Re: [PATCH 2/4] USB: UDC: Implement udc_async_callbacks in dummy-hcd

From: Felipe Balbi <balbi@kernel.org>
To: Alan Stern <stern@rowland.harvard.edu>
Cc: USB mailing list <linux-usb@vger.kernel.org>
Subject: Re: [PATCH 2/4] USB: UDC: Implement udc_async_callbacks in dummy-hcd
Date: Fri, 04 Jun 2021 17:20:54 +0300	[thread overview]
Message-ID: <87h7idbtix.fsf@kernel.org> (raw)
In-Reply-To: <20210604141531.GA1676809@rowland.harvard.edu>

[-- Attachment #1: Type: text/plain, Size: 1027 bytes --]

Alan Stern <stern@rowland.harvard.edu> writes:

> On Fri, Jun 04, 2021 at 08:21:11AM +0300, Felipe Balbi wrote:
>> 
>> Hi,
>> 
>> Alan Stern <stern@rowland.harvard.edu> writes:
>> > @@ -990,7 +1000,6 @@ static int dummy_udc_start(struct usb_ga
>> >  	spin_lock_irq(&dum->lock);
>> >  	dum->devstatus = 0;
>> >  	dum->driver = driver;
>> > -	dum->ints_enabled = 1;
>> 
>> should the matching write of 0 be removed from dummy_udc_stop()?
>
> No, it's okay to leave that one.  In practice it won't make any 
> difference because now the core will always turn off async callbacks 
> before doing udc_stop.  It's there for the sake of thoroughness, and it 
> lets the reader know that emulated interrupts are supposed to be turned 
> off whenever the UDC stops running (just like a driver for a real UDC).
>
> Whereas this line here in dummy_udc_start would be actively wrong if it 
> were to remain.

fair enough :-) I see Greg took the series already ;-) Thanks for
working on this, again.

-- 
balbi

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 511 bytes --]