"config x has an invalid interface number: y but max is z"

Post by BjÃ¸rn Mork
I'm wondering about the usefulness and correctness of some of the
warnings we print.

OK, I went looking for the spec background and found this in=20
"Universal Serial Bus Specification Revision 2.0" section
"9.2.3 Configuration":

"Interfaces are numbered from zero to one less than the number of con=
current interfaces
supported by the configuration."

So I guess the warnings are correct. Given that the non-sequentional
numbers are used and work just fine, I still question the usefulness
though. There are ~15 lines of code in drivers/usb/core/config.c just
for printing these warnings...

Bj=C3=B8rn
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Alan Stern

2013-01-17 18:40:28 UTC

Post by BjÃ¸rn Mork
I'm wondering about the usefulness and correctness of some of the
warnings we print.

The warnings are correct. How useful they are is another matter...

Post by BjÃ¸rn Mork
Some vendors obviously allocate fixed interface numbers for specific
functions and present configurations with large "holes" in the interf=

ace

Post by BjÃ¸rn Mork
numbers.

Doing so violates the USB spec. Section 9.2.3 says: "Interfaces are=20
numbered from zero to one less than the number of concurrent interfaces
supported by the configuration." These devices probably could not pass=
=20
the USB-CV tests.

Post by BjÃ¸rn Mork
This example was just posted on the NM list, and I have some
=20
usb 1-1.5: new high-speed USB device number 7 using ehci_hcd
usb 1-1.5: config 1 has an invalid interface number: 8 but max is 3
usb 1-1.5: config 1 has no interface number 1
usb 1-1.5: config 2 has an invalid interface number: 12 but max is 1
usb 1-1.5: config 2 has an invalid interface number: 13 but max is 1
usb 1-1.5: config 2 has an invalid interface number: 13 but max is 1
usb 1-1.5: config 2 has no interface number 0
usb 1-1.5: config 2 has no interface number 1
=20
Is this useful to anyone?

In theory it would be useful to somebody developing firmware for a USB=20
device. If only such people would test their firmwares under Linux...

Post by BjÃ¸rn Mork
I expect it is just log noise to most users,
and some are probably confused and think there is something wrong her=

Post by BjÃ¸rn Mork
At least when they are having some unrelated problem with the device.
=20
Should we just drop those warnings?

No opinion.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Peter Stuge

2013-01-17 18:55:02 UTC

Post by BjÃ¸rn Mork
Is this useful to anyone?

In theory it would be useful to somebody developing firmware for a USB
device. If only such people would test their firmwares under Linux...

Some do. Maybe it's only hobbyists, but I think it may change, and I
think it's very good to warn about things which are wrong.

Post by BjÃ¸rn Mork
Should we just drop those warnings?

No opinion.

I think they should stay.

//Peter
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Bjørn Mork

2013-01-17 20:52:12 UTC

Post by BjÃ¸rn Mork
Is this useful to anyone?

=20
In theory it would be useful to somebody developing firmware for a U=

SB=20

device. If only such people would test their firmwares under Linux.=

=2E.

Some do. Maybe it's only hobbyists, but I think it may change,=20

The example I posted actually originates from a device with vendor
Linux support. I am pretty sure that some firmware engineer there has
tested the device with Linux and ignored those warnings.

and I
think it's very good to warn about things which are wrong.

I agree in principle, as long as the warning is good for something. In
this case it is not. There is absolutely zilch you and I can do about
the issue, and it does not cause any problems we need to be aware of
either.

Post by BjÃ¸rn Mork
Should we just drop those warnings?

=20
No opinion.

I think they should stay.

OK, but how about demoting them to debug messages instead of warnings?

Bj=C3=B8rn
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Alan Stern

2013-01-18 16:24:49 UTC

Post by BjÃ¸rn Mork
Is this useful to anyone?

=20
In theory it would be useful to somebody developing firmware for a=

USB=20

device. If only such people would test their firmwares under Linu=

x...

Some do. Maybe it's only hobbyists, but I think it may change,=20

=20
The example I posted actually originates from a device with vendor
Linux support. I am pretty sure that some firmware engineer there ha=

Post by BjÃ¸rn Mork
tested the device with Linux and ignored those warnings.
=20

and I
think it's very good to warn about things which are wrong.

=20
I agree in principle, as long as the warning is good for something. =

Post by BjÃ¸rn Mork
this case it is not. There is absolutely zilch you and I can do abou=

Post by BjÃ¸rn Mork
the issue, and it does not cause any problems we need to be aware of
either.
=20

Post by BjÃ¸rn Mork
Should we just drop those warnings?

=20
No opinion.

I think they should stay.

=20
OK, but how about demoting them to debug messages instead of warnings=

?

If you want to make that change, I don't mind.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo-***@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html

Peter Stuge

2013-01-18 16:27:44 UTC

Post by Peter Stuge