linux-kernel - Re: [PATCH v8 07/18] vfio/platform: return info for device memory mapped IO regions

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Wed, 22 Oct 2014 10:46:29 -0600
From:	Alex Williamson <alex.williamson@...hat.com>
To:	Antonios Motakis <a.motakis@...tualopensystems.com>
Cc:	kvm-arm <kvmarm@...ts.cs.columbia.edu>,
	Linux IOMMU <iommu@...ts.linux-foundation.org>,
	Will Deacon <will.deacon@....com>,
	VirtualOpenSystems Technical Team <tech@...tualopensystems.com>,
	Christoffer Dall <christoffer.dall@...aro.org>,
	Eric Auger <eric.auger@...aro.org>,
	Kim Phillips <kim.phillips@...escale.com>,
	Marc Zyngier <marc.zyngier@....com>,
	"open list:VFIO DRIVER" <kvm@...r.kernel.org>,
	open list <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v8 07/18] vfio/platform: return info for device memory
 mapped IO regions

On Wed, 2014-10-22 at 15:54 +0200, Antonios Motakis wrote:
> On Tue, Oct 21, 2014 at 6:34 PM, Alex Williamson
> <alex.williamson@...hat.com> wrote:
> > On Mon, 2014-10-13 at 15:10 +0200, Antonios Motakis wrote:
> >> This patch enables the IOCTLs VFIO_DEVICE_GET_REGION_INFO ioctl call,
> >> which allows the user to learn about the available MMIO resources of
> >> a device.
> >>
> >> Signed-off-by: Antonios Motakis <a.motakis@...tualopensystems.com>
> >> ---
> >>  drivers/vfio/platform/vfio_platform_common.c  | 93 +++++++++++++++++++++++++--
> >>  drivers/vfio/platform/vfio_platform_private.h | 22 +++++++
> >>  2 files changed, 111 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/vfio/platform/vfio_platform_common.c b/drivers/vfio/platform/vfio_platform_common.c
> >> index 1e4073f..8a7e474 100644
> >> --- a/drivers/vfio/platform/vfio_platform_common.c
> >> +++ b/drivers/vfio/platform/vfio_platform_common.c
> >> @@ -27,17 +27,84 @@
> >>
> >>  #include "vfio_platform_private.h"
> >>
> >> +static int vfio_platform_regions_init(struct vfio_platform_device *vdev)
> >> +{
> >> +     int cnt = 0, i;
> >> +
> >> +     while (vdev->get_resource(vdev, cnt))
> >> +             cnt++;
> >> +
> >> +     vdev->regions = kcalloc(cnt, sizeof(struct vfio_platform_region),
> >> +                             GFP_KERNEL);
> >> +     if (!vdev->regions)
> >> +             return -ENOMEM;
> >> +
> >> +     for (i = 0; i < cnt;  i++) {
> >> +             struct resource *res =
> >> +                     vdev->get_resource(vdev, i);
> >> +
> >> +             if (!res)
> >> +                     goto err;
> >> +
> >> +             vdev->regions[i].addr = res->start;
> >> +             vdev->regions[i].size = resource_size(res);
> >> +             vdev->regions[i].flags = 0;
> >> +
> >> +             switch (resource_type(res)) {
> >> +             case IORESOURCE_MEM:
> >> +                     vdev->regions[i].type = VFIO_PLATFORM_REGION_TYPE_MMIO;
> >> +                     break;
> >> +             case IORESOURCE_IO:
> >> +                     vdev->regions[i].type = VFIO_PLATFORM_REGION_TYPE_PIO;
> >> +                     break;
> >
> > Ok, we have support for PIO in platform now (thanks!), does the user
> > know what type of region they're dealing with?  Do they care?  For PCI
> > the user tests the PCI BAR in config space to determine which type it
> > is.  I'm guessing that platform would do something similar against the
> > device tree or ACPI, right?
> 
> Maybe is worthwhile to add an explicit flag in
> VFIO_DEVICE_GET_REGION_INFO for PIO regions. For platform devices I
> don't know if we can always rely on DT or ACPI info to be available.
> For VFIO PCI the BAR is always implemented, and while I have proposed
> an RFC to return DT information, I don't think we can assume how a
> device is described in the host, whether DT, ACPI, or dark magic.

Is this already handled by the fact that vfio-platform is not meant to
be a generic meta driver to the same extent as vfio-pci?  There is no
self-describing config space on platform devices like there is on PCI,
so the user will need to know in advance somehow what the device is and
what resources/irqs it uses.  We do need to make sure though that we
provide a userspace ABI that allows them to match VFIO indexes to the
device in a predictable way.  It's not fully clear to me how that works.
Thanks,

Alex

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/