Hi Sean,

On 5/8/2021 11:14 PM, Sean Anderson wrote:
On 5/8/21 12:57 AM, Tianrui Wei wrote:
Hi Sean,

On 5/7/2021 9:03 PM, Sean Anderson wrote:
On 5/6/21 11:48 PM, Tianrui Wei wrote:

On 5/7/2021 11:41 AM, Sean Anderson wrote:
On 5/6/21 11:28 PM, Tianrui Wei wrote:

On 5/7/2021 11:15 AM, Sean Anderson wrote:
On 5/6/21 11:06 PM, Tianrui Wei wrote:
Hi Sean,


Many thanks again for reviewing our code! We really appreciate it. Will fix the things you're suggesting ;p Though I have a few questions in line in the comment. Also, checkpatch didn't catch any of the identation issues. I was wondering if there are some specific flags to enable some of the checks? I'm running ./utils/checkpatch.pl this.patch right now.


On 5/7/2021 10:32 AM, Sean Anderson wrote:
On 5/5/21 11:42 PM, Tianrui Wei wrote:
From: Tianrui Wei <tianrui-...@outlook.com>
Date: Thu, 6 May 2021 11:30:20 +0800
Subject: [PATCH V4 2/2] riscv: board: Support OpenPiton SoC

This patch add board support for OpenPiton.

Please add some information to the commit message describing the board.


Signed-off-by: Tianrui Wei <tianrui-...@outlook.com>
Signed-off-by: Jonathan Balkind <jbalk...@ucsb.edu>
---

  arch/riscv/Kconfig                      |   4 +
  arch/riscv/dts/Makefile                 |   1 +
  arch/riscv/dts/openpiton-riscv64.dts    | 159 +++++
  board/openpiton/riscv/Kconfig           |  42 ++
  board/openpiton/riscv/MAINTAINERS       |   6 +
  board/openpiton/riscv/Makefile          |   5 +
  board/openpiton/riscv/openpiton-riscv.c |  41 ++
  configs/openpiton_riscv64_defconfig     | 132 ++++
  doc/board/index.rst                     |   1 +
  doc/board/openpiton/index.rst           |   9 +
  doc/board/openpiton/riscv64.rst         | 885 ++++++++++++++++++++++++
  include/configs/openpiton-riscv.h       |  58 ++
  12 files changed, 1343 insertions(+)
  create mode 100644 arch/riscv/dts/openpiton-riscv64.dts
  create mode 100644 board/openpiton/riscv/Kconfig
  create mode 100644 board/openpiton/riscv/MAINTAINERS
  create mode 100644 board/openpiton/riscv/Makefile
  create mode 100644 board/openpiton/riscv/openpiton-riscv.c
  create mode 100644 configs/openpiton_riscv64_defconfig
  create mode 100644 doc/board/openpiton/index.rst
  create mode 100644 doc/board/openpiton/riscv64.rst
  create mode 100644 include/configs/openpiton-riscv.h

diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
index 30b05408..9e7deb34 100644
--- a/arch/riscv/Kconfig
+++ b/arch/riscv/Kconfig
@@ -23,6 +23,9 @@ config TARGET_SIFIVE_FU540
  config TARGET_SIPEED_MAIX
      bool "Support Sipeed Maix Board"
  +config TARGET_OPENPITON_RISCV
+  bool "Support riscv cores on openpiton SoC"
+
  endchoice
    config SYS_ICACHE_OFF
@@ -55,6 +58,7 @@ config SPL_SYS_DCACHE_OFF
  source "board/AndesTech/ax25-ae350/Kconfig"
  source "board/emulation/qemu-riscv/Kconfig"
  source "board/microchip/mpfs_icicle/Kconfig"
+source "board/openpiton/riscv/Kconfig"

Fix indentation.


It shows correct formatting in my client, and there's no warning in checkpatch for some reason

Ignore this, sorry.




  source "board/sifive/fu540/Kconfig"
  source "board/sipeed/maix/Kconfig"
  diff --git a/arch/riscv/dts/Makefile b/arch/riscv/dts/Makefile
index 3a6f96c6..b511cd74 100644
--- a/arch/riscv/dts/Makefile
+++ b/arch/riscv/dts/Makefile
@@ -1,6 +1,7 @@
  # SPDX-License-Identifier: GPL-2.0+
    dtb-$(CONFIG_TARGET_AX25_AE350) += ae350_32.dtb ae350_64.dtb
+dtb-$(CONFIG_TARGET_OPENPITON_RISCV) += openpiton-riscv64.dtb
  dtb-$(CONFIG_TARGET_SIFIVE_FU540) += hifive-unleashed-a00.dtb
  dtb-$(CONFIG_TARGET_SIPEED_MAIX) += k210-maix-bit.dtb
  diff --git a/arch/riscv/dts/openpiton-riscv64.dts b/arch/riscv/dts/openpiton-riscv64.dts
new file mode 100644
index 00000000..ce732b92
--- /dev/null
+++ b/arch/riscv/dts/openpiton-riscv64.dts
@@ -0,0 +1,159 @@
+// SPDX-License-Identifier: (GPL-2.0 OR MIT)
+/* Copyright (c) 2021 Tianrui Wei <tianrui-...@outlook.com> */
+
+/*
+ * This dts is for a dual core instance of OpenPiton+Ariane built
+ * to run on a Digilent Genesys 2 FPGA at 66.67MHz. These files
+ * are automatically generated by the OpenPiton build system and + * this configuration may not be what you need if your configuration
+ * is different from the below.
+ */
+
+/dts-v1/;
+
+/ {
+    #address-cells = <2>;
+    #size-cells = <2>;
+    u-boot,dm-pre-reloc;
+    compatible = "openpiton,ariane";
+
+    chosen {
+       stdout-path = "uart0:115200";
+    };
+
+    aliases {
+        console = &uart0;
+        serial0 = &uart0;
+    };
+
+    cpus {
+        #address-cells = <1>;
+        #size-cells = <0>;
+        u-boot,dm-pre-reloc;

This is unnecessary since riscv_cpu has DM_FLAG_PRE_RELOC.

+ timebase-frequency = <520835>;
+
+        CPU0: cpu@0 {
+            clock-frequency = <66667000>;

Please add a clocks node. If you just have one clock for the whole soc,
you can always use a fixed-clock binding.

+ u-boot,dm-pre-reloc;
+            device_type = "cpu";
+            reg = <0>;
+            status = "okay";
+            compatible = "eth, ariane", "riscv";

There should be no space after the comma. And what is "eth"? Should this
be lowrisc?


Oh eth was ETH Zurich, who developed ariane riscv64 CPU.

As I understand it, ETH is an abbreviation like 'UC' in 'UC Berkeley'.
So perhaps a better compatible string would be 'zurich,ariane' or
'eth-zurich,ariane'.


Thank you for your feed back. We plan to change it to openhwgroup, cva6 as it's new rebranded name






+            riscv,isa = "rv64imafdc";
+            mmu-type = "riscv,sv39";
+            tlb-split;
+            // HLIC - hart local interrupt controller
+            CPU0_intc: interrupt-controller {
+                #interrupt-cells = <1>;
+                interrupt-controller;
+                compatible = "riscv,cpu-intc";
+            };
+        };
+
+        CPU1: cpu@1 {

Same comments as above.

+ clock-frequency = <66667000>;
+            u-boot,dm-pre-reloc;
+            device_type = "cpu";
+            reg = <1>;
+            status = "okay";
+            compatible = "eth, ariane", "riscv";
+            riscv,isa = "rv64imafdc";
+            mmu-type = "riscv,sv39";
+            tlb-split;
+            // HLIC - hart local interrupt controller
+            CPU1_intc: interrupt-controller {
+                #interrupt-cells = <1>;
+                interrupt-controller;
+                compatible = "riscv,cpu-intc";
+            };
+        };
+
+    };
+
+    memory@80000000 {
+        u-boot,dm-pre-reloc;
+        device_type = "memory";
+        reg = < 0x00000000 0x80000000 0x00000000 0x40000000 >;
+    };
+

Everything below this line should be under a soc node.


Will do



+    uart0: uart@fff0c2c000 {
+        u-boot,dm-pre-reloc;

This is unnecessary. ns16550_serial has DM_FLAG_PRE_RELOC.

+        compatible = "ns16550";
+        reg = < 0x000000ff 0xf0c2c000 0x00000000 0x000d4000 >;
+        clock-frequency = <66667000>;

Please add a clocks node.Documentation/devicetree/bindings/serial/serial.yaml

+        current-speed = <115200>;

This is only necessary if the baud rate cannot otherwise be determined.
Please remove this.


Oh this was necessary to hard code

Why is that? It should be possible to determine the current baud rate
from the input clock and the divider.


That makes sense, will do.






+        interrupt-parent = <&PLIC0>;

Please specify this under /soc.

+        interrupts = <1>;
+        reg-shift = <0>;
+        // regs are spaced on 8 bit boundary
+    };
+
+    eth: ethernet@fff0d00000 {
+        compatible = "xlnx,xps-ethernetlite-1.00.a";
+        device_type = "network";
+        reg = < 0x000000ff 0xf0d00000 0x00000000 0x00100000 >;
+        interrupt-parent = <&PLIC0>;
+        interrupts = <2>;
+        local-mac-address = [ 00 18 3E 02 E3 E5 ];
+        phy-handle = <&phy0>;
+        xlnx,duplex = <0x1>;
+        xlnx,include-global-buffers = <0x1>;
+        xlnx,include-internal-loopback = <0x0>;
+        xlnx,include-mdio = <0x1>;
+        xlnx,rx-ping-pong = <0x1>;
+        xlnx,s-axi-id-width = <0x1>;
+        xlnx,tx-ping-pong = <0x1>;
+        xlnx,use-internal = <0x0>;
+        axi_ethernetlite_0_mdio: mdio {
+            #address-cells = <1>;
+            #size-cells = <0>;
+            phy0: phy@1 {
+                compatible = "ethernet-phy-id001C.C915";
+                device_type = "ethernet-phy";
+                reg = <1>;
+            };
+        };
+    };
+
+    debug-controller@fff1000000 {
+        compatible = "riscv,debug-013";

Is this binding needed? Will U-Boot/Linux/whoever need to initialize
this device?


I think

Well, for the moment there is no driver for this in U-Boot or Linux.


Oh then I must be mistaken, will remove.






+ interrupts-extended = < &CPU0_intc 65535
+                                &CPU1_intc 65535 >;

Please align these.

+        reg = < 0x000000ff 0xf1000000 0x00000000 0x00001000 >;
+        reg-names = "control";
+    };
+
+    sdhci_0: sdhci@0xf000000000 {

No 0x prefix please.

+ u-boot,dm-pre-reloc;

Why does this need to be pre-reloc?


We're using it in SPL

Then please use u-boot,dm-spl


Will do






+        status = "okay";
+        compatible = "openpiton,piton-mmc";
+        reg = < 0x000000f0 0x00000000 0x00000000 0x00300000 >;
+    };

Missing newline.

+    clint@fff1020000 {
+        u-boot,dm-pre-reloc;

Again, unnecessary since the driver has the flag set.

+        compatible = "sifive,clint0";
+        interrupts-extended = < &CPU0_intc 3
+                                &CPU0_intc 7
+                                &CPU1_intc 3
+                                &CPU1_intc 7 >;

Please align this.

+        reg = < 0x000000ff 0xf1020000 0x00000000 0x000c0000 >;
+        reg-names = "control";

Please add a clocks property.

+    };
+
+    PLIC0: plic@fff1100000 {
+        u-boot,dm-pre-reloc;

Again, unnecessary.

+        #address-cells = <0>;

Remove this.

+        #interrupt-cells = <1>;
+        compatible = "sifive,plic-1.0.0";
+        interrupt-controller;
+        interrupts-extended = < &CPU0_intc 11
+                                &CPU0_intc 9
+                                &CPU1_intc 11
+                                &CPU1_intc 9 >;

Align these please.

+        reg = < 0x000000ff 0xf1100000 0x00000000 0x04000000 >;
+        riscv,max-priority = <7>;
+        riscv,ndev = <2>;
+    };

In general, please add soc-specific compatible strings (e.g.
"openpiton,ns16550") to allow for forward compatibility.

+
+};
+
diff --git a/board/openpiton/riscv/Kconfig b/board/openpiton/riscv/Kconfig
new file mode 100644
index 00000000..31ae44d5
--- /dev/null
+++ b/board/openpiton/riscv/Kconfig
@@ -0,0 +1,42 @@
+if TARGET_OPENPITON_RISCV
+
+config SYS_BOARD
+    default "riscv"
+
+config SYS_VENDOR
+    default "openpiton"
+
+config SYS_CPU
+    default "generic"
+
+config SYS_CONFIG_NAME
+    default "openpiton-riscv"
+
+config SYS_TEXT_BASE
+    default 0x81000000 if SPL
+    default 0x80000000 if !RISCV_SMODE
+    default 0x81000000 if RISCV_SMODE
+
+config SPL_TEXT_BASE
+    default 0x80000000
+
+config SPL_OPENSBI_LOAD_ADDR
+    default 0x81000000
+
+config BOARD_SPECIFIC_OPTIONS # dummy
+    def_bool y
+    select ARCH_EARLY_INIT_R
+    select SUPPORT_SPL
+    imply CPU_RISCV
+    imply RISCV_TIMER if (RISCV_SMODE || SPL_RISCV_SMODE)
+    imply SIFIVE_CLINT if (RISCV_MMODE || SPL_RISCV_MMODE)
+    imply CMD_CPU
+    imply SPL_CPU_SUPPORT
+    imply SPL_OPENSBI
+    imply SPL_LOAD_FIT
+    imply SPL_SMP
+    imply SPL_MMC

Fix indentation please. Did you run checkpatch?

+    imply SMP
+    imply SPL_RISCV_MMODE
+
+endif
diff --git a/board/openpiton/riscv/MAINTAINERS b/board/openpiton/riscv/MAINTAINERS
new file mode 100644
index 00000000..1db6fb60
--- /dev/null
+++ b/board/openpiton/riscv/MAINTAINERS
@@ -0,0 +1,6 @@
+Openpiton Riscv Bootloader

OpenPiton BOARD

+M:    Tianrui Wei<tianrui-...@outlook.com>
+S:    Maintained
+F:    board/openpiton/riscv/

Will there be non-riscv openpiton boards?

In general, the directory layout here is

board/manufacturer/board_name

So who manufactures OpenPiton?


There will probably be, yes.


OpenPiton is a community project, and will continue to support other isas in the future ( already a few other supported now )



+F: include/configs/openpiton-riscv.h
+F:    configs/openpiton_riscv_defconfig
diff --git a/board/openpiton/riscv/Makefile b/board/openpiton/riscv/Makefile
new file mode 100644
index 00000000..8cc20e7c
--- /dev/null
+++ b/board/openpiton/riscv/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0+
+#
+# Copyright (C) 2021 Tianrui Wei
+# Tianrui Wei <tianrui-...@outlook.com>
+obj-y += openpiton-riscv.o
diff --git a/board/openpiton/riscv/openpiton-riscv.c b/board/openpiton/riscv/openpiton-riscv.c
new file mode 100644
index 00000000..5be407e7
--- /dev/null
+++ b/board/openpiton/riscv/openpiton-riscv.c
@@ -0,0 +1,41 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright (c) 2019 SiFive, Inc
+ * Copyright (c) 2021 Tianrui Wei
+ *
+ *
+ * Authors:
+ *   Pragnesh Patel <pragnesh.pa...@sifive.com>
+ *   Tianrui Wei <tianrui-...@outlook.com>
+ */
+#include <common.h>
+#include <init.h>
+#include <configs/openpiton-riscv.h>
+#include <dm.h>
+#include <spl.h>
+
+#ifdef CONFIG_SPL
+void board_boot_order(u32 *spl_boot_list)
+{
+    u8 i;
+    u32 boot_devices[] = {
+        BOOT_DEVICE_MMC1,
+    };
+
+    for (i = 0; i < ARRAY_SIZE(boot_devices); i++)
+        spl_boot_list[i] = boot_devices[i];
+}
+#endif
+
+#ifdef CONFIG_SPL_LOAD_FIT
+int board_fit_config_name_match(const char *name)
+{
+    /* boot using first FIT config */
+    return 0;
+}
+#endif
+
+int board_init(void)
+{
+        return 0;
+}

Do you need get_effective_memsize?

diff --git a/configs/openpiton_riscv64_defconfig b/configs/openpiton_riscv64_defconfig
new file mode 100644
index 00000000..37aa3c80
--- /dev/null
+++ b/configs/openpiton_riscv64_defconfig
@@ -0,0 +1,132 @@
+CONFIG_RISCV=y
+CONFIG_NR_DRAM_BANKS=1
+CONFIG_SPL=y
+CONFIG_SPL_MMC_SUPPORT=y
+CONFIG_DEFAULT_DEVICE_TREE="openpiton-riscv64"
+CONFIG_TARGET_OPENPITON_RISCV=y
+CONFIG_ARCH_RV64I=y
+CONFIG_RISCV_SMODE=y
+CONFIG_MISC_INIT_R=n
+CONFIG_SYS_MALLOC_F_LEN=0x1000
+CONFIG_SPL_SYS_MALLOC_F_LEN=0x100000
+CONFIG_SPL_PAYLOAD=""
+CONFIG_LIBDISK_SUPPORT=y
+CONFIG_NR_CPUS=32
+CONFIG_CMODEL_MEDANY=y
+CONFIG_EXPORT=n
+CONFIG_HASH=n
+CONFIG_SPL_SYS_MALLOC_SIMPLE=y
+CONFIG_SPL_BANNER_PRINT=n
+CONFIG_SPL_CRC32_SUPPORT=n
+CONFIG_GENERATE_SMBIOS_TABLE=n
+CONFIG_EFI_LOADER=n
+CONFIG_LIB_DATA=y
+CONFIG_SPL_USE_TINY_PRINTF=y
+CONFIG_SPL_TINY_MEMSET=y
+CONFIG_FS_EXT4=y
+CONFIG_DM_RTC=y
+CONFIG_SYS_NS16550=y
+CONFIG_SPL_OPENSBI=n
+CONFIG_SPL_RTC_SUPPORT=y
+CONFIG_SPL_FS_EXT4=y
+CONFIG_SPL_LEGACY_IMAGE_SUPPORT=n
+CONFIG_CMD_NET=n
+CONFIG_EFI_PARTITION=y
+CONFIG_EFI_PARTITION_ENTRIES_NUMBERS=128
+CONFIG_EFI_PARTITION_ENTRIES_OFF=0
+CONFIG_SPL_EFI_PARTITION=y
+CONFIG_PARTITION_UUIDS=y
+CONFIG_OF_EMBED=y

This is not allowed for mainline boards. Please use OF_SEPARATE (for
M-mode) or OF_PRIOR_STAGE (for S-mode).


Interesting. Does it work for SPL also?

For SPL two binaries are produced, one with a DTB and one without. You
can use either in whatever build step follows.


Thank you. As long as SPL has an embedded dtb, we're fine.






+CONFIG_HAVE_BLOCK_DEVICE=y
+CONFIG_CPU=y
+CONFIG_CPU_RISCV=y
+CONFIG_MMC=y
+CONFIG_DM_MMC=y
+CONFIG_SPL_DM_MMC=y
+CONFIG_MMC_PITON=y
+CONFIG_MMC_QUIRKS=y
+CONFIG_DEBUG_UART=y
+CONFIG_DEBUG_UART_NS16550=y
+CONFIG_DEBUG_UART_SHIFT=0
+CONFIG_DEBUG_UART_ANNOUNCE=y
+CONFIG_DEBUG_UART_SKIP_INIT=y
+CONFIG_DEBUG_UART_BASE=0xfff0c2c000
+CONFIG_DEBUG_UART_CLOCK=66667000
+CONFIG_SPL_LIBDISK_SUPPORT=y
+CONFIG_MMC_VERBOSE=n
+CONFIG_MMC_WRITE=n
+CONFIG_MMC_HW_PARTITIONING=n
+CONFIG_SYS_RELOC_GD_ENV_ADDR=n
+CONFIG_SAVEENV=n
+CONFIG_NET=n
+CONFIG_SPL_PARTITION_UUIDS=n
+CONFIG_CMD_EXT4=y
+CONFIG_CMD_FS_GENERIC=y
+CONFIG_CMD_READ=y
+CONFIG_CMD_LSBLK=y
+CONFIG_CMD_MMC=y
+CONFIG_CMD_GPT=y
+CONFIG_CMD_MEMINFO=y
+CONFIG_EXPERT=n
+CONFIG_ENV_VARS_UBOOT_CONFIG=y
+CONFIG_LEGACY_IMAGE_FORMAT=n
+CONFIG_ARCH_FIXUP_FDT_MEMORY=n
+CONFIG_MENU=y
+CONFIG_CMD_BOOTZ=y
+CONFIG_CMD_PART=y
+CONFIG_SHOW_REGS=y
+CONFIG_LOG=y
+CONFIG_LOGLEVEL=9
+CONFIG_SPL_LOGLEVEL=9
+CONFIG_TPL_LOGLEVEL=9
+CONFIG_SPL_LOG=y
+CONFIG_SPL_LOG_MAX_LEVEL=9
+CONFIG_SPL_LOG_CONSOLE=y
+CONFIG_LOG_ERROR_RETURN=y
+CONFIG_CMD_CPU=n
+CONFIG_BOOTM_NETBSD=n
+CONFIG_BOOTM_PLAN9=n
+CONFIG_BOOTM_RTEMS=n
+CONFIG_BOOTM_VXWORKS=n
+CONFIG_CMD_RUN=n
+CONFIG_CMD_IMI=n
+CONFIG_CMD_XIMG=n
+CONFIG_CMD_EXPORTENV=n
+CONFIG_CMD_IMPORTENV=n
+CONFIG_CMD_EDITENV=n
+CONFIG_CMD_SAVEENV=n
+CONFIG_CMD_CRC32=n
+CONFIG_CMD_RANDOM=n
+CONFIG_CMD_LZMADEC=n
+CONFIG_CMD_UNLZ4=n
+CONFIG_CMD_UNZIP=n
+CONFIG_CMD_FLASH=n
+CONFIG_RANDOM_UUID=n
+CONFIG_CMD_LOADB=n
+CONFIG_CMD_LOADS=n
+CONFIG_CMD_ECHO=n
+CONFIG_CMD_ITEST=n
+CONFIG_CMD_SOURCE=n
+CONFIG_CMD_SETEXPR=n
+CONFIG_CMD_BLOCK_CACHE=n
+CONFIG_CMD_DATE=n
+CONFIG_CMD_SLEEP=n
+CONFIG_CMD_SYSBOOT=y
+CONFIG_CMD_FAT=y
+CONFIG_DOS_PARTITION=y
+CONFIG_ISO_PARTITION=y
+CONFIG_DM_ETH=y
+CONFIG_RAM=y
+CONFIG_SPL_RAM=y
+CONFIG_FS_FAT=y
+CONFIG_FS_FAT_MAX_CLUSTSIZE=65536
+CONFIG_FS_SQUASHFS=y
+CONFIG_SHA1=y
+CONFIG_SHA256=y
+CONFIG_MD5=y
+CONFIG_ZLIB_UNCOMPRESS=y
+CONFIG_SPL_GZIP=y
+CONFIG_SPL_ZLIB=y
+CONFIG_GETOPT=y
+CONFIG_OF_LIBFDT_OVERLAY=y
+CONFIG_RAM_SIFIVE=n

Did you generate this with "make savedefconfig"?

diff --git a/doc/board/index.rst b/doc/board/index.rst
index 915f1be8..51e60ac4 100644
--- a/doc/board/index.rst
+++ b/doc/board/index.rst
@@ -17,6 +17,7 @@ Board-specific doc
     google/index
     intel/index
     kontron/index
+   openpiton/index

Fix indentation

     renesas/index
     rockchip/index
     sifive/index
diff --git a/doc/board/openpiton/index.rst b/doc/board/openpiton/index.rst
new file mode 100644
index 00000000..c469102c
--- /dev/null
+++ b/doc/board/openpiton/index.rst
@@ -0,0 +1,9 @@
+.. SPDX-License-Identifier: GPL-2.0+
+
+OpenPiton
+=========
+
+.. toctree::
+   :maxdepth: 2
+
+   riscv64
diff --git a/doc/board/openpiton/riscv64.rst b/doc/board/openpiton/riscv64.rst
new file mode 100644
index 00000000..dc934bb4
--- /dev/null
+++ b/doc/board/openpiton/riscv64.rst
@@ -0,0 +1,885 @@
+.. SPDX-License-Identifier: GPL-2.0+
+
+Openpiton RISC-V SoC
+====================
+
+OpenPiton RISC-V SoC
+--------------------
+OpenPiton is an open source, manycore processor and research platform. It is a tiled manycore framework scalable from one to 1/2 billion cores. It supports a number of ISAs including RISC-V with its P-Mesh cache coherence protocol and networks on chip. It is highly configurable in both core and uncore components. OpenPiton has been verified in both ASIC and multiple Xilinx FPGA prototypes running full-stack Debian linux.
+
+RISCV-V Standard Bootflow
+-------------------------
+Currently, OpenPiton implements RISC-V standard bootflow in the following steps
+mover.S -> u-boot-spl -> opensbi -> u-boot -> Linux
+This board supports S-mode u-boot as well as M-mode SPL
+
+Building OpenPition
+---------------------
+If you'd like to build OpenPiton, please go to OpenPiton github repo to build from the latest changes

Please link to the github.


Will do



+
+Building Images
+---------------------------
+
+SPL
+---
+
+1. Add the RISC-V toolchain to your PATH.
+2. Setup ARCH & cross compilation environment variable:
+
+.. code-block:: none
+
+   export CROSS_COMPILE=<riscv64 toolchain prefix>
+   export ARCH=riscv
+
+3. make openpiton_riscv64_defconfig
+4. make
+
+U-Boot
+------
+
+1. Add the RISC-V toolchain to your PATH.
+2. Setup ARCH & cross compilation environment variable:
+
+.. code-block:: none
+
+   export CROSS_COMPILE=<riscv64 toolchain prefix>
+   export ARCH=riscv
+
+3. make openpiton_riscv64_defconfig
+4. make menuconfig, then change CONFIG_SYS_TEXT_BASE to 0x81020000

Why isn't this the default?


Because u-boot SPL will just to CONFIG)SYS_TEXT_BASE ( where opensbi is ). If I change that, it will just jump to the wrong location.

Can you make this dependent on S-Mode?


I'm not sure making it dependent on S-mode helpful, because in essense we need to generate u-boot and u-boot spl in 2 passes, so it had to be different

U-Boot and SPL should be produced in the same pass. See e.g. SiFive Unleashed.


Hmm that's only possible because they're using FIT, whereas we aren't though.

I don't see what that has to do with anything.


If we're using fit, SPL will jump to any address as indicated by fit.

If not however, SPL will jump to CONFIG_SYS_TEXT_BASE, which is the text base for u-boot proper

Ok, so use FIT. I believe the standard boot flow is to have SPL load a FIT with OpenSBI and U-Boot.












+5. make
+
+
+opensbi
+-------
+
+1. Add the RISC-V toolchain to your PATH.
+2. Setup ARCH & cross compilation environment variable:
+
+.. code-block:: none
+
+   export CROSS_COMPILE=<riscv64 toolchain prefix>
+   export ARCH=riscv
+
+3. Go to OpenSBI directory
+4. Edit platform/fpga/openpiton/config.mk, and change FW_TEXT_START to 0x81000000 +5. make PLATFORM=fpga/openpiton FW_PAYLOAD_PATH=<path to u-boot-nodtb.bin>
+
+
+Using fw_payload.bin with linux
+-------------------------------
+Put the generated fw_payload.bin into the /boot directory on the root filesystem, plug in the SD card, then flash the
bitstream. Linux will boot automatically.
+
+Booting
+-------
+Once you plugin the sdcard and power up, you should see the U-Boot prompt.
+
+Sample Dual-core Debian boot log from OpenPiton
+-----------------------------------------------
+
+.. code-block:: none
+
+  <debug_uart>

Please use a log without debug uart.


So this is the part where it was a little confusing. Disabling debug uart acutally doesn't work for some reason, so we had to keep it open. Will submit another patch if we got it working with debug uart turned off.

This is a bit of a strange request, but can you try adding some nops()
(around 10-30) to some function (e.g. board_init). I've been having
alignment problems in k210, so it could be something similar.


I was wondering if you have any idea what may cause the alignment problems, we're also hitting it constantly and adding nops seems to have no impact so far.


Thanks,

Tianrui




Hmm that was a good idea. Thanks for the suggestion!




diff --git a/include/configs/openpiton-riscv.h b/include/configs/openpiton-riscv.h
new file mode 100644
index 00000000..0f30609b
--- /dev/null
+++ b/include/configs/openpiton-riscv.h
@@ -0,0 +1,58 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+/*
+ * Copyright (c) 2019 Western Digital Corporation or its affiliates.
+ * Copyright (c) 2021 Tianrui Wei
+ *
+ * Authors:
+ *   Anup Patel <anup.pa...@wdc.com>
+ *   Tianrui Wei <tianrui-...@outlook.com>
+ */
+
+#ifndef __OPENPITON_RISCV_CONFIG_H
+#define __OPENPITON_RISCV_CONFIG_H
+
+#include <linux/sizes.h>
+#define DEBUG
+#ifdef CONFIG_SPL
+#define CONFIG_SPL_MAX_SIZE     0x00100000
+#define CONFIG_SPL_BSS_START_ADDR   0x82000000
+#define CONFIG_SPL_BSS_MAX_SIZE     0x00100000
+#define CONFIG_SYS_SPL_MALLOC_START (CONFIG_SPL_BSS_START_ADDR + \
+        CONFIG_SPL_BSS_MAX_SIZE)
+#define CONFIG_SYS_SPL_MALLOC_SIZE  0x0100000
+#define CONFIG_SPL_STACK    (0x80000000 + 0x04000000 - \

Please use SZ_64M here, or just use an absolute address.


Will fix



+ GENERATED_GBL_DATA_SIZE)
+
+#define CONFIG_SPL_FS_LOAD_PAYLOAD_NAME "boot/fw_payload.bin"
+#define CONFIG_SPL_GD_ADDR 0x85000000
+#endif
+
+/* Environment options */
+#define CONFIG_SYS_SDRAM_BASE 0x80000000
+#define CONFIG_SYS_INIT_SP_ADDR (CONFIG_SYS_SDRAM_BASE + SZ_2M)
+#define CONFIG_SYS_LOAD_ADDR (CONFIG_SYS_SDRAM_BASE + SZ_2M)
+#define CONFIG_SYS_MALLOC_LEN       SZ_256M
+#define CONFIG_SYS_BOOTM_LEN        SZ_256M
+
+/* -------------------------------------------------
+ * Environment
+ */
+//Disable persistent environment variable storage
+#define CONFIG_ENV_IS_NOWHERE   1
+
+/* ---------------------------------------------------------------------
+ * Board boot configuration
+ */
+
+#define CONFIG_EXTRA_ENV_SETTINGS "\0"

Please define some of the raw addresses used below to make modifying the
boot process easier. For example,

    fdt_addr_r=0x86000000
    kernel_addr_r=0x80200000
    image=Image
    mmcdev=0
    mmcpart=1


Good idea! Will do!




+
+#define CONFIG_USE_BOOTCOMMAND
+#define CONFIG_BOOTCOMMAND \
+    "fdt addr ${fdtcontroladdr}; " \
+    "fdt move ${fdtcontroladdr} 0x86000000; " \
+    "ext4load mmc 0:1 0x80200000 boot/Image; " \

Can you use "load" from CONFIG_CMD_FS_GENERIC?





+    "booti 0x80200000 - 0x86000000; "

Is there a bootm flow?


No we don't support fit yet for some hardware reasons.

Can you elaborate on that? I wasn't aware of any restrictions in this
area.


So OpenPiton is actually some kind of SoC generator that generates
different SoC on FPGA. The device tree was generated at bitstream
creation time to facilitate different configurations, so each board
can have different device trees. We're aware of any way to do this in
FIT.

Oh, so you're saying that the devicetree is placed at a specific address
in hardware? Does it need any fixups?


No for now we just embed it in SPL, will be converted into the bootrom for the chip with a small ZSBL.

Well, atm you are using the DTS embedded in U-Boot.

I don't know what the correct way to do this is...

+CC Bin, Heinrich: Do you have a comment on this?


Also, now that you mention fixups u-boot spl doesn't work at high address like 0xffffff0000 for now. Is that the case?

What happens why you try that?


It would simply fail to compile with some link errors

Hm, I don't know the specific cause of that. You would have to post the errors.

--Sean



Many thanks for your valuable feedbacks!

Tianrui



--Sean


Reply via email to