Re: [edk2-devel] [PATCH] UefiCpuPkg/CpuDxe: Fix boot error

["Laszlo Ersek" <lersek@redhat.com>]

On 12/02/20 22:38, Guo Dong wrote:
> REF: https://bugzilla.tianocore.org/show_bug.cgi?id=3084
> 
> When DXE drivers are dispatched above 4GB memory and
> the system is already in 64bit mode, the address
> setCodeSelectorLongJump in stack will be override
> by parameter. so change to use 64bit address and
> jump to qword address.
> 
> Signed-off-by: Guo Dong <guo.dong@intel.com>
> ---
>  UefiCpuPkg/CpuDxe/X64/CpuAsm.nasm | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/UefiCpuPkg/CpuDxe/X64/CpuAsm.nasm b/UefiCpuPkg/CpuDxe/X64/CpuAsm.nasm
> index c3489bcc3e..6ad32b49f4 100644
> --- a/UefiCpuPkg/CpuDxe/X64/CpuAsm.nasm
> +++ b/UefiCpuPkg/CpuDxe/X64/CpuAsm.nasm
> @@ -23,8 +23,8 @@ ASM_PFX(SetCodeSelector):
>      sub     rsp, 0x10
>      lea     rax, [setCodeSelectorLongJump]
>      mov     [rsp], rax

I'll let Ray and Eric review this patch, just one question for my
understanding:

for clarity, shouldn't the above MOV have referred to "eax" in the first
place? Because (before this patch) it seems to store 8 bytes (from rax),
and then to overwrite the high-address DWORD of those 8 bytes, with the
next operation (which stores 2 bytes from CX).

IMO, for clarity's sake, the original code should have used EAX. Why
write 8 bytes when only the low-address 4 bytes matter anyway?

Anyhow, this is no longer relevant, because of the present patch; I just
wanted to ask about it so I understand better.

Thanks
Laszlo

> -    mov     [rsp+4], cx
> -    jmp     dword far [rsp]
> +    mov     [rsp+8], cx
> +    jmp     qword far [rsp]
>  setCodeSelectorLongJump:
>      add     rsp, 0x10
>      ret
>