gocryptfs/internal/stupidgcm, branch v2.2.0 Mirror of gocryptfs source code on Github http://nuetzlich.net/cgit/gocryptfs/atom?h=v2.2.0 2021-09-14T16:58:04+00:00 stupidgcm: add CpuHasAES() 2021-09-14T16:58:04+00:00 Jakob Unterwurzacher 2021-09-14T16:47:07+00:00 urn:sha1:61e37b2439f0b8a7c16458e73cb57c7428fe61f2 Makes the code clearer, and will be used in the next commit. cli: drop -forcedecode flag 2021-09-10T10:14:19+00:00 Jakob Unterwurzacher 2021-09-10T10:14:19+00:00 urn:sha1:d023cd6c95fcbc6b5056ba1f425d2ac3df4abc5a The rewritten openssl backend does not support this flag anymore, and it was inherently dangerour. Drop it (ignored for compatibility) stupidgcm: add PreferOpenSSL{AES256GCM,Xchacha20poly1305} 2021-09-08T17:48:13+00:00 Jakob Unterwurzacher 2021-09-08T17:48:13+00:00 urn:sha1:1a5866729387c09eca1cdc9737d1b02c74c25901 Add PreferOpenSSLXchacha20poly1305, rename PreferOpenSSL -> PreferOpenSSLAES256GCM. stupidgcm: normalize constructor naming 2021-09-07T16:15:04+00:00 Jakob Unterwurzacher 2021-09-07T16:11:11+00:00 urn:sha1:85c2beccaf674c69b3e30b0d646f5c11d91ecb9b New() -> NewAES256GCM() Also add missing NewChacha20poly1305 constructor in without_openssl.go. stupidgcm: revamp package documentation 2021-09-07T16:15:04+00:00 Jakob Unterwurzacher 2021-09-07T15:58:42+00:00 urn:sha1:f47e287c202ba92e9b48e65bf95e0771c6c4997c Maybe interesting for people following https://github.com/rfjakob/gocryptfs/issues/452 stupidgcm: unexport stupidGCM struct 2021-09-07T16:15:04+00:00 Jakob Unterwurzacher 2021-09-07T15:48:55+00:00 urn:sha1:d598536709db355366e90870c6df3508c71c5884 No need to have it exported. stupidgcm: allow zero-length input data 2021-09-07T16:15:04+00:00 Jakob Unterwurzacher 2021-09-07T15:47:48+00:00 urn:sha1:3a80db953da93c741ad391ae124121459c1046b0 We used to panic in this case because it is useless. But Go stdlib supports it, so we should as well. stupidgcm: fix build with CGO_ENABLED=1 without_openssl 2021-09-07T16:15:04+00:00 Jakob Unterwurzacher 2021-09-07T10:57:52+00:00 urn:sha1:738d5a2b3a001318064be649d683684630eff4c2 We missed some "// +build" lines stupidgcm: NewChacha20poly1305: avoid slice append 2021-09-07T16:14:58+00:00 Jakob Unterwurzacher 2021-09-07T10:44:42+00:00 urn:sha1:d9510d0c0ba2ef31145499f4d49ab1d10c03de94 I noticed that growslice() shows up in the cpuprofile. Avoiding slice append for the private jey copy gives a 0.6% speedup: gocryptfs/internal/speed$ benchstat old new name old time/op new time/op delta StupidXchacha-4 5.68µs ± 0% 5.65µs ± 0% -0.63% (p=0.008 n=5+5) name old speed new speed delta StupidXchacha-4 721MB/s ± 0% 725MB/s ± 0% +0.63% (p=0.008 n=5+5) stupidgcm: add testConcurrency 2021-09-07T16:14:05+00:00 Jakob Unterwurzacher 2021-09-07T10:41:58+00:00 urn:sha1:39b1070506436dcf35ac83c7f8cee65067856062 Verifies that we don't corrupt data when called concurrently.