Re: [PATCH v7 2/2] schemas: Add some common reserved-memory usages

[Simon Glass]

Hi,

On Tue, 28 Nov 2023 at 13:31, Chiu, Chasel <chasel.chiu@xxxxxxxxx> wrote:
>
>
>
>
> > -----Original Message-----
> > From: Ard Biesheuvel <ardb@xxxxxxxxxx>
> > Sent: Tuesday, November 28, 2023 10:08 AM
> > To: Chiu, Chasel <chasel.chiu@xxxxxxxxx>
> > Cc: Simon Glass <sjg@xxxxxxxxxxxx>; devicetree@xxxxxxxxxxxxxxx; Mark Rutland
> > <mark.rutland@xxxxxxx>; Rob Herring <robh@xxxxxxxxxx>; Tan, Lean Sheng
> > <sheng.tan@xxxxxxxxxxxxx>; lkml <linux-kernel@xxxxxxxxxxxxxxx>; Dhaval
> > Sharma <dhaval@xxxxxxxxxxxx>; Brune, Maximilian
> > <maximilian.brune@xxxxxxxxxxxxx>; Yunhui Cui <cuiyunhui@xxxxxxxxxxxxx>;
> > Dong, Guo <guo.dong@xxxxxxxxx>; Tom Rini <trini@xxxxxxxxxxxx>; ron minnich
> > <rminnich@xxxxxxxxx>; Guo, Gua <gua.guo@xxxxxxxxx>; linux-
> > acpi@xxxxxxxxxxxxxxx; U-Boot Mailing List <u-boot@xxxxxxxxxxxxx>
> > Subject: Re: [PATCH v7 2/2] schemas: Add some common reserved-memory
> > usages
> >
> > You are referring to a 2000 line patch so it is not 100% clear where to look tbh.
> >
> >
> > On Tue, 21 Nov 2023 at 19:37, Chiu, Chasel <chasel.chiu@xxxxxxxxx> wrote:
> > >
> > >
> > > In PR, UefiPayloadPkg/Library/FdtParserLib/FdtParserLib.c, line 268 is for
> > related example code.
> > >
> >
> > That refers to a 'memory-allocation' node, right? How does that relate to the
> > 'reserved-memory' node?
> >
> > And crucially, how does this clarify in which way "runtime-code" and "runtime-
> > data" reservations are being used?
> >
> > Since the very beginning of this discussion, I have been asking repeatedly for
> > examples that describe the wider context in which these reservations are used.
> > The "runtime" into runtime-code and runtime-data means that these regions have
> > a special significance to the operating system, not just to the next bootloader
> > stage. So I want to understand exactly why it is necessary to describe these
> > regions in a way where the operating system might be expected to interpret this
> > information and act upon it.
> >
>
>
> I think runtime code and data today are mainly for supporting UEFI runtime services - some BIOS functions for OS to utilize, OS may follow below ACPI spec to treat them as reserved range:
> https://uefi.org/specs/ACPI/6.5/15_System_Address_Map_Interfaces.html#uefi-memory-types-and-mapping-to-acpi-address-range-types
>
> Like I mentioned earlier, that PR is still in early phase and has not reflected all the required changes yet, but the idea is to build gEfiMemoryTypeInformationGuid HOB from FDT reserved-memory nodes.
> UEFI generic Payload has DxeMain integrated, however Memory Types are platform-specific, for example, some platforms may need bigger runtime memory for their implementation, that's why we want such FDT reserved-memory node to tell DxeMain.
>
> The Payload flow will be like this:
>   Payload creates built-in default MemoryTypes table ->
>     FDT reserved-memory node to override if required (this also ensures the same memory map cross boots so ACPI S4 works) ->
>       Build gEfiMemoryTypeInformationGuid HOB by "platfom specific" MemoryTypes Table ->
>         DxeMain/GCD to consume this MemoryTypes table and setup memory service ->
>           Install memory types table to UEFI system table.Configuration table...
>
> Note: if Payload built-in default MemoryTypes table works fine for the platform, then FDT reserved-memory node does not need to provide such 'usage' compatible strings. (optional)
> This FDT node could allow flexibility/compatibility without rebuilding Payload binary.
>
> Not sure if I answered all your questions, please highlight which area you need more information.

Any more thoughts on this? If not, I would like to see this patch
applied, please.

Regards,
Simon


>
> Thanks,
> Chasel
>
>
> >
> > >
> > > > -----Original Message-----
> > > > From: Chiu, Chasel
> > > > Sent: Tuesday, November 21, 2023 10:34 AM
> > > > To: Ard Biesheuvel <ardb@xxxxxxxxxx>; Simon Glass <sjg@xxxxxxxxxxxx>
> > > > Cc: devicetree@xxxxxxxxxxxxxxx; Mark Rutland <mark.rutland@xxxxxxx>;
> > > > Rob Herring <robh@xxxxxxxxxx>; Tan, Lean Sheng
> > > > <sheng.tan@xxxxxxxxxxxxx>; lkml <linux-kernel@xxxxxxxxxxxxxxx>;
> > > > Dhaval Sharma <dhaval@xxxxxxxxxxxx>; Brune, Maximilian
> > > > <maximilian.brune@xxxxxxxxxxxxx>; Yunhui Cui
> > > > <cuiyunhui@xxxxxxxxxxxxx>; Dong, Guo <guo.dong@xxxxxxxxx>; Tom Rini
> > > > <trini@xxxxxxxxxxxx>; ron minnich <rminnich@xxxxxxxxx>; Guo, Gua
> > > > <gua.guo@xxxxxxxxx>; linux-acpi@xxxxxxxxxxxxxxx; U-Boot Mailing List
> > > > <u- boot@xxxxxxxxxxxxx>; Chiu, Chasel <chasel.chiu@xxxxxxxxx>
> > > > Subject: RE: [PATCH v7 2/2] schemas: Add some common reserved-memory
> > > > usages
> > > >
> > > >
> > > > Hi Ard,
> > > >
> > > > Here is the POC PR for your reference:
> > > > https://github.com/tianocore/edk2/pull/4969/files#diff-
> > > >
> > ccebabae5274b21634723a2111ee0de11bed6cfe8cb206ef9e263d9c5f926a9cR26
> > > > 8
> > > > Please note that this PR is still in early phase and expected to
> > > > have significant changes.
> > > >
> > > > The idea is that payload entry will create
> > > > gEfiMemoryTypeInformationGuid HOB with payload default memory types
> > > > and allow FDT to override if correspond node present.
> > > > Please let me know if you have questions or suggestions.
> > > >
> > > > Thanks,
> > > > Chasel
> > > >
> > > >
> > > > > -----Original Message-----
> > > > > From: Ard Biesheuvel <ardb@xxxxxxxxxx>
> > > > > Sent: Tuesday, November 21, 2023 8:42 AM
> > > > > To: Simon Glass <sjg@xxxxxxxxxxxx>
> > > > > Cc: Chiu, Chasel <chasel.chiu@xxxxxxxxx>;
> > > > > devicetree@xxxxxxxxxxxxxxx; Mark Rutland <mark.rutland@xxxxxxx>;
> > > > > Rob Herring <robh@xxxxxxxxxx>; Tan, Lean Sheng
> > > > > <sheng.tan@xxxxxxxxxxxxx>; lkml <linux-kernel@xxxxxxxxxxxxxxx>;
> > > > > Dhaval Sharma <dhaval@xxxxxxxxxxxx>; Brune, Maximilian
> > > > > <maximilian.brune@xxxxxxxxxxxxx>; Yunhui Cui
> > > > > <cuiyunhui@xxxxxxxxxxxxx>; Dong, Guo <guo.dong@xxxxxxxxx>; Tom
> > > > > Rini <trini@xxxxxxxxxxxx>; ron minnich <rminnich@xxxxxxxxx>; Guo,
> > > > > Gua <gua.guo@xxxxxxxxx>; linux- acpi@xxxxxxxxxxxxxxx; U-Boot
> > > > > Mailing List <u-boot@xxxxxxxxxxxxx>
> > > > > Subject: Re: [PATCH v7 2/2] schemas: Add some common
> > > > > reserved-memory usages
> > > > >
> > > > > On Mon, 20 Nov 2023 at 21:12, Simon Glass <sjg@xxxxxxxxxxxx> wrote:
> > > > > >
> > > > > > Hi,
> > > > > >
> > > > > > On Mon, 13 Nov 2023 at 11:09, Chiu, Chasel <chasel.chiu@xxxxxxxxx>
> > wrote:
> > > > > > >
> > > > > > >
> > > > > > > Hi Ard,
> > > > > > >
> > > > > > > Please see my reply below inline.
> > > > > > >
> > > > > > > Thanks,
> > > > > > > Chasel
> > > > > > >
> > > > > > >
> > > > > > > > -----Original Message-----
> > > > > > > > From: Ard Biesheuvel <ardb@xxxxxxxxxx>
> > > > > > > > Sent: Saturday, November 11, 2023 3:04 AM
> > > > > > > > To: Chiu, Chasel <chasel.chiu@xxxxxxxxx>
> > > > > > > > Cc: Simon Glass <sjg@xxxxxxxxxxxx>;
> > > > > > > > devicetree@xxxxxxxxxxxxxxx; Mark Rutland
> > > > > > > > <mark.rutland@xxxxxxx>; Rob Herring <robh@xxxxxxxxxx>; Tan,
> > > > > > > > Lean Sheng <sheng.tan@xxxxxxxxxxxxx>; lkml
> > > > > > > > <linux-kernel@xxxxxxxxxxxxxxx>; Dhaval Sharma
> > > > > > > > <dhaval@xxxxxxxxxxxx>; Brune, Maximilian
> > > > > > > > <maximilian.brune@xxxxxxxxxxxxx>; Yunhui Cui
> > > > > > > > <cuiyunhui@xxxxxxxxxxxxx>; Dong, Guo <guo.dong@xxxxxxxxx>;
> > > > > > > > Tom Rini <trini@xxxxxxxxxxxx>; ron minnich
> > > > > > > > <rminnich@xxxxxxxxx>; Guo, Gua <gua.guo@xxxxxxxxx>; linux-
> > > > > > > > acpi@xxxxxxxxxxxxxxx; U-Boot Mailing List
> > > > > > > > <u-boot@xxxxxxxxxxxxx>
> > > > > > > > Subject: Re: [PATCH v7 2/2] schemas: Add some common
> > > > > > > > reserved-memory usages
> > > > > > > >
> > > > > > > > On Sat, 11 Nov 2023 at 04:20, Chiu, Chasel
> > > > > > > > <chasel.chiu@xxxxxxxxx>
> > > > wrote:
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > Just sharing some usage examples from UEFI/EDK2 scenario.
> > > > > > > > > To support ACPI S4/Hibernation, memory map must be
> > > > > > > > > consistent before entering and after resuming from S4, in
> > > > > > > > > this case payload may need to know previous memory map
> > > > > > > > > from bootloader (currently generic payload cannot access
> > > > > > > > > platform/bootloader specific non-volatile data, thus could
> > > > > > > > > not save/restore memory map
> > > > > > > > > information)
> > > > > > > >
> > > > > > > > So how would EDK2 reconstruct the entire EFI memory map from
> > > > > > > > just these unannotated /reserved-memory nodes? The EFI
> > > > > > > > memory map contains much more information than that, and all
> > > > > > > > of it has to match the pre-hibernate situation, right? Can you given an
> > example?
> > > > > > >
> > > > > > >
> > > > > > > Here we listed only typically memory types that may change
> > > > > > > cross different
> > > > > platforms.
> > > > > > > Reserved memory type already can be handled by reserved-memory
> > > > > > > node,
> > > > > and rest of the types usually no need to change cross platforms
> > > > > thus currently we could rely on default in generic payload.
> > > > > > > In the future if we see a need to add new memory types we will
> > > > > > > discuss and
> > > > > add it to FDT schema.
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > >
> > > > > > > > > Another usage is to support binary model which generic
> > > > > > > > > payload is a prebuilt
> > > > > > > > binary compatible for all platforms/configurations, however
> > > > > > > > the payload default memory map might not always work for all
> > > > > > > > the configurations and we want to allow bootloader to
> > > > > > > > override payload default
> > > > > memory map without recompiling.
> > > > > > > > >
> > > > > > > >
> > > > > > > > Agreed. But can you explain how a EDK2 payload might make
> > > > > > > > meaningful use of 'runtime-code' regions provided via DT  by
> > > > > > > > the
> > > > > > > > non-EDK2 platform init? Can you give an example?
> > > > > > >
> > > > > > >
> > > > > > > Runtime-code/data is used by UEFI payload for booting UEFI OS
> > > > > > > which
> > > > > required UEFI runtime services.
> > > > > > > Platform Init will select some regions from the usable memory
> > > > > > > and assign it to
> > > > > runtime-code/data for UPL to consume. Or assign same
> > > > > runtime-code/data from previous boot.
> > > > > > > If UEFI OS is not supported, PlatformInit may not need to
> > > > > > > provide runtime-code/data regions to payload. (always
> > > > > > > providing runtime-code/data should be supported too)
> > > > > > >
> > > > > > >
> > > > > > > >
> > > > > > > > > Under below assumption:
> > > > > > > > >         FDT OS impact has been evaluated and taken care by
> > > > > > > > > relevant
> > > > > > > > experts/stakeholders.
> > > > > > > > > Reviewed-by: Chasel Chiu <chasel.chiu@xxxxxxxxx>
> > > > > > > > >
> > > > > > > >
> > > > > > > > I am sorry but I don't know what 'FDT OS impact' means. We
> > > > > > > > are talking about a firmware-to-firmware abstraction that
> > > > > > > > has the potential to leak into the OS visible interface.
> > > > > > > >
> > > > > > > > I am a maintainer in the Tianocore project myself, so it
> > > > > > > > would help if you could explain who these relevant experts
> > > > > > > > and stakeholders are. Was this discussed on the edk2-devel
> > > > > > > > mailing list? If so, apologies for missing it but I may not have been cc'ed
> > perhaps?
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > I'm not familiar with FDT OS, also I do not know if who from
> > > > > > > edk2-devel were
> > > > > supporting FDT OS, I think Simon might be able to connect FDT OS
> > > > > experts/stakeholders.
> > > > > > > We are mostly focusing on payload firmware phase
> > > > > > > implementation in
> > > > > > > edk2 (and other payloads too), however, since we have aligned
> > > > > > > the payload FDT and OS FDT months ago, I'm assuming FDT OS
> > > > > > > impact must be there and we need (or already done?) FDT OS
> > > > > > > experts to support it. (again, maybe Simon could share more
> > > > > > > information about FDT OS)
> > > > > > >
> > > > > > > In edk2 such FDT schema is UefiPayloadPkg internal usage only
> > > > > > > and payload
> > > > > entry will convert FDT into HOB thus we expected the most of the
> > > > > edk2 generic code are no-touch/no impact, that's why we only had
> > > > > small group
> > > > > (UefiPayloadPkg) discussion.
> > > > > > > Ard, if you are aware of any edk2 code that's for supporting
> > > > > > > FDT OS, please let
> > > > > us know and we can discuss if those code were impacted or not.
> > > > > >
> > > > > > We discussed this and just to clarify, 'FDT OS' is not a special
> > > > > > OS, it is just Linux.
> > > > > >
> > > > > > So, with the above, are we all on the same page? Can the patch
> > > > > > be applied, perhaps? If not, what other discussion is needed?
> > > > > >
> > > > >
> > > > > An example of how a platform-init/payload combination would make
> > > > > meaningful use of such runtime-code/data regions.