On Tue, Sep 6, 2016 at 9:49 AM, Dan Williams <[email protected]> wrote: > In pgoff_to_phys() 'pgoff' is already relative to base of the dax > device, so we only need to compare if the current offset is within the > current resource extent. Otherwise, we are double accounting the > resource start offset when translating pgoff to a physical address. > > Cc: <[email protected]> > Signed-off-by: Dan Williams <[email protected]>
On second look this results in the exact same translation, correct in both cases. This is also confirmed by a new ndctl unit test that does data verification by writing through a /dev/pmem device and the verifying via a /dev/dax device associated with the same namespace, so I'm dropping this patch.
