Hi Gerd,
For the PCD, it current has below usage.
1) for 32-bit PEI and 64-bit DXE, this PCD will decide if use 5 level paging in 
DXE.
2) for 64-bit PEI and DXE, reset vector chooses if use 5 level paging in early 
phase. But BIOS can still switch paging mode based on this PCD later.
Since this PCD can be dynamic, we can let user to choose different paging mode 
in BIOS setup menu, and use this paging mode on reboot.

You may want to use one BIOS to support machine with different la57 capability, 
I assume two possible ways:
1) make the PCD as dynamic, and set it based on la57 capability.
2) Remove the ASSERT in DxeIpl

Both are fine for me.

Thanks
Zhiguang

> -----Original Message-----
> From: devel@edk2.groups.io <devel@edk2.groups.io> On Behalf Of Gerd
> Hoffmann
> Sent: Thursday, January 25, 2024 8:17 PM
> To: Ni, Ray <ray...@intel.com>
> Cc: devel@edk2.groups.io; Liming Gao <gaolim...@byosoft.com.cn>; László
> Érsek <ler...@redhat.com>; Oliver Steffen <ostef...@redhat.com>
> Subject: Re: [edk2-devel] [PATCH v2 1/2] MdeModulePkg: rename
> PcdUse5LevelPageTable to PcdEnable5LevelPageTable
> 
> On Thu, Jan 25, 2024 at 12:01:46PM +0000, Ni, Ray wrote:
> > Rename an existing PCD might break lots of platform builds.
> >
> > When 5-level paging capability was added to ResetVector, I also
> > considered to remove the PcdUse5LevelPageTable reference from C code.
> >
> > Let me think about it...
> 
> Purging this completely from the source code is fine with me too.
> 
> From OVMF point of view this is not needed, all backward compatibility
> concerns for old guests without 5-level paging support can be solved by just
> removing la57 capability from the vCPU (qemu -cpu host,la57=off).
> 
> take care,
>   Gerd
> 
> 
> 
> 
> 



-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.
View/Reply Online (#114491): https://edk2.groups.io/g/devel/message/114491
Mute This Topic: https://groups.io/mt/103950404/21656
Group Owner: devel+ow...@edk2.groups.io
Unsubscribe: https://edk2.groups.io/g/devel/unsub [arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-


Reply via email to