Jeff Lucovsky

Continued in #7932

Continued in #7951

Continued in #8356

[This document](https://intel.github.io/hyperscan/dev-reference/performance.html?highlight=scratch) recommends the allocation of one scratch space per scanning context. [Suricata](https://github.com/OISF/suricata) relies on this performance technique. The commit for this change has caused [this Suricata issue](https://redmine.openinfosecfoundation.org/issues/5882). I investigated...

> @jlucovsky Yes you're right. We already notice this issue in Snort integration and Rspamd. A quick roll back patch for this particular issue will be available soon. Sorry for...

Thanks. I've verified proper operation on the Suricata LTS and current (`master`) releases. Thanks!

@hongyang7 Will there be a new release with the update? Many of our users only used "released" software. Thus, could the change be in 5.4.2?

Continued in #11079

> Are you working on getting a green CI ? Yes, that's why this is a draft.

@ralpheastwood - We'd like your help with this PR. Could you validate that it works properly on a Suricata 6.0.x build/deployment?