md_5
255d7fde9a
Ensure native zlib actually loads.
2016-01-16 13:33:09 +11:00
md_5
7907610eeb
Compatability and benchmark fixes for native code.
2016-01-16 13:29:50 +11:00
kamcio96
b6e26e0c09
Load NativeCode manually, not in constructor.
...
Bungee will call load method from main thread
https://github.com/SpigotMC/BungeeCord/blob/master/proxy/src/main/java/net/md_5/bungee/BungeeCord.java#L202
2015-11-19 08:42:01 +11:00
md_5
b4997f6379
Add better exception handling for native code.
2015-02-08 09:11:52 +11:00
md_5
eeaa44e1e7
#1362 : Tighter catch on errors in native lib compilation.
2015-02-08 08:08:20 +11:00
md_5
f4ae511af0
2 megabytes is sufficient for testing purposes.
2015-02-07 14:19:33 +11:00
md_5
32693aeaff
Don't compile with the stdlib, I don't think it is needed.
2015-02-07 14:14:33 +11:00
md_5
0d569ac0d1
Refactor native code and implement our own JNI wrapper around zlib.
...
The previous native cipher code has been refactored so that it may be loaded and used slightly more generically, allowing more native components to be easily added as time goes on.
I have also written a new native code compression module, which wraps around zlib in the same manner that Inflater / Deflater does, however it operates directly on the memory addresses of it's input / output buffers which means that we can save one, or maybe even two copies. To support this, the VarInt decoder has been adjusted to always use a native buffer.
2015-02-07 14:06:41 +11:00
zreed
22084b2c75
Try loading native cipher from java.library.path first
2014-12-16 13:21:25 +11:00
md_5
2cec5f344a
Update versions to 1.8... oops
2014-09-01 16:56:03 +10:00
md_5
c965e60f5e
Cipher thread locals should be static
2014-07-12 19:49:12 +10:00
md_5
02cb1fc65b
Allocate cipher with EVP_CIPHER_CTX_new. Thanks @Adam- for the tip.
2014-07-02 12:57:17 +10:00
md_5
7318750ed0
Update native cipher to make use of the OpenSSL EVP API so that it can actually utilise hardware acceleration and other goodies.
...
Whereas before OpenSSL would often lose benchmarks now it always wins.
2014-07-01 20:13:30 +10:00
md_5
886f7499fb
Update compile-native script.
2014-02-16 11:40:36 +11:00
md_5
8064a3d4fb
Revert "Compile native cipher on ancient CentOS for maximum compatability - closes #880 "
...
This reverts commit 7bfa024c23
.
2014-02-16 11:39:47 +11:00
md_5
5039922fa7
Revert "Compile on CentOS 6.4 with OpenSSL 1.0.0e and glibc 2.12"
...
This reverts commit eb753c8109
.
2014-02-16 11:39:42 +11:00
md_5
eb753c8109
Compile on CentOS 6.4 with OpenSSL 1.0.0e and glibc 2.12
2014-02-15 19:37:53 -05:00
md_5
7bfa024c23
Compile native cipher on ancient CentOS for maximum compatability - closes #880
2014-02-16 00:13:13 -05:00
md_5
56e9e6a245
Delete extracted native code on exit
2014-02-12 19:53:33 +11:00
md_5
90104b03b7
Only allow key sizes of 16 in native cipher
2014-02-12 19:45:28 +11:00
md_5
0b7789035f
Revert as its unstable anyway.
2014-02-12 18:02:02 +11:00
md_5
3f7850dc5a
Clean up JNI code, add stack trace to check up on Jenkins test.
2014-02-12 17:43:49 +11:00
md_5
2e80bf30dd
Various improvements to native cipher so that it now actually destroys Java in terms of speed. Closes #871 - thanks @ninja-
2014-02-12 17:37:08 +11:00
md_5
5adc0000d8
Add small benchmark to the native cipher suite - see #755
2014-02-02 10:18:01 +11:00
md_5
a0cc5d84be
Add explicit casts to make clang support easier - see #755
2014-02-02 10:15:19 +11:00
md_5
94ee61cd35
Native cipher, with more smoke tests!
2013-11-19 07:16:06 +11:00
md_5
897a59254c
Still segfaulting.
2013-11-18 20:00:15 +11:00
md_5
c1ba555553
Hopefully fix native cipher segfault
2013-11-18 07:18:23 +11:00
ninja-
c70006a36c
Implement new, high-performance cipher in native code. Currently available only for Linux-x64, other platforms will fallback to Java cipher.
2013-11-18 07:17:14 +11:00
md_5
12ef019d69
Revert native cipher, causing SIGSEGV
2013-11-17 16:51:11 +11:00
ninja-
fc189e81d5
Implement new, high-performance cipher in native code. Currently available only for Linux-x64, other platforms will fallback to Java cipher.
2013-11-17 15:59:18 +11:00