Not always, I'd say they still do not. Look into any docs with references to 888828.1. For example KB214168 and try link there. It looks like: support.oracle.com/rs?type=doc&... and do not work for me. I hope this will improve soon.

KB153930, no longer easy to remember. Lots of other difficult to find (888.1, 555.1). Funny thing: if you search by old numbers you find other documents which include links to original docs, which as you might guess... no loger works. Bookmarked documents... forget it (feature not implementend :( )

Just verified this in my lab and it solves it. Easiest way to check:
term1: dd if=/dev/dm-12 of=/dev/null status=progress iflag=direct bs=128K
term2: iostat -xdm dm-12 1
Before iostat show 2x what dd shows.
With new kernel - the same results

8 (standard FC/FCP) - but it does not matter. I reproduced this in my lab using iSCSI with only 1 path.

Bug 37931495 (hidden). They just replied it is solved in kernel-uek-5.15.0-310.184.5.2.el8uek.x86_64 - but we did not verified this yet

We notice similar issue recently on OLVM and submitted bug for OL8. Iostat (kernel stats) reports double IOPs and MBs values on multipath devices.

This is indeed strange

ASM has backup header at the end of second allocation unit (Doc ID 3016283.1) which is probably used when first is currupted.
BTW: That is why I prefere "wipefs -a <device>" instead of "dd" - when want to cleanup disk.

Sometimes it understands me better than my wife does. ;)

Just started... ;)