Re: [PATCH v11 8/8] KVM: x86/mmu: Handle non-refcounted pages

[Dmitry Osipenko <dmitry.osipenko@collabora.com>]

Hi David,

On 2/29/24 05:57, David Stevens wrote:
> From: David Stevens <stevensd@chromium.org>
> 
> Handle non-refcounted pages in __kvm_faultin_pfn. This allows the
> host to map memory into the guest that is backed by non-refcounted
> struct pages - for example, the tail pages of higher order non-compound
> pages allocated by the amdgpu driver via ttm_pool_alloc_page.
> 
> Signed-off-by: David Stevens <stevensd@chromium.org>

This patch has a problem on v6.8 kernel. Pierre-Eric of AMD found that
Qemu crashes with "kvm bad address" error when booting Ubuntu 23.10 ISO
with a disabled virtio-gpu and I was able to reproduce it. Pierre-Eric
said this problem didn't exist with v6.7 kernel and using v10 kvm
patches. Could you please take a look at this issue?

To reproduce the bug, run Qemu like this and load the Ubuntu installer:

  qemu-system-x86_64 -boot d -cdrom ubuntu-23.10.1-desktop-amd64.iso -m
4G --enable-kvm -display gtk -smp 1 -machine q35

Qemu fails with "error: kvm run failed Bad address"

On the host kernel there is this warning:

 ------------[ cut here ]------------
 WARNING: CPU: 19 PID: 11696 at mm/gup.c:229 try_grab_page+0x64/0x100
 Call Trace:
  <TASK>
  ? try_grab_page+0x64/0x100
  ? __warn+0x81/0x130
  ? try_grab_page+0x64/0x100
  ? report_bug+0x171/0x1a0
  ? handle_bug+0x3c/0x80
  ? exc_invalid_op+0x17/0x70
  ? asm_exc_invalid_op+0x1a/0x20
  ? try_grab_page+0x64/0x100
  follow_page_pte+0xfa/0x4b0
  __get_user_pages+0xe5/0x6e0
  get_user_pages_unlocked+0xe7/0x370
  hva_to_pfn+0xa2/0x760 [kvm]
  ? free_unref_page+0xf9/0x180
  kvm_faultin_pfn+0x112/0x610 [kvm]
  kvm_tdp_page_fault+0x104/0x150 [kvm]
  kvm_mmu_page_fault+0x298/0x860 [kvm]
  kvm_arch_vcpu_ioctl_run+0xc7d/0x16b0 [kvm]
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? kvm_arch_vcpu_put+0x128/0x190 [kvm]
  ? srso_alias_return_thunk+0x5/0xfbef5
  kvm_vcpu_ioctl+0x199/0x700 [kvm]
  __x64_sys_ioctl+0x94/0xd0
  do_syscall_64+0x86/0x170
  ? kvm_on_user_return+0x64/0x90 [kvm]
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? fire_user_return_notifiers+0x37/0x70
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? syscall_exit_to_user_mode+0x80/0x230
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? do_syscall_64+0x96/0x170
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? do_syscall_64+0x96/0x170
  ? do_syscall_64+0x96/0x170
  ? do_syscall_64+0x96/0x170
  ? srso_alias_return_thunk+0x5/0xfbef5
  ? do_syscall_64+0x96/0x170
  ? srso_alias_return_thunk+0x5/0xfbef5
  entry_SYSCALL_64_after_hwframe+0x6e/0x76
 ---[ end trace 0000000000000000 ]---

-- 
Best regards,
Dmitry