Re: [qemu-s390x] [PATCH v9 5/6] s390x/vfio: ap: Introduce VFIO AP device

[Tony Krowiak]

On 09/27/2018 09:56 AM, Thomas Huth wrote:
> On 2018-09-27 00:54, Tony Krowiak wrote:
> > Introduces a VFIO based AP device. The device is defined via
> > the QEMU command line by specifying:
> >
> >      -device vfio-ap,sysfsdev=<path-to-mediated-matrix-device>
> >
> > There may be only one vfio-ap device configured for a guest.
> >
> > The mediated matrix device is created by the VFIO AP device
> > driver by writing a UUID to a sysfs attribute file (see
> > docs/vfio-ap.txt). The mediated matrix device will be named
> > after the UUID. Symbolic links to the $uuid are created in
> > many places, so the path to the mediated matrix device $uuid
> > can be specified in any of the following ways:
> >
> > /sys/devices/vfio_ap/matrix/$uuid
> > /sys/devices/vfio_ap/matrix/mdev_supported_types/vfio_ap-passthrough/devices/$uuid
> > /sys/bus/mdev/devices/$uuid
> > /sys/bus/mdev/drivers/vfio_mdev/$uuid
> >
> > When the vfio-ap device is realized, it acquires and opens the
> > VFIO iommu group to which the mediated matrix device is
> > bound. This causes a VFIO group notification event to be
> > signaled. The vfio_ap device driver's group notification
> > handler will get called at which time the device driver
> > will configure the the AP devices to which the guest will
> > be granted access.
> >
> > Signed-off-by: Tony Krowiak <address@hidden>
> > ---
> [...]
> > diff --git a/hw/vfio/ap.c b/hw/vfio/ap.c
> > new file mode 100644
> > index 000000000000..429988f23f98
> > --- /dev/null
> > +++ b/hw/vfio/ap.c
> > @@ -0,0 +1,181 @@
> > +/*
> > + * VFIO based AP matrix device assignment
> > + *
> > + * Copyright 2018 IBM Corp.
> > + * Author(s): Tony Krowiak <address@hidden>
> > + *            Halil Pasic <address@hidden>
> > + *
> > + * This work is licensed under the terms of the GNU GPL, version 2 or (at
> > + * your option) any later version. See the COPYING file in the top-level
> > + * directory.
> > + */
> > +
> > +#include <linux/vfio.h>
> > +#include <sys/ioctl.h>
> > +#include "qemu/osdep.h"
> > +#include "qapi/error.h"
> > +#include "hw/sysbus.h"
> > +#include "hw/vfio/vfio.h"
> > +#include "hw/vfio/vfio-common.h"
> > +#include "hw/s390x/ap-device.h"
> > +#include "qemu/error-report.h"
> > +#include "qemu/queue.h"
> > +#include "qemu/option.h"
> > +#include "qemu/config-file.h"
> > +#include "cpu.h"
> > +#include "kvm_s390x.h"
> > +#include "sysemu/sysemu.h"
> > +#include "hw/s390x/ap-bridge.h"
> > +#include "exec/address-spaces.h"
> > +
> > +#define VFIO_AP_DEVICE_TYPE      "vfio-ap"
> > +
> > +typedef struct VFIOAPDevice {
> > +    APDevice apdev;
> > +    VFIODevice vdev;
> > +} VFIOAPDevice;
> > +
> > +static void vfio_ap_compute_needs_reset(VFIODevice *vdev)
> > +{
> > +    vdev->needs_reset = false;
> > +}
> > +
> > +/*
> > + * We don't need vfio_hot_reset_multi and vfio_eoi operations for
> > + * vfio-ap device now.
> > + */
> > +struct VFIODeviceOps vfio_ap_ops = {
> > +    .vfio_compute_needs_reset = vfio_ap_compute_needs_reset,
> > +};
> > +
> > +static void vfio_ap_put_device(VFIOAPDevice *vapdev)
> > +{
> > +    g_free(vapdev->vdev.name);
> > +    vfio_put_base_device(&vapdev->vdev);
> > +}
> > +
> > +static VFIOGroup *vfio_ap_get_group(VFIOAPDevice *vapdev, Error **errp)
> > +{
> > +    char *tmp, group_path[PATH_MAX];
> > +    ssize_t len;
> > +    int groupid;
> > +
> > +    tmp = g_strdup_printf("%s/iommu_group", vapdev->vdev.sysfsdev);
> > +    len = readlink(tmp, group_path, sizeof(group_path));
> > +    g_free(tmp);
> > +
> > +    if (len <= 0 || len >= sizeof(group_path)) {
> > +        error_setg(errp, "%s: no iommu_group found for %s",
> > +                   VFIO_AP_DEVICE_TYPE, vapdev->vdev.sysfsdev);
> > +        return NULL;
> > +    }
> > +
> > +    group_path[len] = 0;
>
> You could maybe use g_file_read_link() instead to avoid the ugliness
> that is needed around readlink().

I will make that change.

> > +    if (sscanf(basename(group_path), "%d", &groupid) != 1) {
> > +        error_setg(errp, "vfio: failed to read %s", group_path);
> > +        return NULL;
> > +    }
> > +
> > +    return vfio_get_group(groupid, &address_space_memory, errp);
> > +}
> > +
> > +static void vfio_ap_realize(DeviceState *dev, Error **errp)
> > +{
> > +    int ret;
> > +    char *mdevid;
> > +    Error *local_err = NULL;
> > +    VFIOGroup *vfio_group;
> > +    APDevice *apdev = DO_UPCAST(APDevice, parent_obj, dev);
>
> IIRC DO_UPCAST should be avoided in new code. So this is now here the
> right place to finally use the AP_DEVICE() macro?

Will do.

> > +    VFIOAPDevice *vapdev = DO_UPCAST(VFIOAPDevice, apdev, apdev);
> > +
> > +    vapdev = DO_UPCAST(VFIOAPDevice, apdev, apdev);
>
> Double assignment to vapdev.

Ah, yes ... it needs to go

> > +    vfio_group = vfio_ap_get_group(vapdev, &local_err);
> > +    if (!vfio_group) {
> > +        goto out_err;
> > +    }
> > +
> > +    vapdev->vdev.ops = &vfio_ap_ops;
> > +    vapdev->vdev.type = VFIO_DEVICE_TYPE_AP;
> > +    mdevid = basename(vapdev->vdev.sysfsdev);
> > +    vapdev->vdev.name = g_strdup_printf("%s", mdevid);
>
> g_strdup instead of g_strdup_printf should be sufficient here, shouldn't it?

Yes, it is sufficient, I'll change it.

> > +    vapdev->vdev.dev = dev;
> > +
> > +    ret = vfio_get_device(vfio_group, mdevid, &vapdev->vdev, &local_err);
> > +    if (ret) {
> > +        goto out_get_dev_err;
> > +    }
> > +
> > +    /* Enable hardware to intepret AP instructions executed on the guest */
> > +    object_property_set_bool(OBJECT(qdev_get_machine()), true, "apie", NULL);
> > +
> > +    return;
> > +
> > +out_get_dev_err:
> > +    vfio_ap_put_device(vapdev);
> > +    vfio_put_group(vfio_group);
> > +out_err:
> > +    error_propagate(errp, local_err);
> > +}
>
>   Thomas