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

Re: [PATCH] x86emul: convert op_bytes/opc checks in SIMD emulation


  • To: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • From: Jan Beulich <jbeulich@xxxxxxxx>
  • Date: Wed, 21 Aug 2024 17:03:35 +0200
  • 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@xxxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • Delivery-date: Wed, 21 Aug 2024 15:03:46 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 21.08.2024 15:37, Andrew Cooper wrote:
> On 21/08/2024 2:30 pm, Jan Beulich wrote:
>> Delivering #UD for an internal shortcoming of the emulator isn't quite
>> right. Similarly BUG() is bigger a hammer than needed.
>>
>> Switch to using EXPECT() instead.
>>
>> Suggested-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> To confirm, this is ASSERT_UNREACHABLE() (which fuzzing will now notice
> as an error), and unhandleable in release builds (which ultimately ends
> up in #UD)?

Yes, at least mostly. What exactly UNHANDLEABLE converts to depends on
the use site, I think.

> I think it would be helpful to at least note the fuzzing aspect in the
> commit message.

I've added "This way even for insns not covered by the test harness
fuzzers will have a chance of noticing issues, should any still exist
or new ones be introduced" to the 2nd paragraph.

>> --- a/xen/arch/x86/x86_emulate/x86_emulate.c
>> +++ b/xen/arch/x86/x86_emulate/x86_emulate.c
>> @@ -8114,13 +8114,13 @@ x86_emulate(
>>      }
>>      else if ( state->simd_size != simd_none )
>>      {
>> -        generate_exception_if(!op_bytes, X86_EXC_UD);
>>          generate_exception_if((vex.opcx && (d & TwoOp) &&
>>                                 (vex.reg != 0xf || (evex_encoded() && 
>> !evex.RX))),
>>                                X86_EXC_UD);
>>  
>> -        if ( !opc )
>> -            BUG();
>> +        EXPECT(op_bytes);
>> +        EXPECT(opc);
> 
> This is the only BUG() in x86_emulate.c, and it's right to get rid of it
> IMO.
> 
> Therefore, we should have a hunk removing it from
> tools/tests/x86_emulator/x86-emulate.h too, which will prevent
> reintroduction.
> 
> Maybe even undef BUG somewhere in x86_emulate/private.h?

Both of these actions can only be taken if the other BUG() in decode.c
also goes away. But yes, what you suggest is probably the best course of
action. I guess I'll do that in yet another patch, though.

Jan



 


Rackspace

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