Hi Paul,
> If you look closely at the databook, you will see that DPTXFSIZn and
> DIEPTXFn are both aliases for the same register address.
Ah, right. Remebmer I don't have the databook, only the register
descriptions from the RT3052 datasheet, which are riddled with typos, so
I had assumed the ide
Whoops, resending as text instead of html.
> From: Matthijs Kooijman [mailto:matth...@stdin.nl]
> Sent: Thursday, March 28, 2013 9:32 AM
>
> Hi Paul,
>
> while continuing this patch, I stumbled upon a bit of code which doesn't
> make sense to me. In dwc2_dump_global_registers is the following bi
Hi Paul,
while continuing this patch, I stumbled upon a bit of code which doesn't
make sense to me. In dwc2_dump_global_registers is the following bit:
if (hsotg->core_params->en_multiple_tx_fifo <= 0) {
ep_num = hsotg->hwcfg4 >> GHWCFG4_NUM_DEV_PERIO_IN_EP_SHIFT &
Hi Paul,
> Overall this seems to be an improvement, so I have no objections. There
> is a pretty high likelihood of something getting broken in the translation,
> however.
>
> For example, this:
> + hw->host_nperio_tx_fifo_size = (readl(hsotg->regs + GNPTXFSIZ) &
> GNPTXFSIZ_NP_TXF_DEP_MASK)
> From: Matthijs Kooijman [mailto:matth...@stdin.nl]
> Sent: Wednesday, March 27, 2013 8:42 AM
> To: Greg KH
>
> Hi Greg,
>
> > But yes, it does seem to be a good idea,
>
> Ok, thanks.
>
> > as long as these options can't change over time.
> It's only read-only registers, or registers whose pow
Hi Greg,
> But yes, it does seem to be a good idea,
Ok, thanks.
> as long as these options can't change over time.
It's only read-only registers, or registers whose power-on values
determine their maximum possible value, so they indeed won't change over
time.
Gr.
Matthijs
--
To unsubscribe fro
On Wed, Mar 27, 2013 at 10:48:17AM +0100, Matthijs Kooijman wrote:
> Hi folks,
>
> while going through the dwc2 driver, I've found that there is a lot of
> places that need info from the hwcfg registers. Currently, each of these
> places do their own bitshifting and masking, but it occurs to me th
