Re: [PATCH 2/2] usb: musb-new: sunxi: clarify the purpose of SRAM initialization

[Sam Edwards <cfsworks@gmail.com>]

Hi Andre,

I've applied most of this feedback (most of which comes as a relief; I 
dislike inventing names for mystery bits) in preparation to send a v2, 
but had two questions:

On 6/9/23 04:13, Andre Przywara wrote:
>> The new comments and function name are not from any official source,
>> but are updated to mirror how the Linux kernel sources treat this
>> mystery magic bit. If we wanted to confirm that this speculation is
>> correct, we could verify that SRAM-D is inaccessible whenever the
>> bit is set, and then try clearing it again while the MUSB is in use
>> to see what debris gets left behind in SRAM-D.
>>
>> This cleanup also adds a TODO comment about runtime discovery
>> of the SYSCON base, per discussion with Andre.
> 
> thanks for sending this, looks good. Some stylistic comments below.

I take it that this (in combination with your review on 1/2) means you 
concur with my speculated purpose of the mystery bit. If so, I'd like to 
rephrase the above paragraph in the commit message to:

"""
The new comments and function name are not from any official source,
but are updated to mirror how the Linux kernel sources treat this
mystery magic bit. This also reflects what's been observed on actual
hardware: SRAM-D is inaccessible by the CPU when the bit is set, and
the MUSB unit crashes when this bit is cleared while USB is in use,
leaving behind debris in SRAM-D from its use as a "scratch space."
"""

Does this accurately reflect what you've seen, particularly (especially) 
that last line, or should I end the commentary at "is in use."?

> I think we should use "non-net" commenting style, with the "/*" on a line
> on its own.

This seems to be an obscure term of art, and searching "non-net comment" 
and "non-net style" on Google isn't finding me any style guide or set of 
rules to check against. (Amusingly: if I search for the "net" style, I 
get a lot of .NET suggestions.)

The main thing I'm trying to figure out is if it also demands */ on its 
own line, which I would intuitively think makes sense (it does look 
better to me that way), but I'm unsure if your lack of critique at the 
closing side of my multiline comments means you would prefer:
/*
  * A block of text that's long enough to become a
  * multiline comment and ends up looking like this */

Thanks for your quick and thorough feedback,
Sam