Re: [RFC 2/2] mt76: introduce mt76_hw_queue data structure

[Lorenzo Bianconi <lorenzo.bianconi@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1179 bytes --]

> On Fri, Mar 01, 2019 at 10:22:04AM +0100, Lorenzo Bianconi wrote:
> > -struct mt76_queue {
> > +struct mt76_hw_queue {
> >  
> >  	u16 first;
> >  	u16 head;
> >  	u16 tail;
> > -	int ndesc;
> > -	int queued;
> > -	int buf_size;
> >  
> > +	dma_addr_t desc_dma;
> > +
> > +	int buf_size;
> >  	u8 buf_offset;
> > +
> >  	u8 hw_idx;
> >  
> > -	dma_addr_t desc_dma;
> > -	struct sk_buff *rx_head;
> >  	struct page_frag_cache rx_page;
> >  	spinlock_t rx_page_lock;
> > +	struct sk_buff *rx_head;
> > +};
> > +
> > +struct mt76_queue {
> > +	struct mt76_hw_queue *hwq;
> > +
> > +	struct list_head swq;
> > +	int swq_queued;
> >  };
> 
> I don't think massive renaming is necessary. It makes patch more complex
> and eventual future backporting of fixes harder. Simpler approach would
> be doing:
> 
> struct mt76_sw_queue {
> 	struct mt76_queue *q;
> 
> 	struct list_head swq;
> 	int swq_queued;
> }

ack, it sounds good to me. Anyway mt76_queue is pretty spread as well :(
I will post a v2 squashing this change.

Regards,
Lorenzo

> 
> what should achieve the same goal without massive changes in the code.
> 
> Stanislaw

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]