[openssl/openssl] 96dd7b: Fix performance regression of ChaCha20 on LoongArch64

[lrzlin]

  Branch: refs/heads/openssl-3.2
  Home:   https://github.com/openssl/openssl
  Commit: 96dd7b23f5f1a54316ab9408be630f7cd61e262e
      https://github.com/openssl/openssl/commit/96dd7b23f5f1a54316ab9408be630f7cd61e262e
  Author: Lin Runze <lrzlin at 163.com>
  Date:   2024-01-17 (Wed, 17 Jan 2024)

  Changed paths:
    M crypto/chacha/asm/chacha-loongarch64.pl

  Log Message:
  -----------
  Fix performance regression of ChaCha20 on LoongArch64

The regression was introduced in PR #22817.

In that pull request, the input length check was moved forward,
but the related ori instruction was missing, and it will cause
input of any length down to the much slower scalar implementation.

Fixes #23300

CLA: trivial

Reviewed-by: Shane Lontis <shane.lontis at oracle.com>
Reviewed-by: Tomas Mraz <tomas at openssl.org>
(Merged from https://github.com/openssl/openssl/pull/23301)

(cherry picked from commit 971028535e6531c89449e06b1f6862c18f04ff91)