diff --git a/README.md b/README.md index 17e1470..cb5d492 100644 --- a/README.md +++ b/README.md @@ -1,38 +1,578 @@ -# libreBMC README # - -This repository will contain the high level documentaion for libreBMC - -**The overall goal of the project.** We want to boot an AC922 using an FPGA (not the AST2500 ASIC) on a DC-SCM card. We want to prove that the FPGA can boot a high power, modern server. - -To do this, we need the following pieces: -1. A DC-SCM card that has an FPGA on it to replace the AST2500 ASIC - * Antmicro built some of these that use Xylinx A7s. They power on, but no bring-up has been done to show they will boot the system. - * Antmicro is building some that have ECP5’s on them, but they are still missing some parts -2. An interposer that will plug into the DC-SCM card and the AC922. https://git.openpower.foundation/librebmc/ac922interposer.git - * I have 6 of these built -3. A Root of Trust bypass jumper https://github.com/antmicro/dc-scm-rot-jumper-card - * Plans for these are available and they are easy to build -4. A softcore (microwatt) running on the FPGA - * **This works, but it needs to be better incorporated with Lite-x** -5. **Full gateware in Lite-x for that FPGA such that the entire OpenBMC stack can run** - * Some modules are done. Some not. https://git.openpower.foundation/librebmc/gateware.git - * **Need build instructions for the full FPGA image** -6. **Kernel Driver updates** - * IPMI-BT done but needs merging. - * LPC-CTRL. Set up the address window. Needs to be written from scratch. -7. **Need the full OpenBMC code running** - * We currently power on our prototype with scripts. OpenBMC is not running. - * https://git.openpower.foundation/librebmc/openbmc_for_AC922.git -8. **Bring-up of the AC922 with the DC-SCM card running OpenBMC on the FPGA** - * Create build instructions for pulling the entire FPGA image together - * Boot the AC922 - * Verify all functions - * Prove there are no performance/timing issues - * Need tests defined - * Video the bringup/running system for display at conferences - -So far we did a prototype where we used the xylinx A7 FPGA on a custom breadboard and ran microwatt and a very stripped down FPGA. We proved it could boot the AC922, but again it was very low function. We need to have it fully functional to really prove it. We did not run OpenBMC on this prototype - -Educational Documentation: -1. See the files in this repo -2. Video from OPF 2021: https://www.youtube.com/watch?v=f6VRuKf4Bcs +# libreBMC + +This repository contains the high-level documentation for libreBMC - an +open source hardware, gateware and software project to implement a fully +open Baseboard Management Controller to boot a high power, modern server. + +This work was conducted as a collaboration between the OpenPOWER foundation +and [Code Construct](https://codeconstruct.com.au); this documentation +is based on the +[original article](https://codeconstruct.com.au/docs/dcscm-openbmc) on the +Code Construct docs site. + +We are building on a lot of prior work from IBM and Antmicro, mainly: + + * The [Antmicro Artix DC-SCM](https://github.com/antmicro/artix-dc-scm) + hardware - an open hardware implementation of the Open Compute DC-SCM + module, using an Artix FPGA as the core + + * The "[AC922 + Interposer](https://git.openpower.foundation/librebmc/ac922interposer)" - + an adapter between the AC922's baseboard and a DC-SCI connector. + + * Anton Blanchard's + [Microwatt](https://github.com/antonblanchard/microwatt) project, + implementing a fully open powerpc core, as gateware suitable for the Artix. + + * System-on-Chip peripherals using gateware from the + [LiteX](https://github.com/enjoy-digital/) project - `litesdcard`, `liteeth`, + and `litedram`. + + * [LPC controller + gateware](https://git.openpower.foundation/librebmc/lpcperipheral), from + Michael Neuling and Anton Blanchard, as part of the OpenPOWER Foundation's + LibreBMC project. + + * A [Linux kernel](https://kernel.org) with a few extra drivers and platform + configuration for Microwatt and the DC-SCM board. + + * An [OpenBMC](https://github.com/openbmc/) userspace, ported from the existing + witherspoon platform definition. + +Our work here was mainly to integrate these parts; a few fixes here and there, +plus a few new features at various places in the stack. + +There are a few other projects implementing a similar platform: + + * The original "boxarty" project from Joel Stanley, Paul Mackerras and Michael + Neuling from IBM, an original prototype replacing the AC922's AST2500-based + BMC board (codenamed "boxelder") with an Arty A7 FPGA prototyping board with + Microwatt gateware. This was running a set of scripts to boot the AC922 host. + + * The [Axiado](https://axiado.com/) / [Antmicro](https://antmicro.com) + boxarty updates, which built the original boxarty into a more reproducible + embedded system using a + [buildroot port](https://github.com/Axiado/boxarty-buildroot/tree/antmicro). + +However, this is the first implementation (that we're aware of!) that runs +a close-to-production OpenBMC stack, and uses a DC-SCM module. + +Finally, this document provides an early snapshot of the project; we will +keep things up-to-date as things progress, and/or as issues are fixed or +the source code is updated. + +# Hardware + +The hardware components for this are: + + * The Antmicro DC-SCM module + * An Antmicro Root of Trust (RoT) [bypass board](https://github.com/antmicro/dc-scm-rot-jumper-card) (we're not using a RoT on + this platform). + * The AC922 interposer module + * The AC922 server itself + +If you're following along at home, you'll also need: + + * Some industrial hearing protection + +\- because we're working on pre-fan-control firmware, and the AC922 fans running +at full speed are, well, _industrial_. + +We needed a few reworks to the hardware to fix a few bugs and incompatibilities. +Those are, in roughly-decreasing order of complexity: + +## Rework 1: LPC connectivity + +The version of the DC-SCM board that we're using does not connect the DC-SCI's +ESPI channel to the FPGA, due to an intentionally non-populated component: a +voltage translator between the 3.3V signals of the FPGA, and the 1.8V signals +of the ESPI bus. + +We have documented this at +[issue #2](https://github.com/antmicro/artix-dc-scm/issues/2) of the DC-SCM +repo. If/when that issue is fixed, this rework will not be necessary. + +On the POWER9 platform, we use LPC rather than ESPI, and re-use those ESPI pins +for the LPC connection to the host. In its current state, we can't interface the +Artix FPGA to the LPC bus due to the unpopulated voltage translator. + +Fortunately, we don't actually need voltage translation here, and can connect +the 3.3V FPGA IOs to the DC-SCI directly. + +This does require a rework though, bypassing the missing translator (refdes U21 +in the design): + +![DCSCM U21 rework detail](images/dc-scm-rework-1-detail.png) + +We have a printable version of the rework instructions at +[dc-scm-rework-1.v1.pdf](resources/dc-scm-rework-1.v1.pdf) if you want to +replicate these modifications, and/or take to a fabricator. + +## Rework 2: DC-SCM i2c5 routing + +There's an off-by-two error in the DC-SCM board, where pins for I2C bus 5 are +out-of-place, and are instead connected to PCIe signals on the Artix. + +We have documented this at +[issue #1](https://github.com/antmicro/artix-dc-scm/issues/1) of the DC-SCM +project, which is now fixed in the design. + +In the meantime, we have cut the two traces for the A11 and A12 lines on the +DC-SCM board, just inboard of the DC-SCI connector. + +![DC-SCM i2c5 rework detail](images/dc-scm-rework-2.png) + +Note that these traces are on the back side of the board; the board diagram +is flipped to show the traces as they appear on this side. + +Since we still need the i2c5 bus connected, we have used some spare IO pins on +the TPM header, and routed those to the i2c5 SDA and SCL signals. There are +corresponding gateware and device tree updates to use those TPM pins as GPIOs +(for i2c5), covered below. + +## Rework 3: Interposer FSI mux signal routing + +The AC922 interposer connects the baseboard's FSI mux input to pin B51 of the +DC-SCI connector, which is a USB signal, and so not routed to the FPGA. + +We have filed [an +issue](https://git.openpower.foundation/librebmc/ac922interposer/issues/2) on +the interposer design for this. + +To work around the issue, we have connected pin E4 of the molex connector to +another FPGA IO on the TPM header. + +However, the USB controller still drives this signal, so we need to isolate +the USB controller device from the trace. There are a couple of options here: +masking the pins on the DC-SCI connector, or isolating a single trace on +the DC-SCM board. + +We went for the former, as it's less intrusive, and the issue is with the +interposer rather than the DC-SCM. Since none of the signals in that same +section of the DC-SCI edge connector are used, we can just mask the entire +section with ESD-safe insulating tape: + +![DCSCM FSI connector masking](images/dc-scm-rework-4.jpg) + +Alternatively, you may isolate this one signal by cutting the USB trace on the +DC-SCM board, near U18. + +![DCSCM FSI mux rework detail](images/dc-scm-rework-3.png) + +This is quite close to the `USB_DP` trace, but that is unused on the AC922 +hardware - so any "collateral damage" from cutting both traces is likely to be +fine - provided you're not planning to use the DC-SCM board with other hardware. + +## Replacement IO + +In implementing the workaround above, we're replacing some existing (incorrect) +connections with other FPGA IO lines. Since the TPM header is available, and +connected directly to the FPIO IO banks, this gives us a convenient spot for our +four replacement IO signals: two for i2c SDA/SCL, and two for the FSI mux/trans +signals. + +We have used the following pins on the board: + +![DCSCM IOs on TPM header](images/tpm-ios.png) + +These can either be routed to corresponding pins on the back of the interposer's +pin grid, or directly to the AC922 baseboard. + +## Other reworks + +We had a couple of issues with missing pins on the AC922 interposer board, +which we have worked-around by using other spare FPGA IO pins (also on the +DC-SCM TPM header), routed directly to the AC922 baseboard. + +These should be contained to our particular hardware, but do check your +own interposer for full pins on the 20x12 connector grid before going too +far with debugging! + +# Gateware + +Microwatt gateware for the FPGA on the DC-SCM board can be found +in the +[`dev/dcscm`](https://github.com/CodeConstruct/microwatt/commits/dev/dcscm) +branch of the Code Construct microwatt repo on github. + +Building microwatt for the Artix FPGA is mainly described in Microwatt's +[README](https://github.com/antonblanchard/microwatt#synthesis-on-xilinx-fpgas-using-vivado). +In short, you'll need to set up `fusesoc`, and install Xilinx Vivado. In our +case, we used Vivado Free Edition 2022.1, but older versions should also work. + +Build with: + +```sh +$ fusesoc run --target=antmicro-artix-dc-scm microwatt --no_bram --memory_size=0 +``` + +The build will take around 10 minutes, and the resultant bitstream will +`build/microwatt_0/antmicro-artix-dc-scm-vivado/microwatt_0.bit`, ready to +load on the board. Alternatively, try our pre-built +bitstream: +[`microwatt-dcscm-405c16a924.bit`](https://codeconstruct.com.au/docs/dcscm-openbmc/microwatt-dcscm-405c16a924.bit). + +The micro-USB connector on the DC-SCM board provides JTAG access to the FPGA, +and a Microwatt serial console. Connect this to your development machine +to upload the bitstream to your DC-SCM board. + +One of the board's 4MB NOR SPI flash chips is used to store the FPGA bitstream +and early software - this can be flashed using OpenOCD for JTAG, wrapped in the +`flash-arty` script to handle configuration. Behind the scenes `flash-arty` +loads a small +[JTAG-SPI proxy bitstream](https://github.com/quartiq/bscan_spi_bitstreams) +on the FPGA, to write to the flash chip. + +```sh +$ ./openocd/flash-arty -c antmicro-artix-dc-scm -f a100 -t bin \ + build/microwatt_0/antmicro-artix-dc-scm-vivado/microwatt_0.bit +``` + +If you specify the `-l` argument, the bitstream will be loaded directly to +the FPGA (rather than to flash), for a one-time boot. Use `-x` for a FPGA reset. + +A serial console will show Microwatt's first stage loader running: +``` +$ screen /dev/serial/by-id/usb-FTDI_Quad_RS232-HS-if02-port0 115200 + +Welcome to Microwatt ! + + Soc signature: f00daa5500010001 + HDL Git SHA1: 405c16a924694a + Soc features: UART DRAM SPIFLASH ETHERNET SDCARD + DRAM: 512 MB +... +Trying flash... +Doesn't look like an elf64 +HDR: ff ff ff ff ff ff ff ff +Copying payload to DRAM... +Booting from DRAM... +``` + +In this case, microwatt is up and running, but can't yet find software to load +from flash - that needs `u-boot`, which we'll add in the later [u-boot +section](#u-boot). + +The changes implemented in our microwatt port are: + +## LiteEth + +We have regenerated the [LiteEth](https://github.com/enjoy-digital/liteeth/) +ethernet with a hardware reset to fix intermittent startup. This also +required patching newer LiteEth to have stable register bank offsets (these are +both ready pull requests). + +The RGMII PHY chip on the DC-SCM also needs a zero delay, so a separate +Microwatt config. + +## Extra GPIOs + +The BMC uses several I2C busses from the host for monitoring and control. We use +the Linux GPIO bitbanging I2C driver for this, taking a SDA and SCL GPIO line +per I2C bus. + +Since there are 32 pins per GPIO bank, we added a second (`GPIOB`) bank at +`0xc0008000` on the wishbone MMIO bus. + +In future, the I2C functionality could be moved to gateware to avoid inefficient +polling of GPIO by the kernel driver. + +The interposer only connects a subset of the platform GPIO signals to the DC-SCI +connector, so there are some input signals that OpenBMC expects and are not +implemented in hardware. To work around this, we have added a few special GPIO +lines with fixed `0` or `1` values, so that certain missing hardware presence +lines can be emulated. + +## LPC gateware + +A LPC bus between host and BMC provides low-level interactions between these +two components, including host firmware load, host serial console, and an +IPMI channel. + +The BMC LPC client is implemented through [LPC +gateware](https://github.com/OpenPOWERFoundation/lpcperipheral), originally +provided by Michael Neuling and Anton Blanchard. + +Amaranth was used to generate the verilog, no changes were required. + +## GPIO interrupts + +OpenBMC monitors GPIO interrupts, so we added support to Microwatt's +GPIO implementation to trigger interrupts on input edges/levels, +configurable by hardware registers. + +## SoC reset + +We have connected the SoC reset signal in Microwatt, triggerable by a syscon +register. + +## LiteSDCard 1-bit data + +SD and MMC cards normally start in 1-bit data width mode and transition to wider +modes after MMC host initialisation. +[LiteSDCard](https://github.com/enjoy-digital/litesdcard) +previously only supported 4 bit mode, so required driver workarounds to send an +early "switch to 4 bit" command. + +We have added support to LiteSDCard for 1-bit mode and switching at runtime, so +the normal SD/MMC infrastructure can be used in u-boot and Linux. This should +also support the 8-bit data width on the DC-SCM's eMMC, pending further testing. + +# Software + +The software architecture for the DC-SCM + OpenBMC setup is based on an +fairly plain OpenBMC image, using the "witherspoon" platform. This is the +production AC922 stack, using an AST2500 BMC. We have modified this for the +Microwatt environment, plus hardware definitions for the DC-SCM board. + +There are a couple of areas of storage available: + + * A small amount of SPI flash; we use this for the FPGA bitstream and the + u-boot binary + + * A larger eMMC device; we use this for the OpenBMC image, including kernel + and PNOR (host firmware) storage + +The boot flow for the device is: + + * The FPGA bitstream is read from the on-board SPI flash + + * The bitstream contains a tiny loader implemented in BRAM (block RAM in the FPGA). This loader + initialises the DRAM code, reads the u-boot image (also from SPI flash) into + DRAM, and jumps to the `u-boot` entry point + + * The `u-boot` binary has a driver for the eMMC device, plus support for reading + a kernel from the `ext4` filesystem + + * `u-boot` loads the Linux kernel `dtbImage` from a pre-defined path on an eMMC + partition, and executes the kernel + + * The Linux kernel has drivers for the eMMC device, which contains the + OpenBMC root filesystem. + + * AC922 host firmware is kept on a separate partition on the eMMC device, + as a simple 64MB flat file. This is loaded during the host power-on + procedure. + +All of the components for this project are available on github: + +| Component | Repository & Branch | +|-----------|--------| +| u-boot | [u-boot `dev/dcscm`](https://github.com/CodeConstruct/u-boot/tree/dev/dcscm) | +| Linux | [linux `dev/dcscm`](https://github.com/CodeConstruct/linux/tree/dev/dcscm) | +| OpenBMC | [openbmc `dev/dcscm`](https://github.com/CodeConstruct/openbmc/tree/dev/dcscm) | + +Note that these branches are still in-development, and may be updated and +rebased. The general trend will be for patches to go upstream, and we will +update this document to correspond to any changes as they happen. + +## u-boot + +The u-boot port for the DC-SCM hardware is based on the general `microwatt` +port, by Paul Mackerras and Joel Stanley. This gives a base microwatt port, +including support for ethernet and SPI peripherals. + +On top of this base, we have added support for the on-board eMMC device (to +allow booting the kernel), added a general platform definition (a device tree +and defconfig), and added some small features to make development easier. + +The code is up at the [`dev/dcscm` +branch](https://github.com/CodeConstruct/u-boot/tree/dev/dcscm) of the +Code Construct `u-boot` tree on github. + +Building should just be a matter of: + +```sh +$ make CROSS_COMPILE=powerpc64le-linux-gnu- O=obj.microwatt \ + microwatt_antmicro_artix_dcscm_defconfig +$ make CROSS_COMPILE=powerpc64le-linux-gnu- O=obj.microwatt -j$(nproc) +``` + +the resulting image will be an ELF binary in `obj.microwatt/u-boot`; the +gateware can load and execute this directly from SPI flash. + +To write this to flash, use the `flash-arty` script in the microwatt repository, +at the base address of `0x300000` + +``` +$ ./openocd/flash-arty -c antmicro-artix-dc-scm -f a100 -a 0x300000 -t bin \ + ../u-boot/obj.microwatt/u-boot +``` + +## Linux kernel + +The kernel port for the DC-SCM board is based on the general microwatt bringup +work, plus a few OpenBMC-specific patches. + +We have published our changes, based on a v6.0 upstream kernel, at the +[`dev/dcscm` branch](https://github.com/CodeConstruct/linux/tree/dev/dcscm) of +the Code Construct `linux` tree on github. + +The kernel `.config` is available here too: +[`linux-v6.0-dcscm.config`](resources/linux-v6.0-dcscm.config). This also includes support +for a NFS root, which can be handy for initial development work - particularly +setting up the initial filesystem on the MMC device. + +To build: + +```sh +$ mkdir obj.microwatt +$ cp linux-v6.0-dcscm.config obj.microwatt/.config +$ make ARCH=powerpc CROSS_COMPILE=powerpc64le-linux-gnu- O=obj.microwatt oldconfig +$ make ARCH=powerpc CROSS_COMPILE=powerpc64le-linux-gnu- O=obj.microwatt -j$(nproc) +``` + +The DC-SCM kernel will build into a combined zImage + device-tree blob (a +`dtbImage`), located at +`obj.microwatt/arch/powerpc/boot/dtbImage.microwatt-dcscm.elf`. + +If you need to alter the boot arguments, they are located in the +`/chosen/bootargs` property of the device tree source file, at +`arch/powerpc/boot/dts/microwatt-dcscm.dts`: + +```dts + chosen { + bootargs = "root=/dev/mmcblk0p1"; + ibm,architecture-vec-5 = [19 00 10 00 00 00 00 00 00 00 00 00 00 00 00 00 + 00 00 00 00 00 00 00 00 40 00 40]; + stdout-path = &UART0; + }; +``` + +for NFS booting, try something like: +```dts + bootargs = "root=/dev/nfs nfsroot=192.168.0.1:/srv/nfs/dcscm ip=dhcp"; +``` + +\- substituting IP and paths for your own environment, of course. + +## OpenBMC + +In order to get OpenBMC running on the DC-SCM board, we've added a new +`witherspoon-microwatt` platform definition, based on the original `witherspoon` +(ie., production AC922) platform. Anton Blanchard had created the base +support, and we have added further porting work to make a usable BMC image. + +However: our main goal here was to prove the DC-SCM + microwatt components; this +is still a long way from an upstreamable OpenBMC implementation. There are +changes that are not suitable for the general repo, as they alter existing +(working) components. So: treat this as reference code, rather than production! + +The OpenBMC implementation is available from the +[`dev/dcscm` branch](https://github.com/CodeConstruct/openbmc/tree/dev/dcscm) of +the Code Construct `openbmc` tree on github. + +The build process is standard for an OpenBMC tree, just using the +`witherspoon-microwat` configuration: + +```sh +$ . setup witherspoon-microwatt +$ bitbake obmc-phosphor-image +``` + +The build infrastructure will download and build the OpenBMC components, and +assemble into an `ext4` filesystem image, at: + +``` +tmp/deploy/images/witherspoon-microwatt/obmc-phosphor-image-witherspoon-microwatt.ext4 +``` + +# Deploying + +From here, we have a few components: + + * The gateware, already flashed to the SPI device on the DC-SCM + * A u-boot build, also flashed to SPI + * A kernel, as a `dtbImage.elf` + * An OpenBMC image, as an `ext4` filesystem image + +The easiest way to assemble this into a full BMC will be to netboot the kernel, +using an embedded initramfs or small NFS root filesystem, and use that to set up +the MMC device. I would suggest a buildroot image for this; Joel Stanley has a +suitable tree in his [`microwatt` branch of +buildroot](https://github.com/shenki/buildroot/tree/microwatt). + +From there, you can: + + * Create two partitions on the eMMC device (`/dev/mmcblk0`): + * One large one for the OpenBMC root filesystem + * One small (say, 200MB) for the host firmware + + * Write the OpenBMC `.ext4` file to the first MMC partition + + * Mount the new OpenBMC partition (`/dev/mmcblk0p1`) + + * Add the new kernel to the new partition: say, as `/boot/dtbImage.elf` + + * Create a new filesystem for host firmware on the second partition: + ``` + mkfs.ext4 -m 0 /dev/mmcblk0p1 + ``` + + * Add a recent witherspoon PNOR to that new filesystem + + * Create a symlink from the OpenBMC `/var/pnor.img` to the PNOR image + on the second partition + +Using the two partitions will save time should you want to re-write the OpenBMC +filesystem after a rebuild. + +Then, set up `u-boot` to boot automatically: + +``` +=> setenv mmcboot 'ext4load mmc 0 0xc000000 /boot/dtbImage.elf; bootelf 0xc000000' +=> setenv bootcmd 'run mmcboot' +=> saveenv +``` + +And you should be good to go! + +Of course, if you have any issues or queries, please get in touch with the +LibreBMC SIG, and/or Jeremy Kerr . + +# Future work + +This project was originally intended as a proof-of-concept of a full BMC +running on the open source microwatt core, on the open DC-SCM hardware. We've +reached that goal, but there are still a few components that would get us +closer to a production-suitable BMC implementation. + +## General cleanups and upstreaming + +As this is mainly a proof-of-concept, some of the changes are a little +rough-around-the-edges, so could benefit from a review and rework. There are +also areas that can be optimised - both in the new code and the underlying +codebases. + +The main components of this work would be suitable for submitting upstream. We +have a number of pull requests pending already, which will reduce the amount +DC-SCM specific code needed. + +## Software optimisation + +In working through the stack for this implementation, we have identified a few +areas that may benefit from some performance optimisation work. Parts of the +OpenBMC stack are certainly designed for a higher-performance machine than +the (current) microwatt core on the Artix FPGA, and there are definitely areas +that would be fairly straightforward to start an optimisation effort on. + +## Gateware additions + +Currently, the FSI and I2C interfaces are implemented in software, so can be +taxing on the CPU when there is traffic over those busses. With the system +running, the full fan control implementation may use around 60%-80% of CPU just +in FSI and I2C code. + +The next step here would be to add hardware/gateware implementations of the FSI +and I2C controllers; that would go a long way to reducing system load. + +## Easier deployment + +The current process requires a lot of manual setup to get an OpenBMC image +constructed on the on-board storage. There could certainly be improvements to +this flow; for example, implementing a small bootable image that automates +flashing and initialising the OpenBMC system, or allowing easier access to the +eMMC device for initial image loads. + +If there's any area of development you would be interested in seeing, or +participating in, please let us know! diff --git a/images/dc-scm-rework-1-detail.png b/images/dc-scm-rework-1-detail.png new file mode 100644 index 0000000..be08be7 Binary files /dev/null and b/images/dc-scm-rework-1-detail.png differ diff --git a/images/dc-scm-rework-2.png b/images/dc-scm-rework-2.png new file mode 100644 index 0000000..1218de4 Binary files /dev/null and b/images/dc-scm-rework-2.png differ diff --git a/images/dc-scm-rework-3.png b/images/dc-scm-rework-3.png new file mode 100644 index 0000000..eb3e3bb Binary files /dev/null and b/images/dc-scm-rework-3.png differ diff --git a/images/dc-scm-rework-4.jpg b/images/dc-scm-rework-4.jpg new file mode 100644 index 0000000..69eff77 Binary files /dev/null and b/images/dc-scm-rework-4.jpg differ diff --git a/images/tpm-ios.png b/images/tpm-ios.png new file mode 100644 index 0000000..0892264 Binary files /dev/null and b/images/tpm-ios.png differ diff --git a/resources/linux-v6.0-dcscm.config b/resources/linux-v6.0-dcscm.config new file mode 100644 index 0000000..f77d287 --- /dev/null +++ b/resources/linux-v6.0-dcscm.config @@ -0,0 +1,207 @@ +# CONFIG_CROSS_MEMORY_ATTACH is not set +CONFIG_HIGH_RES_TIMERS=y +CONFIG_BPF_SYSCALL=y +CONFIG_PREEMPT_VOLUNTARY=y +CONFIG_TICK_CPU_ACCOUNTING=y +CONFIG_IKCONFIG=y +CONFIG_IKCONFIG_PROC=y +CONFIG_LOG_BUF_SHIFT=16 +CONFIG_PRINTK_SAFE_LOG_BUF_SHIFT=12 +CONFIG_CGROUPS=y +CONFIG_CGROUP_BPF=y +CONFIG_NAMESPACES=y +CONFIG_BLK_DEV_INITRD=y +CONFIG_CC_OPTIMIZE_FOR_SIZE=y +CONFIG_KALLSYMS_ALL=y +CONFIG_EMBEDDED=y +CONFIG_PERF_EVENTS=y +CONFIG_PPC64=y +# CONFIG_PPC_KUEP is not set +# CONFIG_PPC_KUAP is not set +CONFIG_CPU_LITTLE_ENDIAN=y +CONFIG_NR_IRQS=64 +CONFIG_PANIC_TIMEOUT=10 +# CONFIG_PPC_POWERNV is not set +# CONFIG_PPC_PSERIES is not set +CONFIG_PPC_MICROWATT=y +# CONFIG_PPC_OF_BOOT_TRAMPOLINE is not set +CONFIG_CPU_FREQ=y +CONFIG_HZ_100=y +CONFIG_KEXEC=y +CONFIG_KEXEC_FILE=y +CONFIG_RELOCATABLE=y +CONFIG_PPC_4K_PAGES=y +# CONFIG_PPC_MEM_KEYS is not set +CONFIG_JUMP_LABEL=y +# CONFIG_SECCOMP is not set +# CONFIG_MQ_IOSCHED_KYBER is not set +# CONFIG_COREDUMP is not set +# CONFIG_SWAP is not set +# CONFIG_SLAB_MERGE_DEFAULT is not set +# CONFIG_COMPAT_BRK is not set +CONFIG_FLATMEM_MANUAL=y +# CONFIG_VM_EVENT_COUNTERS is not set +CONFIG_NET=y +CONFIG_PACKET=y +CONFIG_PACKET_DIAG=y +CONFIG_UNIX=y +CONFIG_UNIX_DIAG=y +CONFIG_INET=y +CONFIG_IP_PNP=y +CONFIG_IP_PNP_DHCP=y +CONFIG_IP_PNP_BOOTP=y +CONFIG_INET_UDP_DIAG=y +CONFIG_INET_RAW_DIAG=y +# CONFIG_IPV6 is not set +CONFIG_BPFILTER=y +# CONFIG_WIRELESS is not set +CONFIG_DEVTMPFS=y +CONFIG_DEVTMPFS_MOUNT=y +# CONFIG_STANDALONE is not set +# CONFIG_PREVENT_FIRMWARE_BUILD is not set +# CONFIG_FW_LOADER is not set +# CONFIG_ALLOW_DEV_COREDUMP is not set +CONFIG_MTD=y +CONFIG_MTD_BLOCK=y +CONFIG_MTD_PARTITIONED_MASTER=y +CONFIG_MTD_SPI_NOR=y +CONFIG_BLK_DEV_LOOP=y +CONFIG_BLK_DEV_RAM=y +CONFIG_EEPROM_AT24=y +CONFIG_NETDEVICES=y +# CONFIG_NET_VENDOR_ALACRITECH is not set +# CONFIG_NET_VENDOR_AMAZON is not set +# CONFIG_NET_VENDOR_AQUANTIA is not set +# CONFIG_NET_VENDOR_ARC is not set +# CONFIG_NET_VENDOR_ASIX is not set +# CONFIG_NET_VENDOR_BROADCOM is not set +# CONFIG_NET_VENDOR_CADENCE is not set +# CONFIG_NET_VENDOR_CAVIUM is not set +# CONFIG_NET_VENDOR_CORTINA is not set +# CONFIG_NET_VENDOR_DAVICOM is not set +# CONFIG_NET_VENDOR_ENGLEDER is not set +# CONFIG_NET_VENDOR_EZCHIP is not set +# CONFIG_NET_VENDOR_FUNGIBLE is not set +# CONFIG_NET_VENDOR_GOOGLE is not set +# CONFIG_NET_VENDOR_HUAWEI is not set +# CONFIG_NET_VENDOR_INTEL is not set +# CONFIG_NET_VENDOR_WANGXUN is not set +CONFIG_LITEX_LITEETH=y +# CONFIG_NET_VENDOR_MARVELL is not set +# CONFIG_NET_VENDOR_MELLANOX is not set +# CONFIG_NET_VENDOR_MICREL is not set +# CONFIG_NET_VENDOR_MICROCHIP is not set +# CONFIG_NET_VENDOR_MICROSEMI is not set +# CONFIG_NET_VENDOR_MICROSOFT is not set +# CONFIG_NET_VENDOR_NI is not set +# CONFIG_NET_VENDOR_NATSEMI is not set +# CONFIG_NET_VENDOR_NETRONOME is not set +# CONFIG_NET_VENDOR_PENSANDO is not set +# CONFIG_NET_VENDOR_QUALCOMM is not set +# CONFIG_NET_VENDOR_RENESAS is not set +# CONFIG_NET_VENDOR_ROCKER is not set +# CONFIG_NET_VENDOR_SAMSUNG is not set +# CONFIG_NET_VENDOR_SEEQ is not set +# CONFIG_NET_VENDOR_SOLARFLARE is not set +# CONFIG_NET_VENDOR_SOCIONEXT is not set +# CONFIG_NET_VENDOR_STMICRO is not set +# CONFIG_NET_VENDOR_SYNOPSYS is not set +# CONFIG_NET_VENDOR_VERTEXCOM is not set +# CONFIG_NET_VENDOR_VIA is not set +# CONFIG_NET_VENDOR_WIZNET is not set +# CONFIG_NET_VENDOR_XILINX is not set +# CONFIG_WLAN is not set +CONFIG_INPUT_EVDEV=y +# CONFIG_KEYBOARD_ATKBD is not set +CONFIG_KEYBOARD_GPIO=y +CONFIG_KEYBOARD_GPIO_POLLED=y +# CONFIG_INPUT_MOUSE is not set +# CONFIG_SERIO is not set +# CONFIG_VT is not set +# CONFIG_LEGACY_PTYS is not set +CONFIG_SERIAL_8250=y +# CONFIG_SERIAL_8250_DEPRECATED_OPTIONS is not set +CONFIG_SERIAL_8250_CONSOLE=y +CONFIG_SERIAL_OF_PLATFORM=y +CONFIG_SERIAL_NONSTANDARD=y +CONFIG_BT_IPMI_BMC=y +# CONFIG_NVRAM is not set +# CONFIG_RANDOM_TRUST_BOOTLOADER is not set +CONFIG_I2C=y +CONFIG_I2C_CHARDEV=y +CONFIG_I2C_GPIO=y +CONFIG_I2C_FSI=y +CONFIG_SPI=y +CONFIG_SPI_DEBUG=y +CONFIG_SPI_BITBANG=y +CONFIG_SPI_MICROWATT=y +CONFIG_SPI_SPIDEV=y +CONFIG_GPIOLIB=y +CONFIG_GPIO_SYSFS=y +CONFIG_GPIO_FTGPIO010=y +CONFIG_SENSORS_IIO_HWMON=y +CONFIG_SENSORS_LM75=y +CONFIG_SENSORS_OCC_P9_SBE=y +CONFIG_PMBUS=y +CONFIG_SENSORS_IBM_CFFPS=y +CONFIG_SENSORS_IR35221=y +CONFIG_SENSORS_MAX31785=y +CONFIG_SENSORS_UCD9000=y +CONFIG_SENSORS_TMP421=y +CONFIG_MFD_SYSCON=y +# CONFIG_USB_SUPPORT is not set +CONFIG_MMC=y +CONFIG_MMC_CQHCI=y +CONFIG_MMC_HSQ=y +CONFIG_MMC_LITEX=y +CONFIG_NEW_LEDS=y +CONFIG_LEDS_CLASS=y +CONFIG_LEDS_PCA955X=y +CONFIG_LEDS_PCA955X_GPIO=y +CONFIG_RTC_CLASS=y +# CONFIG_VIRTIO_MENU is not set +CONFIG_COMMON_CLK=y +# CONFIG_IOMMU_SUPPORT is not set +CONFIG_LITEX_SOC_CONTROLLER=y +CONFIG_IIO=y +CONFIG_BMP280=y +CONFIG_DPS310=y +CONFIG_FSI=y +CONFIG_FSI_MASTER_GPIO=y +CONFIG_FSI_MASTER_HUB=y +CONFIG_FSI_SCOM=y +CONFIG_FSI_SBEFIFO=y +CONFIG_FSI_OCC=y +CONFIG_EXT4_FS=y +# CONFIG_DNOTIFY is not set +CONFIG_AUTOFS_FS=y +CONFIG_TMPFS=y +# CONFIG_MISC_FILESYSTEMS is not set +CONFIG_NFS_FS=y +# CONFIG_NFS_V2 is not set +CONFIG_NFS_V4=y +CONFIG_ROOT_NFS=y +CONFIG_CRYPTO_SHA256=y +# CONFIG_CRYPTO_HW is not set +# CONFIG_XZ_DEC_X86 is not set +# CONFIG_XZ_DEC_IA64 is not set +# CONFIG_XZ_DEC_ARM is not set +# CONFIG_XZ_DEC_ARMTHUMB is not set +# CONFIG_XZ_DEC_SPARC is not set +CONFIG_PRINTK_TIME=y +CONFIG_DYNAMIC_DEBUG=y +# CONFIG_SYMBOLIC_ERRNAME is not set +# CONFIG_DEBUG_MISC is not set +CONFIG_DEBUG_FS=y +# CONFIG_SLUB_DEBUG is not set +# CONFIG_SCHED_DEBUG is not set +CONFIG_SCHEDSTATS=y +CONFIG_FUNCTION_TRACER=y +# CONFIG_FUNCTION_GRAPH_TRACER is not set +# CONFIG_STRICT_DEVMEM is not set +CONFIG_PPC_DISABLE_WERROR=y +CONFIG_XMON=y +CONFIG_XMON_DEFAULT=y +# CONFIG_XMON_DISASSEMBLY is not set +# CONFIG_XMON_DEFAULT_RO_MODE is not set +# CONFIG_RUNTIME_TESTING_MENU is not set