On Thu, 8 Feb 2018 12:42:15 +0100
Auger Eric <eric.auger@xxxxxxxxxx> wrote:
> Hi Alex,
> On 07/02/18 01:26, Alex Williamson wrote:
> > With vfio ioeventfd support, we can program vfio-pci to perform a
> > specified BAR write when an eventfd is triggered. This allows the
> > KVM ioeventfd to be wired directly to vfio-pci, entirely avoiding
> > userspace handling for these events. On the same micro-benchmark
> > where the ioeventfd got us to almost 90% of performance versus
> > disabling the GeForce quirks, this gets us to within 95%.
> >
> > Signed-off-by: Alex Williamson <alex.williamson@xxxxxxxxxx>
> > ---
> > hw/vfio/pci-quirks.c | 42 ++++++++++++++++++++++++++++++++++++------
> > 1 file changed, 36 insertions(+), 6 deletions(-)
> >
> > diff --git a/hw/vfio/pci-quirks.c b/hw/vfio/pci-quirks.c
> > index e739efe601b1..35a4d5197e2d 100644
> > --- a/hw/vfio/pci-quirks.c
> > +++ b/hw/vfio/pci-quirks.c
> > @@ -16,6 +16,7 @@
> > #include "qemu/range.h"
> > #include "qapi/error.h"
> > #include "qapi/visitor.h"
> > +#include <sys/ioctl.h>
> > #include "hw/nvram/fw_cfg.h"
> > #include "pci.h"
> > #include "trace.h"
> > @@ -287,13 +288,27 @@ static VFIOQuirk *vfio_quirk_alloc(int nr_mem)
> > return quirk;
> > }
> >
> > -static void vfio_ioeventfd_exit(VFIOIOEventFD *ioeventfd)
> > +static void vfio_ioeventfd_exit(VFIOPCIDevice *vdev, VFIOIOEventFD *ioeventfd)
> > {
> > + struct vfio_device_ioeventfd vfio_ioeventfd;
> > +
> > QLIST_REMOVE(ioeventfd, next);
> > +
> > memory_region_del_eventfd(ioeventfd->mr, ioeventfd->addr, ioeventfd->size,
> > ioeventfd->match_data, ioeventfd->data,
> > &ioeventfd->e);
> > +
> > qemu_set_fd_handler(event_notifier_get_fd(&ioeventfd->e), NULL, NULL, NULL);
> > +
> > + vfio_ioeventfd.argsz = sizeof(vfio_ioeventfd);
> > + vfio_ioeventfd.flags = ioeventfd->size;
> > + vfio_ioeventfd.data = ioeventfd->data;
> > + vfio_ioeventfd.offset = ioeventfd->region->fd_offset +
> > + ioeventfd->region_addr;
> > + vfio_ioeventfd.fd = -1;
> > +
> > + ioctl(vdev->vbasedev.fd, VFIO_DEVICE_IOEVENTFD, &vfio_ioeventfd);
> > +
> > event_notifier_cleanup(&ioeventfd->e);
> > g_free(ioeventfd);
> > }
> > @@ -315,6 +330,8 @@ static VFIOIOEventFD *vfio_ioeventfd_init(VFIOPCIDevice *vdev,
> > hwaddr region_addr)
> > {
> > VFIOIOEventFD *ioeventfd = g_malloc0(sizeof(*ioeventfd));
> > + struct vfio_device_ioeventfd vfio_ioeventfd;
> > + char vfio_enabled = '+';
> >
> > if (event_notifier_init(&ioeventfd->e, 0)) {
> > g_free(ioeventfd);
> > @@ -329,15 +346,28 @@ static VFIOIOEventFD *vfio_ioeventfd_init(VFIOPCIDevice *vdev,
> > ioeventfd->region = region;
> > ioeventfd->region_addr = region_addr;
> >
> > - qemu_set_fd_handler(event_notifier_get_fd(&ioeventfd->e),
> > - vfio_ioeventfd_handler, NULL, ioeventfd);
> > + vfio_ioeventfd.argsz = sizeof(vfio_ioeventfd);
> > + vfio_ioeventfd.flags = ioeventfd->size;
> > + vfio_ioeventfd.data = ioeventfd->data;
> > + vfio_ioeventfd.offset = ioeventfd->region->fd_offset +
> > + ioeventfd->region_addr;
> > + vfio_ioeventfd.fd = event_notifier_get_fd(&ioeventfd->e);
> > +
> > + if (ioctl(vdev->vbasedev.fd,
> > + VFIO_DEVICE_IOEVENTFD, &vfio_ioeventfd) != 0) {
> > + qemu_set_fd_handler(event_notifier_get_fd(&ioeventfd->e),
> > + vfio_ioeventfd_handler, NULL, ioeventfd);
> > + vfio_enabled = '-';
> > + }
> > +
> > memory_region_add_eventfd(ioeventfd->mr, ioeventfd->addr,
> > ioeventfd->size, ioeventfd->match_data,
> > ioeventfd->data, &ioeventfd->e);
> >
> > info_report("Enabled automatic ioeventfd acceleration for %s region %d, "
> > - "offset 0x%"HWADDR_PRIx", data 0x%"PRIx64", size %u",
> > - vdev->vbasedev.name, region->nr, region_addr, data, size);
> > + "offset 0x%"HWADDR_PRIx", data 0x%"PRIx64", size %u, vfio%c",
> > + vdev->vbasedev.name, region->nr, region_addr, data, size,
> > + vfio_enabled);
> Not sure if this message is really helpful for the end-user to
> understand what happens. Maybe adding a trace event when everything
> happens as it should and an error_report if we failed setting up the
> vfio kernel handler, explaining the sub-optimal performance that can result.
For right now, I think it is useful. Maybe when we get a few kernels
beyond when the vfio support is introduced and we know how different
devices are behaving and what ioeventfds get added, it might make sense
to switch to a trace interface. I don't think we can legitimately
trigger an error_report for a feature which is just an accelerator and
isn't even in upstream kernels yet (though arguably it would be
upstream by the time this gets into QEMU). For now it let's me ask
users to try it and they don't need to learn how to use tracing to send
reports that can be easily verified that both the QEMU and kernel bits
are in place and functional. Thanks,
Alex
