> > This patch introduces 'daxdev_mapping_supported' helper > > which checks if 'MAP_SYNC' is supported with filesystem > > mapping. It also checks if corresponding dax_device is > > synchronous. Virtio pmem device is asynchronous and > > does not not support VM_SYNC. > > > > Suggested-by: Jan Kara <jack@xxxxxxx> > > Signed-off-by: Pankaj Gupta <pagupta@xxxxxxxxxx> > > --- > > include/linux/dax.h | 23 +++++++++++++++++++++++ > > 1 file changed, 23 insertions(+) > > > > diff --git a/include/linux/dax.h b/include/linux/dax.h > > index b896706a5ee9..4a2a60ffec86 100644 > > --- a/include/linux/dax.h > > +++ b/include/linux/dax.h > > @@ -38,6 +38,24 @@ void kill_dax(struct dax_device *dax_dev); > > void dax_write_cache(struct dax_device *dax_dev, bool wc); > > bool dax_write_cache_enabled(struct dax_device *dax_dev); > > bool dax_synchronous(struct dax_device *dax_dev); > > + > > +/* > > + * Callers check if synchronous mapping is enabled for DAX file > > + * and attached dax device is also synchronous. > > + * > > + * dax_synchronous function verifies if dax device is synchronous. > > + * Currently, only virtio pmem device supports asynchronous device > > + * flush. > > + */ > > Thanks for the patch! I'd restructure this comment like: > > /* > * Check if given mapping is supported by the file / underlying device. > */ Sure. > > +static inline bool daxdev_mapping_supported(struct vm_area_struct *vma, > > + struct dax_device *dax_dev) > > +{ > > /* Everyone supports non-sync mappings */ > > + if (!(vma->vm_flags & VM_SYNC)) > > + return true; > /* Sync mappings are supported only for files using DAX */ > > + if (!IS_DAX(file_inode(vma->vm_file))) > > + return false; > /* Underlying device must support persisting through CPU instructions */ > > + return dax_synchronous(dax_dev); > > +} > > #else > > static inline struct dax_device *dax_get_by_host(const char *host) > > { > > @@ -69,6 +87,11 @@ static inline bool dax_synchronous(struct dax_device > > *dax_dev) > > { > > return true; > > } > > +static inline bool daxdev_mapping_supported(struct vm_area_struct *vma, > > + struct dax_device *dax_dev) > > +{ > > + return true; > > This looks wrong. Shouldn't it rather be: > > return !(vma->flags & VM_SYNC); > > ? Right. I will correct this. Thanks, Pankaj > > Honza > -- > Jan Kara <jack@xxxxxxxx> > SUSE Labs, CR >