Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> writes:
> On Fri, May 11, 2012 at 02:21:07PM +0200, Bjørn Mork wrote:
>> This enables the current list of dynamic IDs to be read out through
>> either new_id or remove_id.
>>
>> Signed-off-by: Bjørn Mork <bjorn@xxxxxxx>
>> ---
>>
>> > BTW, another wishlist item is the ability to view the dynamic
>> > device list. Any objections against providing a read method
>> > for the new_id (and maybe also remove_id) file?
>>
>> So, how about something like this? Bad idea? Good idea?
>
> Great idea!
>
>> The background for this is not only my failing memory, but that there
>> are userspace tools (usb_modeswitch is one example) which will add
>> device IDs to drivers on behalf of the user. Being able to read out
>> the list is very useful in such cases.
>
> I agree, I like this a lot. I'll try to test it out on Monday and see
> how it goes. At first glance, it looks good to me, but we will need
> a Documentation/ABI/ file update as well to show the format of what we
> are printing out.
Is something like this sufficient, or should the dates be updated?
diff --git a/Documentation/ABI/testing/sysfs-bus-usb b/Documentation/ABI/testing/sysfs-bus-usb
index 7c22a53..e2a3839 100644
--- a/Documentation/ABI/testing/sysfs-bus-usb
+++ b/Documentation/ABI/testing/sysfs-bus-usb
@@ -135,6 +135,17 @@ Description:
for the device and attempt to bind to it. For example:
# echo "8086 10f5" > /sys/bus/usb/drivers/foo/new_id
+ Reading from this file will list all dynamically added
+ device IDs in the same format, with one entry per
+ line. For example:
+ # cat /sys/bus/usb/drivers/foo/new_id
+ 8086 10f5
+ dead beef 06
+ f00d cafe
+
+ The list will be truncated at PAGE_SIZE bytes due to
+ sysfs restrictions.
+
What: /sys/bus/usb-serial/drivers/.../new_id
Date: October 2011
Contact: linux-usb@xxxxxxxxxxxxxxx
@@ -157,6 +168,10 @@ Description:
match the driver to the device. For example:
# echo "046d c315" > /sys/bus/usb/drivers/foo/remove_id
+ Reading from this file will list the dynamically added
+ device IDs, exactly like reading from the entry
+ "/sys/bus/usb/drivers/.../new_id"
+
What: /sys/bus/usb/device/.../avoid_reset_quirk
Date: December 2009
Contact: Oliver Neukum <oliver@xxxxxxxxxx>
Note that the above was based on this minor adjustment to reflect the
possible bInterfaceClass input as well:
diff --git a/drivers/usb/core/driver.c b/drivers/usb/core/driver.c
index cf26fa1..61f4b7e 100644
--- a/drivers/usb/core/driver.c
+++ b/drivers/usb/core/driver.c
@@ -88,8 +88,13 @@ static ssize_t show_dynids(struct device_driver *driver, char *buf)
size_t count = 0;
list_for_each_entry(dynid, &usb_drv->dynids.list, node)
- count += scnprintf(&buf[count], PAGE_SIZE - count, "%04x %04x\n",
- dynid->id.idVendor, dynid->id.idProduct);
+ if (dynid->id.bInterfaceClass != 0)
+ count += scnprintf(&buf[count], PAGE_SIZE - count, "%04x %04x %02x\n",
+ dynid->id.idVendor, dynid->id.idProduct,
+ dynid->id.bInterfaceClass);
+ else
+ count += scnprintf(&buf[count], PAGE_SIZE - count, "%04x %04x\n",
+ dynid->id.idVendor, dynid->id.idProduct);
return count;
}
Reading the documentation I also realized that the usb-serial new_id
should include the same output as well, to keep the functionality in
sync. Which I guess means that show_dynids() should be exported (after
proper renaming) so that it can be called from drivers/usb/serial/bus.c.
I'll clean all this up and submit as a two patch series.
Bjørn
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
