On Sat, Jan 02, 2021 at 11:13:01AM +0800, Xu Yilun wrote:
> This patch supports the DFL drivers be written in userspace. This is
> realized by exposing the userspace I/O device interfaces.
>
> The driver leverages the uio_pdrv_genirq, it adds the uio_pdrv_genirq
> platform device with the DFL device's resources, and let the generic UIO
> platform device driver provide support to userspace access to kernel
> interrupts and memory locations.
>
> The driver now supports the ether group feature. To support a new DFL
> feature been directly accessed via UIO, its feature id should be added to
> the driver's id_table.
>
> Signed-off-by: Xu Yilun <yilun...@intel.com>
> Reviewed-by: Tom Rix <t...@redhat.com>
> ---
> v2: switch to the new matching algorithem. It matches DFL devices which
> could not be handled by other DFL drivers.
> refacor the code about device resources filling.
> fix some comments.
> v3: split the dfl.c changes out of this patch.
> some minor fixes
> v4: drop the idea of a generic matching algorithem, instead we specify
> each matching device in id_table.
> to make clear that only one irq is supported, the irq handling code
> is refactored.
> v5: refactor the irq resource code.
> ---
> drivers/fpga/Kconfig | 10 +++++
> drivers/fpga/Makefile | 1 +
> drivers/fpga/dfl-uio-pdev.c | 91
> +++++++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 102 insertions(+)
> create mode 100644 drivers/fpga/dfl-uio-pdev.c
>
> diff --git a/drivers/fpga/Kconfig b/drivers/fpga/Kconfig
> index 5ff9438..61445be 100644
> --- a/drivers/fpga/Kconfig
> +++ b/drivers/fpga/Kconfig
> @@ -203,6 +203,16 @@ config FPGA_DFL_NIOS_INTEL_PAC_N3000
> the card. It also instantiates the SPI master (spi-altera) for
> the card's BMC (Board Management Controller).
>
> +config FPGA_DFL_UIO_PDEV
> + tristate "FPGA DFL Driver for Userspace I/O platform devices"
> + depends on FPGA_DFL && UIO_PDRV_GENIRQ
> + help
> + Enable this to allow some DFL drivers be written in userspace. It
> + adds the uio_pdrv_genirq platform device with the DFL feature's
> + resources, and lets the generic UIO platform device driver provide
> + support for userspace access to kernel interrupts and memory
> + locations.
> +
> config FPGA_DFL_PCI
> tristate "FPGA DFL PCIe Device Driver"
> depends on PCI && FPGA_DFL
> diff --git a/drivers/fpga/Makefile b/drivers/fpga/Makefile
> index 18dc9885..8847fe0 100644
> --- a/drivers/fpga/Makefile
> +++ b/drivers/fpga/Makefile
> @@ -45,6 +45,7 @@ dfl-afu-objs := dfl-afu-main.o dfl-afu-region.o
> dfl-afu-dma-region.o
> dfl-afu-objs += dfl-afu-error.o
>
> obj-$(CONFIG_FPGA_DFL_NIOS_INTEL_PAC_N3000) += dfl-n3000-nios.o
> +obj-$(CONFIG_FPGA_DFL_UIO_PDEV) += dfl-uio-pdev.o
>
> # Drivers for FPGAs which implement DFL
> obj-$(CONFIG_FPGA_DFL_PCI) += dfl-pci.o
> diff --git a/drivers/fpga/dfl-uio-pdev.c b/drivers/fpga/dfl-uio-pdev.c
> new file mode 100644
> index 0000000..a4cd581
> --- /dev/null
> +++ b/drivers/fpga/dfl-uio-pdev.c
> @@ -0,0 +1,91 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * DFL driver for Userspace I/O platform devices
> + *
> + * Copyright (C) 2020 Intel Corporation, Inc.
> + */
> +#include <linux/dfl.h>
> +#include <linux/errno.h>
> +#include <linux/kernel.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +#include <linux/slab.h>
> +#include <linux/uio_driver.h>
> +
> +#define DRIVER_NAME "dfl-uio-pdev"
> +
> +static int dfl_uio_pdev_probe(struct dfl_device *ddev)
> +{
> + struct platform_device_info pdevinfo = { 0 };
> + struct uio_info uio_pdata = { 0 };
> + struct platform_device *uio_pdev;
> + struct device *dev = &ddev->dev;
> + unsigned int num_res = 1;
> + struct resource res[2];
> +
> + res[0].parent = &ddev->mmio_res;
> + res[0].flags = IORESOURCE_MEM;
> + res[0].start = ddev->mmio_res.start;
> + res[0].end = ddev->mmio_res.end;
> +
> + if (ddev->num_irqs) {
> + if (ddev->num_irqs > 1)
> + dev_warn(&ddev->dev,
> + "%d irqs for %s, but UIO only supports the
> first one\n",
> + ddev->num_irqs, dev_name(&ddev->dev));
> +
> + res[1].flags = IORESOURCE_IRQ;
> + res[1].start = ddev->irqs[0];
> + res[1].end = ddev->irqs[0];
> + num_res++;
> + }
> +
> + uio_pdata.name = DRIVER_NAME;
> + uio_pdata.version = "0";
> +
> + pdevinfo.name = "uio_pdrv_genirq";
> + pdevinfo.res = res;
> + pdevinfo.num_res = num_res;
> + pdevinfo.parent = &ddev->dev;
> + pdevinfo.id = PLATFORM_DEVID_AUTO;
> + pdevinfo.data = &uio_pdata;
> + pdevinfo.size_data = sizeof(uio_pdata);
> +
> + uio_pdev = platform_device_register_full(&pdevinfo);
> + if (!IS_ERR(uio_pdev))
> + dev_set_drvdata(dev, uio_pdev);
I'm not sure if this is more readable than:
uio_pdev = platform_device_register_full(&pdevinfo);
if (IS_ERR(uio_pdev))
return PTR_ERR(uio_pdev);
dev_set_drvdata(dev, uio_pdev);
return 0;
No strong preference, though ... :)
> +
> + return PTR_ERR_OR_ZERO(uio_pdev);
> +}
> +
> +static void dfl_uio_pdev_remove(struct dfl_device *ddev)
> +{
> + struct platform_device *uio_pdev = dev_get_drvdata(&ddev->dev);
> +
> + platform_device_unregister(uio_pdev);
> +}
> +
> +#define FME_FEATURE_ID_ETH_GROUP 0x10
> +
> +static const struct dfl_device_id dfl_uio_pdev_ids[] = {
> + { FME_ID, FME_FEATURE_ID_ETH_GROUP },
Will you want to always bind FME_FEATURE_ID_ETH_GROUP? If not I'd suggest not
to add it here.
If you want to provide an option to somewhat non-ABI fixed bind things
you could look at what vfio-pci does (provide a module parameter),
otherwise use sysfs 'new_id' or 'bind'.
Does that sound reasonable?
> +
> + /* Add your new id entries here to support uio for more dfl features */
> +
> + { }
> +};
> +MODULE_DEVICE_TABLE(dfl, dfl_uio_pdev_ids);
> +
> +static struct dfl_driver dfl_uio_pdev_driver = {
> + .drv = {
> + .name = DRIVER_NAME,
> + },
> + .id_table = dfl_uio_pdev_ids,
> + .probe = dfl_uio_pdev_probe,
> + .remove = dfl_uio_pdev_remove,
> +};
> +module_dfl_driver(dfl_uio_pdev_driver);
> +
> +MODULE_DESCRIPTION("DFL driver for Userspace I/O platform devices");
> +MODULE_AUTHOR("Intel Corporation");
> +MODULE_LICENSE("GPL v2");
> --
> 2.7.4
>
Thanks,
Moritz
