On Mon Apr 5, 2021 at 2:39 AM CDT, Niels Möller wrote:
> "Christopher M. Riedl" writes:
>
> > An implementation combining AES+GCM _can potentially_ yield significant
> > performance boosts by allowing for increased instruction parallelism,
> > avoiding
> > C-function call overhead, more
On Thu, Apr 1, 2021 at 12:01 AM Maamoun TK
wrote:
> I'll modify the patch of basic AES-128 optimized functions to be built on
> top of the splitted aes functions.
>
Done!
It works as a file-override basis. The patch also passes the testsuite and
yields expected benchmark numbers.
regards,
Hello,
I've added test cases to verify that unwrap fail if the input values are
incorrect [1]. I reuse all the unwrap test cases, changed one ciphertext
byte and expect the unwrap function to return 0.
Le 2021-03-29 à 13 h 32, Niels Möller a écrit :
That's one possibility, but I think it
"Christopher M. Riedl" writes:
> An implementation combining AES+GCM _can potentially_ yield significant
> performance boosts by allowing for increased instruction parallelism, avoiding
> C-function call overhead, more flexibility in assembly fine-tuning, etc. This
> series provides such an
Hi!
This series introduces a mechanism to support arch specific, combined AES+GCM
{en,de}cryption functions. These functions are stubbed by default and will
fall-back to the separate hash and crypt functions if no arch override exists.
The arch override can be provided either at build time via
This implementation is based on the existing, per-algorithm optimized
powerpc64/p8/aes-encrypt-internal.asm and powerpc64/p8/gcm-hash.asm
implementations by Niels Möller and Mamone Tarsha. See the previous
gcm_aes_encrypt() commit for details about major changes.
Signed-off-by: Christopher M.
Currently the AES-GCM crypt and hash parts are performed in two separate
functions. Each can be replaced with an arch-specific optimized assembly
routine. This makes it difficult to introduce an arch-specific routine
implementing the combination of both parts in a single function.
Rework the
This implementation is based on the existing, per-algorithm optimized
powerpc64/p8/aes-encrypt-internal.asm and powerpc64/p8/gcm-hash.asm
implementations by Niels Möller and Mamone Tarsha.
Significant changes:
- Combine AES + GCM into a single function call which does up-to 8x
unrolled AES
Enable runtime override via FAT for gcm_aes_{de,en}crypt() on ppc
ISA 3.0 (P9 and beyond) platforms.
Signed-off-by: Christopher M. Riedl
---
fat-ppc.c | 33 +++
fat-setup.h | 6 +
gcm-internal.h| 14
The AC_ARG_ENABLE(...) macro for --enable-power-altivec is called with
enable_altivec=no as the default when the commandline option is not
given to configure. However, the variable $enable_power_altivec is
actually checked - not $enable_altivec. This doesn't matter in practice
since
Signed-off-by: Christopher M. Riedl
---
configure.ac | 9 -
fat-ppc.c| 12
2 files changed, 20 insertions(+), 1 deletion(-)
diff --git a/configure.ac b/configure.ac
index 253735a7..a0df0cc8 100644
--- a/configure.ac
+++ b/configure.ac
@@ -101,6 +101,10 @@
11 matches
Mail list logo