On 14/09/22(Wed) 15:47, Klemens Nanni wrote: > On 14.09.22 18:55, Mike Larkin wrote: > > On Sun, Sep 11, 2022 at 12:26:31PM +0200, Martin Pieuchot wrote: > > > Diff below adds a minimalist set of assertions to ensure proper locks > > > are held in uvm_mapanon() and uvm_unmap_remove() which are the guts of > > > mmap(2) for anons and munmap(2). > > > > > > Please test it with WITNESS enabled and report back. > > > > > > > Do you want this tested in conjunction with the aiodoned diff or by itself? > > This diff looks like a subset of the previous uvm lock assertion diff > that came out of the previous "unlock mmap(2) for anonymous mappings" > thread[0]. > > https://marc.info/?l=openbsd-tech&m=164423248318212&w=2 > > It didn't land eventually, I **think** syzcaller was a blocker which we > only realised once it was committed and picked up by syzcaller. > > Now it's been some time and more UVM changes landed, but the majority > (if not all) lock assertions and comments from the above linked diff > should still hold true. > > mpi, I can dust off and resend that diff, If you want. > Nothing for release, but perhaps it helps testing your current efforts.
Please hold on, this diff is known to trigger a KASSERT() with witness. I'll send an update version soon. Thank you for disregarding this diff for the moment.