Commit Graph

4 Commits (e1516df0cd528ce9484b292ce2231434199fdb24)

Author SHA1 Message Date
David Bauer 0d727f231f treewide: move commonly disabled symbols to generic config
Move new commonly disabled kernel 5.4 symbols to the generic kernel
configuration.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years ago
Christian Lamparter 91a7c9c27d ipq40xx: 5.4: refresh patches and config
This patch just refreshes the 5.4 patches. It seems as if
070-v4.20-soc-qcom-spm-add-SCM-probe-dependency.patch is
already applied, so drop it. It also does a quick
make kernel_oldconfig to get rid of unneeded symbols.

[Looks like USB and Ethernet need some more work].

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 years ago
Eneas U de Queiroz ee4a0afdcd ipq40xx: use neon crypto drivers
This adds the neon based implementations of AES & SHA256.

For AES, according to the kernel config help:

Use a faster and more secure NEON based implementation of AES in CBC,
CTR and XTS modes.
Bit sliced AES gives around 45% speedup on Cortex-A15 for CTR mode
and for XTS mode encryption, CBC and XTS mode decryption speedup is
around 25%. (CBC encryption speed is not affected by this driver.)
This implementation does not rely on any lookup tables so it is
believed to be invulnerable to cache timing attacks.
...
The observed speedups on ipq40xx are more modest: speedup is around 20%
for CTR mode and for XTS mode encryption, CBC and XTS mode decryption
speedup is around 10%. Measurements were made using tcrypt, with
1024-bytes blocks for CTR & CBC, and 4096-bytes for XTS.
The aes-neon-bs driver uses a fallback for CBC encryption; that fallback
could be either the generic driver written in C, or the scalar arm-asm
one.  Even though aes-arm is 1.9% slower, it is more resilient to timing
attacks (the reason for being slower), so it is being included here.

The neon sha256 module increases performance over the generic module by
33%.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
[Enable only ciphers for now, reorder patch in series to help bisect
as new symbols could lead to build failures, 5.4]
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 years ago
John Crispin 272e0a702a ipq40xx: add v5.4 support
Signed-off-by: John Crispin <john@phrozen.org>
4 years ago