Re: [PATCH v3 1/4] x86/syscall: sign-extend system calls on entry to int

From: H. Peter Anvin
Date: Mon May 17 2021 - 16:24:42 EST

No, this is what the Linux kernel did for a very long time.

On May 16, 2021 12:48:19 AM PDT, Ingo Molnar <mingo@xxxxxxxxxx> wrote:
>* H. Peter Anvin <hpa@xxxxxxxxx> wrote:
>> This is an ABI change, but is in fact a revert to the original x86-64
>> ABI. The original assembly entry code would zero-extend the system
>> call number; this patch uses sign extend to be explicit that this is
>> treated as a signed number (although in practice it makes no
>> difference, of course) and to avoid people getting the idea of
>> "optimizing" it, as has happened on at least two(!) separate
>> occasions.
>The original x86-64 ABI as documented by AMD, as we (probably) never
>this in Linux, right?
>Sounds sensible to do this, assuming nothing relies on the weirdness.
> Ingo

Sent from my Android device with K-9 Mail. Please excuse my brevity.