projects / platform / upstream / lz4.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: f405427) | patch
Call LZ4_memcpy() instead of memcpy()
authorNick Terrell <terrelln@fb.com>
Mon, 3 Aug 2020 18:18:35 +0000 (11:18 -0700)
committerNick Terrell <terrelln@fb.com>
Mon, 3 Aug 2020 18:28:02 +0000 (11:28 -0700)
commitfe2a1b3707d59db8bb5ad599cbed6507ffbd544e
treec2057af66766037abea946b302db501407a70dc7tree | snapshot
parentf4054274faa24f15efedc7431fcb6914d1826220commit | diff
Call LZ4_memcpy() instead of memcpy()

`LZ4_memcpy()` uses `__builtin_memcpy()` to ensure that clang/gcc
can inline the `memcpy()` calls in freestanding mode.

This is necessary for decompressing the Linux Kernel with LZ4.
Without an analogous patch decompression ran at 77 MB/s, and with
the patch it ran at 884 MB/s.
lib/lz4.c diff | blob | history
Domain: System / System Framework;