On Fri, Mar 13, 2015 at 10:18 PM, Shuah Khan <shuahkh@xxxxxxxxxxxxxxx> wrote: > au0828 s_std and s_input are missing queue busy checks. Add > vb2_is_busy() calls to s_std and s_input and return -EBUSY > if found busy. These checks are only needed on devices which support more than a single format (typically for devices which support standards for both 480 and 576 lines). The au0828 only supports 720x480 capture, and thus there are no conditions in which the capture window size can change due to a standard change (since the device only supports NTSC-M). Hans has made clear in the past that it's permitted to toggle inputs when we can be confident that there will be no change in video standards (in fact, devices that are targeted at surveillance prefer this to minimize the time toggling between multiple cameras). A bridge that only supports one video standard and thus input frame size falls into this category. While a patch like this is appropriate for some bridges, it is not needed for au0828. Can you guys kindly please stop trying to cleanup/refactor au0828 and xc5000? First the xc5000 regression that caused arbitrary memory corruption/panics and now this junk with the vbi_buffer_filled (btw, did the submitter even *TRY* that patch to see if it actually worked?). I know patches like this are well intentioned, but poorly tested patches and "cleanup" fixes which are completely untested are doing more harm than good. Devin -- Devin J. Heitmueller - Kernel Labs http://www.kernellabs.com -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
