From: Sascha Hauer <sha@pengutronix.de>
To: Steffen Trumtrar <str@pengutronix.de>
Cc: barebox@lists.infradead.org,
Steffen Trumtrar <s.trumtrar@pengutronix.de>
Subject: Re: [PATCH v3 5/6] ARM: socfpga: add support for Enclustra AA1
Date: Mon, 11 Jul 2022 10:54:15 +0200 [thread overview]
Message-ID: <20220711085415.GM5208@pengutronix.de> (raw)
In-Reply-To: <20220711075209.2377254-5-str@pengutronix.de>
On Mon, Jul 11, 2022 at 09:52:08AM +0200, Steffen Trumtrar wrote:
> From: Steffen Trumtrar <s.trumtrar@pengutronix.de>
>
> Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
> ---
> arch/arm/boards/Makefile | 1 +
> arch/arm/boards/enclustra-aa1/Makefile | 4 +
> arch/arm/boards/enclustra-aa1/board.c | 47 ++++++++
> arch/arm/boards/enclustra-aa1/lowlevel.c | 113 ++++++++++++++++++
> .../enclustra-aa1/pinmux-config-arria10.c | 104 ++++++++++++++++
> .../boards/enclustra-aa1/pll-config-arria10.c | 56 +++++++++
> arch/arm/dts/Makefile | 1 +
> arch/arm/dts/socfpga_arria10_mercury_aa1.dts | 83 +++++++++++++
> arch/arm/mach-socfpga/Kconfig | 4 +
> images/Makefile.socfpga | 12 ++
> 10 files changed, 425 insertions(+)
> create mode 100644 arch/arm/boards/enclustra-aa1/Makefile
> create mode 100644 arch/arm/boards/enclustra-aa1/board.c
> create mode 100644 arch/arm/boards/enclustra-aa1/lowlevel.c
> create mode 100644 arch/arm/boards/enclustra-aa1/pinmux-config-arria10.c
> create mode 100644 arch/arm/boards/enclustra-aa1/pll-config-arria10.c
> create mode 100644 arch/arm/dts/socfpga_arria10_mercury_aa1.dts
>
> diff --git a/arch/arm/boards/Makefile b/arch/arm/boards/Makefile
> index d303999614..9d700cfbda 100644
> --- a/arch/arm/boards/Makefile
> +++ b/arch/arm/boards/Makefile
> @@ -134,6 +134,7 @@ obj-$(CONFIG_MACH_SCB9328) += scb9328/
> obj-$(CONFIG_MACH_SEEED_ODYSSEY) += seeed-odyssey/
> obj-$(CONFIG_MACH_SOCFPGA_ALTERA_SOCDK) += altera-socdk/
> obj-$(CONFIG_MACH_SOCFPGA_EBV_SOCRATES) += ebv-socrates/
> +obj-$(CONFIG_MACH_SOCFPGA_ENCLUSTRA_AA1) += enclustra-aa1/
> obj-$(CONFIG_MACH_SOCFPGA_REFLEX_ACHILLES) += reflex-achilles/
> obj-$(CONFIG_MACH_SOCFPGA_TERASIC_DE0_NANO_SOC) += terasic-de0-nano-soc/
> obj-$(CONFIG_MACH_SOCFPGA_TERASIC_DE10_NANO) += terasic-de10-nano/
> diff --git a/arch/arm/boards/enclustra-aa1/Makefile b/arch/arm/boards/enclustra-aa1/Makefile
> new file mode 100644
> index 0000000000..5678718188
> --- /dev/null
> +++ b/arch/arm/boards/enclustra-aa1/Makefile
> @@ -0,0 +1,4 @@
> +# SPDX-License-Identifier: GPL-2.0-only
> +
> +lwl-y += lowlevel.o
> +obj-y += board.o
> diff --git a/arch/arm/boards/enclustra-aa1/board.c b/arch/arm/boards/enclustra-aa1/board.c
> new file mode 100644
> index 0000000000..5b8e5a5c9f
> --- /dev/null
> +++ b/arch/arm/boards/enclustra-aa1/board.c
> @@ -0,0 +1,47 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +
> +#include <common.h>
> +#include <init.h>
> +#include <io.h>
> +#include <bbu.h>
> +
> +static int aa1_init(void)
> +{
> + int pbl_index = 0;
> + uint32_t flag_barebox1 = 0;
> + uint32_t flag_barebox2 = 0;
> +
> + if (!of_machine_is_compatible("enclustra,mercury-aa1"))
> + return 0;
> +
> + pbl_index = readl(0xFFD06210);
This is ARRIA10_SYSMGR_ROM_INITSWLASTLD, right? Please use it.
> +
> + pr_debug("Current barebox instance %d\n", pbl_index);
> +
> + switch (pbl_index) {
> + case 0:
> + flag_barebox1 |= BBU_HANDLER_FLAG_DEFAULT;
> + break;
> + case 1:
> + flag_barebox2 |= BBU_HANDLER_FLAG_DEFAULT;
> + break;
> + };
> +
> + bbu_register_std_file_update("emmc-barebox1-xload", flag_barebox1,
> + "/dev/mmc0.barebox1-xload",
> + filetype_socfpga_xload);
> +
> + bbu_register_std_file_update("emmc-barebox1", 0,
> + "/dev/mmc0.barebox1",
> + filetype_arm_barebox);
> +
> + bbu_register_std_file_update("emmc-barebox2-xload", flag_barebox2,
> + "/dev/mmc0.barebox2-xload",
> + filetype_socfpga_xload);
> +
> + bbu_register_std_file_update("emmc-barebox2", 0,
> + "/dev/mmc0.barebox2",
> + filetype_arm_barebox);
Should this be turned into something more intelligent like failsafe
update?
> + return 0;
> +}
> +postcore_initcall(aa1_init);
You could turn this into a postcore_platform_driver.
> +#define BAREBOX_PART 0
> +#define BITSTREAM_PART 1
> +#define BAREBOX1_OFFSET SZ_1M
> +#define BAREBOX2_OFFSET BAREBOX1_OFFSET + SZ_512K
> +#define BAREBOX3_OFFSET BAREBOX2_OFFSET + SZ_512K
> +#define BAREBOX4_OFFSET BAREBOX3_OFFSET + SZ_512K
> +#define BITSTREAM1_OFFSET 0x0
>From looking into the device tree I would expect BITSTREAM1_OFFSET to be 0x700000.
> +#define BITSTREAM2_OFFSET BITSTREAM1_OFFSET + SZ_32M
You should add braces around the macro definitions to make them safe
to use.
> +
> +extern char __dtb_z_socfpga_arria10_mercury_aa1_start[];
> +
> +#define ARRIA10_STACKTOP ARRIA10_OCRAM_ADDR + SZ_256K
> +
> +ENTRY_FUNCTION_WITHSTACK(start_socfpga_aa1_xload, ARRIA10_STACKTOP, r0, r1, r2)
> +{
> + int pbl_index = 0;
> + int barebox = 0;
> + int bitstream = 0;
> +
> + arm_cpu_lowlevel_init();
> +
> + relocate_to_current_adr();
> +
> + setup_c();
> +
> + arria10_init(&mainpll_cfg, &perpll_cfg, pinmux);
> +
> + arria10_prepare_mmc(BAREBOX_PART, BITSTREAM_PART);
> +
> + pbl_index = readl(ARRIA10_SYSMGR_ROM_INITSWLASTLD);
> +
> + /* Allow booting from both PBL0 and PBL1 to allow atomic updates.
> + * Bitstreams redundant too and expected to reside in the second
> + * partition.
> + * There is a fixed relation between the PBL/barebox instance and its
> + * bitstream location (offset) that requires to update them together */
> + switch (pbl_index) {
> + case 0:
> + barebox = BAREBOX1_OFFSET;
> + bitstream = BITSTREAM1_OFFSET;
> + break;
> + case 1:
> + barebox = BAREBOX2_OFFSET;
> + bitstream = BITSTREAM1_OFFSET;
> + break;
> + case 2:
> + case 3:
> + /* Left blank for future extension */
> + break;
You should either bail out or use a sane default for unhandled cases.
> + }
> +
> + arria10_load_fpga(bitstream, SZ_32M);
> +
> + arria10_finish_io(&mainpll_cfg, &perpll_cfg, pinmux);
> +
> + arria10_ddr_calibration_sequence();
> +
> + arria10_start_image(barebox);
> +}
> +
> +ENTRY_FUNCTION(start_socfpga_aa1, r0, r1, r2)
> +{
> + void *fdt;
> +
> + fdt = __dtb_z_socfpga_arria10_mercury_aa1_start + get_runtime_offset();
> +
> + barebox_arm_entry(0x0, SZ_2G, fdt);
> +}
> +
> +ENTRY_FUNCTION_WITHSTACK(start_socfpga_aa1_bringup, ARRIA10_STACKTOP, r0, r1, r2)
> +{
> + void *fdt;
> +
> + arm_cpu_lowlevel_init();
> +
> + relocate_to_current_adr();
> + setup_c();
> +
> + arria10_init(&mainpll_cfg, &perpll_cfg, pinmux);
> +
> + /* wait for fpga_usermode */
> + a10_wait_for_usermode(0x1000000);
> +
> + arria10_finish_io(&mainpll_cfg, &perpll_cfg, pinmux);
> +
> + arria10_ddr_calibration_sequence();
> +
> + fdt = __dtb_z_socfpga_arria10_mercury_aa1_start + get_runtime_offset();
> +
> + barebox_arm_entry(0x0, SZ_2G, fdt);
> +}
> diff --git a/arch/arm/dts/socfpga_arria10_mercury_aa1.dts b/arch/arm/dts/socfpga_arria10_mercury_aa1.dts
> new file mode 100644
> index 0000000000..ef3afc9b98
> --- /dev/null
> +&mmc {
> + bus-width = <8>;
> + non-removable;
> + disable-wp;
> + no-sd;
> +
> + partitions {
> + compatible = "fixed-partitions";
> + #size-cells = <1>;
> + #address-cells = <1>;
> +
> + barebox1_xload: partition@100000 {
> + label = "barebox1-xload";
> + reg = <0x100000 0x40000>;
> + };
> +
> + barebox2_xload: partition@140000 {
> + label = "barebox2-xload";
> + reg = <0x140000 0x40000>;
> + };
> +
> + barebox1: partition@200000 {
> + label = "barebox1";
> + reg = <0x200000 0x80000>;
> + };
> +
> + barebox2: partition@280000 {
> + label = "barebox2";
> + reg = <0x280000 0x80000>;
> + };
It might be worth increasing the size to 1MiB. It's easy to make barebox
bigger than 512KiB.
Sascha
--
Pengutronix e.K. | |
Steuerwalder Str. 21 | http://www.pengutronix.de/ |
31137 Hildesheim, Germany | Phone: +49-5121-206917-0 |
Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
next prev parent reply other threads:[~2022-07-11 8:56 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-07-11 7:52 [PATCH v3 1/6] scripts: socfgpa_xml_to_config: document pincfg Steffen Trumtrar
2022-07-11 7:52 ` [PATCH v3 2/6] ARM: socfpga: achilles: remove cargo-cult Steffen Trumtrar
2022-07-11 7:52 ` [PATCH v3 3/6] ARM: socfpga: achilles: use dtbz instead of dtb Steffen Trumtrar
2022-07-11 7:52 ` [PATCH v3 4/6] ARM: socfpga: achilles: use ENTRY_FUNCTION_WITHSTACK Steffen Trumtrar
2022-07-11 7:52 ` [PATCH v3 5/6] ARM: socfpga: add support for Enclustra AA1 Steffen Trumtrar
2022-07-11 8:54 ` Sascha Hauer [this message]
2022-07-12 5:48 ` Steffen Trumtrar
2022-07-11 7:52 ` [PATCH v3 6/6] ARM: socfpga: defconfig: add aa1 Steffen Trumtrar
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20220711085415.GM5208@pengutronix.de \
--to=sha@pengutronix.de \
--cc=barebox@lists.infradead.org \
--cc=s.trumtrar@pengutronix.de \
--cc=str@pengutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox