aboutsummaryrefslogtreecommitdiffstats
path: root/src/hash
Commit message (Collapse)AuthorAgeFilesLines
* The Algorithm_Factory has this logic on looking for an object:lloyd2011-07-051-0/+2
| | | | | | | | | | | | | | | | | | | | - Check the cache; if found, return value - Populate cache, if the value is already there, delete the old object and save the new one. - Recheck the cache value Raja <[email protected]> pointed out on the list that this could race if multiple threads called a lookup function in close succession while the cache was cold. All of them would fail the lookup, then each of them would add it, but the values returned would be deleted by other threads. Instead, declare that first write wins. Then, the cache stays consistent even if there is a race, the only issue is an extra search and delete. Modify GOST and Skein, as their name() function did not roundtrip properly which caused failures otherwise.
* Rename some dirs for shorter object files where configure willlloyd2011-06-033-0/+0
| | | | coalesce the dir and file name if identical.
* Fix building with --via-amalgamation; it wouldn't generate thelloyd2011-06-032-2/+17
| | | | | | | | amalgamation properly, but would happen to work if a previously written amalgamation was around. Also make changes allowing using the SIMD optimized versions of SHA-1 and Serpent to be used in the amalgamation.
* Rename all references of ia32 to x86-32 and amd64 to x86-64.lloyd2011-04-2217-96/+96
| | | | | | Back the reported version from 1.10.0 to 1.9.17 for the time being. Still on the fence if this will be 1.10.0 or another release candidate instead.
* Maintainer mode warning cleanups, mostly for C style casts which Illoyd2011-04-183-4/+16
| | | | added to the flags here.
* My workaround for gcc pr 44174 caused cpuid support to be silentlylloyd2011-04-181-2/+2
| | | | | | | disabled under gcc. Hide asm labels in the x86-32 code; some did not begin with a .L so they were being exported.
* Indentlloyd2011-03-211-6/+6
|
* Use the same mesasge expansion trick in SHA-512. Performance onlloyd2011-03-132-119/+121
| | | | an i7-860 goes from 197 MiB/s to 256 - a 30% improvement!
* Modify SHA-224/SHA-256 so that it does the message expansion inlloyd2011-03-122-107/+101
| | | | | | registers as needed (16 turns before the value is used), with indexes fully precomputed. On a Core i7-860 with GCC 4.5.2, performance went from 152 MiB/s to 167 MiB/s.
* Update Keccak to the round 3 variant announced 2011-01-17lloyd2011-01-202-9/+4
| | | | | Only change is the padding rule. It 'simplifies' the padding by making it less flexible and harder to implement efficiently. :(
* Avoid more VC warningslloyd2010-12-133-4/+8
|
* Do the XOR a word at a time where possible which gets performance backlloyd2010-12-131-0/+17
| | | | | to the level of the version using xor_buf, at least on a little-endian CPU.
* Fix Keccak message XOR, did not work on big-endian CPUslloyd2010-12-131-8/+12
|
* Reset state on finalizationlloyd2010-12-111-3/+2
|
* XOR the input directly into the state. Only slightly faster (about 5%lloyd2010-12-112-19/+22
| | | | | | on motoko), but does also reduce memory overhead. Perhaps doing the XOR by hand and then doing aligned 64-bit XORs where possible would be beneficial here.
* Single update for padding byteslloyd2010-12-111-4/+3
|
* Add a simple but functioning implementation of Keccak. Only thelloyd2010-12-103-0/+230
| | | | proposed SHA-3 parameter sets are supported.
* Mass disable copy constructors and assignment on all algos until proven ↵lloyd2010-11-041-3/+0
| | | | safe/useful
* Update Skein-512 for the new 1.3 specification, just changes the keylloyd2010-11-031-2/+2
| | | | schedule constant to prevent rotational attacks.
* Remove needless destructorslloyd2010-11-031-2/+0
|
* Doxygenlloyd2010-11-025-4/+34
|
* Add new top-level algorithm which provides basic functionality: namelloyd2010-11-012-13/+4
| | | | | | | | query, clearing, and cloning. Applies to ciphers, hashes, MACs, and PBKDFs. May extend to KDFs later as well. A single combined hierarchy in particular will make the algo_factory much simpler.
* Remove BufferedComputation::OUTPUT_LENGTHlloyd2010-10-2925-95/+103
|
* More size_tlloyd2010-10-131-1/+1
|
* More size_t. Document changeslloyd2010-10-133-6/+6
|
* Use buffer.size() directly to avoid lots of virtual calls in core looplloyd2010-10-131-9/+9
|
* Remove HashFunction::HASH_BLOCK_SIZE entirelylloyd2010-10-1310-36/+44
|
* Remove most uses of HASH_BLOCK_SIZElloyd2010-10-1319-44/+49
|
* Use output_length() instead of OUTPUT_LENGTH pseudo-propertylloyd2010-10-1316-23/+23
|
* More size_tlloyd2010-10-131-1/+1
|
* Use size_t for BufferedComputation::add_datalloyd2010-10-1244-161/+165
|
* Split SHA-2 into 32 and 64 bit versions; they are totally independentlloyd2010-09-306-1/+6
| | | | of each other anyway.
* Make configure output more sensible wrt incompatible moduleslloyd2010-09-284-0/+16
|
* Do the prep/unroll phase 4 rounds before it is needed instead of 3;lloyd2010-09-211-97/+92
| | | | tests on Nehalem indicate a small but measurable win there (about 3%).
* Clean up, hide union accesses with a macro to make it easier to testlloyd2010-09-211-40/+92
| | | | alternative methods of getting pieces of the expanded message.
* Implicit conversionslloyd2010-09-141-7/+7
|
* More changes to avoid vector to pointer implicit conversionslloyd2010-09-141-2/+2
|
* Completely remove the second parameter to SecureVector which specifieslloyd2010-09-1418-53/+69
| | | | | | | | | | | | | | | | | | | | the initial/default length of the array, update all users to instead pass the value to the constructor. This is a old vestigal thing from a class (SecureBuffer) that used this compile-time constant in order to store the values in an array. However this was changed way back in 2002 to use the same allocator hooks as the rest of the containers, so the only advantage to using the length field was that the initial length was set and didn't have to be set in the constructor which was midly convenient. However this directly conflicts with the desire to be able to (eventually) use std::vector with a custom allocator, since of course vector doesn't support this. Fortunately almost all of the uses are in classes which have only a single constructor, so there is little to no duplication by instead initializing the size in the constructor.
* Remove more uses of vector to pointer implicit conversionslloyd2010-09-136-18/+38
|
* More vector->pointer conversion removals.lloyd2010-09-131-3/+3
| | | | | | | | | | | Add RandomNumberGenerator::random_vec, which takes an length n and returns a new SecureVector with randomized contents of that size. This nicely covers most of the cases where randomize was being called on a vector, and is a little cleaner in the code as well, instead of vec.resize(length); rng.randomize(&vec[0], vec.size()); we just write vec = rng.random_vec(length);
* Anywhere where we use MemoryRegion::begin to get access to the raw pointerlloyd2010-09-1315-18/+18
| | | | | representation (rather than in an interator context), instead use &buf[0], which works for both MemoryRegion and std::vector
* Big, invasive but mostly automated change, with a further attempt atlloyd2010-09-0715-25/+25
| | | | | | | | | | | | | | harmonising MemoryRegion with std::vector: The MemoryRegion::clear() function would zeroise the buffer, but keep the memory allocated and the size unchanged. This is very different from STL's clear(), which is basically the equivalent to what is called destroy() in MemoryRegion. So to be able to replace MemoryRegion with a std::vector, we have to rename destroy() to clear() and we have to expose the current functionality of clear() in some other way, since vector doesn't support this operation. Do so by adding a global function named zeroise() which takes a MemoryRegion which is zeroed. Remove clear() to ensure all callers are updated.
* Also allow clang with 32-bit assembly code, everything seems to worklloyd2010-08-084-75/+0
| | | | fine with latest SVN.
* Clang understands at least some GCC inline asm syntax as well as whatlloyd2010-08-081-0/+1
| | | | an .S file is, so allow it for x86-64. Tested/works with Clang SVN.
* Consolidate the two engines that provided assembler implementationslloyd2010-07-131-1/+1
| | | | | | (amd64_eng and ia32_eng) into a new asm_engine. This same engine could also be used in the event that asm code for other CPUs was added later on.
* For the SHA-2 classes, don't use inheritence to share a handful oflloyd2010-06-284-85/+106
| | | | | things, just share the compression function via an anon namespace member, and replicate the simple stuff like copy_out.
* Replace "@return a blah" and "@return the blah" with just "@return blah"lloyd2010-06-161-1/+1
|
* Yet more Doxygen commentslloyd2010-06-162-6/+18
|
* More Doxygen commentslloyd2010-06-167-2/+32
|
* Tiger::clone's result always used 3 passeslloyd2010-06-161-1/+5
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-26 11:26:08 +0000