Skip to contentSkip to conversation

USB: Also match device drivers using the ->match vfunc

PATCH In review Bastien Nocera 4 participants 8 messages Aug 18, 2020 – Aug 19, 2020 USB

lore
Share & tools
lore
#0 lore raw

PATCH v8 1/3 USB: Also match device drivers using the ->match vfunc

We only ever used the ID table matching before, but we should also support open-coded match functions.

Fixes 88b7381a939de ("USB: Select better matching USB drivers when available")

---
Changes since v7:
- Reword commit subject

Changes since v2:
- Added Alan's ack

Changes since v1:
- Fixed typo in commit message

 drivers/usb/core/generic.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/usb/core/generic.c +3 -2
index b6f2d4b44754..2b2f1ab6e36a 100644
@@ -205,8 +205,9 @@ static int __check_usb_generic(struct device_driver *drv, void *data)
 	udrv = to_usb_device_driver(drv);
 	if (udrv == &usb_generic_driver)
 		return 0;
-
-	return usb_device_match_id(udev, udrv->id_table) != NULL;
+	if (usb_device_match_id(udev, udrv->id_table) != NULL)
+		return 1;
+	return (udrv->match && udrv->match(udev));
 }
 
 static bool usb_generic_driver_match(struct usb_device *udev)
Signature · 1 line
2.26.2
#1 lore raw

PATCH v8 2/3 USB: Better name for __check_usb_generic()

__check_usb_generic() doesn't explain very well what the function actually does: It checks to see whether the driver is non-generic and matches the device.

Change it to check_for_non_generic_match()

---
No changes since v7 which was the first version based on Alan's comment:
https://marc.info/?l=linux-usb&m=159568870400795&w=2

 drivers/usb/core/generic.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/usb/core/generic.c +2 -2
index 2b2f1ab6e36a..22c887f5c497 100644
@@ -195,7 +195,7 @@ int usb_choose_configuration(struct usb_device *udev)
 }
 EXPORT_SYMBOL_GPL(usb_choose_configuration);
 
-static int __check_usb_generic(struct device_driver *drv, void *data)
+static int __check_for_non_generic_match(struct device_driver *drv, void *data)
 {
 	struct usb_device *udev = data;
 	struct usb_device_driver *udrv;
@@ -219,7 +219,7 @@ static bool usb_generic_driver_match(struct usb_device *udev)
 	 * If any other driver wants the device, leave the device to this other
 	 * driver.
 	 */
-	if (bus_for_each_drv(&usb_bus_type, NULL, udev, __check_usb_generic))
+	if (bus_for_each_drv(&usb_bus_type, NULL, udev, __check_for_non_generic_match))
 		return false;
 
 	return true;
Signature · 1 line
2.26.2
#2 lore raw

PATCH v8 3/3 USB: Fix device driver race

In reply to Bastien Nocera

When a new device with a specialised device driver is plugged in, the new driver will be modprobe()'d but the driver core will attach the "generic" driver to the device.

After that, nothing will trigger a reprobe when the modprobe()'d device driver has finished initialising, as the device has the "generic" driver attached to it.

Trigger a reprobe ourselves when new specialised drivers get registered.

Fixes 88b7381a939d ("USB: Select better matching USB drivers when available")

---
Changes since v7:
- None

Changes since v6:
- Added Alan's ack

Changes since v5:
- Throw error when device_reprobe() fails

Changes since v4:
- Add commit subject to "fixes" section
- Clarify conditional that checks for generic driver
- Remove check duplicated inside the loop

Changes since v3:
- Only reprobe devices that could use the new driver
- Many code fixes

Changes since v2:
- Fix formatting

Changes since v1:
- Simplified after Alan Stern's comments and some clarifications from Benjamin Tissoires.

 drivers/usb/core/driver.c | 40 +++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/usb/core/driver.c +38 -2
index f81606c6a35b..7e73e989645b 100644
@@ -905,6 +905,35 @@ static int usb_uevent(struct device *dev, struct kobj_uevent_env *env)
 	return 0;
 }
 
+static bool is_dev_usb_generic_driver(struct device *dev)
+{
+	struct usb_device_driver *udd = dev->driver ?
+		to_usb_device_driver(dev->driver) : NULL;
+
+	return udd == &usb_generic_driver;
+}
+
+static int __usb_bus_reprobe_drivers(struct device *dev, void *data)
+{
+	struct usb_device_driver *new_udriver = data;
+	struct usb_device *udev;
+	int ret;
+
+	if (!is_dev_usb_generic_driver(dev))
+		return 0;
+
+	udev = to_usb_device(dev);
+	if (usb_device_match_id(udev, new_udriver->id_table) == NULL &&
+	    (!new_udriver->match || new_udriver->match(udev) != 0))
+		return 0;
+
+	ret = device_reprobe(dev);
+	if (ret && ret != -EPROBE_DEFER)
+		dev_err(dev, "Failed to reprobe device (error %d)\n", ret);
+
+	return 0;
+}
+
 /**
  * usb_register_device_driver - register a USB device (not interface) driver
  * @new_udriver: USB operations for the device driver
@@ -934,13 +963,20 @@ int usb_register_device_driver(struct usb_device_driver *new_udriver,
 
 	retval = driver_register(&new_udriver->drvwrap.driver);
 
-	if (!retval)
+	if (!retval) {
 		pr_info("%s: registered new device driver %s\n",
 			usbcore_name, new_udriver->name);
-	else
+		/*
+		 * Check whether any device could be better served with
+		 * this new driver
+		 */
+		bus_for_each_dev(&usb_bus_type, NULL, new_udriver,
+				 __usb_bus_reprobe_drivers);
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",
 			usbcore_name, retval, new_udriver->name);
+	}
 
 	return retval;
 }
Signature · 1 line
2.26.2
#3 lore raw

Re: [PATCH v8 3/3] USB: Fix device driver race

Quoted 12 lines · Bastien Nocera

When a new device with a specialised device driver is plugged in, the new driver will be modprobe()'d but the driver core will attach the "generic" driver to the device.

After that, nothing will trigger a reprobe when the modprobe()'d device driver has finished initialising, as the device has the "generic" driver attached to it.

Trigger a reprobe ourselves when new specialised drivers get registered.

Fixes 88b7381a939d ("USB: Select better matching USB drivers when available")

[...]

Quoted 7 lines
 drivers/usb/core/driver.c | 40 +++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/usb/core/driver.c +0 -0
index f81606c6a35b..7e73e989645b 100644

[...]

Quoted 8 lines
drivers/usb/core/driver.c +1 -1
@@ -934,13 +963,20 @@ int usb_register_device_driver(struct usb_device_driver *new_udriver,
 
 	retval = driver_register(&new_udriver->drvwrap.driver);
 
-	if (!retval)
+	if (!retval) {
 		pr_info("%s: registered new device driver %s\n",
 			usbcore_name, new_udriver->name);

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 977

Quoted 10 lines
drivers/usb/core/driver.c +7 -1
-	else
+		/*
+		 * Check whether any device could be better served with
+		 * this new driver
+		 */
+		bus_for_each_dev(&usb_bus_type, NULL, new_udriver,
+				 __usb_bus_reprobe_drivers);
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",

Unrelated but... hm, this string literal seems weird. GregKH, would it be OK if we fix it?

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 982

Quoted 6 lines
drivers/usb/core/driver.c +1 -0
 			usbcore_name, retval, new_udriver->name);
+	}
 
 	return retval;
 }

MBR, Sergei

#4 lore raw

Re: [PATCH v8 3/3] USB: Fix device driver race

Quoted 27 lines · Sergei Shtylyov
Quoted 12 lines · Bastien Nocera

When a new device with a specialised device driver is plugged in, the new driver will be modprobe()'d but the driver core will attach the "generic" driver to the device.

After that, nothing will trigger a reprobe when the modprobe()'d device driver has finished initialising, as the device has the "generic" driver attached to it.

Trigger a reprobe ourselves when new specialised drivers get registered.

Fixes 88b7381a939d ("USB: Select better matching USB drivers when available")

[...]

Quoted 7 lines
 drivers/usb/core/driver.c | 40 +++++++++++++++++++++++++++++++++++++--
 1 file changed, 38 insertions(+), 2 deletions(-)
drivers/usb/core/driver.c +0 -0
index f81606c6a35b..7e73e989645b 100644

[...]

Quoted 4 lines
drivers/usb/core/driver.c +0 -0
@@ -934,13 +963,20 @@ int usb_register_device_driver(struct usb_device_driver *new_udriver,
 
 	retval = driver_register(&new_udriver->drvwrap.driver);
 
Quoted 16 lines
Quoted 4 lines
drivers/usb/core/driver.c +1 -1
-	if (!retval)
+	if (!retval) {
 		pr_info("%s: registered new device driver %s\n",
 			usbcore_name, new_udriver->name);

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 977

Quoted 10 lines
drivers/usb/core/driver.c +7 -1
-	else
+		/*
+		 * Check whether any device could be better served with
+		 * this new driver
+		 */
+		bus_for_each_dev(&usb_bus_type, NULL, new_udriver,
+				 __usb_bus_reprobe_drivers);
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",

Unrelated but... hm, this string literal seems weird. GregKH, would it be OK if we fix it?

Please do.

#5 lore raw

Re: [PATCH v8 3/3] USB: Fix device driver race

In reply to Sergei Shtylyov
Quoted 9 lines · Sergei Shtylyov

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 977

Quoted 3 lines · Bastien Nocera
drivers/usb/core/driver.c +1 -0
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",

Unrelated but... hm, this string literal seems weird. GregKH, would it be OK if we fix it?

Quoted 1 line

usbcore_name, retval, new_udriver->name);

Indeed, an extra tab character snuck in there by mistake. It has been present ever since 2006, when the routine was originally added by commit 8bb54ab573ec ("usbcore: add usb_device_driver definition").

It's perfectly okay with me if someone wants to remove the extra tab. In fact, nowadays we'd remove the line break entirely.

Alan Stern

#6 lore raw

Re: [PATCH v8 3/3] USB: Fix device driver race

Quoted 11 lines · Alan Stern
Quoted 7 lines

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 977

Quoted 3 lines
drivers/usb/core/driver.c +1 -0
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",

Unrelated but... hm, this string literal seems weird. GregKH, would it be OK if we fix it?

Quoted 1 line

usbcore_name, retval, new_udriver->name);

Indeed, an extra tab character snuck in there by mistake. It has been present ever since 2006, when the routine was originally added by commit 8bb54ab573ec ("usbcore: add usb_device_driver definition").

And meanwhile it got copied to another function, usb_register_driver().
I guess it's OK to fix both w/one patch?

Quoted 2 lines

It's perfectly okay with me if someone wants to remove the extra tab. In fact, nowadays we'd remove the line break entirely.

It seems this wasn't needed even in the 80-column era...

Quoted 1 line

Alan Stern

MBR, Sergei

#7 lore raw

Re: [PATCH v8 3/3] USB: Fix device driver race

Quoted 8 lines · Sergei Shtylyov
Quoted 6 lines · Alan Stern
Quoted 6 lines

drivers/usb/core/driver.c int usb_register_device_driver(struct usb_device_driver *new_udriver, · line 977

Quoted 3 lines
drivers/usb/core/driver.c +1 -0
+	} else {
 		printk(KERN_ERR "%s: error %d registering device "
 			"	driver %s\n",

Unrelated but... hm, this string literal seems weird. GregKH, would it be OK if we fix it?

Quoted 8 lines
Quoted 5 lines
Quoted 1 line
Quoted 1 line

usbcore_name, retval, new_udriver->name);

Indeed, an extra tab character snuck in there by mistake. It has been present ever since 2006, when the routine was originally added by commit 8bb54ab573ec ("usbcore: add usb_device_driver definition").

And meanwhile it got copied to another function, usb_register_driver().
I guess it's OK to fix both w/one patch?

Yes.