Re: [PATCH 1/6] virtio_console: don't tie bufs to a vq

From: Michael S. Tsirkin
Date: Tue Apr 24 2018 - 14:56:42 EST


On Sat, Apr 21, 2018 at 09:30:05AM +0200, Greg Kroah-Hartman wrote:
> On Fri, Apr 20, 2018 at 09:18:01PM +0300, Michael S. Tsirkin wrote:
> > an allocated buffer doesn't need to be tied to a vq -
> > only vq->vdev is ever used. Pass the function the
> > just what it needs - the vdev.
> >
> > Signed-off-by: Michael S. Tsirkin <mst@xxxxxxxxxx>
> > ---
> > drivers/char/virtio_console.c | 14 +++++++-------
> > 1 file changed, 7 insertions(+), 7 deletions(-)
> >
> > diff --git a/drivers/char/virtio_console.c b/drivers/char/virtio_console.c
> > index 468f061..3e56f32 100644
> > --- a/drivers/char/virtio_console.c
> > +++ b/drivers/char/virtio_console.c
> > @@ -422,7 +422,7 @@ static void reclaim_dma_bufs(void)
> > }
> > }
> >
> > -static struct port_buffer *alloc_buf(struct virtqueue *vq, size_t buf_size,
> > +static struct port_buffer *alloc_buf(struct virtio_device *vdev, size_t buf_size,
> > int pages)
> > {
> > struct port_buffer *buf;
> > @@ -445,16 +445,16 @@ static struct port_buffer *alloc_buf(struct virtqueue *vq, size_t buf_size,
> > return buf;
> > }
> >
> > - if (is_rproc_serial(vq->vdev)) {
> > + if (is_rproc_serial(vdev)) {
> > /*
> > * Allocate DMA memory from ancestor. When a virtio
> > * device is created by remoteproc, the DMA memory is
> > * associated with the grandparent device:
> > * vdev => rproc => platform-dev.
> > */
> > - if (!vq->vdev->dev.parent || !vq->vdev->dev.parent->parent)
> > + if (!vdev->dev.parent || !vdev->dev.parent->parent)
> > goto free_buf;
> > - buf->dev = vq->vdev->dev.parent->parent;
> > + buf->dev = vdev->dev.parent->parent;
> >
> > /* Increase device refcnt to avoid freeing it */
> > get_device(buf->dev);
> > @@ -838,7 +838,7 @@ static ssize_t port_fops_write(struct file *filp, const char __user *ubuf,
> >
> > count = min((size_t)(32 * 1024), count);
> >
> > - buf = alloc_buf(port->out_vq, count, 0);
> > + buf = alloc_buf(port->portdev->vdev, count, 0);
> > if (!buf)
> > return -ENOMEM;
> >
> > @@ -957,7 +957,7 @@ static ssize_t port_fops_splice_write(struct pipe_inode_info *pipe,
> > if (ret < 0)
> > goto error_out;
> >
> > - buf = alloc_buf(port->out_vq, 0, pipe->nrbufs);
> > + buf = alloc_buf(port->portdev->vdev, 0, pipe->nrbufs);
> > if (!buf) {
> > ret = -ENOMEM;
> > goto error_out;
> > @@ -1374,7 +1374,7 @@ static unsigned int fill_queue(struct virtqueue *vq, spinlock_t *lock)
> >
> > nr_added_bufs = 0;
> > do {
> > - buf = alloc_buf(vq, PAGE_SIZE, 0);
> > + buf = alloc_buf(vq->vdev, PAGE_SIZE, 0);
> > if (!buf)
> > break;
> >
> > --
> > MST
>
> <formletter>
>
> This is not the correct way to submit patches for inclusion in the
> stable kernel tree. Please read:
> https://www.kernel.org/doc/html/latest/process/stable-kernel-rules.html
> for how to do this properly.
>
> </formletter>


Thanks!
I have some questions about this one:

Cc: <stable@xxxxxxxxxxxxxxx> # 3.3.x: a1f84a3: sched: Check for idle
Cc: <stable@xxxxxxxxxxxxxxx> # 3.3.x: 1b9508f: sched: Rate-limit newidle
Cc: <stable@xxxxxxxxxxxxxxx> # 3.3.x: fd21073: sched: Fix affinity logic
Cc: <stable@xxxxxxxxxxxxxxx> # 3.3.x
Signed-off-by: Ingo Molnar <mingo@xxxxxxx>

1. what does the kernel version mean? can I omit it?
2. so when I rebase to add the tag, this changes commit IDs for
following tags in the same tree, breaking their tags
in the process. Pretty annoying. Any idea how to do it better?

Thanks!

--
MST