> On Dec 23, 2007 9:31 PM, YanBo < dre...@gm...> wrote:
> >
> > On Dec 23, 2007 8:51 PM, far seer <se...@gm...> wrote:
> > > YanBo,
> > > I want to descripte the driver work flow: after user send http request
> via
> > > socket, then agnx_tx deal with the requeset, the result is raising a
> > > interrupt. In interrupt handl, handle_txd_irq is responsed for
> transfering
> > > data. Is it right?
> >
> > yes. totally right.
> >
> >
> > > I find agnx_tx at line 3018 in agnx.dmesg which is http request, and
> >
> > no it get the handle interrupt in the line 3653
> >
> >
> >
> >
> > > following report have no interrupt handle for handle_txd_irq.
> > > Because there is no interrupt reason with AGNX_STAT_TXD.
> > >
> > > BR
> > >
> >
>
>
On Dec 23, 2007 10:26 PM, far seer <se...@gm...> wrote:
> Is this line?
> [ 1821.101561] agnx: Interrupt reason is 0x801
> but AGNX_STAT_TXD can't not been masked by it, since
No matter txm or txd was triggered, all what will see is the interrupt
reason register is 0x801.
> #define AGNX_STAT_TXD 0x10
This is define by myself to distinguish txd and txm, it will no be
know by hardware. There is a comment above this code:
/* Below two interrupt flags will be set by our but not CPU */
lyb
|
