Seeking help addressing maintainer objections

[Ian Pilcher]

On 4/3/24 06:38, Bjørn Mork wrote:
> Ian Pilcher <arequipeno at gmail.com> writes:
> 
>> The maintainer of the LEDs subsystem considers this interface to be
>> "crazy", but he has not suggested any alternative.[3]
> 
> I read a strong suggestion that one of the unlink interfaces is
> dropped.
> 
> You should probably read "crazy" as "unjustified".
> 
>> (I have not been able to find any existing example of configuring
>> many-to-many relationships via sysfs.)
> 
> There's at least one such example, hidden under drivers/usb:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/usb/core/ledtrig-usbport.c
> 
> See the initial commit for justification and some of the background for
> the choices made:
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/core/ledtrig-usbport.c?id=0f247626cbbfa2010d2b86fdee652605e084e248

Thanks! Will take a look at that.

>> The second problem is the use of device file paths (including symbolic
>> link resolution), rather than kernel names, when creating associations.
>> This limitation exists because the block subsystem does not provide an
>> API to open a block device by its kernel name; only device file paths
>> and device numbers (major & minor) are supported.
> 
> So this explains why you can't have link_dev_by_path.  It still does not
> explain why you need unlink_dev_by_name.

It's not absolutely needed, but it does make it easier to unlink an LED
from all devices by using the names of the symlinks in the LED's
linked_devices directory, which will be kernel names.

> And if file name with symlink resolution really is a problem, then why
> can't you use the major:minor for link/unlink?  That's easy for
> userspace to look up whether the input is a device path or a sysfs path.
> And it avoids having to wait for an unrelated and unnecessary device
> path creation.
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/usb/core/ledtrig-usbport.c
> 
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/usb/core/ledtrig-usbport.c?id=0f247626cbbfa2010d2b86fdee652605e084e248

Personally, I don't think that using file paths is a problem, and it
can be useful.  ("/dev/vg_root/lv_root" is probably more useful than
"dm-0".)  OTOH, "sda" is slightly simpler than "/dev/sda", so I think
that the ideal situation would be to have both interfaces available.

I did propose using device numbers.  I never received a response from
the maintainer.

Thanks for taking the time to respond to this!

-- 
========================================================================
If your user interface is intuitive in retrospect ... it isn't intuitive
========================================================================