RE: [PATCH 2/7] Initial skeleton of VFIO support for Device Tree based devices

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 




> -----Original Message-----
> From: Don Dutile [mailto:ddutile@xxxxxxxxxx]
> Sent: Tuesday, October 29, 2013 12:13 PM
> To: Alex Williamson
> Cc: Bhushan Bharat-R65777; Antonios Motakis;
> kvmarm@xxxxxxxxxxxxxxxxxxxxx; linux-samsung-soc@xxxxxxxxxxxxxxx;
> kvm@xxxxxxxxxxxxxxx; agraf@xxxxxxx; Yoder Stuart-B08248;
> iommu@xxxxxxxxxxxxxxxxxxxxxxxxxx; tech@xxxxxxxxxxxxxxxxxxxxxx
> Subject: Re: [PATCH 2/7] Initial skeleton of VFIO support for Device Tree
> based devices
> 
> On 10/29/2013 07:47 AM, Alex Williamson wrote:
> > On Mon, 2013-10-28 at 21:29 -0400, Don Dutile wrote:
> >> On 09/30/2013 11:37 AM, Bhushan Bharat-R65777 wrote:
> >>>
> >>>
> >>>> -----Original Message-----
> >>>> From: iommu-bounces@xxxxxxxxxxxxxxxxxxxxxxxxxx [mailto:iommu-
> >>>> bounces@xxxxxxxxxxxxxxxxxxxxxxxxxx] On Behalf Of Antonios Motakis
> >>>> Sent: Monday, September 30, 2013 8:59 PM
> >>>> To: kvmarm@xxxxxxxxxxxxxxxxxxxxx; alex.williamson@xxxxxxxxxx
> >>>> Cc: linux-samsung-soc@xxxxxxxxxxxxxxx; kvm@xxxxxxxxxxxxxxx;
> agraf@xxxxxxx; Yoder
> >>>> Stuart-B08248; iommu@xxxxxxxxxxxxxxxxxxxxxxxxxx; Antonios Motakis;
> >>>> tech@xxxxxxxxxxxxxxxxxxxxxx
> >>>> Subject: [PATCH 2/7] Initial skeleton of VFIO support for Device
> Tree based
> >>>> devices
> >>>>
> >>>> Platform devices in the Linux kernel are usually managed by the DT
> interface.
> >>>> This patch forms the base to support these kind of devices with
> VFIO.
> >>>>
> >>>> Signed-off-by: Antonios Motakis<a.motakis@xxxxxxxxxxxxxxxxxxxxxx>
> >>>> ---
> >>>>    drivers/vfio/Kconfig         |  11 +++
> >>>>    drivers/vfio/Makefile        |   1 +
> >>>>    drivers/vfio/vfio_platform.c | 187
> +++++++++++++++++++++++++++++++++++++++++++
> >>>>    include/uapi/linux/vfio.h    |   1 +
> >>>>    4 files changed, 200 insertions(+)
> >>>>    create mode 100644 drivers/vfio/vfio_platform.c
> >>>>
> >>>> diff --git a/drivers/vfio/Kconfig b/drivers/vfio/Kconfig index
> 1f84eda..35254b7
> >>>> 100644
> >>>> --- a/drivers/vfio/Kconfig
> >>>> +++ b/drivers/vfio/Kconfig
> >>>> @@ -13,4 +13,15 @@ menuconfig VFIO
> >>>>
> >>>>    	  If you don't know what to do here, say N.
> >>>>
> >>>> +config VFIO_PLATFORM
> >>>> +	tristate "VFIO support for device tree based platform
> devices"
> >>>> +	depends on VFIO&&   EVENTFD&&   OF
> >>>> +	help
> >>>> +	  Support for platform devices with VFIO. This is required to
> make
> >>>> +	  use of platform devices present on device tree nodes using
> the VFIO
> >>>> +	  framework. Devices that are not described in the device
> tree cannot
> >>>> +	  be used by this driver.
> >>>> +
> >>>> +	  If you don't know what to do here, say N.
> >>>> +
> >>>>    source "drivers/vfio/pci/Kconfig"
> >>>> diff --git a/drivers/vfio/Makefile b/drivers/vfio/Makefile index
> >>>> 2398d4a..575c8dd 100644
> >>>> --- a/drivers/vfio/Makefile
> >>>> +++ b/drivers/vfio/Makefile
> >>>> @@ -1,3 +1,4 @@
> >>>>    obj-$(CONFIG_VFIO) += vfio.o
> >>>>    obj-$(CONFIG_VFIO_IOMMU_TYPE1) += vfio_iommu_type1.o
> >>>>    obj-$(CONFIG_VFIO_PCI) += pci/
> >>>> +obj-$(CONFIG_VFIO_PLATFORM) += vfio_platform.o
> >>>> diff --git a/drivers/vfio/vfio_platform.c
> b/drivers/vfio/vfio_platform.c new
> >>>
> >>> We can make this parallel to PCI, something like
> drivers/vfio/platform/platform.c
> >>>
> >> pls, no.  'platform' is too generic, and it really means 'arm-dt' ...
> so can
> >> move it to the arch/arm space, and have it's kconfig conditional on
> ARM&&VFIO.
> >> if kept under drivers/vfio, then use a better directory name that ties
> it to arm-dt.
> >> thanks.
> >
> > The intention is that vfio platform device support is not arm-dt
> > specific.  This is to be used by both arm and embedded ppc.  The
> devices
> > we intend to support with them are known as platform drivers in the
> > kernel, thus the name.  I suppose the question remains whether the
> > interface here is really generic for any "platform" device or whether
> > we're making whether we're making an interface specifically for device
> > tree platform devices, or if those are one in the same.  In any case,
> > arm-dt is certainly not the answer.
> >
> > Alex
> >
> I thought that was the intention until I saw this use in platform.c:
> 	static const struct of_device_id vfio_platform_match[] = {
> 
> So, of_device_id hit me as DT-specific, and thus, the file should have
> a name that implies its not as platform-generic as one may want/expect.
> I agree that the 'arm' part can be dropped, but it's not of-agnostic atm.

That part of the patch needs to be re-worked...and the comment
indicates it is not final.  The vfio-platform driver should not register
to handle specific compatible strings.  Kim Phillips has a 
patch posted that adds 'wildcard' support to platform drivers
so that a driver can bind to any device.

Stuart

--
To unsubscribe from this list: send the line "unsubscribe linux-samsung-soc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux SoC Development]     [Linux Rockchip Development]     [Linux USB Development]     [Video for Linux]     [Linux Audio Users]     [Linux SCSI]     [Yosemite News]

  Powered by Linux