Re: [PATCH] USB: storage: fix Huawei mode switching regression

["Bjørn Mork" <bjorn@mork.no>]

Matthew Dharm <mdharm-usb@one-eyed-alien.net> writes:

> Frankly, I consider it appropriate.

OK, I'll try to cook something up.

> The question is not one of reminding me of what I said earlier....
> it's one of pointing people in the right direction.  Frankly, some of
> the fault for this patch lies with Greg and myself for letting it
> through.  I had just assumed that the Huawei guys had already been in
> touch with usb-modeswitch for some reason, and that this was just an
> optimization of existing logic (not an expansion).  And, frankly, I
> was just a bit tired of fighting this fight over and over again;
> having something in the file which says "here's the right and official
> way to do this" would be good.
>
> I also asked the Huawei guys about the possibility of affecting other
> devices than the one listed.... I guess one of us either wasn't clear
> or mis-understood the request.  The fact that there are devices out
> there failing now illustrates that.  Avoiding breaking existing
> systems is one of the highest priorities....

Well, I am a bit embarrassed by this.  I did notice the patch and even
looked through it when it was posted.  But when I later noticed the
switching failure I didn't think more about it at all.  I just put the
error on my to-be-investigated-further list.  Which tends to erase
itself before I get to it...

> Who is maintaining usb-modeswitch these days, anyway?  The comment in
> the file should point people directly there....

That would be Josua Dietze, who I intended to CC from the beginning of
this discussion but forgot...  Sorry about that.  I'm certainly not
going to complain about anyone else forgetting stuff :)

CCed now. Josh, I assume it's OK for you if we put a pointer to
http://www.draisberghof.de/usb_modeswitch/ in the unusual_devs.h code? 

> And, as of now, I would really like to see as many of these devices
> migrated (albeit slowly) to using usb-modeswitch wherever possible.  I
> know there are a few devices for which that might not be possible, but
> I am DONE dealing with this same issue over and over and over again.
> It will certainly be work to migrate support; maybe we should wrap all
> the relevant unusual_devs.h entires with
> CONFIG_UPDATED_MODESWITCH_INSTALLED_SO_MAKE_THESE_GO_AWAY during a
> transition period?

Sounds good to me, if that is acceptable to the user interface stability
committee.

I guess the real problem will be verifying that all of the entries *can*
go away. This type of hardware tends to get old very fast, but there is
always someone having a really ancient device.

Maybe we could use the same method the netdev people use when they want
to remove stuff: Subtly break it and leave it like that for a few years
before finally deleting it, arguing that noone can be using the feature
because it's broken :)


Bjørn

> On Mon, Mar 4, 2013 at 8:47 AM, Bjørn Mork <bjorn@mork.no> wrote:
>> Ben Hutchings <ben@decadent.org.uk> writes:
>>> On Mon, 2013-03-04 at 14:19 +0100, Bjørn Mork wrote:
>>> [...]
>>>> In-kernel mode switching was deprecated years ago with the
>>>> development of the more user friendly userspace alternatives. The
>>>> existing list of devices in usb-storage was only kept to prevent
>>>> breaking already working systems.  The long term plan is to remove
>>>> the list, not to add to it. Ref:
>>>> http://permalink.gmane.org/gmane.linux.usb.general/28543
>>> [...]
>>>
>>> Can you add a comment to this effect?
>>
>> In the table in unusual_devs.h, you mean?  Sure, I can do that.
>>
>> But it feels a bit strange since I can only quote and/or refer to what
>> Matthew and Greg said about the issue years ago.  Putting a comment in
>> the code to remind the current maintainers about their own statements
>> could be considered out of line?  Or is this appropriate here?
>>
>>
>> Bjørn
>
>
>
> --
> Matthew Dharm
> Maintainer, USB Mass Storage driver for Linux
> --
> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html