Re: [PATCH v3] x86/asm: Force native_apic_mem_read to use mov

[Thomas Gleixner]

On Tue, Feb 06 2024 at 14:36, Adam Dunlap wrote:

Can you please use foo() as notation for functions all over the place
including the subject line, which also wants s/mov/the MOV instruction/
and use MOV instead of mov.

> When done from a virtual machine, instructions that touch APIC memory
> must be emulated. By convention, MMIO access are typically performed via
> io.h helpers such as 'readl()' or 'writeq()' to simplify instruction
> emulation/decoding (ex: in KVM hosts and SEV guests) [0].
>
> Currently, native_apic_mem_read does not follow this convention,
> allowing the compiler to emit instructions other than the mov generated
> by readl(). In particular, when compiled with clang and run as a SEV-ES
> or SEV-SNP guest, the compiler would emit a testl instruction which is
> not supported by the SEV-ES emulator, causing a boot failure in that
> environment. It is likely the same problem would happen in a TDX guest
> as that uses the same instruction emulator as SEV-ES.
>
> To make sure all emulators can emulate APIC memory reads via mov, use
> the readl function in native_apic_mem_read. It is expected that any
> emulator would support mov in any addressing mode it is the most generic
> and is what is ususally emitted currently.
>
> The testl instruction is emitted when native_apic_mem_read
> is inlined into __xapic_wait_icr_idle. The emulator comes from
> insn_decode_mmio in arch/x86/lib/insn-eval.c. It would not be worth it

s/It would/It's/

Either it's a fact or not.

> to extend insn_decode_mmio to support more instructions since, in
> theory, the compiler could choose to output nearly any instruction for
> such reads which would bloat the emulator beyond reason.
>
> An alterative to this approach would be to use inline assembly instead
> of the readl helper, as that is what native_apic_mem_write does. I
> consider using readl to be cleaner since it is documented to be a simple
> wrapper and inline assembly is less readable. native_apic_mem_write
> cannot be trivially updated to use writel since it appears to use custom
> asm to workaround for a processor-specific bug.

How is this paragraph relevant?

> [0] https://lore.kernel.org/all/20220405232939.73860-12-kirill.shutemov@xxxxxxxxxxxxxxx/
>
> Signed-off-by: Adam Dunlap <acdunlap@xxxxxxxxxx>
> Tested-by: Kevin Loughlin <kevinloughlin@xxxxxxxxxx>

Other than the above nit picks:

Reviewed-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>