Skip to contentSkip to conversation

misc: Port USB251xB/xBi Hi-Speed Hub Controller Driver from Linux

PATCH In review Marek Vasut 3 participants 11 messages Apr 10, 2022 – Jul 11, 2022 Char and misc drivers

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 misc: Port USB251xB/xBi Hi-Speed Hub Controller Driver from Linux

This patch adds a driver for configuration of the Microchip USB251xB/xBi USB 2.0 hub controller series with USB 2.0 upstream connectivity, SMBus configuration interface and two to four USB 2.0 downstream ports.

This is ported from Linux as of Linux kernel commit 5c2b9c61ae5d8 ("usb: usb251xb: add boost-up property support")

Cc Bin Meng <[email protected]>
Cc Michal Simek <[email protected]>
Cc Simon Glass <[email protected]>
---
 drivers/misc/Kconfig    |   9 +
 drivers/misc/Makefile   |   1 +
 drivers/misc/usb251xb.c | 605 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 615 insertions(+)
 create mode 100644 drivers/misc/usb251xb.c
drivers/misc/Kconfig +9 -0
index 7029bb7b5c5..48c968316ad 100644
@@ -360,6 +360,15 @@ config TEST_DRV
 	  model. This should only be enabled for testing as it is not useful for
 	  anything else.
 
+config USB_HUB_USB251XB
+	tristate "USB251XB Hub Controller Configuration Driver"
+	depends on I2C
+	help
+	  This option enables support for configuration via SMBus of the
+	  Microchip USB251x/xBi USB 2.0 Hub Controller series. Configuration
+	  parameters may be set in devicetree or platform data.
+	  Say Y or M here if you need to configure such a device via SMBus.
+
 config TWL4030_LED
 	bool "Enable TWL4030 LED controller"
 	help
drivers/misc/Makefile +1 -0
index f22eff601a1..9f25160d29b 100644
@@ -20,6 +20,7 @@ endif
 ifdef CONFIG_$(SPL_)DM_I2C
 ifndef CONFIG_SPL_BUILD
 obj-$(CONFIG_SANDBOX) += i2c_eeprom_emul.o
+obj-$(CONFIG_USB_HUB_USB251XB) += usb251xb.o
 endif
 endif
 ifdef CONFIG_SPL_OF_PLATDATA
drivers/misc/usb251xb.c new file +605 -0

605 lines not on this page · raw message

Signature · 1 line
2.35.1
#1 lore raw
Quoted 6 lines · Marek Vasut

This patch adds a driver for configuration of the Microchip USB251xB/xBi USB 2.0 hub controller series with USB 2.0 upstream connectivity, SMBus configuration interface and two to four USB 2.0 downstream ports.

This is ported from Linux as of Linux kernel commit 5c2b9c61ae5d8 ("usb: usb251xb: add boost-up property support")

The patch is aligned with kernel as you said but I expect this hub will be described on i2c bus which is based on
https://lore.kernel.org/all/CAL_JsqJedhX6typpUKbnzV7CLK6UZVjq3CyG9iY_j5DLPqvVdw@mail.gmail.com/

not where Rob would like to see ends.

It is not a problem with the patch from my point of view I just wanted to let you know that based on him it should be handled differently.

Thanks,
Michal

#2 lore raw
Quoted 10 lines · Michal Simek
Quoted 6 lines · Marek Vasut

This patch adds a driver for configuration of the Microchip USB251xB/xBi USB 2.0 hub controller series with USB 2.0 upstream connectivity, SMBus configuration interface and two to four USB 2.0 downstream ports.

This is ported from Linux as of Linux kernel commit 5c2b9c61ae5d8 ("usb: usb251xb: add boost-up property support")

Quoted 7 lines

The patch is aligned with kernel as you said but I expect this hub will be described on i2c bus which is based on
https://lore.kernel.org/all/CAL_JsqJedhX6typpUKbnzV7CLK6UZVjq3CyG9iY_j5DLPqvVdw@mail.gmail.com/

not where Rob would like to see ends.

It is not a problem with the patch from my point of view I just wanted to let you know that based on him it should be handled differently.

Can you include me in that discussion ?

However, it seems the discussion is not about USB251x , but some other hub which is connected over HSIC, not USB, that's likey why Rob wants it described differently than this hub.

#3 lore raw
In reply to Marek Vasut
Quoted 11 lines · Marek Vasut

This patch adds a driver for configuration of the Microchip USB251xB/xBi USB 2.0 hub controller series with USB 2.0 upstream connectivity, SMBus configuration interface and two to four USB 2.0 downstream ports.

This is ported from Linux as of Linux kernel commit 5c2b9c61ae5d8 ("usb: usb251xb: add boost-up property support")

Cc Bin Meng <[email protected]>
Cc Michal Simek <[email protected]>
Cc Simon Glass <[email protected]>

If there are no comments, I will apply this soon.

#4 lore raw
In reply to Marek Vasut

Hi Marek,

sorry for the late comments but I think the driver doesn't work as expected.

Am 10.04.2022 um 06:27 schrieb Marek Vasut:

Quoted 24 lines

This patch adds a driver for configuration of the Microchip USB251xB/xBi USB 2.0 hub controller series with USB 2.0 upstream connectivity, SMBus configuration interface and two to four USB 2.0 downstream ports.

This is ported from Linux as of Linux kernel commit 5c2b9c61ae5d8 ("usb: usb251xb: add boost-up property support")

Cc Bin Meng <[email protected]>
Cc Michal Simek <[email protected]>
Cc Simon Glass <[email protected]>
---
  drivers/misc/Kconfig    |   9 +
  drivers/misc/Makefile   |   1 +
  drivers/misc/usb251xb.c | 605 ++++++++++++++++++++++++++++++++++++++++
  3 files changed, 615 insertions(+)

create mode 100644 drivers/misc/usb251xb.c

drivers/misc/usb251xb.c new file +0 -0
new file mode 100644
index 00000000000..077edc25045
@@ -0,0 +1,605 @@

[snip]

Quoted 20 lines
drivers/misc/usb251xb.c new file +20 -0
+static int usb251xb_of_to_plat(struct udevice *dev)
+{
+	struct usb251xb_data *data =
+		(struct usb251xb_data *)dev_get_driver_data(dev);
+	struct usb251xb *hub = dev_get_priv(dev);
+	char str[USB251XB_STRING_BUFSIZE / 2];
+	const char *cproperty_char;
+	u32 property_u32 = 0;
+	int len, err;
+
+	if (dev_read_bool(dev, "skip-config"))
+		hub->skip_config = 1;
+	else
+		hub->skip_config = 0;
+
+	err = gpio_request_by_name(dev, "reset-gpios", 0, &hub->gpio_reset,
+				   GPIOD_IS_OUT | GPIOD_IS_OUT_ACTIVE);
+	if (err && err != -ENOENT) {
+		dev_err(dev, "unable to request GPIO reset pin (%d)\n", err);
+		return err;

drivers/misc/usb251xb.c line 401

Quoted 10 lines
drivers/misc/usb251xb.c new file +10 -0
+	}
+
+	if (IS_ENABLED(CONFIG_DM_REGULATOR)) {
+		err = device_get_supply_regulator(dev, "vdd-supply",
+						  &hub->vdd);
+		if (err && err != -ENOENT) {
+			dev_err(dev, "Warning: cannot get power supply\n");
+			return err;
+		}
+	}

Shouldn't the gpio and regulator request be part of the probe?

Quoted 10 lines
drivers/misc/usb251xb.c new file +10 -0
+
+	if (dev_read_u32(dev, "vendor-id", &hub->vendor_id))
+		hub->vendor_id = USB251XB_DEF_VENDOR_ID;
+
+	if (dev_read_u32(dev, "product-id", &hub->product_id))
+		hub->product_id = data->product_id;
+
+	if (dev_read_u32(dev, "device-id", &hub->device_id))
+		hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

drivers/misc/usb251xb.c line 518

Quoted 3 lines
drivers/misc/usb251xb.c new file +3 -0
+	if (dev_read_u32(dev, "language-id", &hub->lang_id))
+		hub->lang_id = USB251XB_DEF_LANGUAGE_ID;
+

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

drivers/misc/usb251xb.c line 520

Quoted 2 lines
drivers/misc/usb251xb.c new file +2 -0
+	if (!dev_read_u32(dev, "boost-up", &hub->boost_up))
+		hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/usb/misc/usb251xb.c
[2] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/usb/usb251xb.txt

Regards,
Stefan

#5 lore raw
Quoted 1 line · Stefan Herbrechtsmeier

Hi Marek,

Hi,

Quoted 1 line

sorry for the late comments but I think the driver doesn't work as expected.

How come ? It works on my board.

[...]

Quoted 11 lines
Quoted 9 lines

+
+    if (IS_ENABLED(CONFIG_DM_REGULATOR)) {
+        err = device_get_supply_regulator(dev, "vdd-supply", +                          &hub->vdd);
+        if (err && err != -ENOENT) {
+            dev_err(dev, "Warning: cannot get power supply\n"); +            return err;
+        }
+    }

Shouldn't the gpio and regulator request be part of the probe?

What makes you think that ?

Quoted 16 lines
Quoted 9 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

drivers/misc/usb251xb.c new file +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

Quoted 6 lines

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

What kind of problem are you running into exactly ?

#6 lore raw

Hi Marek,

Am 13.06.2022 um 18:00 schrieb Marek Vasut:

Quoted 8 lines · Stefan Herbrechtsmeier
Quoted 1 line

Hi Marek,

Hi,

Quoted 1 line

sorry for the late comments but I think the driver doesn't work as expected.

How come ? It works on my board.

Have you test with 16 bit property data?

Quoted 13 lines

[...]

Quoted 10 lines
Quoted 9 lines

+
+    if (IS_ENABLED(CONFIG_DM_REGULATOR)) {
+        err = device_get_supply_regulator(dev, "vdd-supply", +                          &hub->vdd);
+        if (err && err != -ENOENT) {
+            dev_err(dev, "Warning: cannot get power supply\n"); +            return err;
+        }
+    }

Quoted 3 lines
Quoted 1 line

Shouldn't the gpio and regulator request be part of the probe?

What makes you think that ?

It's more a question regarding the device model design:

of_to_plat - convert device tree data to plat

Quoted 19 lines
Quoted 19 lines
Quoted 10 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

drivers/misc/usb251xb.c new file +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

What kind of problem are you running into exactly ?

Have you test the values from device tree binding documentation:

vendor-id = /bits/ 16 <0xNNNN>;
product-id = /bits/ 16 <0xNNNN>;

I get an EOVERFLOW error.

#7 lore raw
Quoted 1 line · Stefan Herbrechtsmeier

Hi Marek,

Hi,

[...]

Quoted 10 lines
Quoted 10 lines
Quoted 10 lines
Quoted 9 lines

+
+    if (IS_ENABLED(CONFIG_DM_REGULATOR)) {
+        err = device_get_supply_regulator(dev, "vdd-supply", +                          &hub->vdd);
+        if (err && err != -ENOENT) {
+            dev_err(dev, "Warning: cannot get power supply\n"); +            return err;
+        }
+    }

Quoted 7 lines
Quoted 3 lines
Quoted 1 line

Shouldn't the gpio and regulator request be part of the probe?

What makes you think that ?

It's more a question regarding the device model design:

of_to_plat - convert device tree data to plat

As far as I can tell, you should be able to obtain reference to a regulator here, but not probe the regulator.

Quoted 25 lines
Quoted 25 lines
Quoted 16 lines
Quoted 9 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

drivers/misc/usb251xb.c new file +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

Quoted 6 lines

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

Quoted 8 lines
Quoted 1 line

What kind of problem are you running into exactly ?

Have you test the values from device tree binding documentation:

vendor-id = /bits/ 16 <0xNNNN>;
product-id = /bits/ 16 <0xNNNN>;

I get an EOVERFLOW error.

No, I only tested foo = <0xNNNN>; .

Can you send a fix ?

#8 lore raw

Am 13.06.2022 um 22:26 schrieb Marek Vasut:

Quoted 1 line · Stefan Herbrechtsmeier

[snip]

Quoted 29 lines
Quoted 29 lines
Quoted 18 lines
Quoted 18 lines
Quoted 9 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

drivers/misc/usb251xb.c new file +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

What kind of problem are you running into exactly ?

Have you test the values from device tree binding documentation:

Quoted 8 lines
Quoted 4 lines

vendor-id = /bits/ 16 <0xNNNN>;
product-id = /bits/ 16 <0xNNNN>;

I get an EOVERFLOW error.

No, I only tested foo = <0xNNNN>; .

Can you send a fix ?

Should I add a function to the driver or a dev_read_u8/16, ofnode_read_u8/16 and friends function?

#9 lore raw
Quoted 37 lines · Stefan Herbrechtsmeier

Am 13.06.2022 um 22:26 schrieb Marek Vasut:

Quoted 1 line · Stefan Herbrechtsmeier

[snip]

Quoted 32 lines
Quoted 32 lines
Quoted 18 lines
Quoted 18 lines
Quoted 9 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

drivers/misc/usb251xb.c new file +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

What kind of problem are you running into exactly ?

Have you test the values from device tree binding documentation:

vendor-id = /bits/ 16 <0xNNNN>;
product-id = /bits/ 16 <0xNNNN>;

Quoted 7 lines
Quoted 5 lines
Quoted 1 line

I get an EOVERFLOW error.

No, I only tested foo = <0xNNNN>; .

Can you send a fix ?

Should I add a function to the driver or a dev_read_u8/16, ofnode_read_u8/16 and friends function?

It seems more like you would have to add the _u8 and _u16 variants into common code, alongside the _u32 variant (should be easy) and then use them in this driver (should also be easy), so two patches.

#10 lore raw

Hi Marek,

Am 14.06.2022 um 09:38 schrieb Marek Vasut:

Quoted 40 lines · Stefan Herbrechtsmeier
Quoted 39 lines

Am 13.06.2022 um 22:26 schrieb Marek Vasut:

Quoted 1 line · Stefan Herbrechtsmeier

[snip]

Quoted 34 lines
Quoted 33 lines
Quoted 18 lines
Quoted 18 lines
Quoted 9 lines

+    if (dev_read_u32(dev, "vendor-id", &hub->vendor_id)) +        hub->vendor_id = USB251XB_DEF_VENDOR_ID;

(quoted diff) +7 -0
+
+    if (dev_read_u32(dev, "product-id", &hub->product_id))
+        hub->product_id = data->product_id;
+
+    if (dev_read_u32(dev, "device-id", &hub->device_id))
+        hub->device_id = USB251XB_DEF_DEVICE_ID;
+

[snip]

Quoted 3 lines

+    if (dev_read_u32(dev, "language-id", &hub->lang_id)) +        hub->lang_id = USB251XB_DEF_LANGUAGE_ID; +

This doesn't work because the ids are 16 bit [1,2] and the dev_read_u32 function checks the size.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

+    if (!dev_read_u32(dev, "boost-up", &hub->boost_up)) +        hub->boost_up = USB251XB_DEF_BOOST_UP;

This looks like a 8 bit value [1].

The dev_read_u32() is also capable of reading 0xNNNN 16bit value from DT.

What kind of problem are you running into exactly ?

Have you test the values from device tree binding documentation:

vendor-id = /bits/ 16 <0xNNNN>;
product-id = /bits/ 16 <0xNNNN>;

I get an EOVERFLOW error.

Quoted 7 lines
Quoted 5 lines
Quoted 3 lines

No, I only tested foo = <0xNNNN>; .

Can you send a fix ?

Should I add a function to the driver or a dev_read_u8/16, ofnode_read_u8/16 and friends function?

It seems more like you would have to add the _u8 and _u16 variants into common code, alongside the _u32 variant (should be easy) and then use them in this driver (should also be easy), so two patches.

https://lists.denx.de/pipermail/u-boot/2022-June/486423.html
https://lists.denx.de/pipermail/u-boot/2022-June/486424.html

Do you have any comments?

Regards
Stefan