Re: [PATCH v3] Add support for PCIe SSD status LED management

[stuart hayes <stuart.w.hayes@gmail.com>]

On 8/18/2021 2:05 AM, Pavel Machek wrote:
> Hi!
> 
>> This patch adds support for the PCIe SSD Status LED Management
>> interface, as described in the "_DSM Additions for PCIe SSD Status LED
>> Management" ECN to the PCI Firmware Specification revision 3.2.
>>
>> It will add (led_classdev) LEDs to each PCIe device that has a supported
>> _DSM interface (one off/on LED per supported state). Both PCIe storage
>> devices, and the ports to which they are connected, can support this
>> interface.
>>
>> Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
> 
> I believe these are not LEDs, right? This is something that displays
> information to the user, but how exactly it is implemented is up to
> BIOS vendor.
> 
> I don't think it is good fit for LED subsystem.
> 
> Best regards,
> 								Pavel
> 

I think it very likely these will be LEDs on most, if not all, systems 
(likely enough that the PCI ECN has "LEDs" in the name)... they have 
been LEDs on every system I've seen.  I would suspect that systems which 
support more than one or two of the states won't have a 1-to-1 mapping 
of logical LEDs to physical LEDs, though, but might instead use 
something like IBPI (SFF-8489) to use fewer LEDs and be easier to read 
at a glance.

I'm not sure I understand why the LED subsystem would be that much worse 
a fit for this than for keyboard (or other) indicator LEDs.  I 
understand that many other indicators are more likely to have each 
single kernel LED mapped to a single physical LED, but functionally they 
are both kernel LEDs controlling on/off indicators which are likely 
displayed on LEDs that are always visible on the system.