Manuel Pégourié-Gonnard
d4588cfb6a
aesni_gcm_mult() now returns void
2013-12-30 13:54:23 +01:00
Manuel Pégourié-Gonnard
80637c7520
Use aesni_gcm_mult() if available
2013-12-26 16:09:58 +01:00
Manuel Pégourié-Gonnard
9d57482280
Add comments on GCM multiplication
2013-12-26 15:51:13 +01:00
Manuel Pégourié-Gonnard
dae7093875
gcm_selftest depends on AES
2013-10-24 15:06:33 +02:00
Manuel Pégourié-Gonnard
4fe9200f47
Fix memory leak in GCM by adding gcm_free()
2013-09-13 13:45:58 +02:00
Paul Bakker
a0558e0484
Check that the cipher GCM receives is a 128-bit-based cipher
2013-09-10 14:25:51 +02:00
Paul Bakker
43aff2aec4
Moved GCM to use cipher layer instead of AES directly
2013-09-09 00:10:27 +02:00
Manuel Pégourié-Gonnard
9241be7ac5
Change cipher prototypes for GCM
2013-08-31 18:07:42 +02:00
Manuel Pégourié-Gonnard
20d6a17af9
Make GCM tag check "constant-time"
2013-08-31 16:37:46 +02:00
Manuel Pégourié-Gonnard
07f8fa5a69
GCM in the cipher layer, step 1
...
- no support for additional data
- no support for tag
2013-08-31 16:08:22 +02:00
Paul Bakker
52cf16caeb
Fixed multiple use of GCM-context bug due to split-up of GCM functions
2013-07-26 13:56:22 +02:00
Paul Bakker
b9d3cfa114
Split up GCM into a start/update/finish cycle
2013-06-26 15:08:29 +02:00
Paul Bakker
b6c5d2e1a6
Cleanup up non-prototyped functions (static) and const-correctness
...
More fixes based on the compiler directives -Wcast-qual -Wwrite-strings
-Wmissing-prototypes -Wmissing-declarations. Not everything with regards
to -Wcast-qual has been fixed as some have unwanted consequences for the
rest of the code.
2013-06-25 16:25:17 +02:00
Paul Bakker
169b7f4a13
Fixed gcm.c formatting (removed redundant spaces)
2013-06-25 15:06:54 +02:00
Paul Bakker
0ecdb23eed
Cleanup of the GCM code
...
Removed unused variable 'v'
orig_len and orig_add_len are now uint64_t to support larger than 2^29
data sizes
2013-04-09 11:36:42 +02:00
Paul Bakker
09d67258a2
Modified to work in-place
2013-03-13 11:46:00 +01:00
Paul Bakker
3d2dc0f8e5
Corrected GCM counter incrementation to use only 32-bits instead of 128-bits
...
Using 32-bits has the possibility to overwrite the IV in the first 12
bytes of the Y variable.
Found by Yawning Angel
2013-02-28 10:55:39 +01:00
Paul Bakker
4a2bd0da0f
- Merged fixes 1394 and 1395 from trunk to PolarSSL 1.2 branch
2012-11-02 11:06:08 +00:00
Paul Bakker
67f9d534ee
- Removed code breaking strict-aliasing
2012-10-23 11:49:05 +00:00
Paul Bakker
5c2364c2ba
- Moved from unsigned long to uint32_t throughout code
2012-10-01 14:41:15 +00:00
Paul Bakker
fc5183cf5d
- Added input checking and more efficient buffer overlap use
2012-04-18 14:17:01 +00:00
Paul Bakker
f9169629c9
- Removed unused variables
2012-03-20 15:05:51 +00:00
Paul Bakker
89e80c9a43
- Added base Galois/Counter mode (GCM) for AES
2012-03-20 13:50:09 +00:00
