After bisecting to commit 4e7babba30d820c4195b1d58cf51dce3c22ecf2b: crypto: testmgr - convert skcipher testing to use testvec_configs and getting a glimpse of the error during one of the bisect runs (otherwise I didn't get any errors, kernel hangs without any warnings before KMS kicks in), HARDENED_USERCOPY_PAGESPAN probably does what it's supposed to do and the system hangs up. Now, it does state that it should be used only to find such use-cases before they eventually get removed and this one is new (or dusted off an old one), I don't have a trace - but I thought I should report it all the same. Should be easy enough to reproduce... Ivan
Adding Herbert, couldn't find Eric with 100% certainty...