Re: [PATCH v3 3/3] x86: Add Kconfig option to require NX bit support

[Jan Beulich <jbeulich@xxxxxxxx>]

On 18.07.2023 15:19, Jan Beulich wrote:
> On 29.06.2023 14:17, Alejandro Vallejo wrote:
>> --- a/xen/arch/x86/boot/head.S
>> +++ b/xen/arch/x86/boot/head.S
>> @@ -123,6 +123,7 @@ multiboot2_header:
>>  .Lbad_ldr_nih: .asciz "ERR: EFI ImageHandle is not provided by bootloader!"
>>  .Lbad_efi_msg: .asciz "ERR: EFI IA-32 platforms are not supported!"
>>  .Lbag_alg_msg: .asciz "ERR: Xen must be loaded at a 2Mb boundary!"
>> +.Lno_nx_msg:   .asciz "ERR: Not an NX-capable CPU!"
>>  
>>          .section .init.data, "aw", @progbits
>>          .align 4
>> @@ -153,6 +154,11 @@ early_error: /* Here to improve the disassembly. */
>>  .Lnot_aligned:
>>          add     $sym_offs(.Lbag_alg_msg), %esi
>>          jmp     .Lget_vtb
>> +#ifdef CONFIG_REQUIRE_NX
>> +.Lno_nx:
>> +        add     $sym_offs(.Lno_nx_msg), %esi
>> +        jmp     .Lget_vtb
>> +#endif
> 
> Since I'm in the process of introducing more such paths (for the x86-64-v<N>
> series), I'm curious: Have you actually had success with getting any output
> from this code path? I see unreadable output come through serial (provided
> it's the normal com1 I/O port location where the serial port is), which
> likely is because baud rate wasn't configured yet, and hence I might have
> success by changing the config of the receiving side. And I see nothing at
> all on the screen. While kind of expected when in graphics mode, I wonder
> whether this ever worked, or whether this has simply bitrotted because of
> never actually coming into play.

Pretty clearly this was broken in the course of adding MB2 support, by
b28044226e1c using %esi as the "relocation base" after already having
clobbered it. I'm working on a fix.

Jan