KVM: arm64: fix the mmio faulting
v2 cover KVM: arm64: fix the mmio faulting
Description of the Reproducer scenario as asked in the thread [1].
Tried to create the reproducer scenario with vfio-pci driver using nvidia GPU in PT mode, As because vfio-pci driver now supports vma faulting (/vfio_pci_mmap_fault) so could create a crude reproducer situation with that.
To create the repro - I did an ugly hack into arm64/kvm/mmu.c. The hack is to make sure that stage2 mapping are not created at the time of vm_init by unsetting VM_PFNMAP flag. This unsetting flag needed because vfio-pci's mmap func(/vfio_pci_mmap) by-default sets the VM_PFNMAP flag for the MMIO region but I want
the remap_pfn_range() func to set the _PFNMAP flag via vfio's fault handler func vfio_pci_mmap_fault().
So with above, when guest access the MMIO region, this will trigger the mmio fault path at arm64-kvm hypervisor layer like below:user_mem_abort() {->...
--> Check the VM_PFNMAP flag, since not set so marks force_pte=false
....
__gfn_to_pfn_memslot()-->
...
handle_mm_fault()-->
do_fault()-->
vfio_pci_mmio_fault()-->
remap_pfn_range()--> Now will set the VM_PFNMAP flag.
}Since the force_pte flag is set to false so will lead to THP oops. By setting the force_pte=true will avoid the THP Oops which was mentioned in the [2] and patch proposition [1/1] fixes that.
hackish change to reproduce scenario:
--->
arch/arm64/kvm/mmu.c +6 -0
@@ -1318,6 +1318,12 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm, vm_start =max (hva, vma->vm_start); vm_end =min (reg_end, vma->vm_end); +/* Hack to make sure stage2 mapping not present, thus trigger +* user_mem_abort for stage2 mapping +*/ +if (vma->vm_flags &VM_PFNMAP ) { + vma->vm_flags = vma->vm_flags & (~VM_PFNMAP ); + }if (vma->vm_flags &VM_PFNMAP ) {gpa_t gpa = mem->guest_phys_addr + (vm_start - mem->userspace_addr);
Thanks.
Santosh
[1] https://lkml.org/lkml/2020/10/23/310
[2] https://lkml.org/lkml/2020/10/21/460
Santosh Shukla (1):
KVM: arm64: Correctly handle the mmio faulting
arch/arm64/kvm/mmu.c | 1 + 1 file changed, 1 insertion(+)
Signature · 1 line
2.7.4
PATCH v2 1/1 KVM: arm64: Correctly handle the mmio faulting
The Commit:6d674e28 introduces a notion to detect and handle the device mapping. The commit checks for the VM_PFNMAP flag is set in vma->flags and if set then marks force_pte to true such that if force_pte is true then ignore the THP function check (/transparent_hugepage_adjust()).
There could be an issue with the VM_PFNMAP flag setting and checking. For example consider a case where the mdev vendor driver register's the vma_fault handler named vma_mmio_fault(), which maps the host MMIO region in-turn calls remap_pfn_range() and maps the MMIO's vma space. Where, remap_pfn_range implicitly sets the VM_PFNMAP flag into vma->flags.
Now lets assume a mmio fault handing flow where guest first access the MMIO region whose 2nd stage translation is not present. So that results to arm64-kvm hypervisor executing guest abort handler, like below:
kvm_handle_guest_abort() -->
user_mem_abort()--> {
...
0. checks the vma->flags for the VM_PFNMAP.
1. Since VM_PFNMAP flag is not yet set so force_pte _is_ false;
2. gfn_to_pfn_prot() -->
__gfn_to_pfn_memslot() -->
fixup_user_fault() -->
handle_mm_fault()-->
__do_fault() -->
vma_mmio_fault() --> // vendor's mdev fault handler
remap_pfn_range()--> // Here sets the VM_PFNMAP
flag into vma->flags.
3. Now that force_pte is set to false in step-2),
will execute transparent_hugepage_adjust() func and
that lead to Oops [4].}
The proposition is to set force_pte=true if kvm_is_device_pfn is true.
[4] THP Oops:
Quoted 10 lines
pc:
kvm_is_transparent_hugepage+0x18/0xb0
...
...user_mem_abort+0x340/0x9b8 kvm_handle_guest_abort+0x248/0x468 handle_exit+0x150/0x1b0 kvm_arch_vcpu_ioctl_run+0x4d4/0x778 kvm_vcpu_ioctl+0x3c0/0x858 ksys_ioctl+0x84/0xb8 __arm64_sys_ioctl+0x28/0x38
Tested on Huawei Kunpeng Taishan-200 arm64 server, Using VFIO-mdev device. Linux-5.10-rc1 tip: 3650b228
---
v2:
- Per Marc's suggestion - setting force_pte=true.
- Rebased and tested for 5.10-rc1 commit: 3650b228
v1: https://lkml.org/lkml/2020/10/21/460
arch/arm64/kvm/mmu.c | 1 +
1 file changed, 1 insertion(+)
arch/arm64/kvm/mmu.c +1 -0
@@ -839,6 +839,7 @@ static int user_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa,if (kvm_is_device_pfn (pfn)) { device = true; + force_pte = true; }else if (logging_active && !write_fault) {/* * Only actually map the page as writable if this was a write
Signature · 1 line
2.7.4
Re: [PATCH v2 1/1] KVM: arm64: Correctly handle the mmio faulting
Hi Santosh,
Quoted 62 lines · Santosh Shukla
The Commit:6d674e28 introduces a notion to detect and handle the device mapping. The commit checks for the
VM_PFNMAPflag is set invma->flagsand if set then marksforce_pteto true such that ifforce_pteis true then ignore the THP function check (/transparent_hugepage_adjust()).There could be an issue with the
VM_PFNMAPflag setting and checking. For example consider a case where the mdev vendor driver register's thevma_faulthandler namedvma_mmio_fault(), which maps the host MMIO region in-turn callsremap_pfn_range()and maps the MMIO's vma space. Where,remap_pfn_rangeimplicitly sets theVM_PFNMAPflag intovma->flags.Now lets assume a mmio fault handing flow where guest first access the MMIO region whose 2nd stage translation is not present. So that results to arm64-kvm hypervisor executing guest abort handler, like below:
kvm_handle_guest_abort()-->
user_mem_abort()--> {... 0. checks the vma->flags for the VM_PFNMAP. 1. Since VM_PFNMAP flag is not yet set so force_pte _is_ false; 2. gfn_to_pfn_prot() --> __gfn_to_pfn_memslot() --> fixup_user_fault() --> handle_mm_fault()--> __do_fault() --> vma_mmio_fault() --> // vendor's mdev fault handler remap_pfn_range()--> // Here sets the VM_PFNMAP flag into vma->flags. 3. Now that force_pte is set to false in step-2), will execute transparent_hugepage_adjust() func and that lead to Oops [4].}
The proposition is to set force_pte=true if
kvm_is_device_pfnis true.[4] THP Oops:
Quoted 10 lines
pc:
kvm_is_transparent_hugepage+0x18/0xb0
...
...user_mem_abort+0x340/0x9b8 kvm_handle_guest_abort+0x248/0x468 handle_exit+0x150/0x1b0 kvm_arch_vcpu_ioctl_run+0x4d4/0x778 kvm_vcpu_ioctl+0x3c0/0x858 ksys_ioctl+0x84/0xb8 __arm64_sys_ioctl+0x28/0x38Tested on Huawei Kunpeng Taishan-200 arm64 server, Using VFIO-mdev device. Linux-5.10-rc1 tip:
3650b228Fixes 6d674e28 ("KVM: arm/arm64: Properly handle faulting of device mappings")Suggested-by Marc Zyngier <[email protected]>Signed-off-by Santosh Shukla <[email protected]>---
v2:
- Per Marc's suggestion - setting force_pte=true.
- Rebased and tested for 5.10-rc1 commit:3650b228
Quoted 5 lines
v1: https://lkml.org/lkml/2020/10/21/460
arch/arm64/kvm/mmu.c| 1 +1 file changed, 1 insertion(+)
arch/arm64/kvm/mmu.c static int user_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa, · line 845
Quoted 13 lines
arch/arm64/kvm/mmu.c +1
-0@@ -839,6 +839,7 @@ static int user_mem_abort(struct kvm_vcpu *vcpu, phys_addr_t fault_ipa,if (kvm_is_device_pfn (pfn)) { device = true; + force_pte = true; }else if (logging_active && !write_fault) {/* * Only actually map the page as writable if this was a write
Cheers,
Gavin
Quoted 8 lines · Santosh Shukla
Description of the Reproducer scenario as asked in the thread [1].
Tried to create the reproducer scenario with vfio-pci driver using nvidia GPU in PT mode, As because vfio-pci driver now supports vma faulting (/vfio_pci_mmap_fault) so could create a crude reproducer situation with that.
[...]
Applied to next, thanks!
[1/1] KVM: arm64: Force PTE mapping on fault resulting in a device mapping
commit: 91a2c34b7d6fadc9c5d9433c620ea4c32ee7cae8
Cheers,
M.
Signature · 1 line
Without deviation from the norm, progress is not possible.