On Tue, Jul 28, 2020 at 06:07:44PM +0800, Yu Xu wrote:
> On 7/28/20 5:39 PM, Catalin Marinas wrote:
> > On Tue, Jul 28, 2020 at 10:22:20AM +0100, Will Deacon wrote:
> > > On Sat, Jul 25, 2020 at 04:58:43PM +0100, Catalin Marinas wrote:
> > > > On Fri, Jul 24, 2020 at 06:29:43PM -0700, Linus Torvalds wrote:
> > > > > For any architecture that guarantees that a page fault will always
> > > > > flush the old TLB entry for this kind of situation, that
> > > > > flush_tlb_fix_spurious_fault() thing can be a no-op.
> > > > >
> > > > > So that's why on x86, we just do
> > > > >
> > > > > #define flush_tlb_fix_spurious_fault(vma, address) do { } while (0)
> > > > >
> > > > > and have no issues.
> > > > >
> > > > > Note that it does *not* need to do any cross-CPU flushing or anything
> > > > > like that. So it's actually wrong (I think) to have that default
> > > > > fallback for
> > > > >
> > > > > #define flush_tlb_fix_spurious_fault(vma, address)
> > > > > flush_tlb_page(vma, address)
> > > > >
> > > > > because flush_tlb_page() is the serious "do cross CPU etc".
> > > > >
> > > > > Does the arm64 flush_tlb_page() perhaps do the whole expensive
> > > > > cross-CPU thing rather than the much cheaper "just local invalidate"
> > > > > version?
> > > >
> > > > I think it makes sense to have a local-only
> > > > flush_tlb_fix_spurious_fault(), but with ptep_set_access_flags() updated
> > > > to still issue the full broadcast TLBI. In addition, I added a minor
> > > > optimisation to avoid the TLB flush if the old pte was not accessible.
> > > > In a read-access fault case (followed by mkyoung), the TLB wouldn't have
> > > > cached a non-accessible pte (not sure it makes much difference to Yang's
> > > > case). Anyway, from ARMv8.1 onwards, the hardware handles the access
> > > > flag automatically.
> > > >
> > > > I'm not sure the first dsb(nshst) below is of much use in this case. If
> > > > we got a spurious fault, the write to the pte happened on a different
> > > > CPU (IIUC, we shouldn't return to user with updated ptes without a TLB
> > > > flush on the same CPU). Anyway, we can refine this if it solves Yang's
> > > > performance regression.
> > > >
> > > > -------------8<-----------------------
> > > > diff --git a/arch/arm64/include/asm/tlbflush.h b/arch/arm64/include/asm/tlbflush.h
> > > > index d493174415db..d1401cbad7d4 100644
> > > > --- a/arch/arm64/include/asm/tlbflush.h
> > > > +++ b/arch/arm64/include/asm/tlbflush.h
> > > > @@ -268,6 +268,20 @@ static inline void flush_tlb_page(struct vm_area_struct *vma,
> > > > dsb(ish);
> > > > }
> > > > +static inline void local_flush_tlb_page(struct vm_area_struct *vma,
> > > > + unsigned long uaddr)
> > > > +{
> > > > + unsigned long addr = __TLBI_VADDR(uaddr, ASID(vma->vm_mm));
> > > > +
> > > > + dsb(nshst);
> > > > + __tlbi(vale1, addr);
> > > > + __tlbi_user(vale1, addr);
> > > > + dsb(nsh);
> > > > +}
> > > > +
> > > > +#define flush_tlb_fix_spurious_fault(vma, address) \
> > > > + local_flush_tlb_page(vma, address)
> > >
> > > Why can't we just have flush_tlb_fix_spurious_fault() be a NOP on arm64?
> >
> > Possibly, as long as any other optimisations only defer the TLB flushing
> > for relatively short time (the fault is transient, it will get a
> > broadcast TLBI eventually).
> >
> > Either way, it's worth benchmarking the above patch but with
> > flush_tlb_fix_spurious_fault() a no-op (we still need flush_tlb_page()
> > in ptep_set_access_flags()). Xu, Yang, could you please give it a try?
>
> If I understand correctly, this should do as good as the patch of Linux
> or Yang in will-it-scale page_fault3 testcase, which both avoid doing
> flush_tlb_fix_spurious_fault(), in case of FAULT_FLAG_TRIED.
That would be a no-op even when FAULT_FLAG_TRIED isn't set. Not sure
whether it makes a difference for the benchmark though.
> Shouldn't we be concerned about data integrity if to have
> flush_tlb_fix_spurious_fault() be a nop on arm64?
Not as long as it's only called from handle_pte_fault(). In the worst
case (TLBI missing on another path that makes the pte dirty/writable),
it would get stuck in a spurious fault loop. That's why your early
report on the 128 threads test getting stuck worried me a bit.
--
Catalin
