Skip to contentSkip to conversation

switch arc-uart to devicetree based probing

PATCH In review Vineet Gupta 4 participants 20 messages Jan 11, 2013 – Feb 14, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover switch arc-uart to devicetree based probing

Hi,

As part of converting ARC Port to devicetree infrastructure, the following series converts the arc-uart driver to DT.

* The first patch is a bug-fix which showed up in the process as DT based platform devices by default have -ve id
* Next two prepare the driver for forthcoming DT changes.
* Last one contains the DT bindings and driver using those.

Couple of points worth mentioning:
* The earlyprintk portion of driver still relies on static platform data we would need some earlyprintk handling in of_fdt_* to clean it up properly
* Two of the three platform data instances are now retrieved from DT. However one still needs to be dynamically passed by platform (using of_dev_auxdata) as we want to run same image in simulator and hardware

Tested on in-works ARC 3.8 port.

P.S. Greg, can this be treated as a bug-fix for 3.8

Thx,
-Vineet

Vineet Gupta (4):
serial/arc-uart: Don't index with -ve platform_device->id serial/arc-uart: split probe from probe_earlyprintk serial/arc-uart: platform_data order changed serial/arc-uart: switch to devicetree based probing

 .../devicetree/bindings/tty/serial/arc-uart.txt    |   26 ++++++
 drivers/tty/serial/arc_uart.c                      |   95 ++++++++++++++------
 2 files changed, 94 insertions(+), 27 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/tty/serial/arc-uart.txt
Signature · 1 line
1.7.4.1
#1 lore raw

PATCH 1/4 serial/arc-uart: Don't index with -ve platform_device->id

probe routine could index into port[] with -ve index. The check in arc_uart_init_one() was too late.

This came to light when trying to port driver to CONFIG_OF, where bydefault of-core code sets -ve platform dev id and in absence of DT serial aliases, driver would use the -ve index.

Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Jiri Slaby <[email protected]>
---
 drivers/tty/serial/arc_uart.c |   29 +++++++++++++++--------------
 1 files changed, 15 insertions(+), 14 deletions(-)
drivers/tty/serial/arc_uart.c +15 -14
index 3e0b3fa..8089dc3 100644
@@ -526,15 +526,11 @@ static struct uart_ops arc_serial_pops = {
 };
 
 static int
-arc_uart_init_one(struct platform_device *pdev, struct arc_uart_port *uart)
+arc_uart_init_one(struct platform_device *pdev, int dev_id)
 {
 	struct resource *res, *res2;
 	unsigned long *plat_data;
-
-	if (pdev->id < 0 || pdev->id >= CONFIG_SERIAL_ARC_NR_PORTS) {
-		dev_err(&pdev->dev, "Wrong uart platform device id.\n");
-		return -ENOENT;
-	}
+	struct arc_uart_port *uart = &arc_uart_ports[dev_id];
 
 	plat_data = ((unsigned long *)(pdev->dev.platform_data));
 	uart->baud = plat_data[0];
@@ -557,7 +553,7 @@ arc_uart_init_one(struct platform_device *pdev, struct arc_uart_port *uart)
 	uart->port.dev = &pdev->dev;
 	uart->port.iotype = UPIO_MEM;
 	uart->port.flags = UPF_BOOT_AUTOCONF;
-	uart->port.line = pdev->id;
+	uart->port.line = dev_id;
 	uart->port.ops = &arc_serial_pops;
 
 	uart->port.uartclk = plat_data[1];
@@ -657,9 +653,14 @@ static struct __initdata console arc_early_serial_console = {
 
 static int arc_serial_probe_earlyprintk(struct platform_device *pdev)
 {
-	arc_early_serial_console.index = pdev->id;
+	int dev_id = pdev->id < 0 ? 0 : pdev->id;
+	int rc;
 
-	arc_uart_init_one(pdev, &arc_uart_ports[pdev->id]);
+	arc_early_serial_console.index = dev_id;
+
+	rc = arc_uart_init_one(pdev, dev_id);
+	if (rc)
+		panic("early console init failed\n");
 
 	arc_serial_console_setup(&arc_early_serial_console, NULL);
 
@@ -675,18 +676,18 @@ static int arc_serial_probe_earlyprintk(struct platform_device *pdev)
 
 static int arc_serial_probe(struct platform_device *pdev)
 {
-	struct arc_uart_port *uart;
-	int rc;
+	int rc, dev_id;
 
 	if (is_early_platform_device(pdev))
 		return arc_serial_probe_earlyprintk(pdev);
 
-	uart = &arc_uart_ports[pdev->id];
-	rc = arc_uart_init_one(pdev, uart);
+	dev_id = pdev->id < 0 ? 0 : pdev->id;
+	rc = arc_uart_init_one(pdev, dev_id);
 	if (rc)
 		return rc;
 
-	return uart_add_one_port(&arc_uart_driver, &uart->port);
+	rc = uart_add_one_port(&arc_uart_driver, &arc_uart_ports[dev_id].port);
+	return rc;
 }
 
 static int arc_serial_remove(struct platform_device *pdev)
Signature · 1 line
1.7.4.1
#2 lore raw

PATCH 2/4 serial/arc-uart: split probe from probe_earlyprintk

In reply to Vineet Gupta

This is in preparation for devicetree based probing, where earlyprintk won't have access to DT serial aliases which the normal probe would absolutely rely on.

Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Jiri Slaby <[email protected]>
---
 drivers/tty/serial/arc_uart.c |   21 +++++++++++----------
 1 files changed, 11 insertions(+), 10 deletions(-)
drivers/tty/serial/arc_uart.c +11 -10
index 8089dc3..9de26ba 100644
@@ -651,7 +651,7 @@ static struct __initdata console arc_early_serial_console = {
 	.index = -1
 };
 
-static int arc_serial_probe_earlyprintk(struct platform_device *pdev)
+static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev)
 {
 	int dev_id = pdev->id < 0 ? 0 : pdev->id;
 	int rc;
@@ -667,20 +667,12 @@ static int arc_serial_probe_earlyprintk(struct platform_device *pdev)
 	register_console(&arc_early_serial_console);
 	return 0;
 }
-#else
-static int arc_serial_probe_earlyprintk(struct platform_device *pdev)
-{
-	return -ENODEV;
-}
 #endif	/* CONFIG_SERIAL_ARC_CONSOLE */
 
 static int arc_serial_probe(struct platform_device *pdev)
 {
 	int rc, dev_id;
 
-	if (is_early_platform_device(pdev))
-		return arc_serial_probe_earlyprintk(pdev);
-
 	dev_id = pdev->id < 0 ? 0 : pdev->id;
 	rc = arc_uart_init_one(pdev, dev_id);
 	if (rc)
@@ -706,6 +698,15 @@ static struct platform_driver arc_platform_driver = {
 };
 
 #ifdef CONFIG_SERIAL_ARC_CONSOLE
+
+static struct platform_driver early_arc_platform_driver = {
+	.probe = arc_serial_probe_earlyprintk,
+	.remove = arc_serial_remove,
+	.driver = {
+		.name = DRIVER_NAME,
+		.owner = THIS_MODULE,
+	 },
+};
 /*
  * Register an early platform driver of "earlyprintk" class.
  * ARCH platform code installs the driver and probes the early devices
@@ -713,7 +714,7 @@ static struct platform_driver arc_platform_driver = {
  * or it could be done independently, for all "earlyprintk" class drivers.
  * [see arch/arc/plat-arcfpga/platform.c]
  */
-early_platform_init("earlyprintk", &arc_platform_driver);
+early_platform_init("earlyprintk", &early_arc_platform_driver);
 
 #endif  /* CONFIG_SERIAL_ARC_CONSOLE */
 
Signature · 1 line
1.7.4.1
#3 lore raw

PATCH 3/4 serial/arc-uart: platform_data order changed

In reply to Vineet Gupta

* is_emulated is now 1st element, rather than last
* also tucked all platform data refs together

Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Jiri Slaby <[email protected]>
---
 drivers/tty/serial/arc_uart.c |   11 ++++++-----
 1 files changed, 6 insertions(+), 5 deletions(-)
drivers/tty/serial/arc_uart.c +6 -5
index 9de26ba..2db6410 100644
@@ -533,7 +533,12 @@ arc_uart_init_one(struct platform_device *pdev, int dev_id)
 	struct arc_uart_port *uart = &arc_uart_ports[dev_id];
 
 	plat_data = ((unsigned long *)(pdev->dev.platform_data));
-	uart->baud = plat_data[0];
+	if (!plat_data)
+		return -ENODEV;
+
+	uart->is_emulated = !!plat_data[0];	/* workaround ISS bug */
+	uart->port.uartclk = plat_data[1];
+	uart->baud = plat_data[2];
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	if (!res)
@@ -556,7 +561,6 @@ arc_uart_init_one(struct platform_device *pdev, int dev_id)
 	uart->port.line = dev_id;
 	uart->port.ops = &arc_serial_pops;
 
-	uart->port.uartclk = plat_data[1];
 	uart->port.fifosize = ARC_UART_TX_FIFO_SIZE;
 
 	/*
@@ -565,9 +569,6 @@ arc_uart_init_one(struct platform_device *pdev, int dev_id)
 	 */
 	uart->port.ignore_status_mask = 0;
 
-	/* Real Hardware vs. emulated to work around a bug */
-	uart->is_emulated = !!plat_data[2];
-
 	return 0;
 }
 
Signature · 1 line
1.7.4.1
#4 lore raw

PATCH 4/4 serial/arc-uart: switch to devicetree based probing

In reply to Vineet Gupta

* DT binding for arc-uart
* With alll the bits in place we can now use DT probing.

Note that there's a bit of kludge right now because earlyprintk portion of driver can't use the DT infrastrcuture to get resoures/plat_data. This requires some infrastructre changes to of_flat_ framework

Cc Grant Likely <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Rob Herring <[email protected]>
Cc Rob Landley <[email protected]>
---
 .../devicetree/bindings/tty/serial/arc-uart.txt    |   26 ++++++++++++
 drivers/tty/serial/arc_uart.c                      |   43 ++++++++++++++++++-
 2 files changed, 66 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/tty/serial/arc-uart.txt
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +26 -0
new file mode 100644
index 0000000..c3bd8f9
@@ -0,0 +1,26 @@
+* Synopsys ARC UART : Non standard UART used in some of the ARC FPGA boards
+
+Required properties:
+- compatible		: "snps,arc-uart"
+- reg			: offset and length of the register set for the device.
+- interrupts		: device interrupt
+- clock-frequency	: the input clock frequency for the UART
+- baud			: baud rate for UART
+
+e.g.
+
+arcuart0: serial@c0fc1000 {
+	compatible = "snps,arc-uart";
+	reg = <0xc0fc1000 0x100>;
+	interrupts = <5>;
+	clock-frequency = <80000000>;
+	baud = <115200>;
+	status = "okay";
+};
+
+Note: Each port should have an alias correctly numbered in "aliases" node.
+
+e.g.
+aliases {
+	serial0 = &arcuart0;
+};
drivers/tty/serial/arc_uart.c +40 -3

74 lines not on this page · raw message

Signature · 1 line
1.7.4.1
Acked-by on v1 4/4
#5 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

Quoted 7 lines · Vineet Gupta

* DT binding for arc-uart
* With alll the bits in place we can now use DT probing.

Note that there's a bit of kludge right now because earlyprintk portion of driver can't use the DT infrastrcuture to get resoures/plat_data. This requires some infrastructre changes to of_flat_ framework

Quoted 10 lines
Cc Grant Likely <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Rob Herring <[email protected]>
Cc Rob Landley <[email protected]>

Quick question about the name though: is this UART only used on ARC, or is it something that synopsys licenses to other parties as well? If the latter is true, we might want to add a more generic "compatible" value for those that use it on another architecture.

Arnd

#6 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

Quoted 18 lines · Arnd Bergmann
Quoted 7 lines · Vineet Gupta

* DT binding for arc-uart
* With alll the bits in place we can now use DT probing.

Note that there's a bit of kludge right now because earlyprintk portion of driver can't use the DT infrastrcuture to get resoures/plat_data. This requires some infrastructre changes to of_flat_ framework

Quoted 10 lines
Cc Grant Likely <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Rob Herring <[email protected]>
Cc Rob Landley <[email protected]>
Quoted 7 lines

Quick question about the name though: is this UART only used on ARC, or is it something that synopsys licenses to other parties as well? If the latter is true, we might want to add a more generic "compatible" value for those that use it on another architecture.

It's not licensed as standalone IP - since its not a standard 8250. It is only (but actively) used on the internal FPGA flows from ARC days.

Thx,
Vineet

Quoted 1 line

Arnd

#7 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

Quoted 11 lines · Vineet Gupta
Quoted 5 lines · Arnd Bergmann

Quick question about the name though: is this UART only used on ARC, or is it something that synopsys licenses to other parties as well? If the latter is true, we might want to add a more generic "compatible" value for those that use it on another architecture.

It's not licensed as standalone IP - since its not a standard 8250. It is only (but actively) used on the internal FPGA flows from ARC days.

Ok, thanks for the confirmation.

Arnd

#8 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

In reply to Vineet Gupta
Quoted 29 lines · Vineet Gupta

* DT binding for arc-uart
* With alll the bits in place we can now use DT probing.

Note that there's a bit of kludge right now because earlyprintk portion of driver can't use the DT infrastrcuture to get resoures/plat_data. This requires some infrastructre changes to of_flat_ framework

Cc Grant Likely <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Alan Cox <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc Rob Herring <[email protected]>
Cc Rob Landley <[email protected]>
---
 .../devicetree/bindings/tty/serial/arc-uart.txt    |   26 ++++++++++++
 drivers/tty/serial/arc_uart.c                      |   43 ++++++++++++++++++-
 2 files changed, 66 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/devicetree/bindings/tty/serial/arc-uart.txt
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +0 -0
new file mode 100644
index 0000000..c3bd8f9

Documentation/devicetree/bindings/tty/serial/arc-uart.txt line 8

Quoted 8 lines
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +8 -0
@@ -0,0 +1,26 @@
+* Synopsys ARC UART : Non standard UART used in some of the ARC FPGA boards
+
+Required properties:
+- compatible		: "snps,arc-uart"
+- reg			: offset and length of the register set for the device.
+- interrupts		: device interrupt
+- clock-frequency	: the input clock frequency for the UART
+- baud			: baud rate for UART

change 'baud' to 'current-speed'. There is already precedence for this with other serial devices.

g.

Documentation/devicetree/bindings/tty/serial/arc-uart.txt static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 700

Quoted 9 lines
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +5 -0
@@ -673,8 +693,18 @@ static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev)
 static int arc_serial_probe(struct platform_device *pdev)
 {
 	int rc, dev_id;
+	struct device_node *np = pdev->dev.of_node;
+
+	/* no device tree device */
+	if (!np)
+		return -ENODEV;

This breaks non-DT users. Is this what you intend? It creates a flag day where users have to switch from non-DT to DT cold-turkey.

Documentation/devicetree/bindings/tty/serial/arc-uart.txt static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 705

Quoted 5 lines
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +5 -0
+	dev_id = of_alias_get_id(np, "serial");
+	if (dev_id < 0) {
+		dev_err(&pdev->dev, "failed to get alias id: %d\n", dev_id);
+		return dev_id;
+	}

Don't fail on this. If you can't get an id then choose one dynamically.

Quoted 28 lines
Documentation/devicetree/bindings/tty/serial/arc-uart.txt new file +7 -1
-	dev_id = pdev->id < 0 ? 0 : pdev->id;
 	rc = arc_uart_init_one(pdev, dev_id);
 	if (rc)
 		return rc;
@@ -689,12 +719,19 @@ static int arc_serial_remove(struct platform_device *pdev)
 	return 0;
 }
 
+static const struct of_device_id arc_uart_dt_ids[] = {
+	{ .compatible = "snps,arc-uart" },
+	{ /* Sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, arc_uart_dt_ids);
+
 static struct platform_driver arc_platform_driver = {
 	.probe = arc_serial_probe,
 	.remove = arc_serial_remove,
 	.driver = {
 		.name = DRIVER_NAME,
 		.owner = THIS_MODULE,
+		.of_match_table  = arc_uart_dt_ids,
 	 },
 };
 
Signature · 1 line
1.7.4.1
Signature · 2 lines
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies, Ltd.
#9 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

Quoted 5 lines · Grant Likely
Quoted 2 lines · Vineet Gupta
+- clock-frequency	: the input clock frequency for the UART
+- baud			: baud rate for UART

change 'baud' to 'current-speed'. There is already precedence for this with other serial devices.

While I'm OK with this - I can only see of_serial.c following the rule :-) More importantly I'm not clear about the logistics of this fix. Obviously this has a bearing on DT files in arch/arc/boot/*. So are such changes (platform + driver) routed thru the subsystem tree or the arch tree or bits from both with bisectability not considered - which feels wrong. We have to also consider the fact that Greg has closed the tty/serial for 3.9. So while I have no objection to your comment, it seems that the it needs to wait till 3.9-rc1 - or is there an alternate way.

Quoted 12 lines

drivers/tty/serial/arc_uart.c static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 700

Quoted 10 lines
drivers/tty/serial/arc_uart.c +5 -0
@@ -673,8 +693,18 @@ static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev)
 static int arc_serial_probe(struct platform_device *pdev)
 {
 	int rc, dev_id;
+	struct device_node *np = pdev->dev.of_node;
+
+	/* no device tree device */
+	if (!np)
+		return -ENODEV;

This breaks non-DT users. Is this what you intend? It creates a flag day where users have to switch from non-DT to DT cold-turkey.

Not supporting non-DT user was not the idea - it just simplifies the code a bit given that it would only even be runtime used in a ARC Linux port based platform - which unconditionally enables OF. Further - the ARC port itself is not yet upstream so there are no "official" user of this in tree driver. FWIW, ARC Linux port was recently reviewed on lkml/arch mailing lists and is now in linux-next for a possible 3.9 merge.

Quoted 6 lines

drivers/tty/serial/arc_uart.c static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 706

Quoted 5 lines
drivers/tty/serial/arc_uart.c +5 -0
+	dev_id = of_alias_get_id(np, "serial");
+	if (dev_id < 0) {
+		dev_err(&pdev->dev, "failed to get alias id: %d\n", dev_id);
+		return dev_id;
+	}

Don't fail on this. If you can't get an id then choose one dynamically.

You mean just assume 0.

Thanks for reviewing.
-Vineet

#10 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

Quoted 7 lines · Vineet Gupta
Quoted 5 lines · Grant Likely
Quoted 2 lines · Vineet Gupta
+- clock-frequency   : the input clock frequency for the UART
+- baud                      : baud rate for UART

change 'baud' to 'current-speed'. There is already precedence for this with other serial devices.

Quoted 8 lines

While I'm OK with this - I can only see of_serial.c following the rule :-) More importantly I'm not clear about the logistics of this fix. Obviously this has a bearing on DT files in arch/arc/boot/*. So are such changes (platform + driver) routed thru the subsystem tree or the arch tree or bits from both with bisectability not considered - which feels wrong. We have to also consider the fact that Greg has closed the tty/serial for 3.9. So while I have no objection to your comment, it seems that the it needs to wait till 3.9-rc1 - or is there an alternate way.

I'd consider this one a bug fix, so while Greg is not accepting any new features for the serial tree, I think it should still get in that way and should not be controversial as an add-on change.

Arnd

#11 lore raw

PATCH 1/1 serial/arc-uart: Urgent DT related update

Hi Greg,

Here's an update to arc-uart driver given Grant's recent review comments. As discussed in the thread, can you consider this a bug fix and apply to tty-next for 3.9. If you don't, please let me know as the corresponding DT file change would have to be to be seperately applied to arch/arc/boot/*.

Patch is Based off linux-next of today.

Thx,
-Vineet

Vineet Gupta (1):
serial/arc-uart: Miscll DT related updates (Grant's review comments)

 .../devicetree/bindings/tty/serial/arc-uart.txt    |    4 ++--
 drivers/tty/serial/arc_uart.c                      |   10 ++++------
 2 files changed, 6 insertions(+), 8 deletions(-)
Signature · 1 line
1.7.4.1
#12 lore raw

PATCH 1/1 serial/arc-uart: Miscll DT related updates (Grant's review comments)

-replace "baud" with "current-speed"
-if uart alias doesn't exist in DT, don't abort, pick 0

Cc Greg Kroah-Hartman <[email protected]>
Cc Grant Likely <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Rob Herring <[email protected]>
Cc Rob Landley <[email protected]>
---
 .../devicetree/bindings/tty/serial/arc-uart.txt    |    4 ++--
 drivers/tty/serial/arc_uart.c                      |   10 ++++------
 2 files changed, 6 insertions(+), 8 deletions(-)
Documentation/devicetree/bindings/tty/serial/arc-uart.txt +2 -2
index c3bd8f9..5cae2eb 100644
@@ -5,7 +5,7 @@ Required properties:
 - reg			: offset and length of the register set for the device.
 - interrupts		: device interrupt
 - clock-frequency	: the input clock frequency for the UART
-- baud			: baud rate for UART
+- current-speed		: baud rate for UART
 
 e.g.
 
@@ -14,7 +14,7 @@ arcuart0: serial@c0fc1000 {
 	reg = <0xc0fc1000 0x100>;
 	interrupts = <5>;
 	clock-frequency = <80000000>;
-	baud = <115200>;
+	current-speed = <115200>;
 	status = "okay";
 };
 
drivers/tty/serial/arc_uart.c +4 -6
index 6f7eadc..d97e194 100644
@@ -547,8 +547,8 @@ arc_uart_init_one(struct platform_device *pdev, int dev_id)
 		}
 		uart->port.uartclk = val;
 
-		if (of_property_read_u32(np, "baud", &val)) {
-			dev_err(&pdev->dev, "baud property NOT set\n");
+		if (of_property_read_u32(np, "current-speed", &val)) {
+			dev_err(&pdev->dev, "current-speed property NOT set\n");
 			return -EINVAL;
 		}
 		uart->baud = val;
@@ -694,10 +694,8 @@ static int arc_serial_probe(struct platform_device *pdev)
 		return -ENODEV;
 
 	dev_id = of_alias_get_id(np, "serial");
-	if (dev_id < 0) {
-		dev_err(&pdev->dev, "failed to get alias id: %d\n", dev_id);
-		return dev_id;
-	}
+	if (dev_id < 0)
+		dev_id = 0;
 
 	rc = arc_uart_init_one(pdev, dev_id);
 	if (rc)
Signature · 1 line
1.7.4.1
#13 lore raw

Re: [PATCH] serial/arc-uart: Urgent DT related update

In reply to Vineet Gupta
Quoted 12 lines · Vineet Gupta

Hi Greg,

Here's an update to arc-uart driver given Grant's recent review comments. As discussed in the thread, can you consider this a bug fix and apply to tty-next for 3.9. If you don't, please let me know as the corresponding DT file change would have to be to be seperately applied to arch/arc/boot/*.

Patch is Based off linux-next of today.

Thx,
-Vineet

Quoted 7 lines

Vineet Gupta (1):
serial/arc-uart: Miscll DT related updates (Grant's review comments)

 .../devicetree/bindings/tty/serial/arc-uart.txt    |    4 ++--
 drivers/tty/serial/arc_uart.c                      |   10 ++++------
 2 files changed, 6 insertions(+), 8 deletions(-)

Greg, I'm assuming this won't make into tty until 3.9-rc1 and hence I need to drop the platform patch from arc-next.

Thx,
-Vineet

#14 lore raw

Re: [PATCH] serial/arc-uart: Urgent DT related update

Quoted 16 lines · Vineet Gupta
Quoted 15 lines · Vineet Gupta

Hi Greg,

Here's an update to arc-uart driver given Grant's recent review comments. As discussed in the thread, can you consider this a bug fix and apply to tty-next for 3.9. If you don't, please let me know as the corresponding DT file change would have to be to be seperately applied to arch/arc/boot/*.

Patch is Based off linux-next of today.

Thx,
-Vineet

Vineet Gupta (1):
serial/arc-uart: Miscll DT related updates (Grant's review comments)

Quoted 7 lines
Quoted 4 lines
 .../devicetree/bindings/tty/serial/arc-uart.txt    |    4 ++--
 drivers/tty/serial/arc_uart.c                      |   10 ++++------
 2 files changed, 6 insertions(+), 8 deletions(-)

Greg, I'm assuming this won't make into tty until 3.9-rc1 and hence I need to drop the platform patch from arc-next.

I've taken this now, it will be sent to Linus for 3.9-rc1.

thanks,

greg k-h

#15 lore raw

Re: [PATCH] serial/arc-uart: Urgent DT related update

Quoted 17 lines · Greg Kroah-Hartman
Quoted 16 lines · Vineet Gupta
Quoted 15 lines · Vineet Gupta

Hi Greg,

Here's an update to arc-uart driver given Grant's recent review comments. As discussed in the thread, can you consider this a bug fix and apply to tty-next for 3.9. If you don't, please let me know as the corresponding DT file change would have to be to be seperately applied to arch/arc/boot/*.

Patch is Based off linux-next of today.

Thx,
-Vineet

Vineet Gupta (1):
serial/arc-uart: Miscll DT related updates (Grant's review comments)

Quoted 7 lines
Quoted 6 lines
Quoted 4 lines
 .../devicetree/bindings/tty/serial/arc-uart.txt    |    4 ++--
 drivers/tty/serial/arc_uart.c                      |   10 ++++------
 2 files changed, 6 insertions(+), 8 deletions(-)

Greg, I'm assuming this won't make into tty until 3.9-rc1 and hence I need to drop the platform patch from arc-next.

I've taken this now, it will be sent to Linus for 3.9-rc1.

Thanks greg. I understand the tree closing rules' purpose, to let things stabilize ahead of merge window - both from mechanical conflicts perspective as well as runtime conflicts. However I'm slightly confused here about the process, specially as a n00b maintainer and was hoping you could clarify it.

Your announcement of Feb 9th said

Quoted 4 lines

With the release of the 3.8-rc7 kernel, I think it's time to close the TTY/serial tree for new features / cleanups for 3.9. So I'm closing my tree, and will only be applying obvious bugfixes or regressions to it until 3.9-rc1 comes out.

So a bug fix sent on say Feb 11th would have been be applied to your tree to end up in 3.9 merge-window pull to Linus- correct or no ? So in that sense, since Arnd/Grant are deeming this patch as bug fix, shouldn't it make into 3.9 merge-window merge. So maybe, it's the criticality of the bug which warrants that. Please note that I'm just trying to understand the process here and not arguing about patch itself.

Quoted 3 lines

You can keep sending me patches for the tree that don't fit the "bugfix" category, but note that I'll be storing them away to wait for 3.9-rc1 comes out, which might be a few weeks, so please be patient.

And once 3.9-rc1 hits, everything becomes 3.10 merge material anyways except for bug fixes.

TIA,
-Vineet

#16 lore raw

Re: [PATCH] serial/arc-uart: Urgent DT related update

Quoted 9 lines · Vineet Gupta
Quoted 1 line

I've taken this now, it will be sent to Linus for 3.9-rc1.

Thanks greg. I understand the tree closing rules' purpose, to let things stabilize ahead of merge window - both from mechanical conflicts perspective as well as runtime conflicts. However I'm slightly confused here about the process, specially as a n00b maintainer and was hoping you could clarify it.

Your announcement of Feb 9th said

Quoted 7 lines
Quoted 4 lines

With the release of the 3.8-rc7 kernel, I think it's time to close the TTY/serial tree for new features / cleanups for 3.9. So I'm closing my tree, and will only be applying obvious bugfixes or regressions to it until 3.9-rc1 comes out.

So a bug fix sent on say Feb 11th would have been be applied to your tree to end up in 3.9 merge-window pull to Linus- correct or no ?

Yes.

Quoted 2 lines

So in that sense, since Arnd/Grant are deeming this patch as bug fix, shouldn't it make into 3.9 merge-window merge.

Yes, and that is where it will be, as I said so above.

Quoted 10 lines

So maybe, it's the criticality of the bug which warrants that. Please note that I'm just trying to understand the process here and not arguing about patch itself.

Quoted 3 lines

You can keep sending me patches for the tree that don't fit the "bugfix" category, but note that I'll be storing them away to wait for 3.9-rc1 comes out, which might be a few weeks, so please be patient.

And once 3.9-rc1 hits, everything becomes 3.10 merge material anyways except for bug fixes.

Yes.

Hope this helps,

greg k-h

#17 lore raw

Re: [PATCH 4/4] serial/arc-uart: switch to devicetree based probing

In reply to Vineet Gupta
Quoted 7 lines · Vineet Gupta
Quoted 5 lines · Grant Likely
Quoted 2 lines · Vineet Gupta
+- clock-frequency	: the input clock frequency for the UART
+- baud			: baud rate for UART

change 'baud' to 'current-speed'. There is already precedence for this with other serial devices.

Quoted 8 lines

While I'm OK with this - I can only see of_serial.c following the rule :-) More importantly I'm not clear about the logistics of this fix. Obviously this has a bearing on DT files in arch/arc/boot/*. So are such changes (platform + driver) routed thru the subsystem tree or the arch tree or bits from both with bisectability not considered - which feels wrong. We have to also consider the fact that Greg has closed the tty/serial for 3.9. So while I have no objection to your comment, it seems that the it needs to wait till 3.9-rc1 - or is there an alternate way.

I would consider it a bug fix. The binding isn't what it should be and it needs to be addressed before appearing in a released kernel. If this patch has already been merged, then write and post a fixup patch.

Quoted 13 lines
Quoted 12 lines

drivers/tty/serial/arc_uart.c static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 700

Quoted 10 lines
drivers/tty/serial/arc_uart.c +5 -0
@@ -673,8 +693,18 @@ static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev)
 static int arc_serial_probe(struct platform_device *pdev)
 {
 	int rc, dev_id;
+	struct device_node *np = pdev->dev.of_node;
+
+	/* no device tree device */
+	if (!np)
+		return -ENODEV;

This breaks non-DT users. Is this what you intend? It creates a flag day where users have to switch from non-DT to DT cold-turkey.

Quoted 6 lines

Not supporting non-DT user was not the idea - it just simplifies the code a bit given that it would only even be runtime used in a ARC Linux port based platform - which unconditionally enables OF. Further - the ARC port itself is not yet upstream so there are no "official" user of this in tree driver. FWIW, ARC Linux port was recently reviewed on lkml/arch mailing lists and is now in linux-next for a possible 3.9 merge.

If there are no users, then nobody is broken. If this driver only supports DT platforms then my comment can be ignored.

Quoted 8 lines
Quoted 6 lines

drivers/tty/serial/arc_uart.c static int __init arc_serial_probe_earlyprintk(struct platform_device *pdev) · line 706

Quoted 5 lines
drivers/tty/serial/arc_uart.c +5 -0
+	dev_id = of_alias_get_id(np, "serial");
+	if (dev_id < 0) {
+		dev_err(&pdev->dev, "failed to get alias id: %d\n", dev_id);
+		return dev_id;
+	}

Don't fail on this. If you can't get an id then choose one dynamically.

You mean just assume 0.

No, I mean dynamically assign an ID from ids that are available. Say you had two of these devices in a system, and neither had an alias; they couldn't both be '0'. :-)

g.

Quoted 3 lines

Thanks for reviewing.
-Vineet

Signature · 2 lines
Grant Likely, B.Sc, P.Eng.
Secret Lab Technologies, Ltd.
maintainer
#18 lore raw
In reply to Vineet Gupta
Quoted 17 lines · Vineet Gupta

Hi,

As part of converting ARC Port to devicetree infrastructure, the following series converts the arc-uart driver to DT.

* The first patch is a bug-fix which showed up in the process as DT based platform devices by default have -ve id
* Next two prepare the driver for forthcoming DT changes.
* Last one contains the DT bindings and driver using those.

Couple of points worth mentioning:
* The earlyprintk portion of driver still relies on static platform data we would need some earlyprintk handling in of_fdt_* to clean it up properly
* Two of the three platform data instances are now retrieved from DT. However one still needs to be dynamically passed by platform (using of_dev_auxdata) as we want to run same image in simulator and hardware

Quoted 3 lines

Tested on in-works ARC 3.8 port.

P.S. Greg, can this be treated as a bug-fix for 3.8

No, because it is not a bugfix, but a new feature, sorry.

greg k-h

#19 lore raw
Quoted 18 lines · Greg KH
Quoted 17 lines · Vineet Gupta

Hi,

As part of converting ARC Port to devicetree infrastructure, the following series converts the arc-uart driver to DT.

* The first patch is a bug-fix which showed up in the process as DT based platform devices by default have -ve id
* Next two prepare the driver for forthcoming DT changes.
* Last one contains the DT bindings and driver using those.

Couple of points worth mentioning:
* The earlyprintk portion of driver still relies on static platform data we would need some earlyprintk handling in of_fdt_* to clean it up properly
* Two of the three platform data instances are now retrieved from DT. However one still needs to be dynamically passed by platform (using of_dev_auxdata) as we want to run same image in simulator and hardware

Quoted 6 lines
Quoted 3 lines

Tested on in-works ARC 3.8 port.

P.S. Greg, can this be treated as a bug-fix for 3.8

No, because it is not a bugfix, but a new feature, sorry.

greg k-h

NP. Thx a bunch for merging these.
-Vineet