- TF-A - lists.trustedfirmware.org

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Julius Werner

Just want to point out that TF-A currently already supports a (very simple) mechanism like this: https://review.trustedfirmware.org/plugins/gitiles/TF-A/trusted-firmware-a/… https://review.trustedfirmware.org/plugins/gitiles/TF-A/trusted-firmware-a/… https://review.trustedfirmware.org/plugins/gitiles/TF-A/trusted-firmware-a/… It's just a linked list of tagged elements. The tag space is split into TF-A-wide generic tags and SiP-specific tags (with plenty of room to spare if more areas need to be defined -- a 64-bit tag can fit a lot). This is currently being used by some platforms that run coreboot in place of BL1/BL2, to pass information from coreboot (BL2) to BL31. I would echo Simon's sentiment of keeping this as simple as possible and avoiding complicated and bloated data structures with UUIDs. You usually want to parse something like this as early as possible in the passed-to firmware stage, particularly if the structure encodes information about the debug console (like it does for the platforms I mentioned above). For example, in BL31 this basically means doing it right after moving from assembly to C in bl31_early_platform_setup2() to get the console up before running anything else. At that point in the BL31 initialization, the MMU and caches are disabled, so data accesses are pretty expensive and you don't want to spend a lot of parsing effort or calculate complicated checksums or the like. You just want something extremely simple where you ideally have to touch every data word only once. On Wed, Mar 24, 2021 at 5:06 PM Simon Glass via TF-A < tf-a(a)lists.trustedfirmware.org> wrote: > Hi Harb, > > On Wed, 24 Mar 2021 at 11:39, Harb Abdulhamid OS < > abdulhamid(a)os.amperecomputing.com> wrote: > >> Hello Folks, >> >> Appreciate the feedback and replies on this. Glad to see that there is >> interest in this topic. 😊 >> >> >> >> I try to address the comments/feedback from Francois and Simon below…. >> >> >> >> @François Ozog <francois.ozog(a)linaro.org> – happy to discuss this on a >> zoom call. I will make that time slot work, and will be available to >> attend April 8, 4pm CT. >> >> >> >> Note that I’m using the term “HOB” here more generically, as there are >> typically vendor specific structures beyond the resource descriptor HOB, >> which provides only a small subset of the information that needs to be >> passed between the boot phases. >> >> >> >> The whole point here is to provide mechanism to develop firmware that we >> can build ARM Server SoC’s that support **any** BL33 payload (e.g. EDK2, >> AptioV, CoreBoot, and maybe even directly boot strapping LinuxBoot at some >> point). In other-words, we are trying to come up with a TF-A that would >> be completely agnostic to the implementation of BL33 (i.e. BL33 is built >> completely independently by a separate entity – e.g. an ODM/OEM). >> >> >> >> Keep in mind, in the server/datacenter market segment we are not building >> vertically integrated systems with a single entity compiling >> firmware/software stacks like most folks in TF-A have become use to. There >> are two categories of higher level firmware code blobs in the >> server/datacenter model: >> >> 1. “SoC” or “silicon” firmware – in TF-A this may map to BL1, BL2, >> BL31, and **possibly** one or more BL32 instances >> 2. “Platform” or “board” firmware – in TF-A this may map to BL33 and * >> *possibly** one or more BL32 instances. >> >> >> >> Even the platform firmware stack could be further fragmented by having >> multiple entities involved in delivering the entire firmware stack: IBVs, >> ODMs, OEMs, CSPs, and possibly even device vendor code. >> >> >> >> To support a broad range of platform designs with a broad range of memory >> devices, we need a crisp and clear contract between the SoC firmware that >> initializes memory (e.g. BL2) and how that platform boot firmware (e.g. >> BL33) gathers information about what memory that was initialized, at what >> speeds, NUMA topology, and many other relevant information that needs to be >> known and comprehended by the platform firmware and eventually by the >> platform software. >> >> >> >> I understand the versatility of DT, but I see two major problems with DT: >> >> - DT requires more complicated parsing to get properties, and even >> more complex to dynamically set properties – this HOB structures may need >> to be generated in boot phases where DDR is not available, and therefore we >> will be extremely memory constrained. >> - DT is probably overkill for this purpose – We really just want a >> list of pointers to simple C structures that code cast (e.g. JEDEC SPD data >> blob) >> >> >> >> I think that we should not mix the efforts around DT/ACPI specs with what >> we are doing here, because those specs and concepts were developed for a >> completely different purpose (i.e. abstractions needed for OS / RTOS >> software, and not necessarily suitable for firmware-to-firmware hand-offs). >> >> >> >> Frankly, I would personally push back pretty hard on defining SMC’s for >> something that should be one way information passing. Every SMC we add is >> another attack vector to the secure world and an increased burden on the >> folks that have to do security auditing and threat analysis. I see no >> benefit in exposing these boot/HOB/BOB structures at run-time via SMC >> calls. >> >> >> >> Please do let me know if you disagree and why. Look forward to >> discussing on this thread or on the call. >> >> >> >> @Simon Glass <sjg(a)chromium.org> - Thanks for the pointer to >> bloblist. I briefly reviewed and it seems like a good baseline for what >> we may be looking for. >> >> >> >> That being said, I would say that there is some benefit in having some >> kind of unique identifiers (e.g. UUID or some unique signature) so that we >> can tie standardized data structures (based on some future TBD specs) to a >> particular ID. For example, if the TPM driver in BL33 is looking for the >> TPM structure in the HOB/BOB list, and may not care about the other data >> blobs. The driver needs a way to identify and locate the blob it cares >> about. >> > > The tag is intended to serve that purpose, although perhaps it should > switch from an auto-allocating enum to one with explicit values for each > entry and a range for 'local' use. > >> >> >> I guess we can achieve this with the tag, but the problem with tag when >> you have eco-system with a lot of parties doing parallel development, you >> can end up with tag collisions and folks fighting about who has rights to >> what tag values. We would need some official process for folks to register >> tags for whatever new structures we define, or maybe some tag range for >> vendor specific structures. This comes with a lot of pain and >> bureaucracy. On the other hand, UUID has been a proven way to make it easy >> to just define your own blobs with **either** standard or vendor >> specific structures without worry of ID collisions between vendors. >> > > True. I think the pain is overstated, though. In this case I think we > actually want something that can be shared between projects and orgs, so > some amount of coordination could be considered a benefit. It could just be > a github pull request. I find the UUID unfriendly and not just to code size > and eyesight! Trying to discover what GUIDs mean or are valid is quite > tricky. E.g. see this code: > > #define FSP_HOB_RESOURCE_OWNER_TSEG_GUID \ > EFI_GUID(0xd038747c, 0xd00c, 0x4980, \ > 0xb3, 0x19, 0x49, 0x01, 0x99, 0xa4, 0x7d, 0x55) > (etc.) > > static struct guid_name { > efi_guid_t guid; > const char *name; > } guid_name[] = { > { FSP_HOB_RESOURCE_OWNER_TSEG_GUID, "TSEG" }, > { FSP_HOB_RESOURCE_OWNER_FSP_GUID, "FSP" }, > { FSP_HOB_RESOURCE_OWNER_SMM_PEI_SMRAM_GUID, "SMM PEI SMRAM" }, > { FSP_NON_VOLATILE_STORAGE_HOB_GUID, "NVS" }, > { FSP_VARIABLE_NV_DATA_HOB_GUID, "Variable NVS" }, > { FSP_GRAPHICS_INFO_HOB_GUID, "Graphics info" }, > { FSP_HOB_RESOURCE_OWNER_PCD_DATABASE_GUID1, "PCD database ea" }, > { FSP_HOB_RESOURCE_OWNER_PCD_DATABASE_GUID2, "PCD database 9b" }, > (never figured out what those two are) > > { FSP_HOB_RESOURCE_OWNER_PEIM_DXE_GUID, "PEIM Init DXE" }, > { FSP_HOB_RESOURCE_OWNER_ALLOC_STACK_GUID, "Alloc stack" }, > { FSP_HOB_RESOURCE_OWNER_SMBIOS_MEMORY_GUID, "SMBIOS memory" }, > { {}, "zero-guid" }, > {} > }; > > static const char *guid_to_name(const efi_guid_t *guid) > { > struct guid_name *entry; > > for (entry = guid_name; entry->name; entry++) { > if (!guidcmp(guid, &entry->guid)) > return entry->name; > } > > return NULL; > } > > Believe it or not it took a fair bit of effort to find just that small > list, with nearly every one in a separate doc, from memory. > > >> >> We can probably debate whether there is any value in GUID/UUID or not >> during the call… but again, boblist seems like a reasonable starting point >> as an alternative to HOB. >> > > Indeed. There is certainly value in both approaches. > > Regards, > Simon > > >> >> >> Thanks, >> >> --Harb >> >> >> >> *From:* François Ozog <francois.ozog(a)linaro.org> >> *Sent:* Tuesday, March 23, 2021 10:00 AM >> *To:* François Ozog <francois.ozog(a)linaro.org>; Ron Minnich < >> rminnich(a)google.com>; Paul Isaac's <paul.isaacs(a)linaro.org> >> *Cc:* Simon Glass <sjg(a)chromium.org>; Harb Abdulhamid OS < >> abdulhamid(a)os.amperecomputing.com>; Boot Architecture Mailman List < >> boot-architecture(a)lists.linaro.org>; tf-a(a)lists.trustedfirmware.org >> *Subject:* Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for >> information passing between boot stages >> >> >> >> +Ron Minnich <rminnich(a)google.com> +Paul Isaac's <paul.isaacs(a)linaro.org> >> >> >> >> >> Adding Ron and Paul because I think this interface should be also >> benefiting LinuxBoot efforts. >> >> >> >> On Tue, 23 Mar 2021 at 11:17, François Ozog via TF-A < >> tf-a(a)lists.trustedfirmware.org> wrote: >> >> Hi, >> >> >> >> I propose we cover the topic at the next Trusted Substrate >> <https://collaborate.linaro.org/display/TS/Trusted+Substrate+Home> zoom >> call <https://linaro-org.zoom.us/j/94563644892> on April 8th 4pm CET. >> >> >> >> The agenda: >> >> ABI between non-secure firmware and the rest of firmware (EL3, S-EL1, >> S-EL2, SCP) to adapt hardware description to some runtime conditions. >> >> runtime conditions here relates to DRAM size and topology detection, >> secure DRAM memory carvings, PSCI and SCMI interface publishing. >> >> >> >> For additional background on existing metadata: UEFI Platform >> Initialization Specification Version 1.7 >> <https://uefi.org/sites/default/files/resources/PI_Spec_1_7_final_Jan_2019.p…> >> , 5.5 Resource Descriptor HOB >> >> Out of the ResourceType we care about is EFI_RESOURCE_SYSTEM_MEMORY. >> >> This HOB lacks memory NUMA attachment or something that could be related >> to fill SRAT table for ACPI or relevant DT proximity domains. >> >> HOB is not consistent accros platforms: some platforms (Arm) lists memory >> from the booting NUMA node, other platforms (x86) lists all memory from all >> NUMA nodes. (At least this is the case on the two platforms I tested). >> >> >> >> There are two proposals to use memory structures from SPL/BLx up to the >> handover function (as defined in the Device Tree technical report >> <https://docs.google.com/document/d/1CLkhLRaz_zcCq44DLGmPZQFPbYHOC6nzPowaL0X…>) >> which can be U-boot (BL33 or just U-Boot in case of SPL/U-Boot scheme) or >> EDK2. >> >> I would propose we also discuss possibility of FF-A interface to actually >> query information or request actions to be done (this is a model actually >> used in some SoCs with proprietary SMC calls). >> >> >> >> Requirements (to be validated): >> >> - ACPI and DT hardware descriptions. >> >> - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2) >> >> - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2, >> TF-A/LinuxBoot) >> >> - at least allows complete DRAM description and "persistent" usage >> (reserved areas for secure world or other usages) >> >> - support secure world device assignment >> >> >> >> Cheers >> >> >> >> FF >> >> >> >> >> >> On Mon, 22 Mar 2021 at 19:56, Simon Glass <sjg(a)chromium.org> wrote: >> >> Hi, >> >> Can I suggest using bloblist for this instead? It is lightweight, >> easier to parse, doesn't have GUIDs and is already used within U-Boot >> for passing info between SPL/U-Boot, etc. >> >> Docs here: >> https://github.com/u-boot/u-boot/blob/master/doc/README.bloblist >> Header file describes the format: >> https://github.com/u-boot/u-boot/blob/master/include/bloblist.h >> >> Full set of unit tests: >> https://github.com/u-boot/u-boot/blob/master/test/bloblist.c >> >> Regards, >> Simon >> >> On Mon, 22 Mar 2021 at 23:58, François Ozog <francois.ozog(a)linaro.org> >> wrote: >> > >> > +Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org> >> > >> > standardization is very much welcomed here and need to accommodate a >> very >> > diverse set of situations. >> > For example, TEE OS may need to pass memory reservations to BL33 or >> > "capture" a device for the secure world. >> > >> > I have observed a number of architectures: >> > 1) pass information from BLx to BLy in the form of a specific object >> > 2) BLx called by BLy by a platform specific SMC to get information >> > 3) BLx called by BLy by a platform specific SMC to perform Device Tree >> > fixups >> > >> > I also imagined a standardized "broadcast" FF-A call so that any >> firmware >> > element can either provide information or "do something". >> > >> > My understanding of your proposal is about standardizing on >> architecture 1) >> > with the HOB format. >> > >> > The advantage of the HOB is simplicity but it may be difficult to >> implement >> > schemes such as pruning a DT because device assignment in the secure >> world. >> > >> > In any case, it looks feasible to have TF-A and OP-TEE complement the >> list >> > of HOBs to pass information downstream (the bootflow). >> > >> > It would be good to start with building the comprehensive list of >> > information that need to be conveyed between firmware elements: >> > >> > information. | authoritative entity | reporting entity | information >> > exchanged: >> > dram | TFA | TFA | >> > <format to be detailed, NUMA topology to build the SRAT table or DT >> > equivalent?> >> > PSCI | SCP | TFA? | >> > SCMI | SCP or TEE-OS | TFA? TEE-OS?| >> > secure SRAM | TFA. | TFA. | >> > secure DRAM | TFA? TEE-OS? | TFA? TEE-OS? | >> > other? | | >> > | >> > >> > Cheers >> > >> > FF >> > >> > >> > On Mon, 22 Mar 2021 at 09:34, Harb Abdulhamid OS via TF-A < >> > tf-a(a)lists.trustedfirmware.org> wrote: >> > >> > > Hello Folks, >> > > >> > > >> > > >> > > I'm emailing to start an open discussion about the adoption of a >> concept >> > > known as "hand-off blocks" or HOB to become a part of the TF-A >> Firmware >> > > Framework Architecture (FFA). This is something that is a pretty >> major >> > > pain point when it comes to the adoption of TF-A in ARM Server SoC’s >> > > designed to enable a broad range of highly configurable datacenter >> > > platforms. >> > > >> > > >> > > >> > > >> > > >> > > What is a HOB (Background)? >> > > >> > > --------------------------- >> > > >> > > UEFI PI spec describes a particular definition for how HOB may be >> used for >> > > transitioning between the PEI and DXE boot phases, which is a good >> > > reference point for this discussion, but not necessarily the exact >> solution >> > > appropriate for TF-A. >> > > >> > > >> > > >> > > A HOB is simply a dynamically generated data structure passed in >> between >> > > two boot phases. This is information that was obtained through >> discovery >> > > and needs to be passed forward to the next boot phase *once*, with no >> API >> > > needed to call back (e.g. no call back into previous firmware phase is >> > > needed to fetch this information at run-time - it is simply passed >> one time >> > > during boot). >> > > >> > > >> > > >> > > There may be one or more HOBs passed in between boot phases. If >> there are >> > > more than one HOB that needs to be passed, this can be in a form of a >> "HOB >> > > table", which (for example) could be a UUID indexed array of pointers >> to >> > > HOB structures, used to locate a HOB of interest (based on UUID). In >> such >> > > cases, instead of passing a single HOB, the boot phases may rely on >> passing >> > > the pointer to the HOB table. >> > > >> > > >> > > >> > > This has been extremely useful concept to employ on highly >> configurable >> > > systems that must rely on flexible discovery mechanisms to initialize >> and >> > > boot the system. This is especially helpful when you have multiple >> > > >> > > >> > > >> > > >> > > >> > > Why do we need HOBs in TF-A?: >> > > >> > > ----------------------------- >> > > >> > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server >> SoC in >> > > a way that is SoC specific *but* platform agnostic. This means that a >> > > single ARM SoC that a SiP may deliver to customers may provide a >> single >> > > TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a >> broad >> > > range of platform designs and configurations in order to boot a >> platform >> > > specific firmware (e.g. BL33 and possibly even BL32 code). In order >> to >> > > achieve this, the platform configuration must be *discovered* instead >> of >> > > statically compiled as it is today in TF-A via device tree based >> > > enumeration. The mechanisms of discovery may differ broadly >> depending on >> > > the relevant industry standard, or in some cases may have rely on SiP >> > > specific discovery flows. >> > > >> > > >> > > >> > > For example: On server systems that support a broad range DIMM memory >> > > population/topologies, all the necessary information required to boot >> is >> > > fully discovered via standard JEDEC Serial Presence Detect (SPD) over >> an >> > > I2C bus. Leveraging the SPD bus, may platform variants could be >> supported >> > > with a single TF-A binary. Not only is this information required to >> > > initialize memory in early boot phases (e.g. BL2), the subsequent boot >> > > phases will also need this SPD info to construct a system physical >> address >> > > map and properly initialize the MMU based on the memory present, and >> where >> > > the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) >> may >> > > need to generate standard firmware tables to the operating systems, >> such as >> > > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. >> SLIT, >> > > SRAT, even NFIT if NVDIMM's are present). >> > > >> > > >> > > >> > > In short, it all starts with a standardized or vendor specific >> discovery >> > > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of >> > > information to the next boot stages (e.g. BL31/BL32/BL33). >> > > >> > > >> > > >> > > Today, every HOB may be a vendor specific structure, but in the future >> > > there may be benefit of defining standard HOBs. This may be useful >> for >> > > memory discovery, passing the system physical address map, enabling >> TPM >> > > measured boot, and potentially many other common HOB use-cases. >> > > >> > > >> > > >> > > It would be extremely beneficial to the datacenter market segment if >> the >> > > TF-A community would adopt this concept of information passing >> between all >> > > boot phases as opposed to rely solely on device tree enumeration. >> This is >> > > not intended to replace device tree, rather intended as an >> alternative way >> > > to describe the info that must be discovered and dynamically >> generated. >> > > >> > > >> > > >> > > >> > > >> > > Conclusion: >> > > >> > > ----------- >> > > >> > > We are proposing that the TF-A community begin pursuing the adoption >> of >> > > HOBs as a mechanism used for information exchange between each boot >> stage >> > > (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term >> we >> > > want to explore standardizing some HOB structures for the BL33 phase >> (e.g. >> > > UEFI HOB structures), but initially would like to agree on this being >> a >> > > useful mechanism used to pass information between each boot stage. >> > > >> > > >> > > >> > > Thanks, >> > > >> > > --Harb >> > > >> > > >> > > >> > > >> > > >> > > >> > > -- >> > > TF-A mailing list >> > > TF-A(a)lists.trustedfirmware.org >> > > https://lists.trustedfirmware.org/mailman/listinfo/tf-a >> > > >> > >> > >> > -- >> > François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* >> > T: +33.67221.6485 >> > francois.ozog(a)linaro.org | Skype: ffozog >> > _______________________________________________ >> > boot-architecture mailing list >> > boot-architecture(a)lists.linaro.org >> > https://lists.linaro.org/mailman/listinfo/boot-architecture >> >> >> >> >> -- >> >> *François-Frédéric Ozog* | *Director Linaro Edge & Fog Computing Group* >> >> T: +33.67221.6485 >> francois.ozog(a)linaro.org | Skype: ffozog >> >> >> >> -- >> TF-A mailing list >> TF-A(a)lists.trustedfirmware.org >> https://lists.trustedfirmware.org/mailman/listinfo/tf-a >> >> >> >> >> -- >> >> *François-Frédéric Ozog* | *Director Linaro Edge & Fog Computing Group* >> >> T: +33.67221.6485 >> francois.ozog(a)linaro.org | Skype: ffozog >> >> >> > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a >

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by François Ozog

+Ron Minnich <rminnich(a)google.com> +Paul Isaac's <paul.isaacs(a)linaro.org> Adding Ron and Paul because I think this interface should be also benefiting LinuxBoot efforts. On Tue, 23 Mar 2021 at 11:17, François Ozog via TF-A < tf-a(a)lists.trustedfirmware.org> wrote: > Hi, > > I propose we cover the topic at the next Trusted Substrate > <https://collaborate.linaro.org/display/TS/Trusted+Substrate+Home> zoom > call <https://linaro-org.zoom.us/j/94563644892> on April 8th 4pm CET. > > The agenda: > ABI between non-secure firmware and the rest of firmware (EL3, S-EL1, > S-EL2, SCP) to adapt hardware description to some runtime conditions. > runtime conditions here relates to DRAM size and topology detection, > secure DRAM memory carvings, PSCI and SCMI interface publishing. > > For additional background on existing metadata: UEFI Platform > Initialization Specification Version 1.7 > <https://uefi.org/sites/default/files/resources/PI_Spec_1_7_final_Jan_2019.p…> > , 5.5 Resource Descriptor HOB > Out of the ResourceType we care about is EFI_RESOURCE_SYSTEM_MEMORY. > This HOB lacks memory NUMA attachment or something that could be related > to fill SRAT table for ACPI or relevant DT proximity domains. > HOB is not consistent accros platforms: some platforms (Arm) lists memory > from the booting NUMA node, other platforms (x86) lists all memory from all > NUMA nodes. (At least this is the case on the two platforms I tested). > > There are two proposals to use memory structures from SPL/BLx up to the > handover function (as defined in the Device Tree technical report > <https://docs.google.com/document/d/1CLkhLRaz_zcCq44DLGmPZQFPbYHOC6nzPowaL0X…>) > which can be U-boot (BL33 or just U-Boot in case of SPL/U-Boot scheme) or > EDK2. > I would propose we also discuss possibility of FF-A interface to actually > query information or request actions to be done (this is a model actually > used in some SoCs with proprietary SMC calls). > > Requirements (to be validated): > - ACPI and DT hardware descriptions. > - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2) > - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2, TF-A/LinuxBoot) > - at least allows complete DRAM description and "persistent" usage > (reserved areas for secure world or other usages) > - support secure world device assignment > > Cheers > > FF > > > On Mon, 22 Mar 2021 at 19:56, Simon Glass <sjg(a)chromium.org> wrote: > >> Hi, >> >> Can I suggest using bloblist for this instead? It is lightweight, >> easier to parse, doesn't have GUIDs and is already used within U-Boot >> for passing info between SPL/U-Boot, etc. >> >> Docs here: >> https://github.com/u-boot/u-boot/blob/master/doc/README.bloblist >> Header file describes the format: >> https://github.com/u-boot/u-boot/blob/master/include/bloblist.h >> >> Full set of unit tests: >> https://github.com/u-boot/u-boot/blob/master/test/bloblist.c >> >> Regards, >> Simon >> >> On Mon, 22 Mar 2021 at 23:58, François Ozog <francois.ozog(a)linaro.org> >> wrote: >> > >> > +Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org> >> > >> > standardization is very much welcomed here and need to accommodate a >> very >> > diverse set of situations. >> > For example, TEE OS may need to pass memory reservations to BL33 or >> > "capture" a device for the secure world. >> > >> > I have observed a number of architectures: >> > 1) pass information from BLx to BLy in the form of a specific object >> > 2) BLx called by BLy by a platform specific SMC to get information >> > 3) BLx called by BLy by a platform specific SMC to perform Device Tree >> > fixups >> > >> > I also imagined a standardized "broadcast" FF-A call so that any >> firmware >> > element can either provide information or "do something". >> > >> > My understanding of your proposal is about standardizing on >> architecture 1) >> > with the HOB format. >> > >> > The advantage of the HOB is simplicity but it may be difficult to >> implement >> > schemes such as pruning a DT because device assignment in the secure >> world. >> > >> > In any case, it looks feasible to have TF-A and OP-TEE complement the >> list >> > of HOBs to pass information downstream (the bootflow). >> > >> > It would be good to start with building the comprehensive list of >> > information that need to be conveyed between firmware elements: >> > >> > information. | authoritative entity | reporting entity | information >> > exchanged: >> > dram | TFA | TFA | >> > <format to be detailed, NUMA topology to build the SRAT table or DT >> > equivalent?> >> > PSCI | SCP | TFA? | >> > SCMI | SCP or TEE-OS | TFA? TEE-OS?| >> > secure SRAM | TFA. | TFA. | >> > secure DRAM | TFA? TEE-OS? | TFA? TEE-OS? | >> > other? | | >> > | >> > >> > Cheers >> > >> > FF >> > >> > >> > On Mon, 22 Mar 2021 at 09:34, Harb Abdulhamid OS via TF-A < >> > tf-a(a)lists.trustedfirmware.org> wrote: >> > >> > > Hello Folks, >> > > >> > > >> > > >> > > I'm emailing to start an open discussion about the adoption of a >> concept >> > > known as "hand-off blocks" or HOB to become a part of the TF-A >> Firmware >> > > Framework Architecture (FFA). This is something that is a pretty >> major >> > > pain point when it comes to the adoption of TF-A in ARM Server SoC’s >> > > designed to enable a broad range of highly configurable datacenter >> > > platforms. >> > > >> > > >> > > >> > > >> > > >> > > What is a HOB (Background)? >> > > >> > > --------------------------- >> > > >> > > UEFI PI spec describes a particular definition for how HOB may be >> used for >> > > transitioning between the PEI and DXE boot phases, which is a good >> > > reference point for this discussion, but not necessarily the exact >> solution >> > > appropriate for TF-A. >> > > >> > > >> > > >> > > A HOB is simply a dynamically generated data structure passed in >> between >> > > two boot phases. This is information that was obtained through >> discovery >> > > and needs to be passed forward to the next boot phase *once*, with no >> API >> > > needed to call back (e.g. no call back into previous firmware phase is >> > > needed to fetch this information at run-time - it is simply passed >> one time >> > > during boot). >> > > >> > > >> > > >> > > There may be one or more HOBs passed in between boot phases. If >> there are >> > > more than one HOB that needs to be passed, this can be in a form of a >> "HOB >> > > table", which (for example) could be a UUID indexed array of pointers >> to >> > > HOB structures, used to locate a HOB of interest (based on UUID). In >> such >> > > cases, instead of passing a single HOB, the boot phases may rely on >> passing >> > > the pointer to the HOB table. >> > > >> > > >> > > >> > > This has been extremely useful concept to employ on highly >> configurable >> > > systems that must rely on flexible discovery mechanisms to initialize >> and >> > > boot the system. This is especially helpful when you have multiple >> > > >> > > >> > > >> > > >> > > >> > > Why do we need HOBs in TF-A?: >> > > >> > > ----------------------------- >> > > >> > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server >> SoC in >> > > a way that is SoC specific *but* platform agnostic. This means that a >> > > single ARM SoC that a SiP may deliver to customers may provide a >> single >> > > TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a >> broad >> > > range of platform designs and configurations in order to boot a >> platform >> > > specific firmware (e.g. BL33 and possibly even BL32 code). In order >> to >> > > achieve this, the platform configuration must be *discovered* instead >> of >> > > statically compiled as it is today in TF-A via device tree based >> > > enumeration. The mechanisms of discovery may differ broadly >> depending on >> > > the relevant industry standard, or in some cases may have rely on SiP >> > > specific discovery flows. >> > > >> > > >> > > >> > > For example: On server systems that support a broad range DIMM memory >> > > population/topologies, all the necessary information required to boot >> is >> > > fully discovered via standard JEDEC Serial Presence Detect (SPD) over >> an >> > > I2C bus. Leveraging the SPD bus, may platform variants could be >> supported >> > > with a single TF-A binary. Not only is this information required to >> > > initialize memory in early boot phases (e.g. BL2), the subsequent boot >> > > phases will also need this SPD info to construct a system physical >> address >> > > map and properly initialize the MMU based on the memory present, and >> where >> > > the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) >> may >> > > need to generate standard firmware tables to the operating systems, >> such as >> > > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. >> SLIT, >> > > SRAT, even NFIT if NVDIMM's are present). >> > > >> > > >> > > >> > > In short, it all starts with a standardized or vendor specific >> discovery >> > > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of >> > > information to the next boot stages (e.g. BL31/BL32/BL33). >> > > >> > > >> > > >> > > Today, every HOB may be a vendor specific structure, but in the future >> > > there may be benefit of defining standard HOBs. This may be useful >> for >> > > memory discovery, passing the system physical address map, enabling >> TPM >> > > measured boot, and potentially many other common HOB use-cases. >> > > >> > > >> > > >> > > It would be extremely beneficial to the datacenter market segment if >> the >> > > TF-A community would adopt this concept of information passing >> between all >> > > boot phases as opposed to rely solely on device tree enumeration. >> This is >> > > not intended to replace device tree, rather intended as an >> alternative way >> > > to describe the info that must be discovered and dynamically >> generated. >> > > >> > > >> > > >> > > >> > > >> > > Conclusion: >> > > >> > > ----------- >> > > >> > > We are proposing that the TF-A community begin pursuing the adoption >> of >> > > HOBs as a mechanism used for information exchange between each boot >> stage >> > > (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term >> we >> > > want to explore standardizing some HOB structures for the BL33 phase >> (e.g. >> > > UEFI HOB structures), but initially would like to agree on this being >> a >> > > useful mechanism used to pass information between each boot stage. >> > > >> > > >> > > >> > > Thanks, >> > > >> > > --Harb >> > > >> > > >> > > >> > > >> > > >> > > >> > > -- >> > > TF-A mailing list >> > > TF-A(a)lists.trustedfirmware.org >> > > https://lists.trustedfirmware.org/mailman/listinfo/tf-a >> > > >> > >> > >> > -- >> > François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* >> > T: +33.67221.6485 >> > francois.ozog(a)linaro.org | Skype: ffozog >> > _______________________________________________ >> > boot-architecture mailing list >> > boot-architecture(a)lists.linaro.org >> > https://lists.linaro.org/mailman/listinfo/boot-architecture >> > > > -- > François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* > T: +33.67221.6485 > francois.ozog(a)linaro.org | Skype: ffozog > > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > -- François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* T: +33.67221.6485 francois.ozog(a)linaro.org | Skype: ffozog

4 years, 7 months

3
3
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Bill Fletcher

Adding Andrea On Wed, 24 Mar 2021 at 13:55, Joanna Farley via TF-A < tf-a(a)lists.trustedfirmware.org> wrote: > Hi Harb and others, > > > > This thread is now multi-mailing list and I can see some broader needs and > opinions on aspects not directly defined by the TF-A project such as > differing information exchange formats. However, this is definitely > something the TF-A project can try and help provide enablement for to help > with the goal of supplying support for single or common TF-A binaries builds > for different images. TF-A already have some limited support in this space > and are considering how this can be extended given some of the needs > expressed here. Folks on the TF-A project are studying the below and will > propose soon some ideas on how TF-A could provide more versatile enablement > in this space shortly. > > > > Thanks > > > > Joanna > > > > *From: *TF-A <tf-a-bounces(a)lists.trustedfirmware.org> on behalf of > François Ozog via TF-A <tf-a(a)lists.trustedfirmware.org> > *Reply to: *François Ozog <francois.ozog(a)linaro.org> > *Date: *Wednesday, 24 March 2021 at 08:34 > *To: *Harb Abdulhamid OS <abdulhamid(a)os.amperecomputing.com> > *Cc: *"tf-a(a)lists.trustedfirmware.org" <tf-a(a)lists.trustedfirmware.org>, > Simon Glass <sjg(a)chromium.org>, Boot Architecture Mailman List < > boot-architecture(a)lists.linaro.org>, Paul Isaac's <paul.isaacs(a)linaro.org>, > Ron Minnich <rminnich(a)google.com> > *Subject: *Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for > information passing between boot stages > > > > > > > > On Tue, 23 Mar 2021 at 23:39, Harb Abdulhamid OS < > abdulhamid(a)os.amperecomputing.com> wrote: > > Hello Folks, > > Appreciate the feedback and replies on this. Glad to see that there is > interest in this topic. 😊 > > > > I try to address the comments/feedback from Francois and Simon below…. > > > > @François Ozog <francois.ozog(a)linaro.org> – happy to discuss this on a > zoom call. I will make that time slot work, and will be available to > attend April 8, 4pm CT. > > > > Note that I’m using the term “HOB” here more generically, as there are > typically vendor specific structures beyond the resource descriptor HOB, > which provides only a small subset of the information that needs to be > passed between the boot phases. > > > > The whole point here is to provide mechanism to develop firmware that we > can build ARM Server SoC’s that support **any** BL33 payload (e.g. EDK2, > AptioV, CoreBoot, and maybe even directly boot strapping LinuxBoot at some > point). In other-words, we are trying to come up with a TF-A that would > be completely agnostic to the implementation of BL33 (i.e. BL33 is built > completely independently by a separate entity – e.g. an ODM/OEM). > > > > Keep in mind, in the server/datacenter market segment we are not building > vertically integrated systems with a single entity compiling > firmware/software stacks like most folks in TF-A have become use to. There > are two categories of higher level firmware code blobs in the > server/datacenter model: > > 1. “SoC” or “silicon” firmware – in TF-A this may map to BL1, BL2, > BL31, and **possibly** one or more BL32 instances > 2. “Platform” or “board” firmware – in TF-A this may map to BL33 and * > *possibly** one or more BL32 instances. > > > > Even the platform firmware stack could be further fragmented by having > multiple entities involved in delivering the entire firmware stack: IBVs, > ODMs, OEMs, CSPs, and possibly even device vendor code. > > > > To support a broad range of platform designs with a broad range of memory > devices, we need a crisp and clear contract between the SoC firmware that > initializes memory (e.g. BL2) and how that platform boot firmware (e.g. > BL33) gathers information about what memory that was initialized, at what > speeds, NUMA topology, and many other relevant information that needs to be > known and comprehended by the platform firmware and eventually by the > platform software. > > > > I understand the versatility of DT, but I see two major problems with DT: > > - DT requires more complicated parsing to get properties, and even > more complex to dynamically set properties – this HOB structures may need > to be generated in boot phases where DDR is not available, and therefore we > will be extremely memory constrained. > - DT is probably overkill for this purpose – We really just want a > list of pointers to simple C structures that code cast (e.g. JEDEC SPD data > blob) > > > > I think that we should not mix the efforts around DT/ACPI specs with what > we are doing here, because those specs and concepts were developed for a > completely different purpose (i.e. abstractions needed for OS / RTOS > software, and not necessarily suitable for firmware-to-firmware hand-offs). > > > > Frankly, I would personally push back pretty hard on defining SMC’s for > something that should be one way information passing. Every SMC we add is > another attack vector to the secure world and an increased burden on the > folks that have to do security auditing and threat analysis. I see no > benefit in exposing these boot/HOB/BOB structures at run-time via SMC > calls. > > > > Please do let me know if you disagree and why. Look forward to discussing > on this thread or on the call. > > > > I am not tied to a particular data representation and using SMC to just > pass data structures is overkill as you say. The SMC model seems useful to > do complex things like device assignment to secure world. Or something else > we don't have yet an idea. > > Let's say there is one board with two eMMCs. This board is used by two > OEMs. One is fine with all eMMCs in non-secure world, the other wants to > assign the eMMC to secure world. > > That's something that is related to inter-firmware component communication > to be authoritative. > > We need to avoid "little arrangements between friends" that exist today, > where the Linux provided DT is pruned from the second eMMC to accommodate > the use case. We need to think the OS as "immutable" across platforms and > adapt to available hardware (not come with its own description of what the > board is). > > May be a hob would contain a DT overlay or ACPI equivalent that would do > the job. > > In that case we do not need SMC. > > What do you think of this use case? > > > > @Simon Glass <sjg(a)chromium.org> - Thanks for the pointer to bloblist. > I briefly reviewed and it seems like a good baseline for what we may be > looking for. > > > > That being said, I would say that there is some benefit in having some > kind of unique identifiers (e.g. UUID or some unique signature) so that we > can tie standardized data structures (based on some future TBD specs) to a > particular ID. For example, if the TPM driver in BL33 is looking for the > TPM structure in the HOB/BOB list, and may not care about the other data > blobs. The driver needs a way to identify and locate the blob it cares > about. > > > > I guess we can achieve this with the tag, but the problem with tag when > you have eco-system with a lot of parties doing parallel development, you > can end up with tag collisions and folks fighting about who has rights to > what tag values. We would need some official process for folks to register > tags for whatever new structures we define, or maybe some tag range for > vendor specific structures. This comes with a lot of pain and > bureaucracy. On the other hand, UUID has been a proven way to make it easy > to just define your own blobs with **either** standard or vendor specific > structures without worry of ID collisions between vendors. > > > > We can probably debate whether there is any value in GUID/UUID or not > during the call… but again, boblist seems like a reasonable starting point > as an alternative to HOB. > > > > Thanks, > > --Harb > > > > *From:* François Ozog <francois.ozog(a)linaro.org> > *Sent:* Tuesday, March 23, 2021 10:00 AM > *To:* François Ozog <francois.ozog(a)linaro.org>; Ron Minnich < > rminnich(a)google.com>; Paul Isaac's <paul.isaacs(a)linaro.org> > *Cc:* Simon Glass <sjg(a)chromium.org>; Harb Abdulhamid OS < > abdulhamid(a)os.amperecomputing.com>; Boot Architecture Mailman List < > boot-architecture(a)lists.linaro.org>; tf-a(a)lists.trustedfirmware.org > *Subject:* Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for > information passing between boot stages > > > > +Ron Minnich <rminnich(a)google.com> +Paul Isaac's <paul.isaacs(a)linaro.org> > > > > Adding Ron and Paul because I think this interface should be also > benefiting LinuxBoot efforts. > > > > On Tue, 23 Mar 2021 at 11:17, François Ozog via TF-A < > tf-a(a)lists.trustedfirmware.org> wrote: > > Hi, > > > > I propose we cover the topic at the next Trusted Substrate > <https://collaborate.linaro.org/display/TS/Trusted+Substrate+Home> zoom > call <https://linaro-org.zoom.us/j/94563644892> on April 8th 4pm CET. > > > > The agenda: > > ABI between non-secure firmware and the rest of firmware (EL3, S-EL1, > S-EL2, SCP) to adapt hardware description to some runtime conditions. > > runtime conditions here relates to DRAM size and topology detection, > secure DRAM memory carvings, PSCI and SCMI interface publishing. > > > > For additional background on existing metadata: UEFI Platform > Initialization Specification Version 1.7 > <https://uefi.org/sites/default/files/resources/PI_Spec_1_7_final_Jan_2019.p…> > , 5.5 Resource Descriptor HOB > > Out of the ResourceType we care about is EFI_RESOURCE_SYSTEM_MEMORY. > > This HOB lacks memory NUMA attachment or something that could be related > to fill SRAT table for ACPI or relevant DT proximity domains. > > HOB is not consistent accros platforms: some platforms (Arm) lists memory > from the booting NUMA node, other platforms (x86) lists all memory from all > NUMA nodes. (At least this is the case on the two platforms I tested). > > > > There are two proposals to use memory structures from SPL/BLx up to the > handover function (as defined in the Device Tree technical report > <https://docs.google.com/document/d/1CLkhLRaz_zcCq44DLGmPZQFPbYHOC6nzPowaL0X…>) > which can be U-boot (BL33 or just U-Boot in case of SPL/U-Boot scheme) or > EDK2. > > I would propose we also discuss possibility of FF-A interface to actually > query information or request actions to be done (this is a model actually > used in some SoCs with proprietary SMC calls). > > > > Requirements (to be validated): > > - ACPI and DT hardware descriptions. > > - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2) > > - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2, > TF-A/LinuxBoot) > > - at least allows complete DRAM description and "persistent" usage > (reserved areas for secure world or other usages) > > - support secure world device assignment > > > > Cheers > > > > FF > > > > > > On Mon, 22 Mar 2021 at 19:56, Simon Glass <sjg(a)chromium.org> wrote: > > Hi, > > Can I suggest using bloblist for this instead? It is lightweight, > easier to parse, doesn't have GUIDs and is already used within U-Boot > for passing info between SPL/U-Boot, etc. > > Docs here: > https://github.com/u-boot/u-boot/blob/master/doc/README.bloblist > Header file describes the format: > https://github.com/u-boot/u-boot/blob/master/include/bloblist.h > > Full set of unit tests: > https://github.com/u-boot/u-boot/blob/master/test/bloblist.c > > Regards, > Simon > > On Mon, 22 Mar 2021 at 23:58, François Ozog <francois.ozog(a)linaro.org> > wrote: > > > > +Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org> > > > > standardization is very much welcomed here and need to accommodate a very > > diverse set of situations. > > For example, TEE OS may need to pass memory reservations to BL33 or > > "capture" a device for the secure world. > > > > I have observed a number of architectures: > > 1) pass information from BLx to BLy in the form of a specific object > > 2) BLx called by BLy by a platform specific SMC to get information > > 3) BLx called by BLy by a platform specific SMC to perform Device Tree > > fixups > > > > I also imagined a standardized "broadcast" FF-A call so that any firmware > > element can either provide information or "do something". > > > > My understanding of your proposal is about standardizing on architecture > 1) > > with the HOB format. > > > > The advantage of the HOB is simplicity but it may be difficult to > implement > > schemes such as pruning a DT because device assignment in the secure > world. > > > > In any case, it looks feasible to have TF-A and OP-TEE complement the > list > > of HOBs to pass information downstream (the bootflow). > > > > It would be good to start with building the comprehensive list of > > information that need to be conveyed between firmware elements: > > > > information. | authoritative entity | reporting entity | information > > exchanged: > > dram | TFA | TFA | > > <format to be detailed, NUMA topology to build the SRAT table or DT > > equivalent?> > > PSCI | SCP | TFA? | > > SCMI | SCP or TEE-OS | TFA? TEE-OS?| > > secure SRAM | TFA. | TFA. | > > secure DRAM | TFA? TEE-OS? | TFA? TEE-OS? | > > other? | | > > | > > > > Cheers > > > > FF > > > > > > On Mon, 22 Mar 2021 at 09:34, Harb Abdulhamid OS via TF-A < > > tf-a(a)lists.trustedfirmware.org> wrote: > > > > > Hello Folks, > > > > > > > > > > > > I'm emailing to start an open discussion about the adoption of a > concept > > > known as "hand-off blocks" or HOB to become a part of the TF-A Firmware > > > Framework Architecture (FFA). This is something that is a pretty major > > > pain point when it comes to the adoption of TF-A in ARM Server SoC’s > > > designed to enable a broad range of highly configurable datacenter > > > platforms. > > > > > > > > > > > > > > > > > > What is a HOB (Background)? > > > > > > --------------------------- > > > > > > UEFI PI spec describes a particular definition for how HOB may be used > for > > > transitioning between the PEI and DXE boot phases, which is a good > > > reference point for this discussion, but not necessarily the exact > solution > > > appropriate for TF-A. > > > > > > > > > > > > A HOB is simply a dynamically generated data structure passed in > between > > > two boot phases. This is information that was obtained through > discovery > > > and needs to be passed forward to the next boot phase *once*, with no > API > > > needed to call back (e.g. no call back into previous firmware phase is > > > needed to fetch this information at run-time - it is simply passed one > time > > > during boot). > > > > > > > > > > > > There may be one or more HOBs passed in between boot phases. If there > are > > > more than one HOB that needs to be passed, this can be in a form of a > "HOB > > > table", which (for example) could be a UUID indexed array of pointers > to > > > HOB structures, used to locate a HOB of interest (based on UUID). In > such > > > cases, instead of passing a single HOB, the boot phases may rely on > passing > > > the pointer to the HOB table. > > > > > > > > > > > > This has been extremely useful concept to employ on highly configurable > > > systems that must rely on flexible discovery mechanisms to initialize > and > > > boot the system. This is especially helpful when you have multiple > > > > > > > > > > > > > > > > > > Why do we need HOBs in TF-A?: > > > > > > ----------------------------- > > > > > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC > in > > > a way that is SoC specific *but* platform agnostic. This means that a > > > single ARM SoC that a SiP may deliver to customers may provide a single > > > TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a broad > > > range of platform designs and configurations in order to boot a > platform > > > specific firmware (e.g. BL33 and possibly even BL32 code). In order to > > > achieve this, the platform configuration must be *discovered* instead > of > > > statically compiled as it is today in TF-A via device tree based > > > enumeration. The mechanisms of discovery may differ broadly depending > on > > > the relevant industry standard, or in some cases may have rely on SiP > > > specific discovery flows. > > > > > > > > > > > > For example: On server systems that support a broad range DIMM memory > > > population/topologies, all the necessary information required to boot > is > > > fully discovered via standard JEDEC Serial Presence Detect (SPD) over > an > > > I2C bus. Leveraging the SPD bus, may platform variants could be > supported > > > with a single TF-A binary. Not only is this information required to > > > initialize memory in early boot phases (e.g. BL2), the subsequent boot > > > phases will also need this SPD info to construct a system physical > address > > > map and properly initialize the MMU based on the memory present, and > where > > > the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) > may > > > need to generate standard firmware tables to the operating systems, > such as > > > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. > SLIT, > > > SRAT, even NFIT if NVDIMM's are present). > > > > > > > > > > > > In short, it all starts with a standardized or vendor specific > discovery > > > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of > > > information to the next boot stages (e.g. BL31/BL32/BL33). > > > > > > > > > > > > Today, every HOB may be a vendor specific structure, but in the future > > > there may be benefit of defining standard HOBs. This may be useful for > > > memory discovery, passing the system physical address map, enabling TPM > > > measured boot, and potentially many other common HOB use-cases. > > > > > > > > > > > > It would be extremely beneficial to the datacenter market segment if > the > > > TF-A community would adopt this concept of information passing between > all > > > boot phases as opposed to rely solely on device tree enumeration. > This is > > > not intended to replace device tree, rather intended as an alternative > way > > > to describe the info that must be discovered and dynamically generated. > > > > > > > > > > > > > > > > > > Conclusion: > > > > > > ----------- > > > > > > We are proposing that the TF-A community begin pursuing the adoption of > > > HOBs as a mechanism used for information exchange between each boot > stage > > > (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term we > > > want to explore standardizing some HOB structures for the BL33 phase > (e.g. > > > UEFI HOB structures), but initially would like to agree on this being a > > > useful mechanism used to pass information between each boot stage. > > > > > > > > > > > > Thanks, > > > > > > --Harb > > > > > > > > > > > > > > > > > > > > > -- > > > TF-A mailing list > > > TF-A(a)lists.trustedfirmware.org > > > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > > > > > > > > > -- > > François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* > > T: +33.67221.6485 > > francois.ozog(a)linaro.org | Skype: ffozog > > _______________________________________________ > > boot-architecture mailing list > > boot-architecture(a)lists.linaro.org > > https://lists.linaro.org/mailman/listinfo/boot-architecture > > > > > -- > > [image: Image removed by sender.] > > *François-Frédéric Ozog* | *Director Linaro Edge & Fog Computing Group* > > T: +33.67221.6485 > francois.ozog(a)linaro.org | Skype: ffozog > > > > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > > > > > -- > > [image: Image removed by sender.] > > *François-Frédéric Ozog* | *Director Linaro Edge & Fog Computing Group* > > T: +33.67221.6485 > francois.ozog(a)linaro.org | Skype: ffozog > > > > > > > -- > > [image: Image removed by sender.] > > *François-Frédéric Ozog* | *Director Linaro Edge & Fog Computing Group* > > T: +33.67221.6485 > francois.ozog(a)linaro.org | Skype: ffozog > > > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > -- [image: Linaro] <http://www.linaro.org/> *Bill Fletcher* | *Field Engineering* T: +44 7833 498336 <+44+7833+498336> bill.fletcher(a)linaro.org | Skype: billfletcher2020

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Joanna Farley

Hi Harb and others, This thread is now multi-mailing list and I can see some broader needs and opinions on aspects not directly defined by the TF-A project such as differing information exchange formats. However, this is definitely something the TF-A project can try and help provide enablement for to help with the goal of supplying support for single or common TF-A binaries builds for different images. TF-A already have some limited support in this space and are considering how this can be extended given some of the needs expressed here. Folks on the TF-A project are studying the below and will propose soon some ideas on how TF-A could provide more versatile enablement in this space shortly. Thanks Joanna From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> on behalf of François Ozog via TF-A <tf-a(a)lists.trustedfirmware.org> Reply to: François Ozog <francois.ozog(a)linaro.org> Date: Wednesday, 24 March 2021 at 08:34 To: Harb Abdulhamid OS <abdulhamid(a)os.amperecomputing.com> Cc: "tf-a(a)lists.trustedfirmware.org" <tf-a(a)lists.trustedfirmware.org>, Simon Glass <sjg(a)chromium.org>, Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org>, Paul Isaac's <paul.isaacs(a)linaro.org>, Ron Minnich <rminnich(a)google.com> Subject: Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages On Tue, 23 Mar 2021 at 23:39, Harb Abdulhamid OS <abdulhamid(a)os.amperecomputing.com<mailto:abdulhamid@os.amperecomputing.com>> wrote: Hello Folks, Appreciate the feedback and replies on this. Glad to see that there is interest in this topic. 😊 I try to address the comments/feedback from Francois and Simon below…. @François Ozog<mailto:francois.ozog@linaro.org> – happy to discuss this on a zoom call. I will make that time slot work, and will be available to attend April 8, 4pm CT. Note that I’m using the term “HOB” here more generically, as there are typically vendor specific structures beyond the resource descriptor HOB, which provides only a small subset of the information that needs to be passed between the boot phases. The whole point here is to provide mechanism to develop firmware that we can build ARM Server SoC’s that support *any* BL33 payload (e.g. EDK2, AptioV, CoreBoot, and maybe even directly boot strapping LinuxBoot at some point). In other-words, we are trying to come up with a TF-A that would be completely agnostic to the implementation of BL33 (i.e. BL33 is built completely independently by a separate entity – e.g. an ODM/OEM). Keep in mind, in the server/datacenter market segment we are not building vertically integrated systems with a single entity compiling firmware/software stacks like most folks in TF-A have become use to. There are two categories of higher level firmware code blobs in the server/datacenter model: 1. “SoC” or “silicon” firmware – in TF-A this may map to BL1, BL2, BL31, and *possibly* one or more BL32 instances 2. “Platform” or “board” firmware – in TF-A this may map to BL33 and *possibly* one or more BL32 instances. Even the platform firmware stack could be further fragmented by having multiple entities involved in delivering the entire firmware stack: IBVs, ODMs, OEMs, CSPs, and possibly even device vendor code. To support a broad range of platform designs with a broad range of memory devices, we need a crisp and clear contract between the SoC firmware that initializes memory (e.g. BL2) and how that platform boot firmware (e.g. BL33) gathers information about what memory that was initialized, at what speeds, NUMA topology, and many other relevant information that needs to be known and comprehended by the platform firmware and eventually by the platform software. I understand the versatility of DT, but I see two major problems with DT: * DT requires more complicated parsing to get properties, and even more complex to dynamically set properties – this HOB structures may need to be generated in boot phases where DDR is not available, and therefore we will be extremely memory constrained. * DT is probably overkill for this purpose – We really just want a list of pointers to simple C structures that code cast (e.g. JEDEC SPD data blob) I think that we should not mix the efforts around DT/ACPI specs with what we are doing here, because those specs and concepts were developed for a completely different purpose (i.e. abstractions needed for OS / RTOS software, and not necessarily suitable for firmware-to-firmware hand-offs). Frankly, I would personally push back pretty hard on defining SMC’s for something that should be one way information passing. Every SMC we add is another attack vector to the secure world and an increased burden on the folks that have to do security auditing and threat analysis. I see no benefit in exposing these boot/HOB/BOB structures at run-time via SMC calls. Please do let me know if you disagree and why. Look forward to discussing on this thread or on the call. I am not tied to a particular data representation and using SMC to just pass data structures is overkill as you say. The SMC model seems useful to do complex things like device assignment to secure world. Or something else we don't have yet an idea. Let's say there is one board with two eMMCs. This board is used by two OEMs. One is fine with all eMMCs in non-secure world, the other wants to assign the eMMC to secure world. That's something that is related to inter-firmware component communication to be authoritative. We need to avoid "little arrangements between friends" that exist today, where the Linux provided DT is pruned from the second eMMC to accommodate the use case. We need to think the OS as "immutable" across platforms and adapt to available hardware (not come with its own description of what the board is). May be a hob would contain a DT overlay or ACPI equivalent that would do the job. In that case we do not need SMC. What do you think of this use case? @Simon Glass<mailto:sjg@chromium.org> - Thanks for the pointer to bloblist. I briefly reviewed and it seems like a good baseline for what we may be looking for. That being said, I would say that there is some benefit in having some kind of unique identifiers (e.g. UUID or some unique signature) so that we can tie standardized data structures (based on some future TBD specs) to a particular ID. For example, if the TPM driver in BL33 is looking for the TPM structure in the HOB/BOB list, and may not care about the other data blobs. The driver needs a way to identify and locate the blob it cares about. I guess we can achieve this with the tag, but the problem with tag when you have eco-system with a lot of parties doing parallel development, you can end up with tag collisions and folks fighting about who has rights to what tag values. We would need some official process for folks to register tags for whatever new structures we define, or maybe some tag range for vendor specific structures. This comes with a lot of pain and bureaucracy. On the other hand, UUID has been a proven way to make it easy to just define your own blobs with *either* standard or vendor specific structures without worry of ID collisions between vendors. We can probably debate whether there is any value in GUID/UUID or not during the call… but again, boblist seems like a reasonable starting point as an alternative to HOB. Thanks, --Harb From: François Ozog <francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org>> Sent: Tuesday, March 23, 2021 10:00 AM To: François Ozog <francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org>>; Ron Minnich <rminnich(a)google.com<mailto:rminnich@google.com>>; Paul Isaac's <paul.isaacs(a)linaro.org<mailto:paul.isaacs@linaro.org>> Cc: Simon Glass <sjg(a)chromium.org<mailto:sjg@chromium.org>>; Harb Abdulhamid OS <abdulhamid(a)os.amperecomputing.com<mailto:abdulhamid@os.amperecomputing.com>>; Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org<mailto:boot-architecture@lists.linaro.org>>; tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Subject: Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages +Ron Minnich<mailto:rminnich@google.com> +Paul Isaac's<mailto:paul.isaacs@linaro.org> Adding Ron and Paul because I think this interface should be also benefiting LinuxBoot efforts. On Tue, 23 Mar 2021 at 11:17, François Ozog via TF-A <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> wrote: Hi, I propose we cover the topic at the next Trusted Substrate<https://collaborate.linaro.org/display/TS/Trusted+Substrate+Home> zoom call<https://linaro-org.zoom.us/j/94563644892> on April 8th 4pm CET. The agenda: ABI between non-secure firmware and the rest of firmware (EL3, S-EL1, S-EL2, SCP) to adapt hardware description to some runtime conditions. runtime conditions here relates to DRAM size and topology detection, secure DRAM memory carvings, PSCI and SCMI interface publishing. For additional background on existing metadata: UEFI Platform Initialization Specification Version 1.7<https://uefi.org/sites/default/files/resources/PI_Spec_1_7_final_Jan_2019.p…>, 5.5 Resource Descriptor HOB Out of the ResourceType we care about is EFI_RESOURCE_SYSTEM_MEMORY. This HOB lacks memory NUMA attachment or something that could be related to fill SRAT table for ACPI or relevant DT proximity domains. HOB is not consistent accros platforms: some platforms (Arm) lists memory from the booting NUMA node, other platforms (x86) lists all memory from all NUMA nodes. (At least this is the case on the two platforms I tested). There are two proposals to use memory structures from SPL/BLx up to the handover function (as defined in the Device Tree technical report<https://docs.google.com/document/d/1CLkhLRaz_zcCq44DLGmPZQFPbYHOC6nzPowaL0X…>) which can be U-boot (BL33 or just U-Boot in case of SPL/U-Boot scheme) or EDK2. I would propose we also discuss possibility of FF-A interface to actually query information or request actions to be done (this is a model actually used in some SoCs with proprietary SMC calls). Requirements (to be validated): - ACPI and DT hardware descriptions. - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2) - agnostic to boot framework (SPL/U-Boot, TF-A/U-Boot, TF-A/EDK2, TF-A/LinuxBoot) - at least allows complete DRAM description and "persistent" usage (reserved areas for secure world or other usages) - support secure world device assignment Cheers FF On Mon, 22 Mar 2021 at 19:56, Simon Glass <sjg(a)chromium.org<mailto:sjg@chromium.org>> wrote: Hi, Can I suggest using bloblist for this instead? It is lightweight, easier to parse, doesn't have GUIDs and is already used within U-Boot for passing info between SPL/U-Boot, etc. Docs here: https://github.com/u-boot/u-boot/blob/master/doc/README.bloblist Header file describes the format: https://github.com/u-boot/u-boot/blob/master/include/bloblist.h Full set of unit tests: https://github.com/u-boot/u-boot/blob/master/test/bloblist.c Regards, Simon On Mon, 22 Mar 2021 at 23:58, François Ozog <francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org>> wrote: > > +Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org<mailto:boot-architecture@lists.linaro.org>> > > standardization is very much welcomed here and need to accommodate a very > diverse set of situations. > For example, TEE OS may need to pass memory reservations to BL33 or > "capture" a device for the secure world. > > I have observed a number of architectures: > 1) pass information from BLx to BLy in the form of a specific object > 2) BLx called by BLy by a platform specific SMC to get information > 3) BLx called by BLy by a platform specific SMC to perform Device Tree > fixups > > I also imagined a standardized "broadcast" FF-A call so that any firmware > element can either provide information or "do something". > > My understanding of your proposal is about standardizing on architecture 1) > with the HOB format. > > The advantage of the HOB is simplicity but it may be difficult to implement > schemes such as pruning a DT because device assignment in the secure world. > > In any case, it looks feasible to have TF-A and OP-TEE complement the list > of HOBs to pass information downstream (the bootflow). > > It would be good to start with building the comprehensive list of > information that need to be conveyed between firmware elements: > > information. | authoritative entity | reporting entity | information > exchanged: > dram | TFA | TFA | > <format to be detailed, NUMA topology to build the SRAT table or DT > equivalent?> > PSCI | SCP | TFA? | > SCMI | SCP or TEE-OS | TFA? TEE-OS?| > secure SRAM | TFA. | TFA. | > secure DRAM | TFA? TEE-OS? | TFA? TEE-OS? | > other? | | > | > > Cheers > > FF > > > On Mon, 22 Mar 2021 at 09:34, Harb Abdulhamid OS via TF-A < > tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> wrote: > > > Hello Folks, > > > > > > > > I'm emailing to start an open discussion about the adoption of a concept > > known as "hand-off blocks" or HOB to become a part of the TF-A Firmware > > Framework Architecture (FFA). This is something that is a pretty major > > pain point when it comes to the adoption of TF-A in ARM Server SoC’s > > designed to enable a broad range of highly configurable datacenter > > platforms. > > > > > > > > > > > > What is a HOB (Background)? > > > > --------------------------- > > > > UEFI PI spec describes a particular definition for how HOB may be used for > > transitioning between the PEI and DXE boot phases, which is a good > > reference point for this discussion, but not necessarily the exact solution > > appropriate for TF-A. > > > > > > > > A HOB is simply a dynamically generated data structure passed in between > > two boot phases. This is information that was obtained through discovery > > and needs to be passed forward to the next boot phase *once*, with no API > > needed to call back (e.g. no call back into previous firmware phase is > > needed to fetch this information at run-time - it is simply passed one time > > during boot). > > > > > > > > There may be one or more HOBs passed in between boot phases. If there are > > more than one HOB that needs to be passed, this can be in a form of a "HOB > > table", which (for example) could be a UUID indexed array of pointers to > > HOB structures, used to locate a HOB of interest (based on UUID). In such > > cases, instead of passing a single HOB, the boot phases may rely on passing > > the pointer to the HOB table. > > > > > > > > This has been extremely useful concept to employ on highly configurable > > systems that must rely on flexible discovery mechanisms to initialize and > > boot the system. This is especially helpful when you have multiple > > > > > > > > > > > > Why do we need HOBs in TF-A?: > > > > ----------------------------- > > > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC in > > a way that is SoC specific *but* platform agnostic. This means that a > > single ARM SoC that a SiP may deliver to customers may provide a single > > TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a broad > > range of platform designs and configurations in order to boot a platform > > specific firmware (e.g. BL33 and possibly even BL32 code). In order to > > achieve this, the platform configuration must be *discovered* instead of > > statically compiled as it is today in TF-A via device tree based > > enumeration. The mechanisms of discovery may differ broadly depending on > > the relevant industry standard, or in some cases may have rely on SiP > > specific discovery flows. > > > > > > > > For example: On server systems that support a broad range DIMM memory > > population/topologies, all the necessary information required to boot is > > fully discovered via standard JEDEC Serial Presence Detect (SPD) over an > > I2C bus. Leveraging the SPD bus, may platform variants could be supported > > with a single TF-A binary. Not only is this information required to > > initialize memory in early boot phases (e.g. BL2), the subsequent boot > > phases will also need this SPD info to construct a system physical address > > map and properly initialize the MMU based on the memory present, and where > > the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) may > > need to generate standard firmware tables to the operating systems, such as > > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. SLIT, > > SRAT, even NFIT if NVDIMM's are present). > > > > > > > > In short, it all starts with a standardized or vendor specific discovery > > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of > > information to the next boot stages (e.g. BL31/BL32/BL33). > > > > > > > > Today, every HOB may be a vendor specific structure, but in the future > > there may be benefit of defining standard HOBs. This may be useful for > > memory discovery, passing the system physical address map, enabling TPM > > measured boot, and potentially many other common HOB use-cases. > > > > > > > > It would be extremely beneficial to the datacenter market segment if the > > TF-A community would adopt this concept of information passing between all > > boot phases as opposed to rely solely on device tree enumeration. This is > > not intended to replace device tree, rather intended as an alternative way > > to describe the info that must be discovered and dynamically generated. > > > > > > > > > > > > Conclusion: > > > > ----------- > > > > We are proposing that the TF-A community begin pursuing the adoption of > > HOBs as a mechanism used for information exchange between each boot stage > > (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term we > > want to explore standardizing some HOB structures for the BL33 phase (e.g. > > UEFI HOB structures), but initially would like to agree on this being a > > useful mechanism used to pass information between each boot stage. > > > > > > > > Thanks, > > > > --Harb > > > > > > > > > > > > > > -- > > TF-A mailing list > > TF-A(a)lists.trustedfirmware.org<mailto:TF-A@lists.trustedfirmware.org> > > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > > > > > -- > François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* > T: +33.67221.6485 > francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org> | Skype: ffozog > _______________________________________________ > boot-architecture mailing list > boot-architecture(a)lists.linaro.org<mailto:boot-architecture@lists.linaro.org> > https://lists.linaro.org/mailman/listinfo/boot-architecture -- [Image removed by sender.] François-Frédéric Ozog | Director Linaro Edge & Fog Computing Group T: +33.67221.6485 francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org> | Skype: ffozog -- TF-A mailing list TF-A(a)lists.trustedfirmware.org<mailto:TF-A@lists.trustedfirmware.org> https://lists.trustedfirmware.org/mailman/listinfo/tf-a -- [Image removed by sender.] François-Frédéric Ozog | Director Linaro Edge & Fog Computing Group T: +33.67221.6485 francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org> | Skype: ffozog -- [Image removed by sender.] François-Frédéric Ozog | Director Linaro Edge & Fog Computing Group T: +33.67221.6485 francois.ozog(a)linaro.org<mailto:francois.ozog@linaro.org> | Skype: ffozog

4 years, 7 months

1
0
0 0

Canceled event: TF-A Tech Forum @ Thu Mar 25, 2021 9am - 10am (MST) (tf-a@lists.trustedfirmware.org)

by Don Harbin

This event has been canceled. Title: TF-A Tech Forum We run an open technical forum call for anyone to participate and it is not restricted to Trusted Firmware project members. It will operate under the guidance of the TF TSC. Feel free to forward this invite to colleagues. Invites are via the TF-A mailing list and also published on the Trusted Firmware website. Details are here: https://www.trustedfirmware.org/meetings/tf-a-technical-forum/Tr… Firmware is inviting you to a scheduled Zoom meeting.Join Zoom Meetinghttps://zoom.us/j/9159704974Meeting ID: 915 970 4974One tap mobile+16465588656,,9159704974# US (New York)+16699009128,,9159704974# US (San Jose)Dial by your location        +1 646 558 8656 US (New York)        +1 669 900 9128 US (San Jose)        877 853 5247 US Toll-free        888 788 0099 US Toll-freeMeeting ID: 915 970 4974Find your local number: https://zoom.us/u/ad27hc6t7h   When: Thu Mar 25, 2021 9am – 10am Mountain Standard Time - Phoenix Calendar: tf-a(a)lists.trustedfirmware.org Who: * Bill Fletcher - creator * marek.bykowski(a)gmail.com * okash.khawaja(a)gmail.com * tf-a(a)lists.trustedfirmware.org Invitation from Google Calendar: https://calendar.google.com/calendar/ You are receiving this courtesy email at the account tf-a(a)lists.trustedfirmware.org because you are an attendee of this event. To stop receiving future updates for this event, decline this event. Alternatively you can sign up for a Google account at https://calendar.google.com/calendar/ and control your notification settings for your entire calendar. Forwarding this invitation could allow any recipient to send a response to the organizer and be added to the guest list, or invite others regardless of their own invitation status, or to modify your RSVP. Learn more at https://support.google.com/calendar/answer/37135#forwarding

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by François Ozog

+Boot Architecture Mailman List <boot-architecture(a)lists.linaro.org> standardization is very much welcomed here and need to accommodate a very diverse set of situations. For example, TEE OS may need to pass memory reservations to BL33 or "capture" a device for the secure world. I have observed a number of architectures: 1) pass information from BLx to BLy in the form of a specific object 2) BLx called by BLy by a platform specific SMC to get information 3) BLx called by BLy by a platform specific SMC to perform Device Tree fixups I also imagined a standardized "broadcast" FF-A call so that any firmware element can either provide information or "do something". My understanding of your proposal is about standardizing on architecture 1) with the HOB format. The advantage of the HOB is simplicity but it may be difficult to implement schemes such as pruning a DT because device assignment in the secure world. In any case, it looks feasible to have TF-A and OP-TEE complement the list of HOBs to pass information downstream (the bootflow). It would be good to start with building the comprehensive list of information that need to be conveyed between firmware elements: information. | authoritative entity | reporting entity | information exchanged: dram | TFA | TFA | <format to be detailed, NUMA topology to build the SRAT table or DT equivalent?> PSCI | SCP | TFA? | SCMI | SCP or TEE-OS | TFA? TEE-OS?| secure SRAM | TFA. | TFA. | secure DRAM | TFA? TEE-OS? | TFA? TEE-OS? | other? | | | Cheers FF On Mon, 22 Mar 2021 at 09:34, Harb Abdulhamid OS via TF-A < tf-a(a)lists.trustedfirmware.org> wrote: > Hello Folks, > > > > I'm emailing to start an open discussion about the adoption of a concept > known as "hand-off blocks" or HOB to become a part of the TF-A Firmware > Framework Architecture (FFA). This is something that is a pretty major > pain point when it comes to the adoption of TF-A in ARM Server SoC’s > designed to enable a broad range of highly configurable datacenter > platforms. > > > > > > What is a HOB (Background)? > > --------------------------- > > UEFI PI spec describes a particular definition for how HOB may be used for > transitioning between the PEI and DXE boot phases, which is a good > reference point for this discussion, but not necessarily the exact solution > appropriate for TF-A. > > > > A HOB is simply a dynamically generated data structure passed in between > two boot phases. This is information that was obtained through discovery > and needs to be passed forward to the next boot phase *once*, with no API > needed to call back (e.g. no call back into previous firmware phase is > needed to fetch this information at run-time - it is simply passed one time > during boot). > > > > There may be one or more HOBs passed in between boot phases. If there are > more than one HOB that needs to be passed, this can be in a form of a "HOB > table", which (for example) could be a UUID indexed array of pointers to > HOB structures, used to locate a HOB of interest (based on UUID). In such > cases, instead of passing a single HOB, the boot phases may rely on passing > the pointer to the HOB table. > > > > This has been extremely useful concept to employ on highly configurable > systems that must rely on flexible discovery mechanisms to initialize and > boot the system. This is especially helpful when you have multiple > > > > > > Why do we need HOBs in TF-A?: > > ----------------------------- > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC in > a way that is SoC specific *but* platform agnostic. This means that a > single ARM SoC that a SiP may deliver to customers may provide a single > TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a broad > range of platform designs and configurations in order to boot a platform > specific firmware (e.g. BL33 and possibly even BL32 code). In order to > achieve this, the platform configuration must be *discovered* instead of > statically compiled as it is today in TF-A via device tree based > enumeration. The mechanisms of discovery may differ broadly depending on > the relevant industry standard, or in some cases may have rely on SiP > specific discovery flows. > > > > For example: On server systems that support a broad range DIMM memory > population/topologies, all the necessary information required to boot is > fully discovered via standard JEDEC Serial Presence Detect (SPD) over an > I2C bus. Leveraging the SPD bus, may platform variants could be supported > with a single TF-A binary. Not only is this information required to > initialize memory in early boot phases (e.g. BL2), the subsequent boot > phases will also need this SPD info to construct a system physical address > map and properly initialize the MMU based on the memory present, and where > the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) may > need to generate standard firmware tables to the operating systems, such as > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. SLIT, > SRAT, even NFIT if NVDIMM's are present). > > > > In short, it all starts with a standardized or vendor specific discovery > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of > information to the next boot stages (e.g. BL31/BL32/BL33). > > > > Today, every HOB may be a vendor specific structure, but in the future > there may be benefit of defining standard HOBs. This may be useful for > memory discovery, passing the system physical address map, enabling TPM > measured boot, and potentially many other common HOB use-cases. > > > > It would be extremely beneficial to the datacenter market segment if the > TF-A community would adopt this concept of information passing between all > boot phases as opposed to rely solely on device tree enumeration. This is > not intended to replace device tree, rather intended as an alternative way > to describe the info that must be discovered and dynamically generated. > > > > > > Conclusion: > > ----------- > > We are proposing that the TF-A community begin pursuing the adoption of > HOBs as a mechanism used for information exchange between each boot stage > (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term we > want to explore standardizing some HOB structures for the BL33 phase (e.g. > UEFI HOB structures), but initially would like to agree on this being a > useful mechanism used to pass information between each boot stage. > > > > Thanks, > > --Harb > > > > > > > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a > -- François-Frédéric Ozog | *Director Linaro Edge & Fog Computing Group* T: +33.67221.6485 francois.ozog(a)linaro.org | Skype: ffozog

4 years, 7 months

2
2
0 0

Re: [TF-A] EHF + OPTEE on ARM64

by Sandeep Tripathy

Hi Peng, 1-Asynchronous preemption of SP: The long route is to make changes in the dispatcher and the corresponding SPD implementation to have synchronous preemption. ie: OP-TEE dispatcher will implement a G1NS (fiq) handler and invoke an entry of OP-TEE synchronously. OP-TEE will save the thread context and return. I did some POC but the complexity and effort to generalise was not justified by our requirement at that point especially envisioning the movement to SPMD in future. 2-Synchronous preemption of SP: ref: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6345 I used this approach instead to unblock OP-TEE work alongside EHF. This serves the purpose without changing the routing model with a limitation that non yielding/fast SMC can not be preempted. And ofcourse OP-TEE can mask G0 interrupt in anycase. But I think this is sufficient for your purpose. Please feedback if the above patch works for you. Thanks Sandeep On Mon, Mar 22, 2021 at 2:43 PM Peng Fan via TF-A <tf-a(a)lists.trustedfirmware.org> wrote: > > Hi Achin, > > > > We are using SDEI for Jailhouse hypervisor to minimize interrupt latency, however we also wanna use OP-TEE when SDEI enabled. > > > > So I wanna how to make both work together. > > > > Thanks, > > Peng. > > > > From: Achin Gupta [mailto:Achin.Gupta@arm.com] > Sent: 2021年3月17日 17:59 > To: Peng Fan <peng.fan(a)nxp.com>; Jens Wiklander <jens.wiklander(a)linaro.org> > Cc: op-tee(a)lists.trustedfirmware.org; tf-a(a)lists.trustedfirmware.org > Subject: Re: EHF + OPTEE on ARM64 > > > > Hi Peng, > > > > +TF-A folk. > > > > My 0.02$. > > > > What is the problem you are trying to solve? Why do you need to run OP-TEE and EHF together? EHF was originally written to support a S-EL0 SP that is managed directly by TF-A in EL3 (TF-A folk can chime in). > > > > The SP could perform RAS error handling for which it needs the EHF. The EHF triages asynchronous exceptions and hands RAS errors to the SP for further handling. > > > > This is just one use case but there is no Trusted OS in these configurations. > > > > So, it would help to understand the requirement. > > > > cheers, > > Achin > > > > ________________________________ > > From: OP-TEE <op-tee-bounces(a)lists.trustedfirmware.org> on behalf of Jens Wiklander via OP-TEE <op-tee(a)lists.trustedfirmware.org> > Sent: 17 March 2021 09:23 > To: Peng Fan <peng.fan(a)nxp.com> > Cc: op-tee(a)lists.trustedfirmware.org <op-tee(a)lists.trustedfirmware.org> > Subject: Re: EHF + OPTEE on ARM64 > > > > On Wed, Mar 17, 2021 at 9:43 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > > > Subject: Re: EHF + OPTEE on ARM64 > > > > > > On Wed, Mar 17, 2021 at 9:02 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > > > > > > > Subject: Re: EHF + OPTEE on ARM64 > > > > > > > > > > On Wed, Mar 17, 2021 at 8:41 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > > > > > > > > > > > Subject: Re: EHF + OPTEE on ARM64 > > > > > > > > > > > > > > On Tue, Mar 16, 2021 at 11:08 AM Peng Fan <peng.fan(a)nxp.com> > > > wrote: > > > > > > > > > > > > > > > > Hi, > > > > > > > > > > > > > > > > In bl31/ehf.c, there are following two lines, per my > > > > > > > > understanding, when cpu is in secure world, the non-secure > > > > > > > > interrupt as FIQ(GICv3) will be directly catched by EL3, not S-EL1 > > > > > > > > /* Route EL3 interrupts when in Secure and Non-secure. > > > */ > > > > > > > > set_interrupt_rm_flag(flags, NON_SECURE); > > > > > > > > set_interrupt_rm_flag(flags, SECURE); > > > > > > > > > > > > > > > > So this will conflict with OP-TEE, because OP-TEE needs catch > > > > > > > > NS-interrupt as FIQ in S-EL1 world. > > > > > > > > > > > > > > In the case of GICv3, OP-TEE is configured to receive the > > > > > > > non-secure interrupts as FIQ and secure interrupts as IRQ. See > > > CFG_ARM_GICV3. > > > > > > > > > > > > But EHF needs NS-interrupt FIQ be catched by EL3 if I understand > > > > > > correct, per " set_interrupt_rm_flag(flags, SECURE);" > > > > > > > > > > > > So currently EHF could not work together with OP-TEE, right? > > > > > > > > > > To be honest, I'm not completely sure what EHF does. From OP-TEE > > > > > point of view we expect to receive the non-secure interrupts as a > > > > > way of doing a controlled exit. This allows OP-TEE to resume > > > > > execution with a different core on re-entry. If EL3 takes the > > > > > non-secure interrupts directly it will have to make sure to only re-enter > > > OP-TEE on this core as a return from exception. > > > > > > > > Is this easy to be achieved? > > > > > > I don't know, it depends on what you intend to do with this non-secure > > > interrupt. If it's handled at EL3 and then there's a return from exception back > > > to S-EL1 there's likely no harm done. But if there's a world switch involved > > > there might be trouble, OP-TEE might not be in a suitable state for a world > > > switch. > > > > > > > > > > > Or by using opteed_sel1_interrupt_handler, could we have similar > > > > behavior to allow the other core resume execution? > > > > > > Only OP-TEE itself can make a controlled exit as there's an internal state to > > > maintain. Currently that's signalled with a non-secure interrupt. > > > > > > Per EHF, https://trustedfirmware-a.readthedocs.io/en/latest/components/exception-han… > > On GICv3 systems, when executing in S-EL1, pending Non-secure interrupts of > > sufficient priority are signalled as FIQs, and therefore will be routed to EL3. > > As a result, S-EL1 software cannot expect to handle Non-secure interrupts at S-EL1. > > Essentially, this deprecates the routing mode described as CSS=0, TEL3=0. > > > > In order for S-EL1 software to handle Non-secure interrupts while having EHF enabled, > > the dispatcher must adopt a model where Non-secure interrupts are received at EL3, > > but are then synchronously handled over to S-EL1. > > > > The issue to me here how to synchronously handled over to S-EL1 and not break optee. > > I understand. OP-TEE is masking interrupts in some critical sections, > while in such a state OP-TEE cannot handle any asynchronous interrupt. > Temporarily masking interrupts is normally a quick operation so we do > it in quite a few places. > So the crux of the problem is to make sure that OP-TEE is in a state > where it can make a controlled exit. I don't have any good ideas for > this right now. > > Cheers, > Jens > > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a

4 years, 7 months

2
1
0 0

Linaro Virtual Connect: March 23-25, 2021

by Joanna Farley

This week a 3 day Linaro Virtual Connect event is being held. There are 60+ sessions many of which may be of interest to the project community including a number of updated TF-A sessions previously presented at the Tech-Forum. The Linaro Virtual Connect schedule is available here<https://connect.linaro.org/schedule/>. Virtual Connect notes: * Free Register here<https://connect.linaro.org/>. * The virtual sessions occur across various time-zones, but all sessions will be recorded and published shortly after the event for you to be able to watch later. Thanks Joanna

4 years, 7 months

1
0
0 0

CANCELLED: TF-A Tech Forum Agenda @ Thr 25th March 2021 16:00-1700 GMT

by Joanna Farley

I am cancelling this weeks TF-A Tech forum Although I had hoped to have a session ready for this week unfortunately that is not the case. However, this week there is a three day Linaro Virtual Connect event March 23-25 where re-runs of a number of previous TF-A Tech Forum sessions are being performed with updated information in a number of cases. Please see the following email with details of the Linaro Virtual Connect event. Joanna

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by raghu.ncstate＠icloud.com

+tf-a list. -----Original Message----- From: raghu.ncstate(a)icloud.com <raghu.ncstate(a)icloud.com> Sent: Monday, March 22, 2021 7:21 AM To: 'Grant Likely' <grant.likely(a)arm.com>; 'Harb Abdulhamid OS' <abdulhamid(a)os.amperecomputing.com>; 'Stuart Yoder' <stuart.yoder(a)arm.com>; 'Jose Marinho' <Jose.Marinho(a)arm.com> Subject: RE: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages I'm also in favor of the proposed method as an alternative(not replace) to the fconf/device tree based method, which works well for vertically integrated systems but not so for systems like Harb has mentioned below. Stuffing/modifying device tree on the fly is awkward even for small pieces of data and not everybody(at least me) would be happy with including something like a device tree library in early boot loader stages and firmware for various reasons(complexity, avoid parsing code for security reasons). Thanks Raghu -----Original Message----- From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> On Behalf Of Grant Likely via TF-A Sent: Monday, March 22, 2021 4:03 AM To: Harb Abdulhamid OS <abdulhamid(a)os.amperecomputing.com>; tf-a(a)lists.trustedfirmware.org; Stuart Yoder <stuart.yoder(a)arm.com>; Jose Marinho <Jose.Marinho(a)arm.com> Subject: Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages Hi Harb, This sounds like a useful abstraction to me. I can see it being useful when we need to pass TPM logs from one stage to another, or to pass on firmware update status. Things that /could/ be stuffed into a single devicetree, but it is awkward to rewrite the devicetree for every piece of dynamic data that gets generated and passed on. It would also be helpful if a common approach can be used regardless of the normal-world firmware (i.e., EDK2, U-Boot, or something else). g. On 22/03/2021 08:34, Harb Abdulhamid OS via TF-A wrote: > Hello Folks, > > I'm emailing to start an open discussion about the adoption of a > concept known as "hand-off blocks" or HOB to become a part of the TF-A > Firmware Framework Architecture (FFA).ï¿½ This is something that is a > pretty major pain point when it comes to the adoption of TF-A in ARM > Server SoCï¿½s designed to enable a broad range of highly configurable > datacenter platforms. > > What is a HOB (Background)? > > --------------------------- > > UEFI PI spec describes a particular definition for how HOB may be used > for transitioning between the PEI and DXE boot phases, which is a good > reference point for this discussion, but not necessarily the exact > solution appropriate for TF-A. > > A HOB is simply a dynamically generated data structure passed in > between two boot phases.ï¿½ This is information that was obtained > through discovery and needs to be passed forward to the next boot > phase *once*, with no API needed to call back (e.g. no call back into > previous firmware phase is needed to fetch this information at > run-time - it is simply passed one time during boot). > > There may be one or more HOBs passed in between boot phases.ï¿½ If > there are more than one HOB that needs to be passed, this can be in a > form of a "HOB table", which (for example) could be a UUID indexed > array of pointers to HOB structures, used to locate a HOB of interest > (based on UUID).ï¿½ In such cases, instead of passing a single HOB, > the boot phases may rely on passing the pointer to the HOB table. > > This has been extremely useful concept to employ on highly > configurable systems that must rely on flexible discovery mechanisms > to initialize and boot the system.ï¿½ This is especially helpful when > you have multiple > > Why do we need HOBs in TF-A?: > > ----------------------------- > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC > in a way that is SoC specific *but* platform agnostic.ï¿½ This means > that a single ARM SoC that a SiP may deliver to customers may provide > a single TF-A binary (e.g. BL1, BL2, BL31) that could be used to > support a broad range of platform designs and configurations in order > to boot a platform specific firmware (e.g. BL33 and possibly even BL32 > code).ï¿½ In order to achieve this, the platform configuration must be > *discovered* instead of statically compiled as it is today in TF-A via > device tree based enumeration.ï¿½ The mechanisms of discovery may > differ broadly depending on the relevant industry standard, or in some > cases may have rely on SiP specific discovery flows. > > For example:ï¿½ On server systems that support a broad range DIMM > memory population/topologies, all the necessary information required > to boot is fully discovered via standard JEDEC Serial Presence Detect > (SPD) over an I2C bus.ï¿½ Leveraging the SPD bus, may platform > variants could be supported with a single TF-A binary.ï¿½ Not only is > this information required to initialize memory in early boot phases > (e.g. BL2), the subsequent boot phases will also need this SPD info to > construct a system physical address map and properly initialize the > MMU based on the memory present, and where the memory may be > present.ï¿½ Subsequent boot phases (e.g. BL33 / UEFI) may need to > generate standard firmware tables to the operating systems, such as > SMBIOS tables describing DIMM topology and various ACPI tables (e.g. > SLIT, SRAT, even NFIT if NVDIMM's are present). > > In short, it all starts with a standardized or vendor specific > discovery flow in an early boot stage (e.g. BL1/BL2), followed by the > passing of information to the next boot stages (e.g. BL31/BL32/BL33). > > Today, every HOB may be a vendor specific structure, but in the future > there may be benefit of defining standard HOBs.ï¿½ This may be useful > for memory discovery, passing the system physical address map, > enabling TPM measured boot, and potentially many other common HOB use-cases. > > It would be extremely beneficial to the datacenter market segment if > the TF-A community would adopt this concept of information passing > between all boot phases as opposed to rely solely on device tree enumeration. > This is not intended to replace device tree, rather intended as an > alternative way to describe the info that must be discovered and > dynamically generated. > > Conclusion: > > ----------- > > We are proposing that the TF-A community begin pursuing the adoption > of HOBs as a mechanism used for information exchange between each boot > stage (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer > term we want to explore standardizing some HOB structures for the BL33 > phase (e.g. UEFI HOB structures), but initially would like to agree on > this being a useful mechanism used to pass information between each > boot stage. > > Thanks, > > --Harb > > -- TF-A mailing list TF-A(a)lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-a

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Grant Likely

Hi Harb, This sounds like a useful abstraction to me. I can see it being useful when we need to pass TPM logs from one stage to another, or to pass on firmware update status. Things that /could/ be stuffed into a single devicetree, but it is awkward to rewrite the devicetree for every piece of dynamic data that gets generated and passed on. It would also be helpful if a common approach can be used regardless of the normal-world firmware (i.e., EDK2, U-Boot, or something else). g. On 22/03/2021 08:34, Harb Abdulhamid OS via TF-A wrote: > Hello Folks, > > I'm emailing to start an open discussion about the adoption of a concept > known as "hand-off blocks" or HOB to become a part of the TF-A Firmware > Framework Architecture (FFA).ï¿½ This is something that is a pretty major > pain point when it comes to the adoption of TF-A in ARM Server SoCï¿½s > designed to enable a broad range of highly configurable datacenter > platforms. > > What is a HOB (Background)? > > --------------------------- > > UEFI PI spec describes a particular definition for how HOB may be used > for transitioning between the PEI and DXE boot phases, which is a good > reference point for this discussion, but not necessarily the exact > solution appropriate for TF-A. > > A HOB is simply a dynamically generated data structure passed in between > two boot phases.ï¿½ This is information that was obtained through > discovery and needs to be passed forward to the next boot phase *once*, > with no API needed to call back (e.g. no call back into previous > firmware phase is needed to fetch this information at run-time - it is > simply passed one time during boot). > > There may be one or more HOBs passed in between boot phases.ï¿½ If there > are more than one HOB that needs to be passed, this can be in a form of > a "HOB table", which (for example) could be a UUID indexed array of > pointers to HOB structures, used to locate a HOB of interest (based on > UUID).ï¿½ In such cases, instead of passing a single HOB, the boot phases > may rely on passing the pointer to the HOB table. > > This has been extremely useful concept to employ on highly configurable > systems that must rely on flexible discovery mechanisms to initialize > and boot the system.ï¿½ This is especially helpful when you have multiple > > Why do we need HOBs in TF-A?: > > ----------------------------- > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC > in a way that is SoC specific *but* platform agnostic.ï¿½ This means that > a single ARM SoC that a SiP may deliver to customers may provide a > single TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a > broad range of platform designs and configurations in order to boot a > platform specific firmware (e.g. BL33 and possibly even BL32 code).ï¿½ In > order to achieve this, the platform configuration must be *discovered* > instead of statically compiled as it is today in TF-A via device tree > based enumeration.ï¿½ The mechanisms of discovery may differ broadly > depending on the relevant industry standard, or in some cases may have > rely on SiP specific discovery flows. > > For example:ï¿½ On server systems that support a broad range DIMM memory > population/topologies, all the necessary information required to boot is > fully discovered via standard JEDEC Serial Presence Detect (SPD) over an > I2C bus.ï¿½ Leveraging the SPD bus, may platform variants could be > supported with a single TF-A binary.ï¿½ Not only is this information > required to initialize memory in early boot phases (e.g. BL2), the > subsequent boot phases will also need this SPD info to construct a > system physical address map and properly initialize the MMU based on the > memory present, and where the memory may be present.ï¿½ Subsequent boot > phases (e.g. BL33 / UEFI) may need to generate standard firmware tables > to the operating systems, such as SMBIOS tables describing DIMM topology > and various ACPI tables (e.g. SLIT, SRAT, even NFIT if NVDIMM's are > present). > > In short, it all starts with a standardized or vendor specific discovery > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of > information to the next boot stages (e.g. BL31/BL32/BL33). > > Today, every HOB may be a vendor specific structure, but in the future > there may be benefit of defining standard HOBs.ï¿½ This may be useful for > memory discovery, passing the system physical address map, enabling TPM > measured boot, and potentially many other common HOB use-cases. > > It would be extremely beneficial to the datacenter market segment if the > TF-A community would adopt this concept of information passing between > all boot phases as opposed to rely solely on device tree enumeration. > This is not intended to replace device tree, rather intended as an > alternative way to describe the info that must be discovered and > dynamically generated. > > Conclusion: > > ----------- > > We are proposing that the TF-A community begin pursuing the adoption of > HOBs as a mechanism used for information exchange between each boot > stage (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer > term we want to explore standardizing some HOB structures for the BL33 > phase (e.g. UEFI HOB structures), but initially would like to agree on > this being a useful mechanism used to pass information between each boot > stage. > > Thanks, > > --Harb > >

4 years, 7 months

1
0
0 0

Re: [TF-A] Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Grant Likely

Hi Harb, This sounds like a useful abstraction to me. I can see it being useful when we need to pass TPM logs from one stage to another, or to pass on firmware update status. Things that /could/ be stuffed into a single devicetree, but it is awkward to rewrite the devicetree for every piece of dynamic data that gets generated and passed on. It would also be helpful if a common approach can be used regardless of the normal-world firmware (i.e., EDK2, U-Boot, or something else). g. On 22/03/2021 08:34, Harb Abdulhamid OS via TF-A wrote: > Hello Folks, > > I'm emailing to start an open discussion about the adoption of a concept > known as "hand-off blocks" or HOB to become a part of the TF-A Firmware > Framework Architecture (FFA).ï¿½ This is something that is a pretty major > pain point when it comes to the adoption of TF-A in ARM Server SoCï¿½s > designed to enable a broad range of highly configurable datacenter > platforms. > > What is a HOB (Background)? > > --------------------------- > > UEFI PI spec describes a particular definition for how HOB may be used > for transitioning between the PEI and DXE boot phases, which is a good > reference point for this discussion, but not necessarily the exact > solution appropriate for TF-A. > > A HOB is simply a dynamically generated data structure passed in between > two boot phases.ï¿½ This is information that was obtained through > discovery and needs to be passed forward to the next boot phase *once*, > with no API needed to call back (e.g. no call back into previous > firmware phase is needed to fetch this information at run-time - it is > simply passed one time during boot). > > There may be one or more HOBs passed in between boot phases.ï¿½ If there > are more than one HOB that needs to be passed, this can be in a form of > a "HOB table", which (for example) could be a UUID indexed array of > pointers to HOB structures, used to locate a HOB of interest (based on > UUID).ï¿½ In such cases, instead of passing a single HOB, the boot phases > may rely on passing the pointer to the HOB table. > > This has been extremely useful concept to employ on highly configurable > systems that must rely on flexible discovery mechanisms to initialize > and boot the system.ï¿½ This is especially helpful when you have multiple > > Why do we need HOBs in TF-A?: > > ----------------------------- > > It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC > in a way that is SoC specific *but* platform agnostic.ï¿½ This means that > a single ARM SoC that a SiP may deliver to customers may provide a > single TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a > broad range of platform designs and configurations in order to boot a > platform specific firmware (e.g. BL33 and possibly even BL32 code).ï¿½ In > order to achieve this, the platform configuration must be *discovered* > instead of statically compiled as it is today in TF-A via device tree > based enumeration.ï¿½ The mechanisms of discovery may differ broadly > depending on the relevant industry standard, or in some cases may have > rely on SiP specific discovery flows. > > For example:ï¿½ On server systems that support a broad range DIMM memory > population/topologies, all the necessary information required to boot is > fully discovered via standard JEDEC Serial Presence Detect (SPD) over an > I2C bus.ï¿½ Leveraging the SPD bus, may platform variants could be > supported with a single TF-A binary.ï¿½ Not only is this information > required to initialize memory in early boot phases (e.g. BL2), the > subsequent boot phases will also need this SPD info to construct a > system physical address map and properly initialize the MMU based on the > memory present, and where the memory may be present.ï¿½ Subsequent boot > phases (e.g. BL33 / UEFI) may need to generate standard firmware tables > to the operating systems, such as SMBIOS tables describing DIMM topology > and various ACPI tables (e.g. SLIT, SRAT, even NFIT if NVDIMM's are > present). > > In short, it all starts with a standardized or vendor specific discovery > flow in an early boot stage (e.g. BL1/BL2), followed by the passing of > information to the next boot stages (e.g. BL31/BL32/BL33). > > Today, every HOB may be a vendor specific structure, but in the future > there may be benefit of defining standard HOBs.ï¿½ This may be useful for > memory discovery, passing the system physical address map, enabling TPM > measured boot, and potentially many other common HOB use-cases. > > It would be extremely beneficial to the datacenter market segment if the > TF-A community would adopt this concept of information passing between > all boot phases as opposed to rely solely on device tree enumeration. > This is not intended to replace device tree, rather intended as an > alternative way to describe the info that must be discovered and > dynamically generated. > > Conclusion: > > ----------- > > We are proposing that the TF-A community begin pursuing the adoption of > HOBs as a mechanism used for information exchange between each boot > stage (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer > term we want to explore standardizing some HOB structures for the BL33 > phase (e.g. UEFI HOB structures), but initially would like to agree on > this being a useful mechanism used to pass information between each boot > stage. > > Thanks, > > --Harb > >

4 years, 7 months

1
0
0 0

Re: [TF-A] EHF + OPTEE on ARM64

by Achin Gupta

Hi Peng, +TF-A folk. My 0.02$. What is the problem you are trying to solve? Why do you need to run OP-TEE and EHF together? EHF was originally written to support a S-EL0 SP that is managed directly by TF-A in EL3 (TF-A folk can chime in). The SP could perform RAS error handling for which it needs the EHF. The EHF triages asynchronous exceptions and hands RAS errors to the SP for further handling. This is just one use case but there is no Trusted OS in these configurations. So, it would help to understand the requirement. cheers, Achin ________________________________ From: OP-TEE <op-tee-bounces(a)lists.trustedfirmware.org> on behalf of Jens Wiklander via OP-TEE <op-tee(a)lists.trustedfirmware.org> Sent: 17 March 2021 09:23 To: Peng Fan <peng.fan(a)nxp.com> Cc: op-tee(a)lists.trustedfirmware.org <op-tee(a)lists.trustedfirmware.org> Subject: Re: EHF + OPTEE on ARM64 On Wed, Mar 17, 2021 at 9:43 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > Subject: Re: EHF + OPTEE on ARM64 > > > > On Wed, Mar 17, 2021 at 9:02 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > > > > > Subject: Re: EHF + OPTEE on ARM64 > > > > > > > > On Wed, Mar 17, 2021 at 8:41 AM Peng Fan <peng.fan(a)nxp.com> wrote: > > > > > > > > > > > Subject: Re: EHF + OPTEE on ARM64 > > > > > > > > > > > > On Tue, Mar 16, 2021 at 11:08 AM Peng Fan <peng.fan(a)nxp.com> > > wrote: > > > > > > > > > > > > > > Hi, > > > > > > > > > > > > > > In bl31/ehf.c, there are following two lines, per my > > > > > > > understanding, when cpu is in secure world, the non-secure > > > > > > > interrupt as FIQ(GICv3) will be directly catched by EL3, not S-EL1 > > > > > > > /* Route EL3 interrupts when in Secure and Non-secure. > > */ > > > > > > > set_interrupt_rm_flag(flags, NON_SECURE); > > > > > > > set_interrupt_rm_flag(flags, SECURE); > > > > > > > > > > > > > > So this will conflict with OP-TEE, because OP-TEE needs catch > > > > > > > NS-interrupt as FIQ in S-EL1 world. > > > > > > > > > > > > In the case of GICv3, OP-TEE is configured to receive the > > > > > > non-secure interrupts as FIQ and secure interrupts as IRQ. See > > CFG_ARM_GICV3. > > > > > > > > > > But EHF needs NS-interrupt FIQ be catched by EL3 if I understand > > > > > correct, per " set_interrupt_rm_flag(flags, SECURE);" > > > > > > > > > > So currently EHF could not work together with OP-TEE, right? > > > > > > > > To be honest, I'm not completely sure what EHF does. From OP-TEE > > > > point of view we expect to receive the non-secure interrupts as a > > > > way of doing a controlled exit. This allows OP-TEE to resume > > > > execution with a different core on re-entry. If EL3 takes the > > > > non-secure interrupts directly it will have to make sure to only re-enter > > OP-TEE on this core as a return from exception. > > > > > > Is this easy to be achieved? > > > > I don't know, it depends on what you intend to do with this non-secure > > interrupt. If it's handled at EL3 and then there's a return from exception back > > to S-EL1 there's likely no harm done. But if there's a world switch involved > > there might be trouble, OP-TEE might not be in a suitable state for a world > > switch. > > > > > > > > Or by using opteed_sel1_interrupt_handler, could we have similar > > > behavior to allow the other core resume execution? > > > > Only OP-TEE itself can make a controlled exit as there's an internal state to > > maintain. Currently that's signalled with a non-secure interrupt. > > > Per EHF, https://trustedfirmware-a.readthedocs.io/en/latest/components/exception-han… > On GICv3 systems, when executing in S-EL1, pending Non-secure interrupts of > sufficient priority are signalled as FIQs, and therefore will be routed to EL3. > As a result, S-EL1 software cannot expect to handle Non-secure interrupts at S-EL1. > Essentially, this deprecates the routing mode described as CSS=0, TEL3=0. > > In order for S-EL1 software to handle Non-secure interrupts while having EHF enabled, > the dispatcher must adopt a model where Non-secure interrupts are received at EL3, > but are then synchronously handled over to S-EL1. > > The issue to me here how to synchronously handled over to S-EL1 and not break optee. I understand. OP-TEE is masking interrupts in some critical sections, while in such a state OP-TEE cannot handle any asynchronous interrupt. Temporarily masking interrupts is normally a quick operation so we do it in quite a few places. So the crux of the problem is to make sure that OP-TEE is in a state where it can make a controlled exit. I don't have any good ideas for this right now. Cheers, Jens

4 years, 7 months

2
1
0 0

Proposal: TF-A to adopt hand-off blocks (HOBs) for information passing between boot stages

by Harb Abdulhamid OS

Hello Folks, I'm emailing to start an open discussion about the adoption of a concept known as "hand-off blocks" or HOB to become a part of the TF-A Firmware Framework Architecture (FFA). This is something that is a pretty major pain point when it comes to the adoption of TF-A in ARM Server SoC's designed to enable a broad range of highly configurable datacenter platforms. What is a HOB (Background)? --------------------------- UEFI PI spec describes a particular definition for how HOB may be used for transitioning between the PEI and DXE boot phases, which is a good reference point for this discussion, but not necessarily the exact solution appropriate for TF-A. A HOB is simply a dynamically generated data structure passed in between two boot phases. This is information that was obtained through discovery and needs to be passed forward to the next boot phase *once*, with no API needed to call back (e.g. no call back into previous firmware phase is needed to fetch this information at run-time - it is simply passed one time during boot). There may be one or more HOBs passed in between boot phases. If there are more than one HOB that needs to be passed, this can be in a form of a "HOB table", which (for example) could be a UUID indexed array of pointers to HOB structures, used to locate a HOB of interest (based on UUID). In such cases, instead of passing a single HOB, the boot phases may rely on passing the pointer to the HOB table. This has been extremely useful concept to employ on highly configurable systems that must rely on flexible discovery mechanisms to initialize and boot the system. This is especially helpful when you have multiple Why do we need HOBs in TF-A?: ----------------------------- It is desirable that EL3 firmware (e.g. TF-A) built for ARM Server SoC in a way that is SoC specific *but* platform agnostic. This means that a single ARM SoC that a SiP may deliver to customers may provide a single TF-A binary (e.g. BL1, BL2, BL31) that could be used to support a broad range of platform designs and configurations in order to boot a platform specific firmware (e.g. BL33 and possibly even BL32 code). In order to achieve this, the platform configuration must be *discovered* instead of statically compiled as it is today in TF-A via device tree based enumeration. The mechanisms of discovery may differ broadly depending on the relevant industry standard, or in some cases may have rely on SiP specific discovery flows. For example: On server systems that support a broad range DIMM memory population/topologies, all the necessary information required to boot is fully discovered via standard JEDEC Serial Presence Detect (SPD) over an I2C bus. Leveraging the SPD bus, may platform variants could be supported with a single TF-A binary. Not only is this information required to initialize memory in early boot phases (e.g. BL2), the subsequent boot phases will also need this SPD info to construct a system physical address map and properly initialize the MMU based on the memory present, and where the memory may be present. Subsequent boot phases (e.g. BL33 / UEFI) may need to generate standard firmware tables to the operating systems, such as SMBIOS tables describing DIMM topology and various ACPI tables (e.g. SLIT, SRAT, even NFIT if NVDIMM's are present). In short, it all starts with a standardized or vendor specific discovery flow in an early boot stage (e.g. BL1/BL2), followed by the passing of information to the next boot stages (e.g. BL31/BL32/BL33). Today, every HOB may be a vendor specific structure, but in the future there may be benefit of defining standard HOBs. This may be useful for memory discovery, passing the system physical address map, enabling TPM measured boot, and potentially many other common HOB use-cases. It would be extremely beneficial to the datacenter market segment if the TF-A community would adopt this concept of information passing between all boot phases as opposed to rely solely on device tree enumeration. This is not intended to replace device tree, rather intended as an alternative way to describe the info that must be discovered and dynamically generated. Conclusion: ----------- We are proposing that the TF-A community begin pursuing the adoption of HOBs as a mechanism used for information exchange between each boot stage (e.g. BL1->BL2, BL2->BL31, BL31->BL32, and BL31->BL33)? Longer term we want to explore standardizing some HOB structures for the BL33 phase (e.g. UEFI HOB structures), but initially would like to agree on this being a useful mechanism used to pass information between each boot stage. Thanks, --Harb

4 years, 7 months

1
0
0 0

Re: [TF-A] Deprecating Arm's sgm775 platform

by Joanna Farley

We already have a Deprecated Interfaces removal policy for core interfaces: https://trustedfirmware-a.readthedocs.io/en/latest/about/release-informatio… although this is not quite the same for platforms the proposed purpose is similar of a managed removing of something that someone has a use case for in the short term. The Arm platforms as mentioned by Manish are targeted as reference platforms for other project users to inspect which is perhaps a little different than other platform providers. So I think it is appropriate to have a transition period as those references platforms become outdated and normally replaced as it is with the sgm775 with the tc0 platform. I can understand for other platform providers they might want a more accelerated or removal approach and that’s fine I think. I’m not suggesting we need to create a formal policy but leave it up to individual platform providers to decide how they manage their platforms as they know best how they are used. Ensuring that the OpenCI is not left in a broken state from a build or test perspective during a deprecation period is of course necessary and the steps suggested by Manish look reasonable to me. Hope this helps on the intent here. Joanna From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> on behalf of Varun Wadekar via TF-A <tf-a(a)lists.trustedfirmware.org> Reply to: Varun Wadekar <vwadekar(a)nvidia.com> Date: Thursday, 11 March 2021 at 22:12 To: Manish Pandey2 <Manish.Pandey2(a)arm.com> Cc: "tf-a(a)lists.trustedfirmware.org" <tf-a(a)lists.trustedfirmware.org> Subject: Re: [TF-A] Deprecating Arm's sgm775 platform Hi, I was wondering if there is any value in keeping non-compiling code in the tree, hence the question. For Tegra it is straight forward, so it would be easy to pull the plug. -Varun From: Manish Pandey2 <Manish.Pandey2(a)arm.com> Sent: Thursday, March 11, 2021 2:44 AM To: Varun Wadekar <vwadekar(a)nvidia.com> Cc: tf-a(a)lists.trustedfirmware.org Subject: Re: Deprecating Arm's sgm775 platform External email: Use caution opening links or attachments Hi Varun, For Arm reference platforms, we are not sure who all are using it. That is why i proposed to keep in repo for around a year before deleting it. But for NV platforms, if you are sure that nobody is going to require it, you can delete it earlier. Also, instead of proposed step 2 earlier we can have an alternate - 2. Don't allow it to be built by default. (introducing PLATFORM_DEPRECATED build macro) + 2. Instead of purposefully failing the build we can print a warning message (It's also quite possible that someday during cooling off period it stops to build naturally) thanks Manish ________________________________ From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Sent: 11 March 2021 02:35 To: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Cc: tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> Subject: RE: Deprecating Arm's sgm775 platform Hi Manish, Just curious, what would be the reason to keep the platform alive for 2 release cycles? I have one Tegra platform that needs to be deprecated and so would like to understand the thought process. -Varun From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org<mailto:tf-a-bounces@lists.trustedfirmware.org>> On Behalf Of Manish Pandey2 via TF-A Sent: Tuesday, March 9, 2021 2:01 PM To: tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Subject: [TF-A] Deprecating Arm's sgm775 platform External email: Use caution opening links or attachments Hi, The purpose of the email is to notify about deprecation of sgm775 platform and proposed process for deprecating a platform. Arm's System Guidance for Mobile(SGM-775) is an old platform and no longer maintained. It is superseded by Total Compute(tc0) platform. Proposal for deprecating a platform: 1. Keep the code in repository. (at least for 2 release cycles) 2. Don't allow it to be built by default. (introducing PLATFORM_DEPRECATED build macro) 3. Disable CI testing. 4. Create appropriate documentation for deprecated platforms. Let me know if you have any suggestions. Thanks Manish P

4 years, 7 months

1
0
0 0

Re: [TF-A] [EXT] RE: NXP Patch-Set for platform lx2160ardb/lx2160aqds/lx2162aqds

by Pankaj Gupta

Hi, Thank you once again for more of your comments. https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> to https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6167 All the patches from the patch-set, are reviewed and all the comments are disposed-off till date. Please share any further action, I need to work on. Looking forward for your continued support. Regards Pankaj From: Varun Wadekar <vwadekar(a)nvidia.com> Sent: Tuesday, February 2, 2021 12:11 AM To: Pankaj Gupta <pankaj.gupta(a)nxp.com>; Manish Pandey2 <Manish.Pandey2(a)arm.com>; Olivier Deprez <Olivier.Deprez(a)arm.com>; javier.almansasobrino(a)arm.com; jimmy.brisson(a)arm.com Cc: Joanna Farley <Joanna.Farley(a)arm.com>; Alexei Fedorov <Alexei.Fedorov(a)arm.com>; Madhukar Pappireddy <Madhukar.Pappireddy(a)arm.com> Subject: [EXT] RE: NXP Patch-Set for platform lx2160ardb/lx2160aqds/lx2162aqds Caution: EXT Email Thanks Pankaj. I don't have further comments. From: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Sent: Sunday, January 31, 2021 11:14 PM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>>; Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>>; Olivier Deprez <Olivier.Deprez(a)arm.com<mailto:Olivier.Deprez@arm.com>>; javier.almansasobrino(a)arm.com<mailto:javier.almansasobrino@arm.com>; jimmy.brisson(a)arm.com<mailto:jimmy.brisson@arm.com> Cc: Joanna Farley <Joanna.Farley(a)arm.com<mailto:Joanna.Farley@arm.com>>; Alexei Fedorov <Alexei.Fedorov(a)arm.com<mailto:Alexei.Fedorov@arm.com>>; Madhukar Pappireddy <Madhukar.Pappireddy(a)arm.com<mailto:Madhukar.Pappireddy@arm.com>> Subject: NXP Patch-Set for platform lx2160ardb/lx2160aqds/lx2162aqds External email: Use caution opening links or attachments Hi all, Thanks to all of you, for your efforts in reviewing the patches. All the patches from the patch-set, are reviewed and all the comments are disposed-off till date. https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> to https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6167<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> Please share any further action, I need to work on. Looking forward for your continued support. Thanks & regards Pankaj From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Sent: Friday, December 4, 2020 10:37 PM To: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>>; Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Cc: Joanna Farley <Joanna.Farley(a)arm.com<mailto:Joanna.Farley@arm.com>>; Alexei Fedorov <Alexei.Fedorov(a)arm.com<mailto:Alexei.Fedorov@arm.com>>; Madhukar Pappireddy <Madhukar.Pappireddy(a)arm.com<mailto:Madhukar.Pappireddy@arm.com>> Subject: RE: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> Caution: EXT Email Thanks Pankaj and Manish. Glad to see us agreeing to a path forward. From: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Sent: Thursday, December 3, 2020 11:39 PM To: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>>; Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Cc: Joanna Farley <Joanna.Farley(a)arm.com<mailto:Joanna.Farley@arm.com>>; Alexei Fedorov <Alexei.Fedorov(a)arm.com<mailto:Alexei.Fedorov@arm.com>>; Madhukar Pappireddy <Madhukar.Pappireddy(a)arm.com<mailto:Madhukar.Pappireddy@arm.com>> Subject: RE: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> External email: Use caution opening links or attachments Hi, Thanks for the email. Your suggestion is good too. But with your suggestion, two lines are need for one line using the macro. My concern is that, there are 6 SoC and 15+ platforms based on these 6 SoC, still pending to be added to this code base. Keeping things simple in soc.mk and platform.mk is of key importance. To achieve it: * Complexity of source file addition is moved away from platform makefile to: o drivers/nxp/driver.mk, and o drivers/nxp/<ip>/<ip>.mk * As a result, the soc.mk & platform.mk is less cluttered. o With the single line addition based on this macro, it is easier to understand, which IP is part of BL2, BL31 or both. o All the complexity about IP files to be included or not is moved to drivers/nxp/<ip>/<ip>.mk. I got your point here. Since it is very much specific to NXP platforms, it is to be moved to "plat/nxp". I have tried this way. It is working and ready for review. Regards Pankaj From: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Sent: Friday, December 4, 2020 6:24 AM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>>; Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Cc: Joanna Farley <Joanna.Farley(a)arm.com<mailto:Joanna.Farley@arm.com>>; Alexei Fedorov <Alexei.Fedorov(a)arm.com<mailto:Alexei.Fedorov@arm.com>>; Madhukar Pappireddy <Madhukar.Pappireddy(a)arm.com<mailto:Madhukar.Pappireddy@arm.com>> Subject: Re: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> Caution: EXT Email Hi Pankaj, I am still trying to understand the complete patchset, also trying to understand different level of makefiles e.g drivers/nxp/drivers.mk(is it really needed?) I think the main concern is complexity added by intermediate mk files, this complexity can be reduced by declaring most of things in specific platform mk file as we know at compile time which all images need a particular driver. For e.g XSPI driver, it has same source files in all the 3 scenarios (BL2/BL31 & COMMON) and there is no usage of IMAGE_BL31, IMAGE_BL2 in xspi driver suggesting that Image type does not alters functionality of driver. So, what we can do is keeping driver mk file simpler and other details in platform mk file. flexspi_xor.mk will be like: XSPI_BOOT_SOURCES += $(FLEXSPI_DRIVERS_PATH)/flexspi_nor.c \ ${FLEXSPI_DRIVERS_PATH}/fspi.c ifeq ($(DEBUG),1) XSPI_BOOT_SOURCES += ${FLEXSPI_DRIVERS_PATH}/test_fspi.c endif Platform mk file will be like: (say only bl2 needs xspi) XSPI_NEEDED := yes bl2_sources += ${XSPI_BOOT_SOURCES} Regarding NEED_BL31/NEED_BL2, these flags tell if binary for given BL stage needs to be generated or not. Finally, I am not saying that your approach is wrong but what i suggested is currently done by most of the platforms. Also, see my comments on your SET_FLAG patch, if we indeed decide to go ahead with your current approach. Thanks Manish ________________________________ From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Sent: 03 December 2020 17:35 To: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Cc: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Subject: RE: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> Hi, Can you point me to a change that uses the newly introduced makefile macro? IIUC, you just need a way to differentiate between a BL2 v BL31 build. Manish can you confirm if NEED_BL31 and NEED_BL2 can be helpful in this case? >> I will replace the SET_FLAG macro with these two flags from entire source tree. I suppose you are alluding to the NXP platform port only. Correct? -Varun From: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Sent: Thursday, December 3, 2020 12:57 AM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Cc: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Subject: RE: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> External email: Use caution opening links or attachments Hi Varun, Thanks for your email. NXP make files are implemented as : 1. Each IP driver has its own .mk file. 2. Each platform has its own "platform.mk". * Every "platform.mk" includes "soc.mk" for the SoC on which this platform is based. 1. Based on the SoC, mandatory IP(s) are included(soc.mk) to BL2 or BL31 or both. * This requires to pass the flag "<BL2 or BL31 or BL_COMM>_<IP>_NEEDED = yes" from soc.mk to mandatory IP(s) .mk. 1. But for certain IP(s), IP file sources inclusions to either BL2 or BL31, is based on optional features. * This also requires to pass the flag to IP(s) .mk. For instance: * If flexspi_nor as a boot-source, This macro sets 2 flags. i. XSPI_NEEDED = yes * XSPI_NEEDED help identify if the xspi.mk needs to be included or not. ii. BL2_ XSPI_NEEDED = yes * XSPI_NEEDED needs to be included in BL2 * In optional feature WARM_RESET, I need to save the last reset_cause in flexspi_nor in BL31. i. Again need two flags: XSPI_NEEDED = yes & BL31_ XSPI_NEEDED = yes ii. XSPI_NEEDED = yes, is still required as it might happen the boot source is SD. * In this case BL2_XSPI_NEEDED, is not set. What I am gaining with this macros is: * Without this macro, I need to add two lines: * <IP>_NEEDED = yes. * To include this driver. * One of the flag to be added depending on the IP source file inclusion to BL2 or BL31 or BL_COMM: * Flag as BL2_<IP>_NEEDED = yes * Flag as BL31_<IP>_NEEDED = yes. * Flag as BLCOMM_<IP>_NEEDED = yes. * This macros helps me add one line $(eval $(call SET_FLAG, <IP>_NEEDED,BL2)), instead of above two lines. If you suggest to remove the SET_FLAG macro, I will replace the SET_FLAG macro with these two flags from entire source tree. Please share your view. I reviewed the usage of NEED_BL31 and NEED_BL2. They are set to 'yes' in ./Makefile, only when BL2_SOURCES & BL31_SOURCES is non-null; They can be over-ridden, but cannot be used for each IP(s) source file. Regards Pankaj From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Sent: Thursday, December 3, 2020 4:17 AM To: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Cc: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Subject: [EXT] RE: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> Caution: EXT Email Hi, >From your description, I think you are including different makefiles from NXP tree for BL31 versus BL2. Can you please help me understand why this decision needs to be taken at the top level makefile? Alternatively, why can't NXP platform decide which files to include? Did you get a chance to review the NEED_BL31, NEED_BL2 makefile variables? >From the gerrit change it is very difficult to understand how the newly introduced macro is used, so trying to suggest already available options to see if they work for you. -Varun From: Pankaj Gupta <pankaj.gupta(a)nxp.com<mailto:pankaj.gupta@nxp.com>> Sent: Wednesday, December 2, 2020 5:21 AM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Cc: Manish Pandey2 <Manish.Pandey2(a)arm.com<mailto:Manish.Pandey2@arm.com>> Subject: https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6122<https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Freview.tr…> External email: Use caution opening links or attachments Hi Varun, As suggested by you, IMAGE_BL31, IMAGE_BL2 etc macros are useful for segregating the code within the same source file. But this will not serve the purpose in our case. Let me try to explain you with an example: -- For one platform; and for a particulate IP driver's source file, if it is true to have in both, then : #if defined(IMAGE_BL2) || #if defined(IMAGE_BL31) -- But for another platform and for the same IP driver's source file, if it is true to have in BL2 only, then: #if defined(IMAGE_BL2) -- And for another SoC and for the same IP driver's source file, it if is true to have in BL31 only, then: #if defined(IMAGE_BL31) It will not be possible to write all the three varying inclusion of code for one IP used across multiple SoC and their platforms. Now, I will explain how this macro helps me with above case: * Taking an example of flexspi_nor as a boot-source: * $(eval $(call SET_FLAG, XSPI_NEEDED,BL2)) * This macro sets 2 flags. * XSPI_NEEDED = yes * XSPI_NEEDED help identify if the xspi.mk needs to be included or not. * BL2_ XSPI_NEEDED = yes * XSPI_NEEDED needs to be included in BL2. * For a conditional feature to enable in BL31, XSPI is to be included in BL31. * BL31_XSPI_NEEDED = yes needs to be set. * The correct solution should be if the feature is enabled, then: $(eval $(call SET_FLAG, XSPI_NEEDED,BL31)) * In this case, I cannot set BL_COMM_XSPI_NEED = yes for all the platforms. I hope, I am able to convey my thoughts to you. This macro is very important for my code orientation. If you think it will not be required by other contributors, then lets rename this macro by prepending it with NXP or any name you suggests. Thanks. Regards Pankaj IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.

4 years, 7 months

1
1
0 0

Re: [TF-A] Deprecating Arm's sgm775 platform

by Varun Wadekar

Hi Manish, Just curious, what would be the reason to keep the platform alive for 2 release cycles? I have one Tegra platform that needs to be deprecated and so would like to understand the thought process. -Varun From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> On Behalf Of Manish Pandey2 via TF-A Sent: Tuesday, March 9, 2021 2:01 PM To: tf-a(a)lists.trustedfirmware.org Subject: [TF-A] Deprecating Arm's sgm775 platform External email: Use caution opening links or attachments Hi, The purpose of the email is to notify about deprecation of sgm775 platform and proposed process for deprecating a platform. Arm's System Guidance for Mobile(SGM-775) is an old platform and no longer maintained. It is superseded by Total Compute(tc0) platform. Proposal for deprecating a platform: 1. Keep the code in repository. (at least for 2 release cycles) 2. Don't allow it to be built by default. (introducing PLATFORM_DEPRECATED build macro) 3. Disable CI testing. 4. Create appropriate documentation for deprecated platforms. Let me know if you have any suggestions. Thanks Manish P

4 years, 7 months

2
2
0 0

Cancelled event: TF-A Tech Forum @ Thu 11 Mar 2021 16:00 - 17:00 (GMT) (tf-a@lists.trustedfirmware.org)

by Bill Fletcher

This event has been cancelled. Title: TF-A Tech Forum We run an open technical forum call for anyone to participate and it is not restricted to Trusted Firmware project members. It will operate under the guidance of the TF TSC. Feel free to forward this invite to colleagues. Invites are via the TF-A mailing list and also published on the Trusted Firmware website. Details are here: https://www.trustedfirmware.org/meetings/tf-a-technical-forum/Tr… Firmware is inviting you to a scheduled Zoom meeting.Join Zoom Meetinghttps://zoom.us/j/9159704974Meeting ID: 915 970 4974One tap mobile+16465588656,,9159704974# US (New York)+16699009128,,9159704974# US (San Jose)Dial by your location        +1 646 558 8656 US (New York)        +1 669 900 9128 US (San Jose)        877 853 5247 US Toll-free        888 788 0099 US Toll-freeMeeting ID: 915 970 4974Find your local number: https://zoom.us/u/ad27hc6t7h   When: Thu 11 Mar 2021 16:00 – 17:00 United Kingdom Time Calendar: tf-a(a)lists.trustedfirmware.org Who: * Bill Fletcher- creator * marek.bykowski(a)gmail.com * okash.khawaja(a)gmail.com * tf-a(a)lists.trustedfirmware.org Invitation from Google Calendar: https://calendar.google.com/calendar/ You are receiving this courtesy email at the account tf-a(a)lists.trustedfirmware.org because you are an attendee of this event. To stop receiving future updates for this event, decline this event. Alternatively, you can sign up for a Google Account at https://calendar.google.com/calendar/ and control your notification settings for your entire calendar. Forwarding this invitation could allow any recipient to send a response to the organiser and be added to the guest list, invite others regardless of their own invitation status or to modify your RSVP. Learn more at https://support.google.com/calendar/answer/37135#forwarding

4 years, 7 months

1
0
0 0

Re: [TF-A] gerrit issues: not able to add additional email

by Yann Gautier

Hi Gyorgy, I've added a new mail address in gerrit last month. I can see both addresses in gerrit Email Addresses. I've chosen the new one as my preferred address. And I only receive gerrit mails on this preferred address. Best regards, Yann On 3/10/21 4:26 PM, Gyorgy Szing via TF-A wrote: > Hi, > > I just tried it and got the notification e-mail. > > Note: I filled the "New email address" text box and pressing "Send Verification" below. If that adds an additional email or replaces the current one is unknown to me. > > /George > > -----Original Message----- > From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> On Behalf Of Igor Opaniuk via TF-A > Sent: 10 March 2021 13:44 > To: tf-a(a)lists.trustedfirmware.org > Subject: [TF-A] gerrit issues: not able to add additional email > > Hi, > > Gerrit for some reason doesn't send verification emails when adding additional email in account settings [1]. > I've tried to add two different emails, in both cases with no success. > Anyone could quickly check if you can receive verification? > > Thanks > > [1] https://review.trustedfirmware.org/settings#EmailAddresses > > -- > Best regards - Freundliche Grüsse - Meilleures salutations > > Igor Opaniuk > > mailto: igor.opaniuk(a)gmail.com > skype: igor.opanyuk > +380 (93) 836 40 67 > http://ua.linkedin.com/in/iopaniuk > -- > TF-A mailing list > TF-A(a)lists.trustedfirmware.org > https://lists.trustedfirmware.org/mailman/listinfo/tf-a >

4 years, 7 months

2
1
0 0

Re: [TF-A] gerrit issues: not able to add additional email

by Gyorgy Szing

Hi, I just tried it and got the notification e-mail. Note: I filled the "New email address" text box and pressing "Send Verification" below. If that adds an additional email or replaces the current one is unknown to me. /George -----Original Message----- From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> On Behalf Of Igor Opaniuk via TF-A Sent: 10 March 2021 13:44 To: tf-a(a)lists.trustedfirmware.org Subject: [TF-A] gerrit issues: not able to add additional email Hi, Gerrit for some reason doesn't send verification emails when adding additional email in account settings [1]. I've tried to add two different emails, in both cases with no success. Anyone could quickly check if you can receive verification? Thanks [1] https://review.trustedfirmware.org/settings#EmailAddresses -- Best regards - Freundliche Grüsse - Meilleures salutations Igor Opaniuk mailto: igor.opaniuk(a)gmail.com skype: igor.opanyuk +380 (93) 836 40 67 http://ua.linkedin.com/in/iopaniuk -- TF-A mailing list TF-A(a)lists.trustedfirmware.org https://lists.trustedfirmware.org/mailman/listinfo/tf-a

4 years, 7 months

1
0
0 0

gerrit issues: not able to add additional email

by Igor Opaniuk

Hi, Gerrit for some reason doesn't send verification emails when adding additional email in account settings [1]. I've tried to add two different emails, in both cases with no success. Anyone could quickly check if you can receive verification? Thanks [1] https://review.trustedfirmware.org/settings#EmailAddresses -- Best regards - Freundliche Grüsse - Meilleures salutations Igor Opaniuk mailto: igor.opaniuk(a)gmail.com skype: igor.opanyuk +380 (93) 836 40 67 http://ua.linkedin.com/in/iopaniuk

4 years, 7 months

1
0
0 0

CANCELLED: TF-A Tech Forum Agenda @ Thr 11th March 2021 16:00-1700 GMT

by Joanna Farley

Apologies for the late notice I am cancelling this weeks TF-A Tech forum tomorrow as we don’t have any subjects ready to present this week. We expect to have subjects for the next two sessions on 25th March and 8th April. Any subjects for future Tech-Forums from the contributor community always welcome so please reach out and I will help schedule. These can be more formal presentations or led discussions on subjects of interest to the TF-A project community. Cancellation of the calendar invite will come from trustedformware.org as I don’t own the invite so it may not appear in your calendars until that is sent out. Thanks Joanna

4 years, 7 months

1
0
0 0

Deprecating Arm's sgm775 platform

by Manish Pandey2

Hi, The purpose of the email is to notify about deprecation of sgm775 platform and proposed process for deprecating a platform. Arm's System Guidance for Mobile(SGM-775) is an old platform and no longer maintained. It is superseded by Total Compute(tc0) platform. Proposal for deprecating a platform: 1. Keep the code in repository. (at least for 2 release cycles) 2. Don't allow it to be built by default. (introducing PLATFORM_DEPRECATED build macro) 3. Disable CI testing. 4. Create appropriate documentation for deprecated platforms. Let me know if you have any suggestions. Thanks Manish P

4 years, 7 months

1
0
0 0

New Defects reported by Coverity Scan for ARM-software/arm-trusted-firmware

by scan-admin＠coverity.com

Hi, Please find the latest report on new defect(s) introduced to ARM-software/arm-trusted-firmware found with Coverity Scan. 1 new defect(s) introduced to ARM-software/arm-trusted-firmware found with Coverity Scan. New defect(s) Reported-by: Coverity Scan Showing 1 of 1 defect(s) ** CID 367340: Control flow issues (MISSING_BREAK) /plat/mediatek/mt8192/drivers/spm/mt_spm_vcorefs.c: 372 in spm_vcorefs_args() ________________________________________________________________________________________________________ *** CID 367340: Control flow issues (MISSING_BREAK) /plat/mediatek/mt8192/drivers/spm/mt_spm_vcorefs.c: 372 in spm_vcorefs_args() 366 uint64_t spm_vcorefs_args(uint64_t x1, uint64_t x2, uint64_t x3, uint64_t *x4) 367 { 368 uint64_t cmd = x1; 369 uint64_t spm_flags; 370 371 switch (cmd) { >>> CID 367340: Control flow issues (MISSING_BREAK) >>> The case for value "VCOREFS_SMC_CMD_INIT" is not terminated by a "break" statement. 372 case VCOREFS_SMC_CMD_INIT: 373 /* vcore_dvfs init + kick */ 374 mmio_write_32(DVFSRC_SW_REQ5, SW_REQ5_INIT_VAL); 375 spm_dvfsfw_init(0ULL, 0ULL); 376 spm_vcorefs_vcore_setting(x3 & 0xF); 377 spm_flags = SPM_FLAG_RUN_COMMON_SCENARIO; ________________________________________________________________________________________________________ To view the defects in Coverity Scan visit, https://u15810271.ct.sendgrid.net/ls/click?upn=HRESupC-2F2Czv4BOaCWWCy7my0P…

4 years, 8 months

1
0
0 0

Re: [TF-A] Adoption of Conventional Commits

by Joanna Farley

I was away last week for the tech forum and not had time to watch the recording which I need to do before joining some of this discussions but I can clarify the purpose of the Changelog and the difference with the git history. The Changelog is part of the release documentation that summarises the main feature and changes that the release delivers. Its not expected every change in the git log history is included it is about recording the main themes of changes grouped together rather than chronologically. Most importantly it meant to be readable and understandable. I would expect as with any such documentation that some manual editing may be needed but the ask was for suggestions to make this task easier. As it stands today creating the Changelog can take days of manually reading the git history and re-writing, infact it’s the most labour intensive single task we do in a release. So guidance, formatting and tooling in creating commit messages so we can reuse them in release documentation is the ask. If this helps consumption of the git log history for other purposes that’s a great bonus. Tooling and automation makes things more efficient for all. Making submitting patches from developers harder is defiantly not the goal, hopefully the tooling makes the effort easier or at least the same level of effort for the developer, if not the solution being sought needs improving. Now we could do away with the release Changelog as it is today but the git log history is not a replacement for user facing release documentation. Before you know it we do away with all documentation and tell consumers to just read the source code 😉 Quote: “… documentation may mean different things to people in different roles.” Joanna From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org> on behalf of Gyorgy Szing via TF-A <tf-a(a)lists.trustedfirmware.org> Reply to: Gyorgy Szing <Gyorgy.Szing(a)arm.com> Date: Wednesday, 3 March 2021 at 13:26 To: Chris Kay <Chris.Kay(a)arm.com>, Varun Wadekar <vwadekar(a)nvidia.com>, "tf-a(a)lists.trustedfirmware.org" <tf-a(a)lists.trustedfirmware.org> Cc: nd <nd(a)arm.com> Subject: Re: [TF-A] Adoption of Conventional Commits Hi, I think the basic question is: what is the difference between the change-log and the git history? Depending on how we draw the line between the release notes and the change log the answer can be: not much. The change log mostly filters and extends the git history. And this filtering and extending needs a lot of manual work currently. But why we wanted to have two change-logs then? The real difference is the presentation format (reST/HTML vs git log), and the tooling you need to be able to read. If the above is true, then the git log -> changelog transformation can be automated, but that needs the git history being machine readable. For developers this creates the requirement to properly format the commit message, and for reviewers adds extra work too. But that can be automated too right? And this is why we need tooling. Tooling on commit message authoring can be optional, but validation tools are mandatory. Otherwise we will end up with badly formatted commit messages (yes, manual validation is boring an error prone), failing automated translation, and the whole effort misses it’s main point. (And as a side effect we also get a git hook framework, which is making a step forward with standardizing distributed automation.) /George From: Chris Kay <Chris.Kay(a)arm.com> Sent: 03 March 2021 03:56 To: Varun Wadekar <vwadekar(a)nvidia.com>; Gyorgy Szing <Gyorgy.Szing(a)arm.com>; tf-a(a)lists.trustedfirmware.org Cc: nd <nd(a)arm.com> Subject: Re: [TF-A] Adoption of Conventional Commits Hi Varun, > I think you just increased the scope of the problem. We should add that as a new requirement – the commit message header should be pretty. I don't think the scope has increased, but perhaps the requirement that we are able to generate the changelog was lacking clarity; it's not necessarily that the commit message headers should be pretty, but that the changelog should remain so to the extent that it can - it is still user-facing documentation, after all. By extension, we gain nothing from using the commit log to generate the changelog if they just mirror one another. > Honestly, we should also check if in an effort to make the changelog “pretty”, are we losing the traditional git log formatting. Honestly, the git log gets used more than the changelog, so your proposal of changing the commit header has a greater impact. I would like to make this low impact to the developers that create patches on a daily basis. The point I'm trying to emphasise is that there is no traditional Git log formatting - as it stands today, our commit guidelines make no mention of tags. As a result, the tags we do see vary drastically, from none at all to generic "TF-A:" tags, to platforms, drivers and sometimes to specific files. Everybody has their own status quo which they obviously want to maintain, but at some point we have to try to bring everybody onto the same page - commit style rules are not particularly rare for the same reasons code style rules aren't. I don't think the CC rules deviate all that much from the styles we most often see today. > Introducing a completely new way of creating the commit message header or introducing more scripts to create that format is a no-no. There are no mandatory scripts involved - you can continue to write your commits as you do today. The only tangible difference is that we are standardising the tag syntax. > Personally, I feel that you are getting the required information from the git log by just adding tags, which to me, seems like a very low impact approach. Incidentally, it was this very disagreement that brought on this investigation - you can see exactly what the v2.4 changelog looked like<https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/6654/1/docs/…> after basic categorisation, which is where it was decided that a straight dump of the commit log did not suffice for user documentation. > Isnt that an easy fix? We just don’t add tags to such commits. I don’t see how “Conventional Commits” is better. You can avoid tagging the revert commit, but you still need to detect whether the probably-tagged commit it reverts was merged before or after last release, and remove it from the log if the latter. I would suggest Conventional Commits is "better" because we don't even have to consider edge cases like these - we've done the configuration, we know it sorts this out for us, and there's nothing more we need to do to make it just work. > As a maintainer, I feel that forcing developers to unlearn the standard way used by almost all other OSS projects, is disruptive. I am all for automating as long as the process does not get in my way every day. But there is no "standard way" - some projects use "component: xyz" (e.g. Linux), others "[component] xyz" (e.g. LLVM), others yet don't use a tag at all (e.g. Mbed TLS), and I would argue most are like us: lax enough that it's largely down to individual contributors to determine their own. This just happens to be one style among many (and, as far as I know, the only one with an entire tooling ecosystem). I appreciate that you have a favoured variant, but I don't think it's any more useful to debate the most popular commit styles than it is to debate the most popular code styles. As we can see today though, TF-A's existing commit guidelines go largely ignored, and it's our intention to rectify that in a way that allows us to do something useful with information that was previously inaccessible. I won't try to argue that enforcing something that wasn't previously enforced takes some initial getting used to, but I think the emphasis on the "extra work for committers" is severely overrepresented here - realistically, it's a minimal change to how we format the tags that we already write, and it's something some of us have already had to get used to (and have, honestly with very little effort). > I think any proposal should be scalable and forward looking. I’m sure we will hit a scenario where someone needs custom tags and this proposal does not allow us that flexibility. It does afford us that flexibility - we can extend the list of supported types, I'm just unsure of why we might. We would not have settled on this solution if we did not believe it to be scalable and, considering it does already see widespread usage, I would argue it's a relatively safe bet that it can handle most, if not all, of what we need now and in the future. Chris ________________________________ From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Sent: 03 March 2021 01:26 To: Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>>; Gyorgy Szing <Gyorgy.Szing(a)arm.com<mailto:Gyorgy.Szing@arm.com>>; tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> Cc: nd <nd(a)arm.com<mailto:nd@arm.com>> Subject: RE: [TF-A] Adoption of Conventional Commits Hi Chris, I think you just increased the scope of the problem. We should add that as a new requirement – the commit message header should be pretty. Honestly, we should also check if in an effort to make the changelog “pretty”, are we losing the traditional git log formatting. Honestly, the git log gets used more than the changelog, so your proposal of changing the commit header has a greater impact. I would like to make this low impact to the developers that create patches on a daily basis. Introducing a completely new way of creating the commit message header or introducing more scripts to create that format is a no-no. Personally, I feel that you are getting the required information from the git log by just adding tags, which to me, seems like a very low impact approach. On the two examples, I don’t see a big difference in the supposedly human readable log you posted. But the proposal to get that is disruptive. >> You can see here that it emphasises the scope for each change for human readability, and also omits both the revert commit and the commit it reverts because neither of them have been part of a release Isnt that an easy fix? We just don’t add tags to such commits. I don’t see how “Conventional Commits” is better. >> I think burdening reviewers with additional work is likely to prove unreliable, and certainly counter-productive if we can both largely automate the problem away and provide rapid feedback to developers before ever even having to push for review As a maintainer, I feel that forcing developers to unlearn the standard way used by almost all other OSS projects, is disruptive. I am all for automating as long as the process does not get in my way every day. >> The tooling I proposed does already offer some flexibility for defining our own types and scopes, though the default set is already pretty extensive I think any proposal should be scalable and forward looking. I’m sure we will hit a scenario where someone needs custom tags and this proposal does not allow us that flexibility. -Varun From: Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>> Sent: Tuesday, March 2, 2021 4:21 PM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>>; Gyorgy Szing <Gyorgy.Szing(a)arm.com<mailto:Gyorgy.Szing@arm.com>>; tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Cc: nd <nd(a)arm.com<mailto:nd@arm.com>> Subject: Re: [TF-A] Adoption of Conventional Commits External email: Use caution opening links or attachments Hi guys, Note: a lot of this email relies on HTML formatting to force monospace fonts and emphasis – it probably won’t show up correctly on the mailing lists archives. One major point of contention with this model is that it’s not immediately clear what goes into the changelog. The obvious first answer is “the commit subject”, but let’s investigate that. Here are the last 17 commits from upstream as of right now: plat/marvell/armada: cleanup MSS SRAM if used for copy plat/marvell: cn913x: allow CP1/CP2 mapping at BLE stage plat/marvell/armada/common/mss: use MSS SRAM in secure mode libc: memset: Fix MISRA issues plat:xilinx:zynqmp: Remove the custom crash implementation lib: cpus: aarch32: sanity check pointers before use Revert "spmd: ensure SIMD context is saved/restored on SPMC entry/exit" plat/arm/css: rename rd_n1e1_edge_scmi_plat_info array docs: stm32mp1: correct formatting issues marvell: uart: a3720: Increase TX FIFO EMPTY timeout from 2ms to 3ms marvell: uart: a3720: Update delay code to be compatible with 1200 MHz CPU marvell: uart: a3720: Fix comments in console_a3700_core_init() function nxp: added the makefile helper macros spmd: ensure SIMD context is saved/restored on SPMC entry/exit nand: stm32_fmc_nand: remove dead code plat/arm: juno: Refactor juno_getentropy() bl32: Enable TRNG service build Here it is if we applied Conventional Commits to it: feat(marvell armada): cleanup MSS SRAM if used for copy feat(marvell cn913x): allow CP1/CP2 mapping at BLE stage feat(marvell armada): use MSS SRAM in secure mode fix(libc): fix MISRA issues refactor(xilinx zynqmp): remove the custom crash implementation refactor(aarch32): add sanity check pointers before use revert: fix(spmd): ensure SIMD context is saved/restored on SPMC entry/exit refactor(arm css): rename rd_n1e1_edge_scmi_plat_info array docs(stm32mp1): correct formatting issues refactor(marvell a3720): increase TX FIFO EMPTY timeout from 2ms to 3ms refactor(marvell a3720): update delay code to be compatible with 1200 MHz CPU fix(marvell a3720): fix comments in console_a3700_core_init() function build(nxp): add Makefile helper macros fix(spmd): ensure SIMD context is saved/restored on SPMC entry/exit refactor(stm32 fmc_nand): remove dead code refactor(arm juno): Refactor juno_getentropy() feat(bl32): Enable TRNG service build Side note: the “screen real estate” concern some raised does not actually seem to manifest itself in any meaningful way – the longest line only increases by 3 characters, and the shortest line is actually reduced by 3 characters. To me, immediately, the single subject style is much less mentally taxing to parse. Without it, there are at least four different schemes at play here that we need to interpret for every commit (and we’re only 17 commits deep!): foo/bar/baz: xyz foo: bar: baz: xyz foo/bar: baz: xyz Revert “xyz” So, if we just forget about trying to read the history manually for a moment, without a standardised subject format we end up with a changelog that looks like this: Features: - plat/marvell/armada: cleanup MSS SRAM if used for copy - plat/marvell: cn913x: allow CP1/CP2 mapping at BLE stage - plat/marvell/armada/common/mss: use MSS SRAM in secure mode - bl32: Enable TRNG service build Bug Fixes: - libc: memset: Fix MISRA issues - marvell: uart: a3720: Fix comments in console_a3700_core_init() function - spmd: ensure SIMD context is saved/restored on SPMC entry/exit Build System: - nxp: added the makefile helper macros Code Refactoring: - plat:xilinx:zynqmp: Remove the custom crash implementation - lib: cpus: aarch32: sanity check pointers before use - plat/arm/css: rename rd_n1e1_edge_scmi_plat_info array - marvell: uart: a3720: Increase TX FIFO EMPTY timeout from 2ms to 3ms - marvell: uart: a3720: Update delay code to be compatible with 1200 MHz CPU - nand: stm32_fmc_nand: remove dead code - plat/arm: juno: Refactor juno_getentropy() Documentation: - docs: stm32mp1: correct formatting issues Reverts: - Revert "spmd: ensure SIMD context is saved/restored on SPMC entry/exit" This, in my opinion, still suffers from the same problem: as a human, it’s difficult to interpret. Compare that to how we expect that to look with Conventional Commits: Features: - bl32: enable TRNG service build - marvell armada: cleanup MSS SRAM if used for copy - marvell armada: use MSS SRAM in secure mode - marvell cn913x: allow CP1/CP2 mapping at BLE stage Bug Fixes: - libc: fix MISRA issues - marvell a3720: fix comments in console_a3700_core_init() function Build System: - nxp: add Makefile helper macros Code Refactoring: - aarch32: add sanity check pointers before use - arm css: rename rd_n1e1_edge_scmi_plat_info array - arm juno: refactor juno_getentropy() - marvell a3720: increase TX FIFO EMPTY timeout from 2ms to 3ms - marvell a3720: update delay code to be compatible with 1200 MHz CPU - stm32 fmc_nand: remove dead code - xilinx zynqmp: remove the custom crash implementation Documentation: - stm32mp1: correct formatting issues … and I feel like the value prop of using robust tooling becomes more obvious – this tooling is intended not just to categorise commits, but to understand them. You can see here that it emphasises the scope for each change for human readability, and also omits both the revert commit and the commit it reverts because neither of them have been part of a release. Additionally, without a way to enforce this, we’re not necessarily solving one of the current fundamental problems: our changelogs do not accurately and reliably reflect the changes to the project. I think burdening reviewers with additional work is likely to prove unreliable, and certainly counter-productive if we can both largely automate the problem away and provide rapid feedback to developers before ever even having to push for review. Just a quick note on one of your points: 3. Flexibility to define project specific tags The tooling I proposed does already offer some flexibility for defining our own types and scopes, though the default set is already pretty extensive: * build (Build System) * ci (Continuous Integration) * docs (Documentation) * feat (Features) * fix (Bug Fixes) * perf (Performance Improvements) * refactor (Code Refactoring) * revert (Reverts) * style (Styles) * test (Tests) Chris From: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>> Date: Tuesday, 2 March 2021 at 22:31 To: Gyorgy Szing <Gyorgy.Szing(a)arm.com<mailto:Gyorgy.Szing@arm.com>>, Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>>, tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> Cc: nd <nd(a)arm.com<mailto:nd@arm.com>>, nd <nd(a)arm.com<mailto:nd@arm.com>> Subject: RE: [TF-A] Adoption of Conventional Commits Hi George, Few clarifications. >> it would need significant investments on tooling [VW] I am not sure why you say that. The only expectation form tooling perspective is to run the ‘git log’ command before the release. >> There is no tool which could help developers crafting the commit message in the right format [VW] I don’t think we need a tool to generate commit messages with the right tags. I expect developers to add the tag manually as the footer. Maintainers will have to check that tags exist as part of the reviews. >> Possibly the easiest would be to modify the javascript machinery available for Conventional Commits [VW] I don’t think we need any tools from the “Conventional Commits” toolbox for this to work. My proposal was from the requirements I heard from Chris in the meeting. If there are any implicit or obvious requirements that I missed, I propose we freeze them first. A solution can only work if the requirements are frozen. -Varun From: Gyorgy Szing <Gyorgy.Szing(a)arm.com<mailto:Gyorgy.Szing@arm.com>> Sent: Sunday, February 28, 2021 11:14 PM To: Varun Wadekar <vwadekar(a)nvidia.com<mailto:vwadekar@nvidia.com>>; Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>>; tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Cc: nd <nd(a)arm.com<mailto:nd@arm.com>>; nd <nd(a)arm.com<mailto:nd@arm.com>> Subject: RE: [TF-A] Adoption of Conventional Commits External email: Use caution opening links or attachments Hi Varun, I really like your proposal, but it would need significant investments on tooling. There is no tool which could help developers crafting the commit message in the right format, there is no tool, which can validate the format (and be used i.e. as a git-hook), and there is no tool, which can generate the change history document from git history. Can you please extend the proposal and turn it to be an end-to-end solution? Can you contribute tooling for commit message editing and validation, and for change log document generation? Possibly the easiest would be to modify the javascript machinery available for Conventional Commits. Can you contribute the needed changes? /George From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org<mailto:tf-a-bounces@lists.trustedfirmware.org>> On Behalf Of Varun Wadekar via TF-A Sent: 26 February 2021 00:39 To: Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>> Cc: tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Subject: Re: [TF-A] Adoption of Conventional Commits Hi, I really like the idea of using tags in the commit message, but the rigidity of the spec puts me off. Frankly, I believe we just need a way to identify commits and their intent. So, I would like to propose an approach that builds on the “Conventional Commits” spec. The approach would be 1. Add an identifier (e.g. Tags: fix) to the commit message footer. 2. At the start of the release window run “git log”* to print a list of features, bug fixes, performance improvements, deprecations etc. 3. Either update the main changelog manually or use a script to append individual sections. *git log v2.4...HEAD --no-merges --pretty='- %s (%C(auto)%h)' --grep "Tags: fix"> ‘git log’ can be easily modified to look for other metadata as long as we agree to add it to the commit message. Advantages 1. Light(er) 2. No impact to the subject header 3. Flexibility to define project specific tags 4. Training needs at par with “Conventional Commits” proposal Thoughts? -Varun From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org<mailto:tf-a-bounces@lists.trustedfirmware.org>> On Behalf Of Chris Kay via TF-A Sent: Thursday, February 25, 2021 9:31 AM To: tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org> Subject: Re: [TF-A] Adoption of Conventional Commits External email: Use caution opening links or attachments Thanks to all those who attended the Tech Forum today! It’s become apparent that the initial 2 week deadline for alternative proposals or implementations is too short, so – as agreed – we’ll push the deadline for the investigation period to the end of March. This period is dedicated to evaluating the changelog automation proposal made, or to identifying alternative solutions. If you have an alternative proposal, any proof-of-concept tooling would be highly appreciated so we can get a clear idea of what sort of work and maintenance is going to be involved. If you do find a solution you wish to propose, please give it just a short name (e.g. “Update it manually”) and make it obvious you want to propose it formally – I’ll collect up the proposals made on the mailing list thread at the end of March and set up a Wiki poll so we can get a clear picture of where the community wants to take this. Chris From: TF-A <tf-a-bounces(a)lists.trustedfirmware.org<mailto:tf-a-bounces@lists.trustedfirmware.org>> on behalf of Chris Kay via TF-A <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> Reply to: Chris Kay <Chris.Kay(a)arm.com<mailto:Chris.Kay@arm.com>> Date: Thursday, 11 February 2021 at 13:59 To: "tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>" <tf-a(a)lists.trustedfirmware.org<mailto:tf-a@lists.trustedfirmware.org>> Subject: [TF-A] Adoption of Conventional Commits Hi all, Recently we had an internal discussion on the merits of introducing semantics to commit messages pushed to the main TF-A repository, the conclusion being that we would look to adopting the Conventional Commits<https://www.conventionalcommits.org/en/v1.0.0/> specification in the near future. There was one major reason for this, which was to help us in automating the changelog in future releases, but it might also help us to dramatically reduce the overall amount of work needed to make a formal release in the future. This requires some buy-in (or buy-out, in this case) from maintainers because - even though it’s to only a relatively minor extent - it does involve an adjustment to everybody’s workflow. Notably, commit messages will be expected to adopt the structure defined by the specification, which will be enforced by the CI. Most commits that go upstream today adhere to “something that looks like Conventional Commits”, so the change is not exactly sweeping, but any change has the potential be an inconvenience. With that in mind, I propose the following: * We collectively adopt the specification, enforced only for @arm.com contributors until such a time that the majority of maintainers are familiar with the new demands * We suggest - in the prerequisites documentation - the installation of two helper tools: * Commitizen<https://github.com/commitizen/cz-cli> * Commitlint<https://github.com/conventional-changelog/commitlint> Installation of these tools will be optional, but I believe they can help with the transition. In the patches currently in review, they are installed as Git hooks automatically upon execution of npm install, so it requires no manual installation or configuration (other than a relatively up-to-date Node.js installation). You’ll find the patches here<https://review.trustedfirmware.org/q/topic:%22ck%252Fconventional-commits%2…>, and specifically the changes to the prerequisites documentation here<https://review.trustedfirmware.org/c/TF-A/trusted-firmware-a/+/8224/1/docs/…>. Feel free to review these changes if you have comments specifically on their implementation. Let me know if you have any questions or concerns. If everybody’s on board, we can look to have this upstreamed shortly. Chris

4 years, 8 months

3
2
0 0

2025

2024

2023

2022

2021

2020

2019

2018

TF-A