Commit Graph

308 Commits

Author SHA1 Message Date
Greg Tucker
e1f5284ff8 igzip: Add sse optimized adler32 checksum
Change-Id: Id07727b8a8da4b41aa983b487ca881552d5190ee
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-06-26 04:01:29 -04:00
Greg Tucker
3025e83b91 igzip: Add avx2 optimized adler32 checksum
Change-Id: I019a38cf98836e3e6c7215a6914b85abb9399e33
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-06-26 03:56:49 -04:00
Roy Oursler
f4a5b303e2 igzip: Remove BITBUF8 and BITBUF_ELSE compile options
Change-Id: Iad3b2e6f9a32473b6e59910494c75d82558fc28e
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 06:03:51 -04:00
Roy Oursler
edacadc8fb igzip: Modify encode_df_04 to behave more like encode_df_06 algorithm
Change-Id: I39c5d0d8182efb0fe8aa6bea97d9361df4ee8ddf
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 06:02:13 -04:00
Roy Oursler
5a55e3096c igzip: Avx512 version for encdode_df
Change-Id: I1625a3d7e016805791cfd09e31909562f432fd71
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 05:56:05 -04:00
Roy Oursler
b3c09b9b7c igzip: Improve random data generation in igzip-rand_test
Change-Id: I4835a9e376b4fa24080d765255703a959389487d
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 05:47:04 -04:00
Roy Oursler
64143a741e igzip: Implement type 0 blocks for level 1 compress
Change-Id: If55ab161623d29fa6fb08df3bc813e654918e592
2017-06-21 05:46:38 -04:00
Roy Oursler
87652b4489 igzip: Modify inflate to optionally calculate adler32 hash
Change-Id: I314617b89b59d53608e464c7d2cf299faa3528b5
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 05:42:12 -04:00
Roy Oursler
0cc3d93758 igzip: Modify test to test zlib compression
Change-Id: I52979c9e572ef9703995adf8d2163ba1797b8f53
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 05:41:47 -04:00
Roy Oursler
4259169107 igzip: Implement zlib compression format
Change-Id: I3d3cca425a494ac629cea230de74e3d32fcaea79
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-06-21 05:41:05 -04:00
Hailiang Wang
52f644d3ff igzip:use fgetpos() to replace ftell() to get file's length.
Change-Id: Ia1f3c06e92c01da9d22b3d70b2b05fe4808c9f2c
Signed-off-by: Hailiang Wang <hailiangx.e.wang@intel.com>
2017-06-21 05:22:07 -04:00
Hailiang Wang
ff9c0c1842 igzip:change the default max file size for 32-bit builds.
Change-Id: Ifab108250cfd06211843b5eccb1f1f0482669426
Signed-off-by: Hailiang Wang <hailiangx.e.wang@intel.com>
2017-06-21 05:13:24 -04:00
Greg Tucker
7e1a337433 igzip: Fix warnings from nasm 2.13
A few legitimate warnings got masked by previous nasm bugs.  This cleans last as
of nasm 2.13rc20.

Change-Id: Iaa7e6148e0e506222cc207685263103d62bdd015
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-04-19 10:17:24 -07:00
Greg Tucker
c28be0d306 igzip: Remove unnecessary casts causing warnings in 32bit build
Change-Id: I9fb85c097c9ade8fdc7907e9d6533d1997ccd406
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 20:22:36 -07:00
Greg Tucker
8270237457 igzip: Add decode_huffman_code_block_stateless to base aliases
Change-Id: I60558f5c09df354a1e7608fe479182ca4e10efb6
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 18:23:55 -07:00
Greg Tucker
dbab2ddad7 igzip: Make crc32_gzip_base standalone c function
Change-Id: Iffb55919fb51e9e3d74c5c5cb06a3011bd19e99b
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 18:22:20 -07:00
Greg Tucker
ec6e5de665 igzip: Move build_heap base functions to own file
Change-Id: I0161cd65c71df00fadad9dd69e207e9fb29a54ef
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 18:20:09 -07:00
Roy Oursler
f80a1ed62b igzip: Create base functions for build_heap and build_huff_tree
Change-Id: I19c2d7bbf1ac8270458b165a385c385a60f1cadc
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 18:18:14 -07:00
Greg Tucker
f316e96217 igzip: Add base function aliases and group src by arch
Change-Id: I4b6c8f62a09545d1ed4f48dc16b3fe8c8a5c72ea
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 18:16:36 -07:00
Roy Oursler
871ad43fad igzip: Fix bug in isal_deflate where processed is not calculated correctly.
Change-Id: I61e15a18ebe3130e73010337d2d41f59e2227f08
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:20:32 -07:00
Roy Oursler
be09e56b36 igzip: Modify igzip_example to provide example of higher compression levels
Change-Id: Iccd0528ac088e1eec3921aa5fde99769d75e3334
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:19:54 -07:00
Roy Oursler
13692d61a3 igzip: Fix bug when 0 length buffer is passed
Avoid unsigned addition overflow when avail_in = 0 and next_in = -1 causing a
comparison to be correct when it should not be.

Change-Id: I3ff7123a89867317f383931cc95c8738ba2f8d56
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:17:04 -07:00
Greg Tucker
de7d639ab1 igzip: Change zero length array and typos found in windows build
Change-Id: Ia185b600af3c5f3b34bb9411daae6877f99b05d7
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 17:15:57 -07:00
Greg Tucker
cd2146de95 igzip: Add missing file to dist list
Change-Id: I6dc920bbae9bd1800ba12e7174cdf6bc98f75489
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2017-03-30 17:15:14 -07:00
Roy Oursler
9992cc1920 igzip: Implement static header on level 1 compress
Change-Id: I0fe61eb6d3994a0977a4486a2a4cf21af38dc250
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:13:12 -07:00
Roy Oursler
e38ed4b54e igzip: Increase isal_mod_hist size to stop histogram overflow
Change-Id: I1c651c9625d0fb543cd89e53e2b78b391176ef68
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:11:51 -07:00
Roy Oursler
761b207376 igzip: delete generate_constant_block_header since it has not function
Change-Id: Id236f7a847f660b35165f9936850b3ff73609017
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 17:09:13 -07:00
Roy Oursler
91fef2d39c igzip: Fix generate_custom_huffcodes
Reimplement generate_custom_hufftables.c to avoid internal dependencies. Remove
the symbol subset feature since a default incomplete huffman code is dangerous
and that functionaly is available when generating custom huffman codes.

Change-Id: I014d8e127b49583fe7d6ac9ce861cc3138ffec46
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:58:37 -07:00
Roy Oursler
5d15580467 igzip: Modify stateless_file_perf to test compression levels
Change-Id: If628e7f949da6862b99cebe6a0a59c6e6d5443e4
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:56:13 -07:00
Roy Oursler
a41d352a03 igzip: Modify assembly functions and structs to work on Windows
Change-Id: Icd106c3330dec72601e6f03340c07c6e1d1b5794
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:55:14 -07:00
Roy Oursler
f2c35a9fd4 igzip: Fix underflow bugs whe null pointer is passed in igzip base functions
Change-Id: Ia24326dcde6f8be9ee019690b8206d2779ddebb2
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:54:35 -07:00
Roy Oursler
7c7272e89c igzip: Optimize two pass first pass
Change-Id: I45efa8304f929d91238c8173c0fac00ec64ab323
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:53:26 -07:00
Roy Oursler
4725626847 igzip: Setup to allow encode_df to decode lit lit pairs.
Change-Id: Ic93797a09c5a908fc1c6005b40439978d7484773
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:52:21 -07:00
Roy Oursler
8a05e7d780 igzip: vecotorize encode_df_asm.asm
Change-Id: I11e005556e150a5bf8c6cc9410fa7d98196847f1
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:50:53 -07:00
Roy Oursler
fb13462fac igzip: Use SHLX and SHRX in encode_df_asm
Change-Id: Ic3165579587c905d8e347b35efa6cbedb5dbf5f3
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:50:53 -07:00
Roy Oursler
01dfbcc484 igzip: implement igzip two pass
Change-Id: I9564b2da251a02197b39cab5f141e7aff1ae8439
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:50:48 -07:00
Roy Oursler
43d1029b81 igzip: Modify sync_flush to not worry about writing eob
Change-Id: If3e7d5ff628574d715be348d96cdb82645985c25
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2017-03-30 16:06:47 -07:00
Greg Tucker
9d0fda20fd igzip: Add copyright notice
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 21:06:10 -07:00
Greg Tucker
d301837657 igzip: Add file perf test for semi-dynamic
Change-Id: Ief045d7ca30d160128359b0dce3c63d9034c1016
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 16:20:33 -07:00
Greg Tucker
077b31b560 igzip: Fix feof usage in examples
Change-Id: Ic5f2f3cf62c57b685be6bc4f68a8c8e246452023
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 16:18:54 -07:00
Roy Oursler
dfa4fd0940 igzip: Fix inflate bug with long when invalid huffman codes.
Change-Id: I52bfed0afa208462ed4a8614dd46cda5a5dc968d
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:18:32 -07:00
Roy Oursler
3bc0efa6c2 igzip: Add inflate std vector test to igzip_rand_test
Change-Id: If84de68e9f096e225b3b7e8b695f84c3ab6fe778
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:18:00 -07:00
Roy Oursler
89fedafcc9 igzip: Fix make_huff_code_lookup_table initialization bug.
Change-Id: I192ce6c907271d52bfda5bf4be6d5b2a35a9d459
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:17:17 -07:00
Roy Oursler
f0a39fea2a igzip: Fix make_inflate_huff_code bug causing out of bounds writes
Change-Id: I7dcdb22031cb7b506f76578ee6d080924d917476
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:16:55 -07:00
Roy Oursler
7fefb53938 igzip: Add a hufftable for the static Huffman code in Deflate
Change-Id: If8aadce05bcb705e608fe3ed6028be9a7448fa59
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:15:57 -07:00
Roy Oursler
e1bcebc93a igzip: Fix igzip to work with huffman codes with small headers.
Change-Id: Icd1236ada831c027f97cd4a6cabdeb15022b56be
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 16:15:32 -07:00
Roy Oursler
600d887143 igzip: Fix bug in inflate when exactly one distance is coded
Initialize the lookup table so invalid lookups do not cause segfaults.

Change-Id: Ibb88c4c473fa233f8821c35b3e32ff62814c2b15
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 13:59:04 -07:00
Roy Oursler
d822c79d85 igzip: Modify multibinary so pcmulqdq does not cause error on Nehalem
Change-Id: Ibc68219afa240f5979954e895fd433b6a6da75ce
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 13:02:41 -07:00
Roy Oursler
4809b9ba83 igzip: Create flag signifying existence of buffer history
Create a flag signifying the existence of buffer history so a literal does not
have to be encoded on every call of isal_deflate_body and isal_deflate_finish.

Change-Id: I3750cac7e5f19fbaa190e19fb929172c16289723
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:22:22 -07:00
Roy Oursler
e4c2106049 igzip: Fix naming conflict in igzip_rand_test when compiling with Makefile.unx
Change-Id: I312d276d2fc5d3ddc1e9dcf54674f7246c9e2a72
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:21:40 -07:00
Roy Oursler
886297517b igzip: Delete igzip_check.c and set igzip_rand_test.c to a check test
Change-Id: Ie8127e2aa5f1dcc69501c9e86f83497ddf50b070
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:19:24 -07:00
Roy Oursler
266ac91a6e igzip: Allow user to write a custom gzip header before compressing with gzip
Modify gzip_flag so that if set to IGZIP_GZIP_NO_HDR, no gzip hdr is written.

Change-Id: I79dbd5287cf471777b10412c17d23dc7f5a57bf6
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:17:33 -07:00
Roy Oursler
d4c6067d28 igzip: Fix bug in default histogram generation
Change-Id: Ib1976633c2fbf6f48aeb4b38e73fd75d41c62be5
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:17:15 -07:00
Roy Oursler
e569ff70e7 igzip: Create flag for inflate to return crc of the decompressed data
Change-Id: I2d966f2d4ca135e3c938a9501bf5d2e11d8e92fe
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:16:34 -07:00
Roy Oursler
4ea4f81138 igzip: Make gzip format a run time option.
Change-Id: Ia986d378763146b33c733751652bff788ec3cb9d
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:16:07 -07:00
Roy Oursler
77e43ef6cf igzip: Set isal_deflate_stateless to use same assembly functions as isal_deflate
Change-Id: I813b620286c2f74fabff6937fbceddb65377f40c
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 11:07:08 -07:00
Roy Oursler
d941b66162 igzip: Improve igzip stateful on large buffers
Change-Id: I3e9e56958e8ea3e636df3369b29c3d4b94dce9d8
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 10:59:52 -07:00
Roy Oursler
547e8e1893 igzip: Allow performance check with small input buffers
Add the SMALL_IN_BUFFER compile time condition to igzip_file_perf to perform
perfomance checking when data is separated into small input buffers.

Change-Id: Ic7a6451f23a46773e100b17ce86f84e41e4df48a
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 10:07:06 -07:00
Greg Tucker
bda088b31a igzip: Fix MS incorrect reg compare for igzip_update_histogram
Early exit case for zero length update would compare the wrong register in
windows.  Could cause improper null exit.

Change-Id: I77539556463813a783d8d351df5a20c99299d100
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 10:06:28 -07:00
Greg Tucker
88192ce59f igzip: Add missing slver for isal_create_hufftables_subset
Change-Id: I331c3edacaeb4c3a86839d7f1e099e189aed8200
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 10:05:56 -07:00
Greg Tucker
bd4873c8ea igzip: Fix missing paren in generate hufftable tool
Change-Id: I1b6d771bfaffd8353e260b7ae286d9abef5e4590
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 09:56:28 -07:00
Greg Tucker
352719d62d igzip: Fix missing nx stack section on new file
Change-Id: I9442c6748a1e1c4919ea4e55b6d7df7596d96f49
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-06 09:53:35 -07:00
Roy Oursler
8fe5cbeeed igzip: Decrease stack size of update_histogram_base.
Change-Id: I7329fc0b560d706ff05e1766f4f3e8f9cb1cc306
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:13:41 -07:00
Roy Oursler
7391f27cca igzip: Fix small bugs in igzip_rand_test on test failure.
Change-Id: I7677c3f28ddd9feee22632cb0f117ba2a4dfbc65
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:12:45 -07:00
Roy Oursler
c17f2959d4 igzip: Fix bug read_header causing false error return values
Change-Id: I7b2fc0d7fa1ea78fab6c8788618d3e48ece56cfd
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:12:11 -07:00
Roy Oursler
0bac63c2e6 igzip: Add test for fuzz testing
Change-Id: I9b1bbdbb986f81094004caedde609c72042015c2
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:11:49 -07:00
Roy Oursler
b95c60cd1b igzip: Add check for non-valid decoded symbols
Change-Id: Ie95c28ca87b76ba8751fa5b07be345d3b00d5134
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:11:12 -07:00
Roy Oursler
fb62881057 igzip: Fix bugs in setup dynamic header when invalid input is supplied
Change-Id: I8df868e259397b289387ae7eb25d0805a22e33ea
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:10:34 -07:00
Roy Oursler
8c5fa7010a igzip: Fix incorrect block_state for inflate_stateless
Change-Id: Id26b528dca540ac8e177af59f9ace797fbb0a752
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-06 09:08:40 -07:00
Greg Tucker
a578be3e70 igzip: New default hufftables
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
Reviewed-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-05 17:18:53 -07:00
Roy Oursler
e2b0640fd9 igzip: Make inline functions in igzip_inflate.c static
Make inline functions in igzip_inflate.c static to fix compiling issue on Mac.

Change-Id: Id4fdc0b8fe9ea9655bc3368bade579aebc8c2a72
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
2016-12-05 16:44:31 -07:00
Roy Oursler
88f95d8594 igzip: Merge inflate.h and igzip_lib.h and fix external naming
Merge inflate.h and igzip_lib.h and rename many defines to avoid conflicts,
along with general cleanup of the new header.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:40:43 -07:00
Roy Oursler
be8e83b633 igzip: Fix bug in inflate perf test and modify inflate lookup table sizes
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:37:57 -07:00
Roy Oursler
9d53af0c7c igzip: Improve LARGE_WINDOW and change that option to be default.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:36:31 -07:00
Roy Oursler
ee2e2bce4f igzip: Improving igzip_update_histogram
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:31:23 -07:00
Roy Oursler
22bd9eef65 igzip: Optimize igzip
Remove conditional move and reorder operations to increase speed.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:30:56 -07:00
Roy Oursler
d06e14b937 igzip: Port update_histogram and decode_huffman_code_block to windows
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:17:54 -07:00
Roy Oursler
340e0f99df igzip: Implement stateful inflate.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:17:07 -07:00
Roy Oursler
6b2bedfe28 igzip: Fix to run on SSE
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:16:44 -07:00
Roy Oursler
a1981da94e igzip: Create stateless FULL_FLUSH mode
Add flush modes to isal_deflate_stateless. IGZIP_USE_GZIP_FORMAT is not
supported because of limitations imposed by the statelessness.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:16:11 -07:00
Roy Oursler
b25ef61a57 igzip: Fix warning for igzip_update_histogram when compiled with LARGE_WINDOW.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:15:51 -07:00
Roy Oursler
511a5dea5b igzip: Add missing include to igzip_check.c
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:14:03 -07:00
Roy Oursler
869712d389 igzip: Rename inflate.asm and fix Makefile.am so that Makefile.unx works
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:12:53 -07:00
Roy Oursler
60d9a30f4e igzip: Rename small_code_lookup
Rename small_code_lookup to short_code_lookup to match naming of
long_code_lookup.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:01:46 -07:00
Roy Oursler
464476003a igzip: Separate inflate_huffman_code into a large and small datastructure.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:01:21 -07:00
Roy Oursler
e64bc2ed37 igzip: Improve make_inflate_huffcode algorithm.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 16:00:57 -07:00
Roy Oursler
456be4cfc1 igzip: Rearrange inflate_state
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:53:34 -07:00
Roy Oursler
5a7c97a930 igzip: Document inflate_huff_code data structure.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:52:56 -07:00
Roy Oursler
40b5104397 igzip: More optimizations by speeding up rarely taken branch
For some reason optimizing the rarely taken branch speeds up the program.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:51:22 -07:00
Roy Oursler
84ffaead82 igzip: Optimize inflate more
Some general optimizations, including speculatively loading the distance huffman
code and interleaving inflate_in_load with other instructions.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:48:47 -07:00
Roy Oursler
6fc9029f83 inflate: Optimize inflate_stateless
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:47:38 -07:00
Roy Oursler
c233946b0a igzip: Add some compile time options to igzip_inflate_perf
Allow for running igzip_inflate_perf with zlib -1, zlib -9, and igzip
compression. Also add a parameter to allow comparing igzip inflate's performance
to zlib inflate's performance.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:47:01 -07:00
Roy Oursler
09a5a243bf igzip: Implement statelesss inflate in assembly
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 12:45:53 -07:00
Roy Oursler
17dac9f641 igzip: Implement xmm loads to preload data in update_histogram
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 11:07:28 -07:00
Roy Oursler
4d40cd360d igzip: Speed up hash table initialization in igzip_update_histogram.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 11:02:23 -07:00
Roy Oursler
39ce31de30 igzip: Fix minor bug in igzip_rand_test
Stop igzip_rand_test from trying to free a null pointer.

Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 11:01:40 -07:00
Roy Oursler
af9c0c0f46 igzip: update update_hash to match stateless and stateful compression.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-05 11:01:01 -07:00
Roy Oursler
f97de75fc6 igzip: Port improvements to stateless compress to stateful compress
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 18:40:27 -07:00
Roy Oursler
eb1b7788d0 igzip: Move code in igzip_stateless to hide latencies more in ivybridge.
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:23:29 -07:00
Roy Oursler
cf30138c7b igzip: Load memory into xmm in stateless registers
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:22:52 -07:00
Roy Oursler
4d1fe78bfa igzip: Modify stateless hash updating to match the length 4 short match
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:19:15 -07:00
Roy Oursler
31814483c0 igzip: Create assembly version of isal_update_histogram
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:16:55 -07:00
Roy Oursler
7c91df5e50 igzip: add macros for shrx, shlx, and bzhi instructions on old architectures
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:11:06 -07:00
Roy Oursler
bbc886cf01 igzip: Modify igzip to ignore matches which are shorter than 4
Signed-off-by: Roy Oursler <roy.j.oursler@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:07:45 -07:00
Greg Tucker
45311ea249 Change common and igzip multibinary check from sse4.1 to sse4.2
SSE optimized compression function is using crc32 instruction that is in SSE 4.2
not SSE4.1 as stated.  Fixes incorect choice on core 2 duo that is SSE 4.1 only.

Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 17:05:23 -07:00
Xiaodong Liu
b34cb054fd build: Fix an include path to be srcdir relative
Allows configure to again build in an external directory.  When building ISAL in
an external path, assembler or compiler needs relative include paths.

Signed-off-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Greg Tucker <greg.b.tucker@intel.com>
2016-12-02 16:54:40 -07:00
Greg Tucker
1abf68b7db igzip: Fix missing .note.gnu-stack for non-executable stack
One asm file in the source list for compression missed an include that adds
.note.gnu-stack.  Without it an executable could be marked has having an
executable stack and miss some hardware protections.

Reported-by: Ondřej Nový <novy@ondrej.org>
Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-07-11 14:42:25 -07:00
Greg Tucker
660f49b02d Add data compression unit
Include fast DEFLATE compatable compression functions.

Signed-off-by: Greg Tucker <greg.b.tucker@intel.com>
2016-06-10 17:03:38 -07:00