[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] x86/hvm: Unilaterally inject #UD for unknown VMExits


  • To: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Mon, 1 Dec 2025 10:02:18 +0100
  • Autocrypt: addr=jbeulich@xxxxxxxx; keydata= xsDiBFk3nEQRBADAEaSw6zC/EJkiwGPXbWtPxl2xCdSoeepS07jW8UgcHNurfHvUzogEq5xk hu507c3BarVjyWCJOylMNR98Yd8VqD9UfmX0Hb8/BrA+Hl6/DB/eqGptrf4BSRwcZQM32aZK 7Pj2XbGWIUrZrd70x1eAP9QE3P79Y2oLrsCgbZJfEwCgvz9JjGmQqQkRiTVzlZVCJYcyGGsD /0tbFCzD2h20ahe8rC1gbb3K3qk+LpBtvjBu1RY9drYk0NymiGbJWZgab6t1jM7sk2vuf0Py O9Hf9XBmK0uE9IgMaiCpc32XV9oASz6UJebwkX+zF2jG5I1BfnO9g7KlotcA/v5ClMjgo6Gl MDY4HxoSRu3i1cqqSDtVlt+AOVBJBACrZcnHAUSuCXBPy0jOlBhxPqRWv6ND4c9PH1xjQ3NP nxJuMBS8rnNg22uyfAgmBKNLpLgAGVRMZGaGoJObGf72s6TeIqKJo/LtggAS9qAUiuKVnygo 3wjfkS9A3DRO+SpU7JqWdsveeIQyeyEJ/8PTowmSQLakF+3fote9ybzd880fSmFuIEJldWxp Y2ggPGpiZXVsaWNoQHN1c2UuY29tPsJgBBMRAgAgBQJZN5xEAhsDBgsJCAcDAgQVAggDBBYC AwECHgECF4AACgkQoDSui/t3IH4J+wCfQ5jHdEjCRHj23O/5ttg9r9OIruwAn3103WUITZee e7Sbg12UgcQ5lv7SzsFNBFk3nEQQCACCuTjCjFOUdi5Nm244F+78kLghRcin/awv+IrTcIWF hUpSs1Y91iQQ7KItirz5uwCPlwejSJDQJLIS+QtJHaXDXeV6NI0Uef1hP20+y8qydDiVkv6l IreXjTb7DvksRgJNvCkWtYnlS3mYvQ9NzS9PhyALWbXnH6sIJd2O9lKS1Mrfq+y0IXCP10eS FFGg+Av3IQeFatkJAyju0PPthyTqxSI4lZYuJVPknzgaeuJv/2NccrPvmeDg6Coe7ZIeQ8Yj t0ARxu2xytAkkLCel1Lz1WLmwLstV30g80nkgZf/wr+/BXJW/oIvRlonUkxv+IbBM3dX2OV8 AmRv1ySWPTP7AAMFB/9PQK/VtlNUJvg8GXj9ootzrteGfVZVVT4XBJkfwBcpC/XcPzldjv+3 HYudvpdNK3lLujXeA5fLOH+Z/G9WBc5pFVSMocI71I8bT8lIAzreg0WvkWg5V2WZsUMlnDL9 mpwIGFhlbM3gfDMs7MPMu8YQRFVdUvtSpaAs8OFfGQ0ia3LGZcjA6Ik2+xcqscEJzNH+qh8V m5jjp28yZgaqTaRbg3M/+MTbMpicpZuqF4rnB0AQD12/3BNWDR6bmh+EkYSMcEIpQmBM51qM EKYTQGybRCjpnKHGOxG0rfFY1085mBDZCH5Kx0cl0HVJuQKC+dV2ZY5AqjcKwAxpE75MLFkr wkkEGBECAAkFAlk3nEQCGwwACgkQoDSui/t3IH7nnwCfcJWUDUFKdCsBH/E5d+0ZnMQi+G0A nAuWpQkjM1ASeQwSHEeAWPgskBQL
  • Cc: Roger Pau Monné <roger.pau@xxxxxxxxxx>, Xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • Delivery-date: Mon, 01 Dec 2025 09:02:26 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 28.11.2025 18:47, Andrew Cooper wrote:
> While we do this for unknown user mode exits, crashing for supervisor mode
> exits is unhelpful.  Intel in particular expect the unknown case to be #UD
> because they do introduce new instructions with new VMEXIT_* codes without
> other enablement controls.  e.g. MSRLIST, USER_MSR, MSR_IMM, but AMD have
> RDPRU and SKINIT as examples too.

USER-MSR has MSR_USER_MSR_CTL, so doesn't fully fit here? (It's still not us
to directly control exposure of the insns, but an OS would need to use the
MSR to do so, and hence we need to properly handle writes to that MSR for
the respective exits to become possible.)

MSRLIST has a dedicated exec control, so whether the exits can occur is
under our control.

RDPRU and SKINIT have dedicated intercepts, without use of which the
respective exit can't occur aiui.

IOW it looks to be really only MSR-IMM which is special.

> @@ -3083,8 +3067,13 @@ void asmlinkage svm_vmexit_handler(void)
>          gprintk(XENLOG_ERR, "Unexpected vmexit: reason %#"PRIx64", "
>                  "exitinfo1 %#"PRIx64", exitinfo2 %#"PRIx64"\n",
>                  exit_reason, vmcb->exitinfo1, vmcb->exitinfo2);
> -    crash_or_fault:
> -        svm_crash_or_fault(v);
> +        fallthrough;
> +    case VMEXIT_MONITOR:
> +    case VMEXIT_MWAIT:
> +    case VMEXIT_SKINIT:
> +    case VMEXIT_RDPRU:
> +    inject_ud:
> +        hvm_inject_hw_exception(X86_EXC_UD, X86_EVENT_NO_EC);
>          break;
>      }
>  

Should this be brought more in line with respective VMX code (kept) below,
in never bypassing the gprintk() by any of the case labels? Basically
meaning that the case labels you move could simply be dropped for the time
being (or else, like the INVCPID one visible in context below, would want
re-inserting a few lines earlier).

Jan

> @@ -4902,14 +4903,9 @@ void asmlinkage vmx_vmexit_handler(struct 
> cpu_user_regs *regs)
>      case EXIT_REASON_INVPCID:
>      /* fall through */
>      default:
> -    exit_and_crash:
> +    unexpected_vmexit:
>          gprintk(XENLOG_ERR, "Unexpected vmexit: reason %lu\n", exit_reason);
> -
> -        if ( vmx_get_cpl() )
> -            hvm_inject_hw_exception(X86_EXC_UD,
> -                                    X86_EVENT_NO_EC);
> -        else
> -            domain_crash(v->domain);
> +        hvm_inject_hw_exception(X86_EXC_UD, X86_EVENT_NO_EC);
>          break;
>      }
>  
> 
> base-commit: 117a46287427db2a6f5fe219eb2031d7ca39b603




 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.