fix WOLFSSL_SHA_CTX for OpenSSL w/Espressif HW hash

[gojimmypi]

Description

This PR takes into account the size of the WC_ESP32SHA struct that is added to the wolfSSL WOLFSSL_SHA_CTX struct for the openSSL holder[] in openssl/sha.h for the Espressif ESP-IDF, fixing #6028.

Note that there are currently other errors when enabling OpenSSL and hardware acceleration on the ESP32, specifically the SHA224 failure noted in #6059 as well as SHA512/224 noted in #6028 (comment). I have a fix for those as well, but the days have turned into weeks for my major update in my ED25519_SHA2_fix branch, so I'm creating this separate PR now to address this specific compile issue.

Fixes zd# n/a

Testing

Successful benchmark and wolfssl_test when See #6028 (comment)

Checklist

• added tests
• updated/added doxygen
• updated appropriate READMEs
• Updated manual and documentation

[dgarske]

Fixed in https://github.com/wolfSSL/wolfssl/pull/5938/files/b801a96f8ce445bf8f78c2e7c794096eaa383380..121ee7a6df766ee240f0cb49306daea224ff0e57#diff-b8e6753fbe152cbd7f361ce7515fd023cb544c4e3b912217ee581461b1ce6a8aR106

[gojimmypi]

@dgarske although the #5938 update certainly looked like it would have addresses this issue, it did not.

The sizeof(WC_ESP32SHA) continues to be problematic as noted in 3223fa9.

I'm reopening this issue with the same change suggestion.

To see the error, turn on #define OPENSSL_EXTRA for the ESP32. The error itself is manifested in ssl.c:

[gojimmypi]

ah, apparently once I squash/merged my upstream, I can no longer reopen this issue:

I'll create a fresh PR.