Re: [PATCH V2] mmc: sdhci-acpi: Add 64-bit DMA support

[Arnd Bergmann <arnd@arndb.de>]

On Tuesday 28 October 2014 17:14:25 Adrian Hunter wrote:
> On 28/10/2014 5:08 p.m., Arnd Bergmann wrote:
> > On Tuesday 28 October 2014 16:14:46 Adrian Hunter wrote:
> >> On 28/10/2014 3:54 p.m., Arnd Bergmann wrote:
> >>> On Tuesday 28 October 2014 13:41:30 Adrian Hunter wrote:
> >>>> On 28/10/14 12:18, Arnd Bergmann wrote:
> >>>>> On Tuesday 28 October 2014 12:05:30 Adrian Hunter wrote:
> >>>>>> On 28/10/14 11:43, Arnd Bergmann wrote:
> >>>>>>> On Tuesday 28 October 2014 10:37:20 Adrian Hunter wrote:
> >>>>>>>>    static int sdhci_acpi_enable_dma(struct sdhci_host *host)
> >>>>>>>>    {
> >>>>>>>> -       return 0;
> >>>>>>>> +       struct sdhci_acpi_host *c = sdhci_priv(host);
> >>>>>>>> +       struct device *dev = &c->pdev->dev;
> >>>>>>>> +       int err = -1;
> >>>>>>>> +
> >>>>>>>> +       if (c->dma_setup)
> >>>>>>>> +               return 0;
> >>>>>>>> +
> >>>>>>>> +       if (host->flags & SDHCI_USE_64_BIT_DMA) {
> >>>>>>>> +               if (host->quirks2 & SDHCI_QUIRK2_BROKEN_64_BIT_DMA) {
> >>>>>>>> +                       host->flags &= ~SDHCI_USE_64_BIT_DMA;
> >>>>>>>> +               } else {
> >>>>>>>> +                       err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(64));
> >>>>>>>> +                       if (err)
> >>>>>>>> +                               dev_warn(dev, "Failed to set 64-bit DMA mask\n");
> >>>>>>>> +               }
> >>>>>>>> +       }
> >>>>>>>> +
> >>>>>>>> +       if (err)
> >>>>>>>> +               err = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
> >>>>>>>> +
> >>>>>>>> +       c->dma_setup = !err;
> >>>>>>>> +
> >>>>>>>> +       return err;
> >>>>>>>>    }
> >>>>>>>>
> >>>>>>>
> >>>>>>> I don't think it's worth a dev_warn() message (maybe dev_info), there is nothing
> >>>>>>
> >>>>>> It is worth a dev_warn because 32-bit DMA can allocate memory for bounce
> >>>>>> buffers which jeopardizes memory reclaim.
> >>>>>
> >>>>> Then you should also warn if SDHCI_USE_64_BIT_DMA isn't or if
> >>>>> SDHCI_QUIRK2_BROKEN_64_BIT_DMA is set I guess.
> >>>>
> >>>> The warning is for when the controller supports 64-bit, not when it doesn't.
> >>>
> >>> But why warn about a feature of the controller being present? You just
> >>> said it's a problem for memory reclaim if 64-bit DMA is not supported.
> >>
> >> The warning is for when the controller supports 64-bit but it can't
> >> get a 64-bit DMA mask, and might therefore need to bounce things.
> >
> > What does "can't get a 64-bit DMA mask" mean? This is just a different
> > way to say it doesn't support 64-bit for some reason.
> 
> The host controller advertises whether it is capable of 64-bit DMA.  If
> it is 64-bit capable but the driver cannot get a 64-bit DMA mask it issues
> a warning.

But the host controller doesn't know if it's 64-bit capable, nor should
it know. All the host controller knows is that it has registers to perform
64-bit DMA, but the platform code (ACPI in your case) is the only thing that
knows how many of those address lines are connected to an upstream bus.

	Arnd