The encoder uses atomic operations on 64-bit counters. On riscv32 these
become __atomic_*_8 libcalls, and libx265 is not linked with libatomic,
so linking the x265 CLI fails:
ld.lld: error: undefined reference: __atomic_fetch_or_8
>>> referenced by libx265.so.217 (disallowed by --no-allow-shlib-undefined)
Add a patch that checks in CMake whether 64-bit atomics need libatomic
and adds it to the platform libraries only in that case.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Add SYSTEMD_SERVICE variable to register wireplumber.service with
systemd to enable automatic service startup and management.
Signed-off-by: Paresh Bhagat <p-bhagat@ti.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
avahi is included in PACKAGECONFIG if zeroconf is in DISTRO_FEATURES.
The update to 0.3.63 enabled it unconditionally a second time.
Fixes: c4d52888a0 ("pipewire: update 0.3.62 -> 0.3.63")
Signed-off-by: Einar Jon Gunnarsson <tolvupostur@gmail.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The riscv64 ADDPTR assembler macro invokes "stride_shift" while the
macro is defined as STRIDE_SHIFT. GNU as matches macro names
case-insensitively, but the LLVM integrated assembler does not, so
building sad-a.S with high bit depth (the default 10bit PACKAGECONFIG)
fails with clang:
error: unrecognized instruction mnemonic
stride_shift a6, a7, t5, t5, 1, 1
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
log.c uses open() and O_* flags without including <fcntl.h>, which
glibc pulls in indirectly but musl does not, so clang fails with
-Wimplicit-function-declaration.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
64-bit time was turned off for all of libcamera because v4l2-compat.
libcamera fixed this upstream in v0.3.0 by also
undefining _TIME_BITS for v4l2-compat (commit 705601 "v4l2:
Unset _TIME_BITS in addition to _FILE_OFFSET_BITS"), so the
workaround is no longer needed.
v4l2-compat.so still uses 32-bit APIs on purpose, so keep the
32bit-time QA skip for it.
Signed-off-by: Prasanth R <prasanth.r@lineo.co.jp>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
The GPU software ISP renders with GLES 2 through EGL and never uses
desktop GL, so depend on virtual/libgles2 instead of virtual/libgl.
Requiring virtual/libgl breaks machines whose GPU stack only provides
EGL and GLES, e.g. StarFive JH7110 with the PowerVR mesa-pvr.
Also carry a patch so libcamera doesn't redefine EGL_NO_X11 when an
older Mesa's egl.pc already passes -DEGL_NO_X11, which fails the build
with -Werror.
Signed-off-by: Joel Winarske <joel.winarske@linux.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
dvd_filesystem.h unconditionally uses the glibc LFS64 off64_t name.
musl's off_t is already 64-bit, so typedef it the same way the other
non-glibc branches already do. This also fixes libdvdnav, which
consumes this staged header.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
sofprobeclient uses the bare 'uint' typedef, which is a glibc/BSD
extension musl doesn't provide. Use 'unsigned int' instead.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Switch to the alternate URI from videolan. Upgrade to release 4.2:
- Threaded Motion Estimation (Experimental feature)-It uses a
dedicated threadpool to precompute Motion Estimation in parallel.
Improves encoding speed upto 1.5x for 1080p & lower resolution on
multi core machines with low frequency setting. On high frequency
systems or on machines with low number of cores, the overhead of
additional Motion estimation work may outweigh parallelism
- Add new Levels 6.3 to 7.2 specified in ITU-T H.265 (V9) (09/2023)
and above
- Improve Slices feature with check zeroMv
- Enable frame parallelism with MCSTF feature
- Updated support to signal AOM FGM params
- Improve quality with SBRC feature 6.Updated DolbyVision P5 VUI
defaults
- API Support to enable Threaded Motion Estimation(-threaded-me)
- RISC V optimizations including SAD, SATD, DCT, IDCT, block copy,
pixel utilities, SAO, loopfilter, transpose kernels resulting in
2x encoding speed.
- ARM SIMD optimizations including the use of NEON and SVE
instruction set extensions. The following algorithms now have
optimized SIMD implementations: DST, IDCT, SSE, SSD,
intra_pred_planar, pelFilterLumaStrong, interpolation, planecopy,
dequant_normal, blockcopy, pixel variance resulting in 8% faster
encoding speed compared to v4.1
- Fix memory leaks (no command line option, SEI buffer, analysis
save/load)
- Fix chroma qp offset for non yuv444 inputs
- Fix max supported input resolution
- Fix bugs with ARM SIMD optimizations
- Fix Alpha and Multiview feature flag support in x265_config
- Fix test harness issues, CMake errors
- Fix inconsistent output with aq-motion
- Fix crash with hist-scenecut on high bit-depth builds
- Fix lookahead concurrency bug
- Fix shared link issue (R_X86_64_PC32), yuv recon output issue,
rd-refine and dynamic-refine issue, inputs for Windows named
pipe, weighted prediction delta_chroma_offset, crf and vbv issue
in abr-ladder, psnr and ssim reported with MCSTF feature,
internally overflowed VBV variables
This work was sponsored by GOVCERT.LU.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Upgrade to release 1.23.5 with the following security issues:
- CVE-2026-XXXXX (GHSA-v8qw-hwjv-44hw) Memory exhaustion through a
mismatch between the container and the bitstream image size. A
crafted image can declare a small size in its ispe property while
the bitstream declares a much larger coded frame. The
container-level checks used the ispe size, so the oversized
bitstream reached the decoder, which allocated a frame buffer for
the in-band size before libheif rejected the mismatch. The
advisory demonstrated this for AV1 with the libaom backend (a
351-byte AVIF declaring 64x64 but coding up to 27648x27648,
allocating hundreds of MB to more than 10 GB), but the same class
affects every codec whose real frame size lives in the bitstream.
The coded size is now checked against max_image_size_pixels in
the codec-independent decode path, before any bytes reach a
decoder plugin: all AV1 sequence headers, all HEVC/AVC/VVC SPS
NAL units (including those carried in the item data, not only the
ones in the configuration record), the JPEG SOF marker and the
JPEG 2000 SIZ reference grid are scanned for the largest coded
size. (high)
- CVE-2026-XXXXX (GHSA-qwpf-5wf7-r996) Heap use-after-free and
double free when encoding an image that carries a TAI timestamp,
including transcoding a file with an itai property.
ImageDescription shallow-copied its raw heif_tai_timestamp_packet
pointer, and a temporary in ImageItem::encode_to_bitstream_and_boxes()
freed the packet while the item and the source image still held
it. The timestamp is now stored by value. (medium)
- CVE-2026-XXXXX (GHSA-9c75-9g8r-4728) Memory amplification through
a JPEG 2000 pclr box declaring zero palette columns. The
entry-count bound was skipped for zero columns, so an 11-byte box
allocated 65,535 empty palette entries, and nested j2kH containers
could repeat this within the child and nesting limits: a 3 KB file
reached about 330 MB RSS, none of it charged to max_total_memory.
Zero columns are rejected (ISO/IEC 15444-1 requires 1 to 255),
the byte bound is unconditional, and the palette storage is
charged to the memory limits. (medium)
- CVE-2026-XXXXX (GHSA-r7gr-2xm2-23wf) Heap out-of-bounds read in
alpha compositing for uncompressed (unci) images whose colour
planes have different bit depths. Op_flatten_alpha_plane read
every plane through the sample type of the first colour plane, so
an 8-bit blue plane next to 16-bit red and green planes was read
with a halved stride past its end, and the bytes ended up in the
composited output. ColorState now tracks one bit depth per plane,
and the operator declines mixed sample widths at planning time.
(medium)
- CVE-2026-XXXXX (GHSA-q492-cfcm-895h) The OpenJPEG decoder
plugin's pre-decode size check bounded the JPEG 2000 window span
(x1-x0)*(y1-y0) but not the absolute reference-grid coordinates,
so a codestream with a 17-pixel window on a grid near the 32-bit
boundary reached opj_decode(). Against OpenJPEG 2.3.1 this
produced a heap-buffer-overflow write inside OpenJPEG (the class
of CVE-2020-6851); OpenJPEG 2.5.4 rejects the input. The
reference-grid area is now bounded as well. (low)
- (GHSA-qfj5-c4pq-q998) Heap out-of-bounds read in the uncompressed
encoder when an application attached a separate alpha plane to an
image with an interleaved chroma format. The interleaved encoders
took their component list from the chroma format (three entries)
but decided whether to write alpha from the presence of an alpha
plane, and indexed the list at [3]. heif_image_add_plane() now
rejects a separate alpha plane on interleaved images, and the
encoders derive both decisions from the chroma format. Only
reachable through the public API; decoding never produces such an
image. (low)
- (GHSA-7pwf-qh74-p35w) The caller's heif_security_limits were not
applied when parsing a mini box (the MIAF minimized image format)
or the av1C/hvcC blob embedded in it; the built-in defaults were
used instead. An application that tightened the limits got no
enforcement of its max_memory_block_size or max_total_memory on
such files. The allocations are bounded by the bytes present in
the box, so this could not amplify memory use. (low)
This work was sponsored by GOVCERT.LU.
Signed-off-by: Leon Anavi <leon.anavi@konsulko.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
test-dbus-connection and test-reserve-device spawn a private bus through
GLib's GTestDBus, which executes dbus-daemon. When VIRTUAL-RUNTIME_dbus
is dbus-broker, dbus-daemon is not installable (dbus-broker RCONFLICTS
dbus) and both tests abort with:
GLib-GIO-FATAL-ERROR: Error spawning dbus-daemon: Failed to spawn
child process "dbus-daemon" (No such file or directory)
Enable the dbus-tests meson option only when dbus provides the message
bus, and make the ptest package depend on dbus in that case instead of
relying on it being pulled in indirectly.
AI-Generated: Uses Claude Code
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Mycroft made its last release in 2021 and was formally abandoned in 2024:
Mycroft core is no longer maintaiend and probably likely not work on
your computer anymore. Open Voice OS and Neon-core are both spiritual
successors to Mycroft. (And some of the old code may live on there.)
The version we packaged was from 2019 and nobody cared enought to upgrade
it to any of the sixteen later releases, so I don't believe anyone is
using it.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
0.5.16 is skipped because it contains a regression that broke object
destruction in the monitor scripts; upstream recommends to avoid it.
WirePlumber 0.5.17
~~~~~~~~~~~~~~~~~~
Fixes:
- Fixed factory-created device, node and link objects to stop sharing
ownership with their registry global, fixing a regression from the
``WpClientContext`` introduced in 0.5.16 that broke proper object
destruction in the monitor scripts (#999; !890, !891)
- Improved ``default-nodes`` to also rank on the object.serial of a node's
device and on the node's own object.serial, so that route priorities are
only compared between nodes of the same card, and the election no longer
changes for no apparent reason when all other keys are equal (!893)
- Fixed ``m-lua-scripting`` to include the offending Lua type name in POD
build errors (!889)
WirePlumber 0.5.16
~~~~~~~~~~~~~~~~~~
Additions & Enhancements:
- Added ``WpClientContext``, a second PipeWire connection running on its own
thread that hosts WirePlumber's in-process media objects (loopback and
filter-chain modules, ``LocalNode()``, ``SpaDevice()``), so that slow Lua
event hooks on the main thread no longer stall their control path; the
export core is retired in favor of this new client context (!873)
- Added new ``WpDynamicRules`` class for matching rules with conditions, and
updated ``filter-graph.lua`` to use it, allowing filter-graphs to be loaded
and unloaded dynamically depending on whether other objects exist or not
(!872)
- Added ``wp_impl_module_unload()`` and ``LocalModule:unload()``, giving
scripts control to unload implemented modules at any time (!881)
- Added a ``volume-set`` action to ``module-mpris`` to allow adjusting the
volume of a remote MPRIS2 player from Lua (!860)
- Added ``--yes`` flag to ``wpctl reset`` to skip the confirmation prompt
(!850)
- Added ``device.form-factor`` ALSA node property for use in ALSA rules, and
made HDMI node descriptions include the name of the connected display when
available (!862, !874)
- Refactored the Bluetooth monitor to use event hooks for handling devices
and nodes, matching the design already used by the v4l2 and libcamera
monitors (#916; !803)
- Overhauled the documentation: filled in every stub and missing Lua API and
C API page, added a getting started guide, new ``wireplumber(1)`` and
``wpexec(1)`` man pages, and removed stale pages describing features
renamed or removed in the 0.4 to 0.5 transition (!877)
Fixes:
- Fixed monitors to always activate all device and node features, and made
monitors wait for successful device activation before storing the devices
and nodes as a managed objects (#986, #996; !883, !886, !876)
- Fixed ``find-preferred-profile`` to skip a configured preferred profile
that is not available, instead of leaving no profile selected (#993; !884)
- Fixed ``permission-manager`` to not include destroyed globals when
rebuilding the permissions array on ``objects-changed``, fixing some
pipewire warnings in the log (!882, !888)
- Fixed a memory leak in ``WpSpaDevice`` by only collecting params from our
own requests (#988; !879)
- Fixed ``filter-graph`` to correctly apply graphs defined for the same node
across multiple configuration files (!870)
- Fixed several nil-related crashes and correctness issues in Lua scripts
(#972; !861, !864, !866, !869)
- Improved ``apply-access`` to directly attach an already-active permission
manager to the client instead of wastefully re-running activation on it
(#973; !868)
- Improved ``state-stream`` change detection to avoid spurious state writes
from volume rounding and from unset or empty properties (#974; !867)
- Fixed ``default-nodes`` to ignore smart filters when selecting the best
default node (!857)
- Removed the ``node.filter.forward-format`` setting and its script, which
only worked with 0.4 and never really worked with 0.5 (!878)
AI-Generated: Uses Claude Code (Claude Opus 5)
Signed-off-by: Markus Volk <f_l_k@t-online.de>
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
clang 23 enables -Wunused-template as part of -Wall (via -Wmost ->
-Wunused), where previously neither -Wall nor -Wextra turned it on.
libcamera builds with -Werror, so this now breaks the build:
In file included from include/libcamera/ipa/core_ipa_serializer.h:18:
include/libcamera/internal/ipa_data_serializer.h:57:3: error: unused
function template 'readPOD' [-Werror,-Wunused-template]
57 | T readPOD(std::vector<uint8_t> &vec, size_t pos)
| ^~~~~~~
The POD helpers sit in an anonymous namespace, so they have internal
linkage and clang diagnoses them per translation unit. appendPOD() and
the iterator overload of readPOD() are referenced by the
IPADataSerializer specialisations in the header itself, so they count as
used everywhere. The std::vector overload of readPOD() is not: nothing
in libcamera calls it, its only callers are the IPA proxies and proxy
workers generated at build time from the mojom templates. So it is
genuinely unused in every translation unit that is not generated proxy
code, which is what [[maybe_unused]] exists for.
Signed-off-by: Khem Raj <khem.raj@oss.qualcomm.com>
Khem Raj
Paresh Bhagat
Einar Gunnarsson
Ankur Tyagi
prasanth
Joel Winarske
Leon Anavi
Ross Burton
Markus Volk
Jaipaul Cheernam