On Fri, 24 Apr 2009 07:49:10 -0700, Dan Williams <dan.j.williams@xxxxxxxxx> wrote: > > + /* > > + * We use dma_unmap_page() regardless of how the buffers were > > + * mapped before they were submitted... > > + */ > > This will be caught by the new dma-mapping debug infrastructure that > went in for 2.6.30, but I would not hold up merging this driver for > this issue. Once Maciej's fix [1] goes upstream you can fix up your > driver to use the new "unmap single" flags. OK, I will follow that when the flag is merged. Thanks. --- Atsushi Nemoto
