| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |
|
| |
|
|
|
|
| |
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
make sure dnsmasq listens to all addresses specified in BRIDGE_ADDR,
but not the ones in BRIDGE_ADDR_EXTRA. That way it is possible to have
another daemon listening to the extra address.
Also make sure we remove the addresses on service stop.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
1) Recognize s390x and mips64* as 64-bit architectures to enable 64-bit code path.
2) Enable direct unaligned loads/stores for s390x. See "Integral Boundaries"
in SA22-7832-xx.
3) Enable unaligned access using LWL/LWR/etc instructions on pre-R6 mips*.
This doubles "brotli -1" compression speed on P5600 (~30% for "brotli --best")
and makes the brotli executable smaller by ~10%.
4) Use gcc/clang __builtin_bswapXX on big-endian architectures.
On s390x this reduces byte-swapped unaligned loads/stores to single LRVx/STRVx
instruction. On current big-endian mips* (pre-R2) this changes nothing however
allows for shorter byte swap sequences on R2+.
Test suite still passes on x86_64, mipseln8hf (hardware) and under qemu-s390x.
|
|
|
|
| |
Check host_arch not target_arch to decide whether -latomic is necessary.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
1) Enable generation of LWL/LWR/etc instructions on pre-R6 mips*.
This gives 5x (yes, five times) compression speed-up on P5600
(and additional ~10% for decompression).
2) Recognize s390x as unaligned access capable architecture.
See "Integral Boundaries" in SA22-7832-xx.
3) Force use of gcc/clang __builtin_bswapXX for big-endian code path
in order to avoid suboptimal generic functions from <byteswap.h>.
Test suite still passes on x86_64, mipseln8hf (hardware) and under qemu-s390x.
|
|
|
|
|
|
|
|
|
| |
./CHANGES claims "many security related and stability fixes".
While here, enable built-in test suite (passes on x86_64 and mipseln8hf
at least) and make sure libpng is not pulled accidentally during build.
Pull the source from github.com as referenced on the project's web page.
|
|
|
|
| |
including patch required for compilation.
|
|
|
|
|
|
|
| |
Add xpresent, for low-latency vertical sync.
Add startup-notification, to prevent temporary black screen on xorg
start.
Prevent ambiguity of whether default CFLAGS are included, e.g. -Os.
|
|
|
|
| |
Public header ffitarget.h wants <asm/sgidefs.h> on mips*.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- properly require openssl (status quo, just use a correct parameter)
- explicitly require system libpcap (status quo)
- explicitly require system libpcre [1]
- explicitly require zlib (status quo)
- explicitly require system libssh2 [2]
- explicitly require system lua5.3 [3]
- adjust makedepends accordingly
[1] System pcre is newer and properly configured for mips*.
[2] This is the ./configure default, just make sure the choice is stable.
[3] This is the ./configure default, just let it find the system version.
check() passes on x86_64 and mipseln8hf at least, however this commit does
not re-enable it due to unclear aarch64 / arm* status.
|
|
|
|
|
|
|
| |
upgrade to 0.27.0 version
build with openssl instead of libressl
remove paxmark build dependence
fix test specs according to advice from upstream
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
gcc 6.4.0 is a required dependency and current pkgrel is version 8
|
|
|
|
| |
make: *** [Makefile:129: check] Bus error
|
|
|
|
|
|
|
|
|
|
| |
We already support multiple addresses for the bridge device, but they
would all use the BRIDGE_NETMASK. Fix so you can either specify the
netmask with the address or fallback to BRIDGE_NETMASK.
For example:
BRIDGE_ADDR="192.168.0.1 10.0.0.1/16"
BRIDGE_NETMASK="255.255.255.0"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Extended tests fail on x86_64 builders with:
rm: 'tmp-lower-bindir' is a directory
rm: 'tmp-lower-includedir' is a directory
rm: 'tmp-lower-libdir' is a directory
rm: 'tmp-lower-man1dir' is a directory
rm: 'tmp-lower-mandir' is a directory
rm: 'tmp-lower-pkgconfigdir' is a directory
rm: 'tmp-upper-bindir' is a directory
rm: 'tmp-upper-includedir' is a directory
rm: 'tmp-upper-libdir' is a directory
rm: 'tmp-upper-man1dir' is a directory
rm: 'tmp-upper-mandir' is a directory
rm: 'tmp-upper-pkgconfigdir' is a directory
make[1]: *** [Makefile:352: test-lz4-essentials] Error 1
|
|
|
|
|
|
|
|
|
|
|
| |
3.18.1 upgrade
scons still requires python2, and some of the gps
utilities no longer work with python2
The 0001-workaround-for-scons-using-python2.patch
explicitly uses python3 as the interpreter
3.18.1 insists on putting the python utilites in
/usr/local - we manually move them as
part of the _py subpackage
|
|
|
|
|
|
| |
This upstream patch fixes multi.rf64 test failure on mipsel{n8hf} at least.
Note the bug was reported for armv7hl so arm* check() should probably be re-enabled.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is exactly the same memcpy()-in-hot-path xxHash problem, see
community/xxhash commit log for details.
The bug makes "lz4 -1" or "lz4 -t" twice slower on x86_64 however
"lz4 -9" is barely affected (do not use "lz4 -b" for benchmarking
this change since it does not calculate the checksum).
Other aport changes:
- use short and fast "make check" (unlike extensive "make test")
for pre-v7 ARM architectures.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is the same memcpy()-in-hot-path xxHash problem as described
in community/xxhash commit message.
The bug makes "zstd -1" or "zstd -t" ~15% slower on x86_64 though
higher levels (>= 9) are almost unaffected.
Other aport changes:
- make sure zlib/lzma/lz4 support is not compiled in automatically
(status quo, just make it explicit);
- re-enable armv7 since it is certainly well-supported by upstream,
was disabled by some script in the first place and should not be
affected by "Bus error" (i.e. some alignment problem) mentioned in
commit 2bae66907...
- use faster and shorter "make check" (unlike extensive "make test")
on arm* builders.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Yes, it is 20 times slower on x86* than it should be because xxhash.c always
uses "safe" memcpy()-based methods for unaligned memory access (XXH_readXX)
irregardless of input alignment due to x86-default XXH_FORCE_ALIGN_CHECK=0.
This ends up with real memcpy() calls in hot path (with -O2 too).
The bug affects Alpine x86* (not just edge, but at least 3.8 too -- i.e. this
is not something introduced in 0.6.5) for aligned and unaligned inputs. Other
architectures are severely affected for unaligned inputs only.
The fix lifts the XXH_FORCE_MEMORY_ACCESS=1 condition to enable XXH_readXX
methods based on __attribute__((__packed__)) usage everywhere except ARMv6
(which is covered by its own case earlier).
This is safe and fast because the compiler will either:
- use direct storage access instructions on capable architectures such as
aarch64, armv7, ppc64le, s390x, x86* irregardless of input alignment;
- or use relatively fast LWL/LWR instructions on mips* with unaligned input;
- or use byte load/stores and shifts/ors on armel with unaligned input which
is still faster then memcpy() call.
All aports that use xxhash.c are likely affected. For example, community/zstd
suffers too though not so grave (~15% difference for "zstd -t" on big archive)
and main/lz4 is twice slower on basic compression levels.
Other aport changes:
- modernize;
- enable check(); it is short and fast so suitable for slow builders too.
The python part is left intact though newer version exists.
|
|
|
|
|
|
|
|
|
|
| |
Provide some basic mips* support just to allow the build succeed.
Disabling on mips* is not an option cause the arch= list becomes
too long and ugly.
Do not bump pkgrel since no previously built mips* apks exist and
change does nothing for other architectures.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
While at it, drop depend that was not needed after all
|
|
|
|
| |
needed by diffoscope
|
| |
|
|
|
|
|
| |
1) Disable use of CFC1 (FPU instruction) on all soft-float mips*.
2) Enable MIPS32 ISA (default is MIPS III) on 32-bit mips{el}.
|