[openib-general] [PATCH v3 1/7] AMSO1100 Low Level Driver.

[Arjan van de Ven]

On Fri, 2006-06-23 at 08:56 -0500, Steve Wise wrote:
> On Fri, 2006-06-23 at 15:48 +0200, Arjan van de Ven wrote:
> > > > > +	/* Tell HW to xmit */
> > > > > +	__raw_writeq(cpu_to_be64(mapaddr), elem->hw_desc + C2_TXP_ADDR);
> > > > > +	__raw_writew(cpu_to_be16(maplen), elem->hw_desc + C2_TXP_LEN);
> > > > > +	__raw_writew(cpu_to_be16(TXP_HTXD_READY), elem->hw_desc + C2_TXP_FLAGS);
> > > > 
> > > > or here
> > > > 
> > > 
> > > No need here.  This logic submits the packet for transmission.  We don't
> > > assume it is transmitted until we (after a completion interrupt usually)
> > > read back the HTXD entry and see the TXP_HTXD_DONE bit set (see
> > > c2_tx_interrupt()). 
> > 
> > ... but will that interrupt happen at all if these 3 writes never hit
> > the hardware?
> > 
> 
> I thought the posted write WILL eventually get to adapter memory.  Not
> stall forever cached in a bridge.  I'm wrong?

I'm not sure there is a theoretical upper bound.... 

(and if it's several msec per bridge, then you have a lot of latency
anyway)