This series fixes problems introduced by commit ec51fbd1b8a2 ("r8152:
add USB device driver for config selection") where the r8152 device
would stop functioning if you deauthorized it (by writing 0 to the
"authorized" field in sysfs) and then reauthorized it (by writing a
1).
In v1 this was just a single patch [1], but it's now a 3-patch series
and solves the problem in a much cleaner way, as suggested by Alan
Stern.
Since these three patches straddle the USB subsystem and the
networking subsystem then maintainers will (obviously) need to work
out a way for them to land. I don't have any strong suggestions here
so I'm happy to let the maintainers propose what they think will work
best.
[1] https://lore.kernel.org/r/20231130154337.1.Ie00e07f07f87149c9ce0b27ae4e26991d307e14b@changeid
Changes in v2:
- ("Don't force USB generic_subclass drivers to define ...") new for v2.
- ("Allow subclassed USB drivers to override ...") new for v2.
- ("Choose our USB config with choose_configuration()...) new for v2.
Douglas Anderson (3):
usb: core: Don't force USB generic_subclass drivers to define probe()
usb: core: Allow subclassed USB drivers to override
usb_choose_configuration()
r8152: Choose our USB config with choose_configuration() rather than
probe()
drivers/net/usb/r8152.c | 16 +++++-----------
drivers/usb/core/driver.c | 5 ++++-
drivers/usb/core/generic.c | 7 +++++++
include/linux/usb.h | 6 ++++++
4 files changed, 22 insertions(+), 12 deletions(-)
--
2.43.0.rc2.451.g8631bc7472-goog
There's no real reason that subclassed USB drivers _need_ to define
probe() since they might want to subclass for some other reason. Make
it optional to define probe() if we're a generic_subclass.
Signed-off-by: Douglas Anderson <[email protected]>
---
Changes in v2:
- ("Don't force USB generic_subclass drivers to define ...") new for v2.
drivers/usb/core/driver.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/usb/core/driver.c b/drivers/usb/core/driver.c
index f58a0299fb3b..1dc0c0413043 100644
--- a/drivers/usb/core/driver.c
+++ b/drivers/usb/core/driver.c
@@ -290,7 +290,10 @@ static int usb_probe_device(struct device *dev)
* specialised device drivers prior to setting the
* use_generic_driver bit.
*/
- error = udriver->probe(udev);
+ if (udriver->probe)
+ error = udriver->probe(udev);
+ else if (!udriver->generic_subclass)
+ error = -EINVAL;
if (error == -ENODEV && udriver != &usb_generic_driver &&
(udriver->id_table || udriver->match)) {
udev->use_generic_driver = 1;
--
2.43.0.rc2.451.g8631bc7472-goog
If you deauthorize the r8152 device (by writing 0 to the "authorized"
field in sysfs) and then reauthorize it (by writing a 1) then it no
longer works. This is because when you do the above we lose the
special configuration that we set in rtl8152_cfgselector_probe().
Deauthorizing causes the config to be set to -1 and then reauthorizing
runs the default logic for choosing the best config.
I made an attempt to fix it so that the config is kept across
deauthorizing / reauthorizing [1] but it was a bit ugly.
Let's instead use the new USB core feature to override
choose_configuration().
This patch relies upon the patches ("usb: core: Don't force USB
generic_subclass drivers to define probe()") and ("usb: core: Allow
subclassed USB drivers to override usb_choose_configuration()")
[1] https://lore.kernel.org/r/20231130154337.1.Ie00e07f07f87149c9ce0b27ae4e26991d307e14b@changeid
Fixes: ec51fbd1b8a2 ("r8152: add USB device driver for config selection")
Suggested-by: Alan Stern <[email protected]>
Signed-off-by: Douglas Anderson <[email protected]>
---
Changes in v2:
- ("Choose our USB config with choose_configuration()...) new for v2.
drivers/net/usb/r8152.c | 16 +++++-----------
1 file changed, 5 insertions(+), 11 deletions(-)
diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
index 2c5c1e91ded6..0da723d11326 100644
--- a/drivers/net/usb/r8152.c
+++ b/drivers/net/usb/r8152.c
@@ -10053,7 +10053,7 @@ static struct usb_driver rtl8152_driver = {
.disable_hub_initiated_lpm = 1,
};
-static int rtl8152_cfgselector_probe(struct usb_device *udev)
+static int rtl8152_cfgselector_choose_configuration(struct usb_device *udev)
{
struct usb_host_config *c;
int i, num_configs;
@@ -10080,19 +10080,13 @@ static int rtl8152_cfgselector_probe(struct usb_device *udev)
if (i == num_configs)
return -ENODEV;
- if (usb_set_configuration(udev, c->desc.bConfigurationValue)) {
- dev_err(&udev->dev, "Failed to set configuration %d\n",
- c->desc.bConfigurationValue);
- return -ENODEV;
- }
-
- return 0;
+ return c->desc.bConfigurationValue;
}
static struct usb_device_driver rtl8152_cfgselector_driver = {
- .name = MODULENAME "-cfgselector",
- .probe = rtl8152_cfgselector_probe,
- .id_table = rtl8152_table,
+ .name = MODULENAME "-cfgselector",
+ .choose_configuration = rtl8152_cfgselector_choose_configuration,
+ .id_table = rtl8152_table,
.generic_subclass = 1,
.supports_autosuspend = 1,
};
--
2.43.0.rc2.451.g8631bc7472-goog
On Fri, Dec 01, 2023 at 10:29:50AM -0800, Douglas Anderson wrote:
> There's no real reason that subclassed USB drivers _need_ to define
> probe() since they might want to subclass for some other reason. Make
> it optional to define probe() if we're a generic_subclass.
>
> Signed-off-by: Douglas Anderson <[email protected]>
> ---
Reviewed-by: Alan Stern <[email protected]>
> Changes in v2:
> - ("Don't force USB generic_subclass drivers to define ...") new for v2.
>
> drivers/usb/core/driver.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/core/driver.c b/drivers/usb/core/driver.c
> index f58a0299fb3b..1dc0c0413043 100644
> --- a/drivers/usb/core/driver.c
> +++ b/drivers/usb/core/driver.c
> @@ -290,7 +290,10 @@ static int usb_probe_device(struct device *dev)
> * specialised device drivers prior to setting the
> * use_generic_driver bit.
> */
> - error = udriver->probe(udev);
> + if (udriver->probe)
> + error = udriver->probe(udev);
> + else if (!udriver->generic_subclass)
> + error = -EINVAL;
> if (error == -ENODEV && udriver != &usb_generic_driver &&
> (udriver->id_table || udriver->match)) {
> udev->use_generic_driver = 1;
> --
> 2.43.0.rc2.451.g8631bc7472-goog
>
On Fri, Dec 1, 2023 at 10:31 AM Douglas Anderson <[email protected]> wrote:
>
> There's no real reason that subclassed USB drivers _need_ to define
> probe() since they might want to subclass for some other reason. Make
> it optional to define probe() if we're a generic_subclass.
>
> Signed-off-by: Douglas Anderson <[email protected]>
Reviewed-by: Grant Grundler <[email protected]>
Thanks for pursuing this Doug!
cheers,
grant
> ---
>
> Changes in v2:
> - ("Don't force USB generic_subclass drivers to define ...") new for v2.
>
> drivers/usb/core/driver.c | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/core/driver.c b/drivers/usb/core/driver.c
> index f58a0299fb3b..1dc0c0413043 100644
> --- a/drivers/usb/core/driver.c
> +++ b/drivers/usb/core/driver.c
> @@ -290,7 +290,10 @@ static int usb_probe_device(struct device *dev)
> * specialised device drivers prior to setting the
> * use_generic_driver bit.
> */
> - error = udriver->probe(udev);
> + if (udriver->probe)
> + error = udriver->probe(udev);
> + else if (!udriver->generic_subclass)
> + error = -EINVAL;
> if (error == -ENODEV && udriver != &usb_generic_driver &&
> (udriver->id_table || udriver->match)) {
> udev->use_generic_driver = 1;
> --
> 2.43.0.rc2.451.g8631bc7472-goog
>
On Fri, Dec 1, 2023 at 10:31 AM Douglas Anderson <[email protected]> wrote:
>
> If you deauthorize the r8152 device (by writing 0 to the "authorized"
> field in sysfs) and then reauthorize it (by writing a 1) then it no
> longer works. This is because when you do the above we lose the
> special configuration that we set in rtl8152_cfgselector_probe().
> Deauthorizing causes the config to be set to -1 and then reauthorizing
> runs the default logic for choosing the best config.
>
> I made an attempt to fix it so that the config is kept across
> deauthorizing / reauthorizing [1] but it was a bit ugly.
>
> Let's instead use the new USB core feature to override
> choose_configuration().
>
> This patch relies upon the patches ("usb: core: Don't force USB
> generic_subclass drivers to define probe()") and ("usb: core: Allow
> subclassed USB drivers to override usb_choose_configuration()")
>
> [1] https://lore.kernel.org/r/20231130154337.1.Ie00e07f07f87149c9ce0b27ae4e26991d307e14b@changeid
>
> Fixes: ec51fbd1b8a2 ("r8152: add USB device driver for config selection")
> Suggested-by: Alan Stern <[email protected]>
> Signed-off-by: Douglas Anderson <[email protected]>
Reviewed-by: Grant Grundler <[email protected]>
> ---
>
> Changes in v2:
> - ("Choose our USB config with choose_configuration()...) new for v2.
>
> drivers/net/usb/r8152.c | 16 +++++-----------
> 1 file changed, 5 insertions(+), 11 deletions(-)
>
> diff --git a/drivers/net/usb/r8152.c b/drivers/net/usb/r8152.c
> index 2c5c1e91ded6..0da723d11326 100644
> --- a/drivers/net/usb/r8152.c
> +++ b/drivers/net/usb/r8152.c
> @@ -10053,7 +10053,7 @@ static struct usb_driver rtl8152_driver = {
> .disable_hub_initiated_lpm = 1,
> };
>
> -static int rtl8152_cfgselector_probe(struct usb_device *udev)
> +static int rtl8152_cfgselector_choose_configuration(struct usb_device *udev)
> {
> struct usb_host_config *c;
> int i, num_configs;
> @@ -10080,19 +10080,13 @@ static int rtl8152_cfgselector_probe(struct usb_device *udev)
> if (i == num_configs)
> return -ENODEV;
>
> - if (usb_set_configuration(udev, c->desc.bConfigurationValue)) {
> - dev_err(&udev->dev, "Failed to set configuration %d\n",
> - c->desc.bConfigurationValue);
> - return -ENODEV;
> - }
> -
> - return 0;
> + return c->desc.bConfigurationValue;
> }
>
> static struct usb_device_driver rtl8152_cfgselector_driver = {
> - .name = MODULENAME "-cfgselector",
> - .probe = rtl8152_cfgselector_probe,
> - .id_table = rtl8152_table,
> + .name = MODULENAME "-cfgselector",
> + .choose_configuration = rtl8152_cfgselector_choose_configuration,
> + .id_table = rtl8152_table,
> .generic_subclass = 1,
> .supports_autosuspend = 1,
> };
> --
> 2.43.0.rc2.451.g8631bc7472-goog
>
On Fri, 1 Dec 2023 10:29:49 -0800 Douglas Anderson wrote:
> Since these three patches straddle the USB subsystem and the
> networking subsystem then maintainers will (obviously) need to work
> out a way for them to land. I don't have any strong suggestions here
> so I'm happy to let the maintainers propose what they think will work
> best.
No strong preference here, on a quick read it seems more like a USB
change than networking change, tho, so I'll defer to Greg unless told
otherwise.
--
pw-bot: not-applicable
On Fri, 1 Dec 2023 10:29:52 -0800 Douglas Anderson wrote:
> If you deauthorize the r8152 device (by writing 0 to the "authorized"
> field in sysfs) and then reauthorize it (by writing a 1) then it no
> longer works. This is because when you do the above we lose the
> special configuration that we set in rtl8152_cfgselector_probe().
> Deauthorizing causes the config to be set to -1 and then reauthorizing
> runs the default logic for choosing the best config.
>
> I made an attempt to fix it so that the config is kept across
> deauthorizing / reauthorizing [1] but it was a bit ugly.
>
> Let's instead use the new USB core feature to override
> choose_configuration().
>
> This patch relies upon the patches ("usb: core: Don't force USB
> generic_subclass drivers to define probe()") and ("usb: core: Allow
> subclassed USB drivers to override usb_choose_configuration()")
Acked-by: Jakub Kicinski <[email protected]>
On Mon, Dec 04, 2023 at 06:27:27PM -0800, Jakub Kicinski wrote:
> On Fri, 1 Dec 2023 10:29:49 -0800 Douglas Anderson wrote:
> > Since these three patches straddle the USB subsystem and the
> > networking subsystem then maintainers will (obviously) need to work
> > out a way for them to land. I don't have any strong suggestions here
> > so I'm happy to let the maintainers propose what they think will work
> > best.
>
> No strong preference here, on a quick read it seems more like a USB
> change than networking change, tho, so I'll defer to Greg unless told
> otherwise.
I took these in my tree already, sorry for not saying anything here.
thanks,
greg k-h