On Thu, 2023-01-19 at 23:11 +0000, Sean Christopherson wrote:
> On Thu, Jan 19, 2023, Huang, Kai wrote:
> > On Thu, 2023-01-19 at 21:36 +0000, Sean Christopherson wrote:
> > > On Thu, Jan 19, 2023, Huang, Kai wrote:
> > > > On Thu, 2023-01-19 at 15:37 +0000, Sean Christopherson wrote:
> > > > > On Thu, Jan 19, 2023, Huang, Kai wrote:
> > > > > > On Tue, 2023-01-17 at 21:01 +0000, Sean Christopherson wrote:
> > > > > > > On Tue, Jan 17, 2023, Sean Christopherson wrote:
> > > > > > > > On Tue, Jan 17, 2023, Zhi Wang wrote:
> > > > > > > Oh, the other important piece I forgot to mention is that dropping mmu_lock deep
> > > > > > > in KVM's MMU in order to wait isn't always an option. Most flows would play nice
> > > > > > > with dropping mmu_lock and sleeping, but some paths, e.g. from the mmu_notifier,
> > > > > > > (conditionally) disallow sleeping.
> > > > > >
> > > > > > Could we do something similar to tdp_mmu_iter_cond_resched() but not simple busy
> > > > > > retrying "X times", at least at those paths that can release mmu_lock()?
> > > > >
> > > > > That's effectively what happens by unwinding up the stak with an error code.
> > > > > Eventually the page fault handler will get the error and retry the guest.
> > > > >
> > > > > > Basically we treat TDX_OPERAND_BUSY as seamcall_needbreak(), similar to
> > > > > > rwlock_needbreak(). I haven't thought about details though.
> > > > >
> > > > > I am strongly opposed to that approach. I do not want to pollute KVM's MMU code
> > > > > with a bunch of retry logic and error handling just because the TDX module is
> > > > > ultra paranoid and hostile to hypervisors.
> > > >
> > > > Right. But IIUC there's legal cases that SEPT SEAMCALL can return BUSY due to
> > > > multiple threads trying to read/modify SEPT simultaneously in case of TDP MMU.
> > > > For instance, parallel page faults on different vcpus on private pages. I
> > > > believe this is the main reason to retry.
> > >
> > > Um, crud. I think there's a bigger issue. KVM always operates on its copy of the
> > > S-EPT tables and assumes the the real S-EPT tables will always be synchronized with
> > > KVM's mirror. That assumption doesn't hold true without serializing SEAMCALLs in
> > > some way. E.g. if a SPTE is zapped and mapped at the same time, we can end up with:
> > >
> > > vCPU0 vCPU1
> > > ===== =====
> > > mirror[x] = xyz
> > > old_spte = mirror[x]
> > > mirror[x] = REMOVED_SPTE
> > > sept[x] = REMOVED_SPTE
> > > sept[x] = xyz
> >
> > IIUC this case cannot happen, as the two steps in the vcpu0 are within read
> > lock, which prevents from vcpu1, which holds the write lock during zapping SPTE.
>
> Zapping SPTEs can happen while holding mmu_lock for read, e.g. see the bug fixed
> by commit 21a36ac6b6c7 ("KVM: x86/mmu: Re-check under lock that TDP MMU SP hugepage
> is disallowed").
Sorry, I now recall that I once had patch to change write path to hold
write_lock() for TDX guest to avoid such problem. I thought it was upstream
behaviour :)
