Re: [PATCH net-next v3 01/13] virtio: add packed ring types and macros

From: Tiwei Bie
Date: Fri Nov 30 2018 - 10:39:31 EST

Next message: Jeffrey Hugo: "Re: [PATCH 1/3] clk: qcom: gcc-msm8998: Mark hmss_dvm and lpass_at critical"
Previous message: Dan Murphy: "Re: [PATCH 1/2] dt-bindings: trivial: add ti,lm3630a binding"
In reply to: Michael S. Tsirkin: "Re: [PATCH net-next v3 01/13] virtio: add packed ring types and macros"
Next in thread: Michael S. Tsirkin: "Re: [PATCH net-next v3 01/13] virtio: add packed ring types and macros"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Fri, Nov 30, 2018 at 08:52:42AM -0500, Michael S. Tsirkin wrote:
> On Fri, Nov 30, 2018 at 02:01:06PM +0100, Maxime Coquelin wrote:
> > On 11/30/18 1:47 PM, Michael S. Tsirkin wrote:
> > > On Fri, Nov 30, 2018 at 05:53:40PM +0800, Tiwei Bie wrote:
> > > > On Fri, Nov 30, 2018 at 04:10:55PM +0800, Jason Wang wrote:
> > > > >
> > > > > On 2018/11/21 äå6:03, Tiwei Bie wrote:
> > > > > > Add types and macros for packed ring.
> > > > > >
> > > > > > Signed-off-by: Tiwei Bie <tiwei.bie@xxxxxxxxx>
> > > > > > ---
> > > > > > include/uapi/linux/virtio_config.h | 3 +++
> > > > > > include/uapi/linux/virtio_ring.h | 52 ++++++++++++++++++++++++++++++++++++++
> > > > > > 2 files changed, 55 insertions(+)
> > > > > >
> > > > > > diff --git a/include/uapi/linux/virtio_config.h b/include/uapi/linux/virtio_config.h
> > > > > > index 449132c76b1c..1196e1c1d4f6 100644
> > > > > > --- a/include/uapi/linux/virtio_config.h
> > > > > > +++ b/include/uapi/linux/virtio_config.h
> > > > > > @@ -75,6 +75,9 @@
> > > > > > */
> > > > > > #define VIRTIO_F_IOMMU_PLATFORM 33
> > > > > > +/* This feature indicates support for the packed virtqueue layout. */
> > > > > > +#define VIRTIO_F_RING_PACKED 34
> > > > > > +
> > > > > > /*
> > > > > > * Does the device support Single Root I/O Virtualization?
> > > > > > */
> > > > > > diff --git a/include/uapi/linux/virtio_ring.h b/include/uapi/linux/virtio_ring.h
> > > > > > index 6d5d5faa989b..2414f8af26b3 100644
> > > > > > --- a/include/uapi/linux/virtio_ring.h
> > > > > > +++ b/include/uapi/linux/virtio_ring.h
> > > > > > @@ -44,6 +44,13 @@
> > > > > > /* This means the buffer contains a list of buffer descriptors. */
> > > > > > #define VRING_DESC_F_INDIRECT 4
> > > > > > +/*
> > > > > > + * Mark a descriptor as available or used in packed ring.
> > > > > > + * Notice: they are defined as shifts instead of shifted values.
> > > > >
> > > > >
> > > > > This looks inconsistent to previous flags, any reason for using shifts?
> > > >
> > > > Yeah, it was suggested to use shifts, as _F_ should be a bit
> > > > number, not a shifted value:
> > > >
> > > > https://patchwork.ozlabs.org/patch/942296/#1989390
> > >
> > > But let's add a _SPLIT_ variant that uses shifts consistently.
> >
> > Maybe we could avoid adding SPLIT and PACKED, but define as follow:
> >
> > #define VRING_DESC_F_INDIRECT_SHIFT 2
> > #define VRING_DESC_F_INDIRECT (1ull << VRING_DESC_F_INDIRECT_SHIFT)
> >
> > #define VRING_DESC_F_AVAIL_SHIFT 7
> > #define VRING_DESC_F_AVAIL (1ull << VRING_DESC_F_AVAIL_SHIFT)
> >
> > I think it would be better for consistency.
>
> I don't think that will help. the problem is that
> most of the existing virtio code consistently uses _F_ as shifts.
> So we just need to do something about these 5 being inconsistent:
>
> include/uapi/linux/virtio_ring.h:#define VRING_DESC_F_NEXT 1
> include/uapi/linux/virtio_ring.h:#define VRING_DESC_F_WRITE 2
> include/uapi/linux/virtio_ring.h:#define VRING_DESC_F_INDIRECT 4
> include/uapi/linux/virtio_ring.h:#define VRING_USED_F_NO_NOTIFY 1
> include/uapi/linux/virtio_ring.h:#define VRING_AVAIL_F_NO_INTERRUPT 1
>
> I do not want all of virtio to become verbose with _SHIFT, ergo
> we need to change the above 5 to have names which are with _F_ and
> have the bit number.

How about something like this:

#define VRING_COMM_DESC_F_NEXT 0
#define VRING_COMM_DESC_F_WRITE 1
#define VRING_COMM_DESC_F_INDIRECT 2

#define VRING_SPLIT_USED_F_NO_NOTIFY 0
#define VRING_SPLIT_AVAIL_F_NO_INTERRUPT 0

or

#define VRING_SPLIT_DESC_F_NEXT 0
#define VRING_SPLIT_DESC_F_WRITE 1
#define VRING_SPLIT_DESC_F_INDIRECT 2

#define VRING_SPLIT_USED_F_NO_NOTIFY 0
#define VRING_SPLIT_AVAIL_F_NO_INTERRUPT 0

#define VRING_PACKED_DESC_F_NEXT 0
#define VRING_PACKED_DESC_F_WRITE 1
#define VRING_PACKED_DESC_F_INDIRECT 2

>
>
>
> > >
> > >
> > > > >
> > > > >
> > > > > > + */
> > > > > > +#define VRING_PACKED_DESC_F_AVAIL 7
> > > > > > +#define VRING_PACKED_DESC_F_USED 15
> > > > > > +
> > > > > > /* The Host uses this in used->flags to advise the Guest: don't kick me when
> > > > > > * you add a buffer. It's unreliable, so it's simply an optimization. Guest
> > > > > > * will still kick if it's out of buffers. */
> > > > > > @@ -53,6 +60,23 @@
> > > > > > * optimization. */
> > > > > > #define VRING_AVAIL_F_NO_INTERRUPT 1
> > > > > > +/* Enable events in packed ring. */
> > > > > > +#define VRING_PACKED_EVENT_FLAG_ENABLE 0x0
> > > > > > +/* Disable events in packed ring. */
> > > > > > +#define VRING_PACKED_EVENT_FLAG_DISABLE 0x1
> > > > > > +/*
> > > > > > + * Enable events for a specific descriptor in packed ring.
> > > > > > + * (as specified by Descriptor Ring Change Event Offset/Wrap Counter).
> > > > > > + * Only valid if VIRTIO_RING_F_EVENT_IDX has been negotiated.
> > > > > > + */
> > > > > > +#define VRING_PACKED_EVENT_FLAG_DESC 0x2
> > > > >
> > > > >
> > > > > Any reason for using _FLAG_ instead of _F_?
> > > >
> > > > Yeah, it was suggested to not use _F_, as these are values,
> > > > should not have _F_:
> > > >
> > > > https://patchwork.ozlabs.org/patch/942296/#1989390
> > > >
> > > > Regards,
> > > > Tiwei
> > > >
> > > > >
> > > > > Thanks
> > > > >
> > > > >
> > > > > > +
> > > > > > +/*
> > > > > > + * Wrap counter bit shift in event suppression structure
> > > > > > + * of packed ring.
> > > > > > + */
> > > > > > +#define VRING_PACKED_EVENT_F_WRAP_CTR 15
> > > > > > +
> > > > > > /* We support indirect buffer descriptors */
> > > > > > #define VIRTIO_RING_F_INDIRECT_DESC 28
> > > > > > @@ -171,4 +195,32 @@ static inline int vring_need_event(__u16 event_idx, __u16 new_idx, __u16 old)
> > > > > > return (__u16)(new_idx - event_idx - 1) < (__u16)(new_idx - old);
> > > > > > }
> > > > > > +struct vring_packed_desc_event {
> > > > > > + /* Descriptor Ring Change Event Offset/Wrap Counter. */
> > > > > > + __le16 off_wrap;
> > > > > > + /* Descriptor Ring Change Event Flags. */
> > > > > > + __le16 flags;
> > > > > > +};
> > > > > > +
> > > > > > +struct vring_packed_desc {
> > > > > > + /* Buffer Address. */
> > > > > > + __le64 addr;
> > > > > > + /* Buffer Length. */
> > > > > > + __le32 len;
> > > > > > + /* Buffer ID. */
> > > > > > + __le16 id;
> > > > > > + /* The flags depending on descriptor type. */
> > > > > > + __le16 flags;
> > > > > > +};
> > > > > > +
> > > > > > +struct vring_packed {
> > > > > > + unsigned int num;
> > > > > > +
> > > > > > + struct vring_packed_desc *desc;
> > > > > > +
> > > > > > + struct vring_packed_desc_event *driver;
> > > > > > +
> > > > > > + struct vring_packed_desc_event *device;
> > > > > > +};
> > > > > > +
> > > > > > #endif /* _UAPI_LINUX_VIRTIO_RING_H */

Next message: Jeffrey Hugo: "Re: [PATCH 1/3] clk: qcom: gcc-msm8998: Mark hmss_dvm and lpass_at critical"
Previous message: Dan Murphy: "Re: [PATCH 1/2] dt-bindings: trivial: add ti,lm3630a binding"
In reply to: Michael S. Tsirkin: "Re: [PATCH net-next v3 01/13] virtio: add packed ring types and macros"
Next in thread: Michael S. Tsirkin: "Re: [PATCH net-next v3 01/13] virtio: add packed ring types and macros"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]