Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1360391

Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT

Show all headers | View raw

On Friday 18 March 2016 09:54:14 Peter Chen wrote:
> > 
> > I don't think it's a very widespread problem, there are only very few
> > developers that intentionally use this method, and some use the
> > platform_device_register_full() call to create a device with a known
> > mask, which is generally ok for the limited case where the driver
> > is only ever going to run on a single platform, but not in the
> > more general case that of_dma_configure is designed to handle.
> 
> Even only for qualcomm platforms, it may be possible have different
> DMA masks at ARM64 platforms, so we may can't use a fixed value
> at glue layer driver. So, using of_dma_configure is suitable choice
> for DT platforms for this case, right?

Yes.

> > I think we should fix the drivers to consistently use the device
> > that was created by the platform (DT or ACPI or board file)
> > to pass that into the DMA API, anything else will just cause
> > more subtle bugs.
> > 
> 
> Although I don't know what kinds of bugs it may have, it may be
> met before, otherwise, why most of platform drivers need to call
> dma_set_coherent_mask or dma_coerce_mask_and_coherent explicitly

Any driver that wants to do 64-bit addressing on DMA should call
dma_set_mask()/dma_set_coherent_mask() on its device and check the
return code.

No driver should call dma_coerce_mask_and_coherent() on its own
device, it's basically always a bug and we named the function
to make that more obvious. The problem with dma_coerce_mask_and_coherent()
is that it just overrides whatever the platform knows about the
device when the driver thinks it knows better. 

The reason for having those calls in a lot of drivers is that
traditionally, ARM platforms booting with DT did not set up any DMA
mask and the drivers worked around it by manually setting up a mask
that happened to work for them (almost all 32-bit ARM devices need
a 32-bit mask without coherency or offset or iommu, so that's easy).
We now call of_dma_configure() for all devices that get probed from
DT, so we should be removing those calls.

	Arnd

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Li Yang <leoli@freescale.com> - 2016-03-08 21:00 +0100
  Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Bjorn Andersson <bjorn.andersson@linaro.org> - 2016-03-09 04:50 +0100
    Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Li Yang <leoli@freescale.com> - 2016-03-10 00:50 +0100
      Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Peter Chen <hzpeterchen@gmail.com> - 2016-03-14 12:00 +0100
        Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Arnd Bergmann <arnd@arndb.de> - 2016-03-17 17:00 +0100
          Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Peter Chen <hzpeterchen@gmail.com> - 2016-03-18 03:10 +0100
            RE: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Rajesh Bhagat <rajesh.bhagat@nxp.com> - 2016-03-18 04:50 +0100
            Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Arnd Bergmann <arnd@arndb.de> - 2016-03-18 08:30 +0100
            Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Russell King - ARM Linux <linux@arm.linux.org.uk> - 2016-03-18 12:00 +0100
    Re: [PATCH] usb: chipidea: Configure DMA properties and ops from DT Peter Chen <hzpeterchen@gmail.com> - 2016-03-25 05:10 +0100
      Re: [PATCH 1/1] usb: chipidea: add DMA mask configuration API kbuild test robot <lkp@intel.com> - 2016-03-25 05:50 +0100

csiph-web