Re: [BUG?] X86 arch_tlbbatch_flush() seems to be lacking mm_tlb_flush_nested() integration

[Nadav Amit <nadav.amit@gmail.com>]

On Oct 14, 2022, at 9:19 PM, Jann Horn <jannh@google.com> wrote:

> Hi!
> 
> I haven't actually managed to reproduce this behavior, so maybe I'm
> just misunderstanding how this works; but I think the
> arch_tlbbatch_flush() path for batched TLB flushing in vmscan ought to
> have some kind of integration with mm_tlb_flush_nested().
> 
> I think that currently, the following race could happen:
> 
> [initial situation: page P is mapped into a page table of task B, but
> the page is not referenced, the PTE's A/D bits are clear]
> A: vmscan begins
> A: vmscan looks at P and P's PTEs, and concludes that P is not currently in use
> B: reads from P through the PTE, setting the Accessed bit and creating
> a TLB entry
> A: vmscan enters try_to_unmap_one()
> A: try_to_unmap_one() calls should_defer_flush(), which returns true
> A: try_to_unmap_one() removes the PTE and queues a TLB flush
> (arch_tlbbatch_add_mm())
> A: try_to_unmap_one() returns, try_to_unmap() returns to shrink_folio_list()
> B: calls munmap() on the VMA that mapped P
> B: no PTEs are removed, so no TLB flush happens

Unless I am missing something, flush_tlb_batched_pending() is would be
called and do the flushing at this point, no?

IIUC the scenario, we had some similar cases in the past [1]. Discussing
these scenarios required too many arguments for my liking, and I would’ve
preferred an easier-to-reason batching coordination between the batching
mechanisms. I proposed some schemes in the past, but to be fair, I think
all of them would have some extra overhead.

[1] https://lore.kernel.org/linux-mm/69BBEB97-1B10-4229-9AEF-DE19C26D8DFF@gmail.com/T/#u