App layer error close txs 4318 v6

[catenacyber]

Link to redmine ticket: https://redmine.openinfosecfoundation.org/issues/4318

Describe changes:

• app-layer: clean and "close" all txs if protocol reaches error state

Still need to test this, and check current app-layer parser returning error

Modifies #7243 with adding ticket number in message

[codecov[bot]]

Codecov Report

Merging #7513 (bea2797) into master (8377b9d) will decrease coverage by 0.09%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #7513      +/-   ##
==========================================
- Coverage   75.83%   75.74%   -0.10%     
==========================================
  Files         655      657       +2     
  Lines      186236   186373     +137     
==========================================
- Hits       141239   141161      -78     
- Misses      44997    45212     +215     

Flag	Coverage Δ
fuzzcorpus	59.97% <100.00%> (-0.11%)	:arrow_down:
suricata-verify	52.16% <88.88%> (-0.09%)	:arrow_down:
unittests	60.77% <88.88%> (-0.04%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

[suricata-qa]

Information: QA ran without warnings.

Pipeline 7759

[catenacyber]

One example where I am not sure how it should be dealt with :

if a HTP2 client does not send the magic banner, the app-layer parser returns an error. Should it return an error ? or Should it try to parse further frames ?

It looks like we have two cases here :

• both client and server were recognized using the app-layer protocol
• or client and server do not talk the same protocol

PS : the banner can be split over multiple TCP packets, and the server side may have created multiple transactions before getting the error

[catenacyber]

Another case : SSH where a record has its length <= 1

[suricata-qa]

Information: QA ran without warnings.

Pipeline 7759

[catenacyber]

Not sure of this