Re: [RFC 2/4] vduse: Add file operation for mmap

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

 



On Wed, Jun 28, 2023 at 4:08 PM Jason Wang <jasowang@xxxxxxxxxx> wrote:
>
> On Wed, Jun 28, 2023 at 2:59 PM Cindy Lu <lulu@xxxxxxxxxx> wrote:
> >
> > From: Your Name <you@xxxxxxxxxxx>
> >
> > Add the operation for mmap, The user space APP will
> > use this function to map the pages to userspace
>
> Please be specific in the log. E.g why and what the main goal for this mmap.
>
> >
> > Signed-off-by: Cindy Lu <lulu@xxxxxxxxxx>
> > ---
> >  drivers/vdpa/vdpa_user/vduse_dev.c | 49 ++++++++++++++++++++++++++++++
> >  1 file changed, 49 insertions(+)
> >
> > diff --git a/drivers/vdpa/vdpa_user/vduse_dev.c b/drivers/vdpa/vdpa_user/vduse_dev.c
> > index f845dc46b1db..1b833bf0ae37 100644
> > --- a/drivers/vdpa/vdpa_user/vduse_dev.c
> > +++ b/drivers/vdpa/vdpa_user/vduse_dev.c
> > @@ -1313,6 +1313,54 @@ static struct vduse_dev *vduse_dev_get_from_minor(int minor)
> >         return dev;
> >  }
> >
> > +
> > +static vm_fault_t vduse_vm_fault(struct vm_fault *vmf)
> > +{
> > +       struct vduse_dev *dev = vmf->vma->vm_file->private_data;
> > +       struct vm_area_struct *vma = vmf->vma;
> > +       u16 index = vma->vm_pgoff;
> > +
> > +       struct vdpa_reconnect_info *info;
> > +       info = &dev->reconnect_info[index];
> > +
> > +       vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
> > +       if (remap_pfn_range(vma, vmf->address & PAGE_MASK, PFN_DOWN(info->addr),
> > +                           PAGE_SIZE, vma->vm_page_prot))
>
> I'm not sure if this can work e.g do we want to use separate pages for
> each virtqueue (I think the answer is yes).
>
yes, this map the separate pages per vq, beads on my test this works
> > +               return VM_FAULT_SIGBUS;
> > +       return VM_FAULT_NOPAGE;
> > +}
> > +
> > +static const struct vm_operations_struct vduse_vm_ops = {
> > +       .fault = vduse_vm_fault,
> > +};
> > +
> > +static int vduse_mmap(struct file *file, struct vm_area_struct *vma)
> > +{
> > +       struct vduse_dev *dev = file->private_data;
> > +       struct vdpa_reconnect_info *info;
> > +       unsigned long index = vma->vm_pgoff;
> > +
> > +       if (vma->vm_end - vma->vm_start != PAGE_SIZE)
> > +               return -EINVAL;
> > +       if ((vma->vm_flags & VM_SHARED) == 0)
> > +               return -EINVAL;
> > +
> > +       if (index > 65535)
> > +               return -EINVAL;
> > +
> > +       info = &dev->reconnect_info[index];
> > +       if (info->addr & (PAGE_SIZE - 1))
> > +               return -EINVAL;
> > +       if (vma->vm_end - vma->vm_start != info->size) {
> > +               return -ENOTSUPP;
> > +       }
>
> How can userspace know the correct size (info->size) here?
>
I had hard code the size in userpace , I will add the new ioctl of get
the map size
Thanks
cindy
> > +
> > +       vm_flags_set(vma, VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_DONTDUMP);
>
> Why do you need VM_IO, VM_PFNMAP and VM_DONTDUMP here?
>
> Thanks
>
> > +       vma->vm_ops = &vduse_vm_ops;
> > +
> > +       return 0;
> > +}
> > +
> >  static int vduse_dev_open(struct inode *inode, struct file *file)
> >  {
> >         int ret;
> > @@ -1345,6 +1393,7 @@ static const struct file_operations vduse_dev_fops = {
> >         .unlocked_ioctl = vduse_dev_ioctl,
> >         .compat_ioctl   = compat_ptr_ioctl,
> >         .llseek         = noop_llseek,
> > +       .mmap           = vduse_mmap,
> >  };
> >
> >  static struct vduse_dev *vduse_dev_create(void)
> > --
> > 2.34.3
> >
>





[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux