This patchset covers various places where GEM objects are dirtied by means of CPU writes. The first patch covers cases where only one page is actually written; here we can mark just the specific page in the pagecache dirty. This applies to regular (shmfs-backed) objects only. The second patch covers situations where a subrange that is not limited to a single page is modified, or a whole object is filled with data via CPU writes. In either case, the object is now dirty (i.e. backing store is out-of-date w.r.t. current contents) and must be marked so or risk losing its contents if evicted. For the whole-object cases, marking the individual pages at the point of writing would not be a win; instead put_pages() will propagate the object-dirty flag to each page iff the object is ever evicted. _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
