Messages in this thread | | | Date | Fri, 8 Sep 2017 00:13:35 -0700 | From | Christoph Hellwig <> | Subject | Re: [PATCH v6 04/11] swiotlb: Map the buffer if it was unmapped by XPFO |
| |
On Thu, Sep 07, 2017 at 12:44:14PM -0600, Tycho Andersen wrote: > On Thu, Sep 07, 2017 at 11:10:15AM -0700, Christoph Hellwig wrote: > > > - if (PageHighMem(pfn_to_page(pfn))) { > > > + if (PageHighMem(page) || xpfo_page_is_unmapped(page)) { > > > > Please don't sprinkle xpfo details over various bits of code. > > > > Just add a helper with a descriptive name, e.g. > > > > page_is_unmapped() > > > > that also includes the highmem case, as that will easily document > > what this check is doing. > > Will do, thanks. > > Patch 7 has a similar feel to this one, I can add a wrapper around > __clean_dcache_area_pou() if that makes sense?
That one is in low-level ARM code so I'm not that worried. But in general it seems like we should simply have one interface to check if a page has a kernel mapping or not, nad map/unmap it if not instead of distinguishing between highmem and xpfo.
| |