Re: [RFC PATCH v4 1/3] vfio: platform: add device properties skeleton and user API

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

 





On Wed, Sep 9, 2015 at 10:48 PM, Alex Williamson <alex.williamson@xxxxxxxxxx> wrote:
On Wed, 2015-09-09 at 11:17 +0200, Baptiste Reynal wrote:
> From: Antonios Motakis <a.motakis@xxxxxxxxxxxxxxxxxxxxxx>
>
> This patch introduces an API that allows to return device properties (OF
> or ACPI) of a device bound to the vfio-platform/vfio-amba driver and the
> skeleton of the implementation for VFIO_PLATFORM. Information about any
> device node bound by VFIO_PLATFORM should be queried via the introduced
> ioctl VFIO_DEVICE_GET_DEV_PROPERTY.
>
> The user needs to know the name and the data type of the property he is
> accessing.
>
> Signed-off-by: Antonios Motakis <a.motakis@xxxxxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Baptiste Reynal <b.reynal@xxxxxxxxxxxxxxxxxxxxxx>
>
> ---
> v3 -> v4:
>  - added flags placeholder in vfio_dev_properties
>  - ioctl returns -E2BIG if the buffer is too small
>  - details VFIO_DEVICE_GET_DEV_PROPERTY documentation
> ---
>  drivers/vfio/platform/Makefile                |  3 +-
>  drivers/vfio/platform/properties.c            | 77 +++++++++++++++++++++++++++
>  drivers/vfio/platform/vfio_platform_common.c  | 35 ++++++++++++
>  drivers/vfio/platform/vfio_platform_private.h |  7 +++
>  include/uapi/linux/vfio.h                     | 31 +++++++++++
>  5 files changed, 152 insertions(+), 1 deletion(-)
>  create mode 100644 drivers/vfio/platform/properties.c
>
> diff --git a/drivers/vfio/platform/Makefile b/drivers/vfio/platform/Makefile
> index 9ce8afe..37cf5ed 100644
> --- a/drivers/vfio/platform/Makefile
> +++ b/drivers/vfio/platform/Makefile
> @@ -1,5 +1,6 @@
>
> -vfio-platform-y := vfio_platform.o vfio_platform_common.o vfio_platform_irq.o
> +vfio-platform-y := vfio_platform.o vfio_platform_common.o vfio_platform_irq.o \
> +                properties.o
>
>  obj-$(CONFIG_VFIO_PLATFORM) += vfio-platform.o
>  obj-$(CONFIG_VFIO_PLATFORM) += reset/
> diff --git a/drivers/vfio/platform/properties.c b/drivers/vfio/platform/properties.c
> new file mode 100644
> index 0000000..98754c2
> --- /dev/null
> +++ b/drivers/vfio/platform/properties.c
> @@ -0,0 +1,77 @@
> +/*
> + * Copyright (C) 2015 - Virtual Open Systems
> + * Authors: Antonios Motakis <a.motakis@xxxxxxxxxxxxxxxxxxxxxx>
> + *          Baptiste Reynal <b.reynal@xxxxxxxxxxxxxxxxxxxxxx>
> + *
> + * This program is free software; you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License, version 2, as
> + * published by the Free Software Foundation.
> + *
> + * This program is distributed in the hope that it will be useful,
> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
> + * GNU General Public License for more details.
> + */
> +
> +#include <linux/slab.h>
> +#include <linux/vfio.h>
> +#include <linux/property.h>
> +#include "vfio_platform_private.h"
> +
> +static int dev_property_get_strings(struct device *dev, uint32_t *flags,
> +                                 char *name, unsigned *lenp,
> +                                 void __user *datap, unsigned long datasz)
> +{
> +     return -EINVAL;
> +}
> +
> +static int dev_property_get_uint(struct device *dev, uint32_t *flags,
> +                              char *name, uint32_t type, unsigned *lenp,
> +                              void __user *datap, unsigned long datasz)
> +{
> +     return -EINVAL;
> +}
> +
> +int vfio_platform_dev_properties(struct device *dev, uint32_t *flags,
> +                              uint32_t type, unsigned *lenp,
> +                              void __user *datap, unsigned long datasz)
> +{
> +     char *name;
> +     long namesz;
> +     int ret;
> +
> +     namesz = strnlen_user(datap, datasz);
> +     if (!namesz)
> +             return -EFAULT;
> +     if (namesz > datasz)
> +             return -EINVAL;
> +
> +     name = kzalloc(namesz, GFP_KERNEL);

What prevents the user from passing an arbitrarily large string here?

Nothing, should I put an arbitrary limit ?
 

> +     if (!name)
> +             return -ENOMEM;
> +     if (strncpy_from_user(name, datap, namesz) <= 0) {
> +             kfree(name);
> +             return -EFAULT;
> +     }
> +
> +     switch (type) {
> +     case VFIO_DEV_PROPERTY_TYPE_STRINGS:
> +             ret = dev_property_get_strings(dev, flags, name, lenp,
> +                             datap, datasz);
> +             break;
> +
> +     case VFIO_DEV_PROPERTY_TYPE_U64:
> +     case VFIO_DEV_PROPERTY_TYPE_U32:
> +     case VFIO_DEV_PROPERTY_TYPE_U16:
> +     case VFIO_DEV_PROPERTY_TYPE_U8:
> +             ret = dev_property_get_uint(dev, flags, name, type, lenp,
> +                                         datap, datasz);
> +             break;
> +
> +     default:
> +             ret = -EINVAL;
> +     }
> +
> +     kfree(name);
> +     return ret;
> +}
> diff --git a/drivers/vfio/platform/vfio_platform_common.c b/drivers/vfio/platform/vfio_platform_common.c
> index e43efb5..44ba22c 100644
> --- a/drivers/vfio/platform/vfio_platform_common.c
> +++ b/drivers/vfio/platform/vfio_platform_common.c
> @@ -20,6 +20,7 @@
>  #include <linux/types.h>
>  #include <linux/uaccess.h>
>  #include <linux/vfio.h>
> +#include <linux/property.h>
>
>  #include "vfio_platform_private.h"
>
> @@ -302,6 +303,34 @@ static long vfio_platform_ioctl(void *device_data,
>                       return vdev->reset(vdev);
>               else
>                       return -EINVAL;
> +     } else if (cmd == VFIO_DEVICE_GET_DEV_PROPERTY) {
> +             struct vfio_dev_property info;
> +             void __user *datap;
> +             unsigned long datasz;
> +             int ret;
> +
> +             if (!vdev->dev)
> +                     return -EINVAL;
> +
> +             minsz = offsetofend(struct vfio_dev_property, length);
> +
> +             if (copy_from_user(&info, (void __user *)arg, minsz))
> +                     return -EFAULT;
> +
> +             if (info.argsz < minsz)
> +                     return -EINVAL;
> +
> +             datap = (void __user *) arg + minsz;
> +             datasz = info.argsz - minsz;
> +
> +             ret = vfio_platform_dev_properties(vdev->dev, &info.flags,

Why the address of flags, I don't see anyone modifying it and there's
nothing defined in the ioctl description for flags being modified on
return.  In fact, there are no flags defined yet, it's just a
future-proofing mechanism of the ioctl, so why propagate it at all right
now?

I tried to use them to retrieve some informations about the property, then forgot to remove the propagation. Will be removed on next release.
 

> +                             info.type, &info.length, datap, datasz);
> +
> +             if (copy_to_user((void __user *)arg, &info, minsz))
> +                     ret = -EFAULT;
> +
> +             return ret;
> +
>       }
>
>       return -ENOTTY;
> @@ -553,6 +582,12 @@ int vfio_platform_probe_common(struct vfio_platform_device *vdev,
>
>       vfio_platform_get_reset(vdev, dev);
>
> +     /* add device properties flag */
> +     if (device_property_present(dev, "name")) {
> +             vdev->dev = dev;

It seems precarious to leave ->dev dangling in the else case.  This is a
future bug waiting to happen.

Ok, wil be fixed on next release.
 

> +             vdev->flags |= VFIO_DEVICE_FLAGS_DEV_PROPERTIES;
> +     }
> +
>       mutex_init(&vdev->igate);
>
>       return 0;
> diff --git a/drivers/vfio/platform/vfio_platform_private.h b/drivers/vfio/platform/vfio_platform_private.h
> index 1c9b3d5..c75b089 100644
> --- a/drivers/vfio/platform/vfio_platform_private.h
> +++ b/drivers/vfio/platform/vfio_platform_private.h
> @@ -56,6 +56,7 @@ struct vfio_platform_device {
>       u32                             num_irqs;
>       int                             refcnt;
>       struct mutex                    igate;
> +     struct device                   *dev;
>
>       /*
>        * These fields should be filled by the bus specific binder
> @@ -89,4 +90,10 @@ extern int vfio_platform_set_irqs_ioctl(struct vfio_platform_device *vdev,
>                                       unsigned start, unsigned count,
>                                       void *data);
>
> +/* device properties support in properties.c */
> +extern int vfio_platform_dev_properties(struct device *dev, uint32_t *flags,
> +                                     uint32_t type, unsigned *lenp,
> +                                     void __user *datap,
> +                                     unsigned long datasz);
> +
>  #endif /* VFIO_PLATFORM_PRIVATE_H */
> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
> index 9fd7b5d..a1a0eaa 100644
> --- a/include/uapi/linux/vfio.h
> +++ b/include/uapi/linux/vfio.h
> @@ -164,12 +164,43 @@ struct vfio_device_info {
>  #define VFIO_DEVICE_FLAGS_PCI        (1 << 1)        /* vfio-pci device */
>  #define VFIO_DEVICE_FLAGS_PLATFORM (1 << 2)  /* vfio-platform device */
>  #define VFIO_DEVICE_FLAGS_AMBA  (1 << 3)     /* vfio-amba device */
> +#define VFIO_DEVICE_FLAGS_DEV_PROPERTIES (1 << 4) /* Device properties */
>       __u32   num_regions;    /* Max region index + 1 */
>       __u32   num_irqs;       /* Max IRQ index + 1 */
>  };
>  #define VFIO_DEVICE_GET_INFO         _IO(VFIO_TYPE, VFIO_BASE + 7)
>
>  /**
> + * VFIO_DEVICE_GET_DEV_PROPERTY - _IOR(VFIO_TYPE, VFIO_BASE + 17,
> + *                                           struct vfio_devtree_info)
> + *
> + * Retrieve a device property, e.g. from a HW description (device tree or ACPI)
> + * if available.
> + * Caller will initialize data[] with a single string with the requested
> + * devicetree property name, and type depending on whether an array of strings
> + * or an array of u32 values is expected. On success, data[] will be filled
> + * with the requested information, either as an array of u32, or with a list
> + * of strings separated by the NULL terminating character.
> + * Return: 0 on success, -errno on failure.
> + * Errors:
> + *   E2BIG: the property is too big to fit in the data[] buffer (the required
> + *          size is then written into the length field).

I don't know which is more correct, but the
VFIO_DEVICE_GET_PCI_HOT_RESET_INFO uses ENOSPC when the buffer size is
too small.  If there's no compelling reason to be different, let's use
it here too.

I found E2BIG in others ioctls (KVM_GET_MSR_INDEX_LIST), but let's use ENOSPC.
 

> + */
> +struct vfio_dev_property {
> +     __u32   argsz;
> +     __u32   flags;  /* Placeholder for future extension */
> +     __u32   type;
> +#define VFIO_DEV_PROPERTY_TYPE_STRINGS       0
> +#define VFIO_DEV_PROPERTY_TYPE_U8    1
> +#define VFIO_DEV_PROPERTY_TYPE_U16   2
> +#define VFIO_DEV_PROPERTY_TYPE_U32   3
> +#define VFIO_DEV_PROPERTY_TYPE_U64   4
> +     __u32   length; /* Size of data[] */
> +     __u8    data[];
> +};
> +#define VFIO_DEVICE_GET_DEV_PROPERTY _IO(VFIO_TYPE, VFIO_BASE + 17)
> +
> +/**
>   * VFIO_DEVICE_GET_REGION_INFO - _IOWR(VFIO_TYPE, VFIO_BASE + 8,
>   *                                  struct vfio_region_info)
>   *




_______________________________________________
kvmarm mailing list
kvmarm@xxxxxxxxxxxxxxxxxxxxx
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm

[Index of Archives]     [Linux KVM]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux