On Tue, Mar 19, 2019 at 04:26:54PM -0400, Vivek Goyal wrote:
> On Sun, Mar 17, 2019 at 08:35:21AM +0800, Liu Bo wrote:
> > While doing memremap from pci_dev's system bus address to kernel virtual
> > address, we assign a wrong value to the %end of pgmap.res, which ends up
> > with a wrong resource size in the process of memremap, and that further
> > prevent the second virtiofs pci device from being probed successfully.
> >
> > Signed-off-by: Liu Bo <bo.liu@xxxxxxxxxxxxxxxxx>
>
> Hi Liu Bo,
>
> Thanks for the fix. This seems right. I will fix it in my internal
> branch. These patches are not upstream yet. Will fold this into
> existing patch for the next posting.
>
Feel free to fold it, and I'm looking forward to the 1.0 release.
thanks,
-liubo
> Thanks
> Vivek
>
> > ---
> > fs/fuse/virtio_fs.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c
> > index 88b00055589b..7abf2187d85f 100644
> > --- a/fs/fuse/virtio_fs.c
> > +++ b/fs/fuse/virtio_fs.c
> > @@ -713,7 +713,7 @@ static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs)
> > pgmap->res = (struct resource){
> > .name = "virtio-fs dax window",
> > .start = phys_addr,
> > - .end = phys_addr + cache_len,
> > + .end = phys_addr + cache_len - 1,
> > };
> >
> > fs->window_kaddr = devm_memremap_pages(&pci_dev->dev, pgmap);
> > --
> > 2.20.1.2.gb21ebb6
> >
