Re: kvm_read_guest_page() missing kvm->srcu read lock?

Hi, On 11/05/18 17:44, Paolo Bonzini wrote: > On 11/05/2018 15:25, Andre Przywara wrote: >> - I couldn't find any protection for the usage in >> arch/powerpc/kvm/powerpc.c, but the call chain is quite convoluted >> there, so I might have missed something. It would be good if someone >> more famili

Re: kvm_read_guest_page() missing kvm->srcu read lock?

On 11/05/2018 15:25, Andre Przywara wrote: > - I couldn't find any protection for the usage in > arch/powerpc/kvm/powerpc.c, but the call chain is quite convoluted > there, so I might have missed something. It would be good if someone > more familiar with this code would take a look. I also didn't

Re: kvm_read_guest_page() missing kvm->srcu read lock?

Hi, On 11/05/18 12:43, Paolo Bonzini wrote: > On 11/05/2018 13:02, Andre Przywara wrote: >> Hi Paolo, >> >> thanks for the answer! >> Took me a bit, but I think you are right (see below). >> >> On 10/05/18 18:43, Paolo Bonzini wrote: >>> On 10/05/2018 19:41, Andre Przywara wrote: Hi, >>>

Re: kvm_read_guest_page() missing kvm->srcu read lock?

On 11/05/2018 13:02, Andre Przywara wrote: > Hi Paolo, > > thanks for the answer! > Took me a bit, but I think you are right (see below). > > On 10/05/18 18:43, Paolo Bonzini wrote: >> On 10/05/2018 19:41, Andre Przywara wrote: >>> Hi, >>> >>> Jan posted an lockdep splat complaining about a suspi

Re: kvm_read_guest_page() missing kvm->srcu read lock?

Hi Paolo, thanks for the answer! Took me a bit, but I think you are right (see below). On 10/05/18 18:43, Paolo Bonzini wrote: > On 10/05/2018 19:41, Andre Przywara wrote: >> Hi, >> >> Jan posted an lockdep splat complaining about a suspicious >> rcu_dereference_check: >> https://lists.cs.columbi

Re: kvm_read_guest_page() missing kvm->srcu read lock?

On 10/05/2018 19:41, Andre Przywara wrote: > Hi, > > Jan posted an lockdep splat complaining about a suspicious > rcu_dereference_check: > https://lists.cs.columbia.edu/pipermail/kvmarm/2018-May/031116.html > > The gist of that is: > ... > [ 1025.695517] dump_stack+0x9c/0xd4 > [ 1025.695524] lo

kvm_read_guest_page() missing kvm->srcu read lock?

Hi, Jan posted an lockdep splat complaining about a suspicious rcu_dereference_check: https://lists.cs.columbia.edu/pipermail/kvmarm/2018-May/031116.html The gist of that is: ... [ 1025.695517] dump_stack+0x9c/0xd4 [ 1025.695524] lockdep_rcu_suspicious+0xcc/0x118 [ 1025.695537] gfn_to_memslot+