating a Java ADM instead.
The end goal is to make injection mandatory, and never creating ADMs.
This is one step on the way, and will allow us to clean up the Legacy
ADM code.
Bug: webrtc:7452
Change-Id: Ib99adc50346fe6b748f9435d2fc6321a50c3ee4e
Reviewed-on: https://webrtc-review.googlesource.com/c/123887
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Henrik Andreassson <henrika@webrtc.org>
Commit-Queue: Paulina Hensman <phensman@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26949}
There has been some crashes due to frames having illegal sizes, most
likely 0x0. Probably these frames are created as a workaround for
something.
It would be best to stop 0x0 frames from being created in the first
place, but a reasonable quick fix is to just not draw those frames.
Bug: webrtc:10367
Change-Id: Ib93057c4de7285773c99614b4e7d9bd4b099c4dc
Reviewed-on: https://webrtc-review.googlesource.com/c/124988
Commit-Queue: Paulina Hensman <phensman@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26897}
In the Discussion part of
https://developer.apple.com/documentation/metal/mtltexture/1515679-replaceregion
it seems like we should sync the calls to replaceRegion (inside
setupTexturesForFrame) in RTCMTLRenderer and not just the command
buffer.
This is a speculative fix for the linked bug, but we don't have any
clear repro case. Have done basic testing in AppRTCMobile and don't
see any obvious regressions, so might be worth trying.
Bug: webrtc:10024
Change-Id: Id6848691129fba8845f38c3dfe0ba53b9e5a27ce
Reviewed-on: https://webrtc-review.googlesource.com/c/123766
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Commit-Queue: Anders Carlsson <andersc@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26819}
It was previously possible to escape the sandbox by calling
rtc::Thread::SetAllowBlockingCalls(true).
This CL only removes the loophole on non-Android builds, because we
still have old Android code that relies on it. We expect that code to
go away soon-ish, though.
Bug: webrtc:9987
Change-Id: Ida96400d0abe430af4c2046284795d37d64f6613
Reviewed-on: https://webrtc-review.googlesource.com/c/123523
Commit-Queue: Karl Wiberg <kwiberg@webrtc.org>
Reviewed-by: Tommi <tommi@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26792}
Added audio format field and set method to Builder. - WebRTCAudioRecord. Added audio format field, added to constructor. Default audio format value AudioFormat.ENCODING_PCM_16BIT. initRecord(), added how to calculate bytesPerFrame, depends on audioFormat.
First commit and contribution, updated AUTHORS file
Bug: None
Change-Id: I16f660d42350ec9ce2e329b239bd7f6324e76dfe
Reviewed-on: https://webrtc-review.googlesource.com/c/122302
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Artem Titov <titovartem@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26775}
This CL adds an API for injecting video processing after the WebRTC
CPU and QP scaling step.
Bug: webrtc:10247
Change-Id: I776498e1e9113f50e953ee411bbb31f181863312
Reviewed-on: https://webrtc-review.googlesource.com/c/119953
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26740}
New Robolectric version doesn't allow Surface to be constructed with a
null SurfaceTexture.
Bug: webrtc:10323
Change-Id: Ib6991d40b12b81d16ecb04787945cc4045e99b40
Reviewed-on: https://webrtc-review.googlesource.com/c/123236
Reviewed-by: Mirko Bonadei <mbonadei@webrtc.org>
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26734}
Previously, the use of pass-by-value caused an issue in
ObjCVideoDecoder::Decode, where the EncodedImage was being copied upon
calling initWithNativeEncodedImage, which then created an NSData using
the copy's pointer; then the copy was destroyed, invalidating that
pointer.
Bug: webrtc:9378
Change-Id: Iac28b890c9902108ffc5ec54a607a99034159153
Reviewed-on: https://webrtc-review.googlesource.com/c/121922
Reviewed-by: Kári Helgason <kthelgason@webrtc.org>
Reviewed-by: Niels Moller <nisse@webrtc.org>
Commit-Queue: Rasmus Brandt <brandtr@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26598}
AndroidVideoTrackSource::OnFrameCaptured currently does adaptation
before passing frames on. We want to add video processing between
adaptation and delivering the frame to the rest WebRTC C++. This
CL prepares for that by splitting OnFrameCaptured() into a separate
adaptation step and delivery step.
Bug: webrtc:10247
Change-Id: Iab759bac7f3072d4552ece80d0b81fc3e634c64c
Reviewed-on: https://webrtc-review.googlesource.com/c/119952
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26571}
This CL attempts to do separation of concerns by introducing a simple
class that only handles JNI wrapping of a C++ AndroidVideoTrackSource.
This layer can be easiliy mocked out in Java unit tests.
Bug: webrtc:10247
Change-Id: Idbdbfde6d3e00b64f3f310f76505801fa496580d
Reviewed-on: https://webrtc-review.googlesource.com/c/121562
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26556}
This layer is not needed since the methods are thread safe, and the
classes those method touches (VideoBroadcaster, cricket::VideoAdapter)
are thread safe.
Bug: webrtc:10247
Change-Id: Id4e309de4ac1b9669052aaa60d3bd1ed965aaa29
Reviewed-on: https://webrtc-review.googlesource.com/c/120801
Reviewed-by: Niels Moller <nisse@webrtc.org>
Reviewed-by: Sami Kalliomäki <sakal@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26543}
JNI_COMMIT doesn't actually free the buffer.
From JNI docs:
0: copy back the content and free the elems buffer
JNI_COMMIT: copy back the content but do not free the elems buffer
JNI_ABORT: free the buffer without copying back the possible changes
Also introduces helper methods to help avoid this problem in the
future.
Bug: webrtc:10132
Change-Id: I769df286d3bd186fdf39ee2363e9002f36454509
Reviewed-on: https://webrtc-review.googlesource.com/c/120600
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Sami Kalliomäki <sakal@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26529}
This CL hooks up the recently added native stack trace functionality to
the existing PeerConnectionFactory API.
Bug: webrtc:10168
Change-Id: I16189d2988b1359fc53f9a4d0b3d06f34e2a8fd5
Reviewed-on: https://webrtc-review.googlesource.com/c/118600
Reviewed-by: Tommi <tommi@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26344}
This is a reland of dc32cc00e80972223767be339f9318b803f4a2ae
Relanding because this CL was not the culprit for the Chrome bot.
(This code shouldn't even be executed in Chrome).
Original change's description:
> Android: Add helper methods for printing native stack traces
>
> This CL adds utility functions to unwind the stack for a given thread on
> Android ARM devices. This works on top of unwind.h and unwinds native
> (C++) stack traces only. Unwinding a thread from another thread is done
> by overriding the signal handler with a custom function and then
> interrupting the specific thread.
>
> Bug: webrtc:10168
> Change-Id: If5adffd3a6bb57bf502168743e09a7eefc292bf3
> Reviewed-on: https://webrtc-review.googlesource.com/c/118141
> Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
> Reviewed-by: Tommi <tommi@webrtc.org>
> Cr-Commit-Position: refs/heads/master@{#26328}
TBR=tommi
Bug: webrtc:10168
Change-Id: I4c33c2c147cf10c0172c98a55d32dd35a08517c8
Reviewed-on: https://webrtc-review.googlesource.com/c/118704
Reviewed-by: Magnus Jedvert <magjed@webrtc.org>
Commit-Queue: Magnus Jedvert <magjed@webrtc.org>
Cr-Commit-Position: refs/heads/master@{#26341}
