Re: [PATCH v3 3/5] MdePkg/DxeServicesLib: introduce AllocatePeiAccessiblePages routine

["Yao, Jiewen" <jiewen.yao@intel.com>]

Thanks.
I guess we need it even for X64 PEI/X64 DXE case, because a platform may only allocates 4G even in X64 PEI.

BTW, I don't mind if you want to use a separate patch to handle capsule case.
Current API is good enough and I do not want to block the check-in. :-)

I like this simple solution.
Patch series: Reviewed-by: Jiewen.Yao@intel.com

Thank you
Yao Jiewen

> -----Original Message-----
> From: edk2-devel [mailto:edk2-devel-bounces@lists.01.org] On Behalf Of Ard
> Biesheuvel
> Sent: Tuesday, May 29, 2018 9:20 AM
> To: Kinney, Michael D <michael.d.kinney@intel.com>
> Cc: edk2-devel@lists.01.org; Yao, Jiewen <jiewen.yao@intel.com>; Zeng, Star
> <star.zeng@intel.com>
> Subject: Re: [edk2] [PATCH v3 3/5] MdePkg/DxeServicesLib: introduce
> AllocatePeiAccessiblePages routine
> 
> On 29 May 2018 at 18:09, Kinney, Michael D <michael.d.kinney@intel.com>
> wrote:
> > Jiewen,
> >
> > I see what you mean.  It is not the submitting of
> > capsules you are referring to.  It is the processing
> > if capsules in the PEI phase that depends on some
> > things being setup in DXE phase and DXE phase needs
> > to know if PEI is in IA32 mode.
> >
> > So I agree that the commit message could add capsule
> > processing to the list of features that can use this
> > new service.
> >
> > The logic in that file is using PcdDxeIplSwitchToLongMode.
> > That PCD is TRUE when PEI is IA32 and DXE is X64.  But it
> > is ignored when  PEI and DXE are both X64 and could be
> > TRUE or FALSE.  Is there a logic issue here with using that
> > PCD when PEI and DXE are both X64?
> >
> 
> Hi all,
> 
> The reason I disregarded this particular case in my series is that
> this code is already specific to X64. However, given the discussion
> you are having, I guess the logic of looking at EfiFreeMemoryTop [at
> runtime] combined with the various [build time] PCD values could help
> to refine this case as well.
> _______________________________________________
> edk2-devel mailing list
> edk2-devel@lists.01.org
> https://lists.01.org/mailman/listinfo/edk2-devel