Re: [PATCH v5 07/12] x86/traps: Add #VE support for TDX guest

[Dave Hansen]

On 8/24/21 10:32 AM, Kuppuswamy, Sathyanarayanan wrote:
>>> +    if (fixup_exception(regs, X86_TRAP_VE, error_code, 0))
>>
>> There are exception table entries which can trigger a #VE?
> 
> It is required to handle #VE exceptions raised by unhandled MSR
> read/writes.

Is that really the *complete* set of reasons that a #VE can be triggered
from the kernel?

Just off the top of my head, I could imagine the kernel doing a
copy_{to,from}_user() which touched user-mapped MMIO causing a #VE.