mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: Michael Tretter <m.tretter@pengutronix.de>
To: "Thomas Hämmerle" <Thomas.Haemmerle@wolfvision.net>
Cc: "barebox@lists.infradead.org" <barebox@lists.infradead.org>
Subject: Re: [PATCH 3/5] firmware: zynqmp-fpga: introduce driver to load bitstream to FPGA
Date: Thu, 24 Oct 2019 14:56:55 +0200	[thread overview]
Message-ID: <20191024145655.1c939d37@litschi.hi.pengutronix.de> (raw)
In-Reply-To: <1571912781-17152-4-git-send-email-thomas.haemmerle@wolfvision.net>

On Thu, 24 Oct 2019 10:26:51 +0000, Thomas Hämmerle wrote:
> From: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
> 
> The driver provides functionalities to check and load a bitstream to FPGA.
> A boolean parameter to check if FPGA is already programmed is
> added.
> 
> Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
> ---
>  arch/arm/configs/zynqmp_defconfig                  |   1 +
>  .../arm/mach-zynqmp/include/mach/firmware-zynqmp.h |   2 +
>  drivers/firmware/Kconfig                           |   5 +
>  drivers/firmware/Makefile                          |   1 +
>  drivers/firmware/zynqmp-fpga.c                     | 315 +++++++++++++++++++++
>  5 files changed, 324 insertions(+)
>  create mode 100644 drivers/firmware/zynqmp-fpga.c
> 
> diff --git a/arch/arm/configs/zynqmp_defconfig b/arch/arm/configs/zynqmp_defconfig
> index 4dea964..834212e 100644
> --- a/arch/arm/configs/zynqmp_defconfig
> +++ b/arch/arm/configs/zynqmp_defconfig
> @@ -35,4 +35,5 @@ CONFIG_CMD_OFTREE=y
>  CONFIG_CMD_TIME=y
>  CONFIG_DRIVER_SERIAL_CADENCE=y
>  # CONFIG_SPI is not set
> +CONFIG_FIRMWARE_ZYNQMP_PL=y
>  CONFIG_DIGEST=y
> diff --git a/arch/arm/mach-zynqmp/include/mach/firmware-zynqmp.h b/arch/arm/mach-zynqmp/include/mach/firmware-zynqmp.h
> index f19d73d..6fcfba5 100644
> --- a/arch/arm/mach-zynqmp/include/mach/firmware-zynqmp.h
> +++ b/arch/arm/mach-zynqmp/include/mach/firmware-zynqmp.h
> @@ -23,6 +23,8 @@
>  #define ZYNQMP_FPGA_BIT_ENC_DEV_KEY	BIT(4)
>  #define ZYNQMP_FPGA_BIT_ONLY_BIN	BIT(5)
>  
> +#define ZYNQMP_PCAP_STATUS_FPGA_DONE	BIT(3)
> +

Should be part of the previous patch.

>  enum pm_ioctl_id {
>  	IOCTL_SET_PLL_FRAC_MODE = 8,
>  	IOCTL_GET_PLL_FRAC_MODE,
> diff --git a/drivers/firmware/Kconfig b/drivers/firmware/Kconfig
> index 710b500..3113f40 100644
> --- a/drivers/firmware/Kconfig
> +++ b/drivers/firmware/Kconfig
> @@ -14,4 +14,9 @@ config FIRMWARE_ALTERA_SOCFPGA
>  	bool "Altera SoCFPGA fpga loader"
>  	depends on ARCH_SOCFPGA
>  	select FIRMWARE
> +	
> +config FIRMWARE_ZYNQMP_FPGA
> +	bool "Xilinx ZynqMP FPGA loader"
> +	depends on ARCH_ZYNQMP
> +	select FIRMWARE
>  endmenu
> diff --git a/drivers/firmware/Makefile b/drivers/firmware/Makefile
> index c3a3c34..b162b08 100644
> --- a/drivers/firmware/Makefile
> +++ b/drivers/firmware/Makefile
> @@ -1,2 +1,3 @@
>  obj-$(CONFIG_FIRMWARE_ALTERA_SERIAL) += altera_serial.o
>  obj-$(CONFIG_FIRMWARE_ALTERA_SOCFPGA) += socfpga.o
> +obj-$(CONFIG_FIRMWARE_ZYNQMP_FPGA) += zynqmp-fpga.o
> diff --git a/drivers/firmware/zynqmp-fpga.c b/drivers/firmware/zynqmp-fpga.c
> new file mode 100644
> index 0000000..6a32d28
> --- /dev/null
> +++ b/drivers/firmware/zynqmp-fpga.c
> @@ -0,0 +1,315 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + * Xilinx Zynq MPSoC PL loading
> + *
> + * Copyright (c) 2018 Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
> + *
> + * based on U-Boot zynqmppl code
> + *
> + * (C) Copyright 2015 - 2016, Xilinx, Inc,
> + * Michal Simek <michal.simek@xilinx.com>
> + * Siva Durga Prasad <siva.durga.paladugu@xilinx.com> *
> + */
> +
> +#include <firmware.h>
> +#include <common.h>
> +#include <init.h>
> +#include <dma.h>
> +#include <mach/firmware-zynqmp.h>
> +
> +#define ZYNQMP_PM_VERSION_1_1			((1 << 16) | 1)

A macro ZYNQMP_PM_VERSION(MAJOR, MINOR) would be more intuitive.

> +
> +#define ZYNQMP_PM_FEATURE_BYTE_ORDER_IRREL	BIT(0)
> +
> +#define ZYNQMP_PM_VERSION_1_0_FEATURES	0
> +#define ZYNQMP_PM_VERSION_1_1_FEATURES  ZYNQMP_PM_FEATURE_BYTE_ORDER_IRREL
> +
> +#define DUMMY_WORD			0xFFFFFFFF
> +
> +#define XILINX_BYTE_ORDER_BIT		0
> +#define XILINX_BYTE_ORDER_BIN		1
> +
> +struct fpgamgr {
> +	struct firmware_handler fh;
> +	struct device_d dev;
> +	const struct zynqmp_eemi_ops *eemi_ops;
> +	int programmed;
> +	char *buf;
> +	size_t size;
> +	u32 features;
> +};
> +
> +/* Xilinx binary format header */
> +static const u32 bin_format[] = {
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	0x000000bb,
> +	0x11220044,
> +	DUMMY_WORD,
> +	DUMMY_WORD,
> +	0xaa995566,
> +};
> +
> +static void copy_words_swapped(u32 *dst, const u32 *src, size_t size)
> +{
> +	int i;
> +
> +	for (i = 0; i < size; i++)
> +		dst[i] = __swab32(src[i]);
> +}
> +
> +static int get_byte_order(const u32 *buf, size_t size)
> +{
> +	u32 buf_check[ARRAY_SIZE(bin_format)];
> +
> +	memcpy(buf_check, buf, ARRAY_SIZE(buf_check));
> +	if (memcmp(buf_check, bin_format, sizeof(buf_check)) == 0)
> +		return XILINX_BYTE_ORDER_BIT;
> +
> +	copy_words_swapped(buf_check, buf, ARRAY_SIZE(buf_check));
> +	if (memcmp(buf_check, bin_format, sizeof(buf_check)) == 0)
> +		return XILINX_BYTE_ORDER_BIN;
> +
> +	return -EINVAL;
> +}

The function does not only find the byte order, but also checks if the
header is valid. I would prefer, if we could have a function for
getting the byte order (no copy necessary, just check for one of the
values that allow to distinguish the byte order) and another function
for checking if the header is valid. 

> +
> +static int get_header_length(const char *buf, size_t size)
> +{
> +	u32 *buf_u32;
> +	int p;
> +
> +	for (p = 0; p < size; p++) {
> +		buf_u32 = (u32 *)&buf[p];
> +		if (*buf_u32 == DUMMY_WORD)
> +			return p;
> +	}
> +	return -EINVAL;
> +}
> +
> +static int fpgamgr_program_finish(struct firmware_handler *fh)
> +{
> +	struct fpgamgr *mgr = container_of(fh, struct fpgamgr, fh);
> +	char *buf_aligned;
> +	u32 *buf_size = NULL;
> +	int header_length = 0, byte_order = 0;
> +	u64 addr;
> +	int status = 0, xilfpga_old = 1;
> +	u8 flags = 0;
> +
> +	if (!mgr->buf) {
> +		status = -ENOBUFS;
> +		dev_err(&mgr->dev, "buffer is NULL\n");
> +		goto err_free;
> +	}
> +
> +	header_length = get_header_length(mgr->buf, mgr->size);
> +	if (header_length < 0) {
> +		status = header_length;
> +		goto err_free;
> +	}
> +
> +	byte_order = get_byte_order((u32 *)&mgr->buf[header_length],
> +			mgr->size - header_length);

I think helper variables u32 *body and size_t body_size would probably make
this a lot more readable.

> +	if (byte_order < 0) {
> +		status = byte_order;
> +		goto err_free;
> +	}
> +
> +	if (mgr->features & ZYNQMP_PM_FEATURE_BYTE_ORDER_IRREL) {
> +		/* PMUFW version > 1.0 accepts both byte orders*/
> +		byte_order = 0;
> +		xilfpga_old = 0;
> +	} else {
> +		buf_size = dma_alloc_coherent(sizeof(*buf_size),
> +		DMA_ADDRESS_BROKEN);
> +		if (!buf_size) {
> +			status = -ENOBUFS;
> +			goto err_free;
> +		}
> +		*buf_size = mgr->size - header_length;
> +	}
> +
> +	buf_aligned = dma_alloc_coherent(mgr->size - header_length,
> +			DMA_ADDRESS_BROKEN);
> +	if (!buf_aligned) {
> +		status = -ENOBUFS;
> +		goto err_free;
> +	}
> +
> +	if (byte_order == XILINX_BYTE_ORDER_BIN)
> +		copy_words_swapped((u32 *)buf_aligned,
> +				(u32 *)(&(mgr->buf[header_length])),
> +				(mgr->size - header_length) / sizeof(u32));
> +	else
> +		memcpy((u32 *)buf_aligned, (u32 *)(&(mgr->buf[header_length])),
> +			(mgr->size - header_length));
> +
> +	addr = (u64)buf_aligned;
> +
> +	/* we do not provide a header */
> +	flags |= ZYNQMP_FPGA_BIT_ONLY_BIN;
> +
> +	if (xilfpga_old && buf_size) {

The xilfpga_old variable is pretty hard to follow. Maybe always check
for the feature flag or test the feature flag once and set a local
variable (with a proper name) that is used whenever the feature flag is
relevant.

Michael

> +		status = mgr->eemi_ops->fpga_load(addr,
> +				(u32)(uintptr_t)buf_size,
> +				flags);
> +		dma_free_coherent(buf_size, 0, sizeof(*buf_size));
> +	} else {
> +		status = mgr->eemi_ops->fpga_load(addr,
> +				(u32)(mgr->size - header_length),
> +				flags);
> +	}
> +
> +	if (status < 0)
> +		dev_err(&mgr->dev, "unable to load fpga\n");
> +
> +	dma_free_coherent(buf_aligned, 0, mgr->size - header_length);
> +
> + err_free:
> +	free(mgr->buf);
> +
> +	return status;
> +}
> +
> +static int fpgamgr_program_write_buf(struct firmware_handler *fh,
> +		const void *buf, size_t size)
> +{
> +	struct fpgamgr *mgr = container_of(fh, struct fpgamgr, fh);
> +
> +	/* Since write() is called by copy_file, we only receive chuncks with
> +	 * size RW_BUF_SIZE of the bitstream.
> +	 * Buffer the chunks here and handle it in close()
> +	 */
> +
> +	mgr->buf = realloc(mgr->buf, mgr->size + size);
> +	if (!mgr->buf)
> +		return -ENOBUFS;
> +
> +	memcpy(&(mgr->buf[mgr->size]), buf, size);
> +	mgr->size += size;
> +
> +	return 0;
> +}
> +
> +static int fpgamgr_program_start(struct firmware_handler *fh)
> +{
> +	struct fpgamgr *mgr = container_of(fh, struct fpgamgr, fh);
> +
> +	mgr->size = 0;
> +	mgr->buf = NULL;
> +
> +	return 0;
> +}
> +
> +static int programmed_get(struct param_d *p, void *priv)
> +{
> +	struct fpgamgr *mgr = priv;
> +	u32 status = 0x00;
> +	int ret = 0;
> +
> +	ret = mgr->eemi_ops->fpga_getstatus(&status);
> +	if (ret)
> +		return ret;
> +
> +	mgr->programmed = !!(status & ZYNQMP_PCAP_STATUS_FPGA_DONE);
> +
> +	return 0;
> +}
> +
> +static int zynqmp_fpga_probe(struct device_d *dev)
> +{
> +	struct fpgamgr *mgr;
> +	struct firmware_handler *fh;
> +	const char *alias = of_alias_get(dev->device_node);
> +	const char *model = NULL;
> +	struct param_d *p;
> +	u32 api_version;
> +	int ret;
> +
> +	mgr = xzalloc(sizeof(*mgr));
> +	fh = &mgr->fh;
> +
> +	if (alias)
> +		fh->id = xstrdup(alias);
> +	else
> +		fh->id = xstrdup("zynqmp-fpga-manager");
> +
> +	fh->open = fpgamgr_program_start;
> +	fh->write = fpgamgr_program_write_buf;
> +	fh->close = fpgamgr_program_finish;
> +	of_property_read_string(dev->device_node, "compatible", &model);
> +	if (model)
> +		fh->model = xstrdup(model);
> +	fh->dev = dev;
> +
> +	mgr->eemi_ops = zynqmp_pm_get_eemi_ops();
> +
> +	ret = mgr->eemi_ops->get_api_version(&api_version);
> +	if (ret) {
> +		dev_err(&mgr->dev, "could not get API version\n");
> +		goto out;
> +	}
> +
> +	mgr->features = 0;
> +
> +	if (api_version >= ZYNQMP_PM_VERSION_1_1)
> +		mgr->features |= ZYNQMP_PM_FEATURE_BYTE_ORDER_IRREL;
> +
> +	dev_dbg(dev, "Registering ZynqMP FPGA programmer\n");
> +	mgr->dev.id = DEVICE_ID_SINGLE;
> +	dev_set_name(&mgr->dev, "zynqmp_fpga");
> +	mgr->dev.parent = dev;
> +	ret = register_device(&mgr->dev);
> +	if (ret)
> +		goto out;
> +
> +	p = dev_add_param_bool(&mgr->dev, "programmed", NULL, programmed_get,
> +			&mgr->programmed, mgr);
> +	if (IS_ERR(p)) {
> +		ret = PTR_ERR(p);
> +		goto out_unreg;
> +	}
> +
> +	fh->dev = &mgr->dev;
> +	ret = firmwaremgr_register(fh);
> +	if (ret != 0) {
> +		free(mgr);
> +		goto out_unreg;
> +	}
> +
> +	return 0;
> +out_unreg:
> +	unregister_device(&mgr->dev);
> +out:
> +	free(fh->id);
> +	free(mgr);
> +
> +	return ret;
> +}
> +
> +static struct of_device_id zynqmpp_fpga_id_table[] = {
> +	{
> +		.compatible = "xlnx,zynqmp-pcap-fpga",
> +	},
> +};
> +
> +static struct driver_d zynqmp_fpga_driver = {
> +	.name = "zynqmp_fpga_manager",
> +	.of_compatible = DRV_OF_COMPAT(zynqmpp_fpga_id_table),
> +	.probe = zynqmp_fpga_probe,
> +};
> +device_platform_driver(zynqmp_fpga_driver);

_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox

  reply	other threads:[~2019-10-24 12:56 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-24 10:26 [PATCH 0/5] ARM: zynqmp: add support for bitstream loading Thomas Hämmerle
2019-10-24 10:26 ` [PATCH 1/5] ARM: zynqmp: dts: move firmware node to src tree Thomas Hämmerle
2019-10-24 12:56   ` Michael Tretter
2019-10-24 10:26 ` [PATCH 2/5] firmware-zynqmp: extend driver with fpga relavant functions Thomas Hämmerle
2019-10-24 12:56   ` Michael Tretter
2019-10-24 10:26 ` [PATCH 3/5] firmware: zynqmp-fpga: introduce driver to load bitstream to FPGA Thomas Hämmerle
2019-10-24 12:56   ` Michael Tretter [this message]
2019-10-24 10:26 ` [PATCH 5/5] firmware: zynqmp-fpga: print Xilinx bitstream header Thomas Hämmerle
2019-10-24 10:26 ` [PATCH 4/5] ARM: zynqmp: dts: move pcap node to src tree Thomas Hämmerle
2019-10-25 12:55 ` [PATCH v2 0/4] ARM: zynqmp: add support for bitstream loading Thomas Hämmerle
2019-10-25 12:55   ` [PATCH v2 1/4] firmware-zynqmp: add macros for PMU and trustzone firmware versions Thomas Hämmerle
2019-10-25 12:55   ` [PATCH v2 2/4] firmware-zynqmp: extend driver with fpga relavant functions Thomas Hämmerle
2019-10-25 12:55   ` [PATCH v2 4/4] firmware: zynqmp-fpga: print Xilinx bitstream header Thomas Hämmerle
2019-10-25 12:55   ` [PATCH v2 3/4] firmware: zynqmp-fpga: introduce driver to load bitstream to FPGA Thomas Hämmerle
2019-10-28 11:00   ` [PATCH v2 0/4] ARM: zynqmp: add support for bitstream loading Sascha Hauer

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=20191024145655.1c939d37@litschi.hi.pengutronix.de \
    --to=m.tretter@pengutronix.de \
    --cc=Thomas.Haemmerle@wolfvision.net \
    --cc=barebox@lists.infradead.org \
    /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