2024-04-09 07:43:47

by Guanbing Huang

[permalink] [raw]
Subject: [PATCH v6 0/3] serial: 8250_pnp: Support configurable reg shift property

From: Guanbing Huang <[email protected]>

The 16550a serial port based on the ACPI table requires obtaining the
reg-shift attribute. In the ACPI scenario, If the reg-shift property
is not configured like in DTS, the 16550a serial driver cannot read or
write controller registers properly during initialization.

To address the issue of configuring the reg-shift property, the
__uart_read_properties() universal interface is called to implement it.
Adaptation of PNP devices is done in the __uart_read_properties() function.

Guanbing Huang (3):
PNP: Add dev_is_pnp() macro
serial: port: Add support of PNP IRQ to __uart_read_properties()
serial: 8250_pnp: Support configurable reg shift property

drivers/tty/serial/8250/8250_pnp.c | 40 +++++++++++++++++++-----------
drivers/tty/serial/serial_port.c | 7 +++++-
include/linux/pnp.h | 2 ++
3 files changed, 33 insertions(+), 16 deletions(-)

--
2.17.1



2024-04-09 07:49:29

by Guanbing Huang

[permalink] [raw]
Subject: [PATCH v6 1/3] PNP: Add dev_is_pnp() macro

From: Guanbing Huang <[email protected]>

Add dev_is_pnp() macro to determine whether the device is a PNP device.

Signed-off-by: Guanbing Huang <[email protected]>
Suggested-by: Andy Shevchenko <[email protected]>
Reviewed-by: Bing Fan <[email protected]>
Tested-by: Linheng Du <[email protected]>
---
v5 -> v6: fix the issue that the cover letter is not chained with the patch series
v4 -> v5: change "pnp" in the commit message to uppercase

include/linux/pnp.h | 2 ++
1 file changed, 2 insertions(+)

diff --git a/include/linux/pnp.h b/include/linux/pnp.h
index ddbe7c3ca4ce..792921c06594 100644
--- a/include/linux/pnp.h
+++ b/include/linux/pnp.h
@@ -502,6 +502,8 @@ static inline void pnp_unregister_driver(struct pnp_driver *drv) { }

#endif /* CONFIG_PNP */

+#define dev_is_pnp(d) ((d)->bus == &pnp_bus_type)
+
/**
* module_pnp_driver() - Helper macro for registering a PnP driver
* @__pnp_driver: pnp_driver struct
--
2.17.1


2024-04-09 07:50:50

by Guanbing Huang

[permalink] [raw]
Subject: [PATCH v6 3/3] serial: 8250_pnp: Support configurable reg shift property

From: Guanbing Huang <[email protected]>

The 16550a serial port based on the ACPI table requires obtaining the
reg-shift attribute. In the ACPI scenario, If the reg-shift property
is not configured like in DTS, the 16550a serial driver cannot read or
write controller registers properly during initialization.

Signed-off-by: Guanbing Huang <[email protected]>
Suggested-by: Andy Shevchenko <[email protected]>
Reviewed-by: Bing Fan <[email protected]>
Tested-by: Linheng Du <[email protected]>
---
v5 -> v6: fix the issue that the cover letter is not chained with the patch series
v4 -> v5: modify to obtain the value of mapsize through the pnp_mem_len function,
add annotations for the iotype variable, delete excess uart.port.flags operation
v3 -> v4: dependent on two pre patches: "pnp: Add dev_is_pnp() macro" and
"serial: 8250_port: Add support of pnp irq to __uart_read_properties()",
the iotype is reserved, the mapsize is initialized, fix the UPF_SHARE_IRQ
flag, check for IRQ being absent
v2 -> v3: switch to use uart_read_port_properties(), change "Signed-off-by" to "Reviewed-by" and "Tested-by"
v1 -> v2: change the names after "Signed off by" to the real names

drivers/tty/serial/8250/8250_pnp.c | 40 +++++++++++++++++++-----------
1 file changed, 25 insertions(+), 15 deletions(-)

diff --git a/drivers/tty/serial/8250/8250_pnp.c b/drivers/tty/serial/8250/8250_pnp.c
index 1974bbadc975..8f72a7de1d1d 100644
--- a/drivers/tty/serial/8250/8250_pnp.c
+++ b/drivers/tty/serial/8250/8250_pnp.c
@@ -435,6 +435,7 @@ serial_pnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id)
{
struct uart_8250_port uart, *port;
int ret, line, flags = dev_id->driver_data;
+ unsigned char iotype;

if (flags & UNKNOWN_DEV) {
ret = serial_pnp_guess_board(dev);
@@ -443,37 +444,46 @@ serial_pnp_probe(struct pnp_dev *dev, const struct pnp_device_id *dev_id)
}

memset(&uart, 0, sizeof(uart));
- if (pnp_irq_valid(dev, 0))
- uart.port.irq = pnp_irq(dev, 0);
if ((flags & CIR_PORT) && pnp_port_valid(dev, 2)) {
uart.port.iobase = pnp_port_start(dev, 2);
- uart.port.iotype = UPIO_PORT;
+ iotype = UPIO_PORT;
} else if (pnp_port_valid(dev, 0)) {
uart.port.iobase = pnp_port_start(dev, 0);
- uart.port.iotype = UPIO_PORT;
+ iotype = UPIO_PORT;
} else if (pnp_mem_valid(dev, 0)) {
uart.port.mapbase = pnp_mem_start(dev, 0);
- uart.port.iotype = UPIO_MEM;
+ uart.port.mapsize = pnp_mem_len(dev, 0);
+ iotype = UPIO_MEM;
uart.port.flags = UPF_IOREMAP;
} else
return -ENODEV;

- dev_dbg(&dev->dev,
- "Setup PNP port: port %#lx, mem %#llx, irq %u, type %u\n",
- uart.port.iobase, (unsigned long long)uart.port.mapbase,
- uart.port.irq, uart.port.iotype);
+ uart.port.uartclk = 1843200;
+ uart.port.dev = &dev->dev;
+ uart.port.flags |= UPF_SKIP_TEST | UPF_BOOT_AUTOCONF;
+
+ ret = uart_read_port_properties(&uart.port);
+ /* no interrupt -> fall back to polling */
+ if (ret == -ENXIO)
+ ret = 0;
+ if (ret)
+ return ret;
+
+ /*
+ * The previous call may not set iotype correctly when reg-io-width
+ * property is absent and it doesn't support IO port resource.
+ */
+ uart.port.iotype = iotype;

if (flags & CIR_PORT) {
uart.port.flags |= UPF_FIXED_PORT | UPF_FIXED_TYPE;
uart.port.type = PORT_8250_CIR;
}

- uart.port.flags |= UPF_SKIP_TEST | UPF_BOOT_AUTOCONF;
- if (pnp_irq_flags(dev, 0) & IORESOURCE_IRQ_SHAREABLE)
- uart.port.flags |= UPF_SHARE_IRQ;
- uart.port.uartclk = 1843200;
- device_property_read_u32(&dev->dev, "clock-frequency", &uart.port.uartclk);
- uart.port.dev = &dev->dev;
+ dev_dbg(&dev->dev,
+ "Setup PNP port: port %#lx, mem %#llx, size %#llx, irq %u, type %u\n",
+ uart.port.iobase, (unsigned long long)uart.port.mapbase,
+ (unsigned long long)uart.port.mapsize, uart.port.irq, uart.port.iotype);

line = serial8250_register_8250_port(&uart);
if (line < 0 || (flags & CIR_PORT))
--
2.17.1


2024-04-09 12:53:30

by Andy Shevchenko

[permalink] [raw]
Subject: Re: [PATCH v6 0/3] serial: 8250_pnp: Support configurable reg shift property

On Tue, Apr 09, 2024 at 03:43:20PM +0800, Guanbing Huang wrote:
> From: Guanbing Huang <[email protected]>
>
> The 16550a serial port based on the ACPI table requires obtaining the
> reg-shift attribute. In the ACPI scenario, If the reg-shift property
> is not configured like in DTS, the 16550a serial driver cannot read or
> write controller registers properly during initialization.
>
> To address the issue of configuring the reg-shift property, the
> __uart_read_properties() universal interface is called to implement it.
> Adaptation of PNP devices is done in the __uart_read_properties() function.

You either forgot or deliberately not added my tag. Can you elaborate?

--
With Best Regards,
Andy Shevchenko



2024-04-10 02:49:44

by Guanbing Huang

[permalink] [raw]
Subject: Re: [PATCH v6 0/3] serial: 8250_pnp: Support configurable reg shift property


在 2024/4/9 20:53, Andy Shevchenko 写道:
> On Tue, Apr 09, 2024 at 03:43:20PM +0800, Guanbing Huang wrote:
>> From: Guanbing Huang <[email protected]>
>>
>> The 16550a serial port based on the ACPI table requires obtaining the
>> reg-shift attribute. In the ACPI scenario, If the reg-shift property
>> is not configured like in DTS, the 16550a serial driver cannot read or
>> write controller registers properly during initialization.
>>
>> To address the issue of configuring the reg-shift property, the
>> __uart_read_properties() universal interface is called to implement it.
>> Adaptation of PNP devices is done in the __uart_read_properties() function.
> You either forgot or deliberately not added my tag. Can you elaborate?

I'm very sorry, this is my first time submitting a kernel patch. My
understanding

of the submission specification is not comprehensive and profound enough,

and I didn't intentionally not add tags. I hope you can forgive my
operational mistake.

Should I just add a "Reviewed-by tag", or do I need any other tags? Thanks.


With Best Regards.

>

2024-04-10 13:48:32

by Andy Shevchenko

[permalink] [raw]
Subject: Re: [PATCH v6 0/3] serial: 8250_pnp: Support configurable reg shift property

On Wed, Apr 10, 2024 at 10:49:29AM +0800, albanhuang wrote:
> 在 2024/4/9 20:53, Andy Shevchenko 写道:
> > On Tue, Apr 09, 2024 at 03:43:20PM +0800, Guanbing Huang wrote:
> > > From: Guanbing Huang <[email protected]>
> > >
> > > The 16550a serial port based on the ACPI table requires obtaining the
> > > reg-shift attribute. In the ACPI scenario, If the reg-shift property
> > > is not configured like in DTS, the 16550a serial driver cannot read or
> > > write controller registers properly during initialization.
> > >
> > > To address the issue of configuring the reg-shift property, the
> > > __uart_read_properties() universal interface is called to implement it.
> > > Adaptation of PNP devices is done in the __uart_read_properties() function.
> > You either forgot or deliberately not added my tag. Can you elaborate?
>
> I'm very sorry, this is my first time submitting a kernel patch. My
> understanding
>
> of the submission specification is not comprehensive and profound enough,
>
> and I didn't intentionally not add tags. I hope you can forgive my
> operational mistake.
>
> Should I just add a "Reviewed-by tag", or do I need any other tags? Thanks.

Understood. So if you are is one who is sending a new version, you should take
care about any given tags (such as Reviewed-by) and carry them, in case the
code is not drastically changed. I.o.w. if you don't, you have to explain why.

Anyways, it seems the patch still has a flaw as per LKP, so fix that by
providing probably two macros in the pnp.h header for both cases
(CONFIG_PNP=y/n), and I will review it again.

#ifdef CONFIG_PNP
...
#define dev_is_pnp(...) ...
...
#else
...
#define dev_is_pnp(...) false
...
#endif

--
With Best Regards,
Andy Shevchenko



2024-04-09 07:44:34

by Guanbing Huang

[permalink] [raw]
Subject: [PATCH v6 2/3] serial: port: Add support of PNP IRQ to __uart_read_properties()

From: Guanbing Huang <[email protected]>

The function __uart_read_properties doesn't cover PNP devices, so add IRQ
processing for PNP devices in the branch.

Signed-off-by: Guanbing Huang <[email protected]>
Suggested-by: Andy Shevchenko <[email protected]>
Reviewed-by: Bing Fan <[email protected]>
Tested-by: Linheng Du <[email protected]>
---
v5 -> v6: fix the issue that the cover letter is not chained with the patch series
v4 -> v5: change "pnp" and "irq" in the commit message to uppercase, modify the subject
to start with "serial: port:", modify the location of the header file pnp.h

drivers/tty/serial/serial_port.c | 7 ++++++-
1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/tty/serial/serial_port.c b/drivers/tty/serial/serial_port.c
index 22b9eeb23e68..8504bae1d2c9 100644
--- a/drivers/tty/serial/serial_port.c
+++ b/drivers/tty/serial/serial_port.c
@@ -11,6 +11,7 @@
#include <linux/of.h>
#include <linux/platform_device.h>
#include <linux/pm_runtime.h>
+#include <linux/pnp.h>
#include <linux/property.h>
#include <linux/serial_core.h>
#include <linux/spinlock.h>
@@ -221,7 +222,11 @@ static int __uart_read_properties(struct uart_port *port, bool use_defaults)

if (dev_is_platform(dev))
ret = platform_get_irq(to_platform_device(dev), 0);
- else
+ else if (dev_is_pnp(dev)) {
+ ret = pnp_irq(to_pnp_dev(dev), 0);
+ if (ret < 0)
+ ret = -ENXIO;
+ } else
ret = fwnode_irq_get(dev_fwnode(dev), 0);
if (ret == -EPROBE_DEFER)
return ret;
--
2.17.1