Re-wind the loops entirely on kernels optimized for code size. This is
really not good at all performance-wise. But on m68k, it shaves off 4k
of code size, which is apparently important.
Cc: Geert Uytterhoeven <geert@xxxxxxxxxxxxxx>
Cc: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Cc: Ard Biesheuvel <ardb@xxxxxxxxxx>
Signed-off-by: Jason A. Donenfeld <Jason@xxxxxxxxx>
---
lib/crypto/blake2s-generic.c | 30 ++++++++++++++++++------------
1 file changed, 18 insertions(+), 12 deletions(-)
diff --git a/lib/crypto/blake2s-generic.c b/lib/crypto/blake2s-generic.c
index 75ccb3e633e6..990f000e22ee 100644
--- a/lib/crypto/blake2s-generic.c
+++ b/lib/crypto/blake2s-generic.c
@@ -46,7 +46,7 @@ void blake2s_compress_generic(struct blake2s_state *state, const u8 *block,
{
u32 m[16];
u32 v[16];
- int i;
+ int i, j;
WARN_ON(IS_ENABLED(DEBUG) &&
(nblocks > 1 && inc != BLAKE2S_BLOCK_SIZE));
@@ -86,17 +86,23 @@ void blake2s_compress_generic(struct blake2s_state *state, const u8 *block,
G(r, 6, v[2], v[ 7], v[ 8], v[13]); \
G(r, 7, v[3], v[ 4], v[ 9], v[14]); \
} while (0)
- ROUND(0);
- ROUND(1);
- ROUND(2);
- ROUND(3);
- ROUND(4);
- ROUND(5);
- ROUND(6);
- ROUND(7);
- ROUND(8);
- ROUND(9);
-
+ if (IS_ENABLED(CONFIG_CC_OPTIMIZE_FOR_SIZE)) {
+ for (i = 0; i < 10; ++i) {
+ for (j = 0; j < 8; ++j)
+ G(i, j, v[j % 4], v[((j + (j / 4)) % 4) + 4], v[((j + 2 * (j / 4)) % 4) + 8], v[((j + 3 * (j / 4)) % 4) + 12]);
+ }
+ } else {
+ ROUND(0);
+ ROUND(1);
+ ROUND(2);
+ ROUND(3);
+ ROUND(4);
+ ROUND(5);
+ ROUND(6);
+ ROUND(7);
+ ROUND(8);
+ ROUND(9);
+ }
#undef G
#undef ROUND
--
2.34.1
