On Thu, Sep 04, 2014 at 07:44:51PM +0200, Paolo Bonzini wrote:
Il 04/09/2014 17:05, Gleb Natapov ha scritto:
If you do that, KVM gets down to the if (writeback) and writes the
ctxt-eip from L2 into the L1 EIP.
Heh, that's a bummer. We should not write back if an instruction caused a
On Tue, Sep 02, 2014 at 05:13:49PM +0200, Paolo Bonzini wrote:
This is required for the following patch to work correctly. If a nested page
fault happens during emulation, we must inject a vmexit, not a page fault.
Luckily we already have the required machinery: it is enough to return
of treating nested page
faults like other nested vmexits during emulation (which is what this patch
does).
If I included this patch, I could then remove kvm_propagate_fault
like (I think) this:
diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
index 92493e10937c..e096db566ac2 100644
--- a/arch
-arch.pio.in) {
/* FIXME: return into emulator if single-stepping. */
But I'm not sure how to test it, and I like the idea of treating nested page
faults like other nested vmexits during emulation (which is what this patch
does).
IMO exits due to instruction intercept
Il 04/09/2014 17:05, Gleb Natapov ha scritto:
if (ctxt-have_exception) {
inject_emulated_exception(vcpu);
- r = EMULATE_DONE;
+ return EMULATE_DONE;
If there was no vmexit we still want to writeback. Perhaps:
writeback =
Il 04/09/2014 17:05, Gleb Natapov ha scritto:
If you do that, KVM gets down to the if (writeback) and writes the
ctxt-eip from L2 into the L1 EIP.
Heh, that's a bummer. We should not write back if an instruction caused a
vmexit.
You're right, that works.
Paolo
-- 8
This is required for the following patch to work correctly. If a nested page
fault happens during emulation, we must inject a vmexit, not a page fault.
Luckily we already have the required machinery: it is enough to return
X86EMUL_INTERCEPTED instead of X86EMUL_PROPAGATE_FAULT.
Reported-by:
and nested EPT/NPT,
which was reported by Valentine.
Reviews are very welcome, I'm walking on thin ice here...
Paolo
Paolo Bonzini (4):
KVM: x86: reserve bit 8 of non-leaf PDPEs and PML4Es in 64-bit mode on AMD
KVM: nSVM: propagate the NPF EXITINFO to the guest
KVM: x86: inject nested page
Hi Paolo,
On 02.09.2014 21:13, Paolo Bonzini wrote:
Patches 3 and 4 fix the interaction between emulator and nested EPT/NPT,
which was reported by Valentine.
I can confirm the initial bug I observed is fixed with these patches
(applied to 3.16.1).
All tests in kvm-unit-test's master also
Il 02/09/2014 18:02, Valentine Sinitsyn ha scritto:
I can confirm the initial bug I observed is fixed with these patches
(applied to 3.16.1).
All tests in kvm-unit-test's master also pass, except for ioio which is
(probably) affected by another (unrelated) bug fixed by Jan back in June
but
] = pf_interception,
Which ends up in tdp_page_fault(). This function calls the generic
__direct_map() handler (that is used for real-mode paging too).
Note that nested page faults are different from guest page faults. Guest
page faults occur for guest virtual addresses (and are handled
Hi to all,
I am working on NPT and have some doubts.
Please correct me if I am wrong,
In NPT only the first time a page fault happens it is trapped by the host.
The following page faults are not trapped by the host in order to
reduce page walk overhead.
My question is...
Which Kvm function
On 14.05.2010, at 15:15, Matteo Signorini wrote:
Hi to all,
I am working on NPT and have some doubts.
Please correct me if I am wrong,
In NPT only the first time a page fault happens it is trapped by the host.
The following page faults are not trapped by the host in order to
reduce
13 matches
Mail list logo
