On Oct 27, 2022, at 12:27 AM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote: >> One alternative may be using mm_tlb_flush_pending() when setting a new PTE >> to check for pending flushes and flushing the TLB if that is the case. This >> is somewhat similar to what ptep_clear_flush() does. Anyhow, I guess this >> might induce some overheads. As noted before, it is possible to track >> pending TLB flushes in VMA/page-table granularity, with different tradeoffs >> of overheads. > > Right; I just don't believe in VMAs for this, they're *waaay* to big. Well, I did it for VMA in an RFC only because I was pushed. I thought and do think that page-table granularity is the right one.
