> -----Original Message-----
> From: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> Sent: Friday, August 28, 2020 1:44 PM
> To: kuldip dwivedi <kuldip.dwivedi@xxxxxxxxxxxxxxxx>
> Cc: Jiri Slaby <jirislaby@xxxxxxxxxx>; Dmitry Safonov
<0x7f454c46@xxxxxxxxx>;
> linux-serial@xxxxxxxxxxxxxxx; Vabhav Sharma <vabhav.sharma@xxxxxxx>;
Pankaj
> Bansal <pankaj.bansal@xxxxxxx>; Varun Sethi <V.Sethi@xxxxxxx>
> Subject: Re: [PATCH v2] serial: 8250_fsl: Add ACPI support
>
> On Tue, Aug 25, 2020 at 07:01:35PM +0530, kuldip dwivedi wrote:
> > This adds support for ACPI enumerated FSL 16550 UARTs.
> > For supporting ACPI, I added a wrapper so that this driver can be used
> > if firmware has exposed the HID "NXP0018" in DSDT table.
> >
> > This will be built as object file if config "SERIAL_8250_FSL"
> > is enabled which depends on config "SERIAL_8250_CONSOLE".
> >
> > Signed-off-by: kuldip dwivedi <kuldip.dwivedi@xxxxxxxxxxxxxxxx>
> > ---
> >
> > Notes:
> > 1. revert the license change from v1
> > 2. revert the file header format change from v1
> > 3. remove extra headers from v1
> > 4. this is tested on layerscape platform LS1046A
> > with the firmware having ACPI support.
> >
> > This will be built as object file if config " SERIAL_8250_FSL"
> > is enabled which depends on config " SERIAL_8250_CONSOLE ".
> > this driver will be probed if firmware has exposed HID "
> > NXP0018" in DSDT table
> >
> > drivers/tty/serial/8250/8250_fsl.c | 98
> > ++++++++++++++++++++++++++++++
> > 1 file changed, 98 insertions(+)
> >
> > diff --git a/drivers/tty/serial/8250/8250_fsl.c
> > b/drivers/tty/serial/8250/8250_fsl.c
> > index 0d0c80905c58..62e5e2d33f69 100644
> > --- a/drivers/tty/serial/8250/8250_fsl.c
> > +++ b/drivers/tty/serial/8250/8250_fsl.c
> > @@ -1,11 +1,16 @@
> > // SPDX-License-Identifier: GPL-2.0
> > +#include <linux/acpi.h>
> > #include <linux/serial_reg.h>
> > #include <linux/serial_8250.h>
> >
> > #include "8250.h"
> >
> > +#define DRIVER_NAME "fsl-ns16550-uart"
>
> Why is this needed?
As this is not standalone module so not required. I added MODULE_ALIAS
but I don't think it is required so I will remove this in v3 PATCH.
>
>
> > +
> > /*
> > * Freescale 16550 UART "driver", Copyright (C) 2011 Paul Gortmaker.
> > + * Copyright 2020 NXP
> > + * Copyright 2020 Puresoftware Ltd.
>
> Copyright info goes at the top of the file, not down here.
Old Copyright was already there so I just added our copyright but I will
move
Copyright and file description at top in v3
>
> > *
> > * This isn't a full driver; it just provides an alternate IRQ
> > * handler to deal with an errata. Everything else is just @@ -20,6
> > +25,10 @@
> > * IRQ event to the next one.
> > */
> >
> > +struct fsl8250_data {
> > + int line;
> > +};
> > +
> > int fsl8250_handle_irq(struct uart_port *port) {
> > unsigned char lsr, orig_lsr;
> > @@ -79,3 +88,92 @@ int fsl8250_handle_irq(struct uart_port *port)
> > return 1;
> > }
> > EXPORT_SYMBOL_GPL(fsl8250_handle_irq);
> > +
> > +static int fsl8250_acpi_probe(struct platform_device *pdev) {
> > + struct fsl8250_data *data;
> > + struct uart_8250_port port8250;
> > + struct device *dev = &pdev->dev;
> > + struct resource *regs;
> > +
> > + int ret, irq;
> > +
> > + regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> > + if (!regs) {
> > + dev_err(dev, "no registers defined\n");
> > + return -EINVAL;
> > + }
> > +
> > + irq = platform_get_irq(pdev, 0);
> > + if (irq < 0) {
> > + if (irq != -EPROBE_DEFER)
> > + dev_err(dev, "cannot get irq\n");
> > + return irq;
> > + }
> > +
> > + memset(&port8250, 0, sizeof(port8250));
> > +
> > + ret = device_property_read_u32(dev, "clock-frequency",
> > + &port8250.port.uartclk);
> > + if (ret)
> > + return ret;
> > +
> > + spin_lock_init(&port8250.port.lock);
> > +
> > + port8250.port.mapbase = regs->start;
> > + port8250.port.irq = irq;
> > + port8250.port.handle_irq = fsl8250_handle_irq;
> > + port8250.port.type = PORT_16550A;
> > + port8250.port.flags = UPF_SHARE_IRQ |
UPF_BOOT_AUTOCONF
> > + | UPF_FIXED_PORT |
> UPF_IOREMAP
> > + | UPF_FIXED_TYPE;
> > + port8250.port.dev = dev;
> > + port8250.port.mapsize = resource_size(regs);
> > + port8250.port.iotype = UPIO_MEM;
> > + port8250.port.irqflags = IRQF_SHARED;
> > +
> > + port8250.port.membase = devm_ioremap(dev, port8250.port.mapbase,
> > +
port8250.port.mapsize);
> > + if (!port8250.port.membase)
> > + return -ENOMEM;
> > +
> > + data = devm_kzalloc(dev, sizeof(*data), GFP_KERNEL);
> > + if (!data)
> > + return -ENOMEM;
> > +
> > + data->line = serial8250_register_8250_port(&port8250);
> > + if (data->line < 0)
> > + ret = data->line;
> > +
> > + platform_set_drvdata(pdev, data);
> > + return 0;
> > +}
> > +
> > +static int fsl8250_acpi_remove(struct platform_device *pdev) {
> > + struct fsl8250_data *data = platform_get_drvdata(pdev);
> > +
> > + serial8250_unregister_port(data->line);
> > + return 0;
> > +}
> > +
> > +static const struct acpi_device_id fsl_8250_acpi_id[] = {
> > + { "NXP0018", 0 },
> > + { },
> > +};
> > +MODULE_DEVICE_TABLE(acpi, fsl_8250_acpi_id);
>
> Will this now break the build on non-acpi systems, which the original
driver first
> was written for?
No this will not break on non-acpi systems as irq handler is already
exported
As "fsl8250_handle_irq" and I have tested this change with ACPI as well
as DT boot.
Safe side I can put acpi specific changes in acpi config macro
"CONFIG_ACPI".
>
> Don't you need some sort of dependancy on ACPI now?
Can you please explain more ?
>
> thanks,
>
> greg k-h
