Em 22-06-2010 00:27, Devin Heitmueller escreveu: > Hello, > > Please PULL from > http://www.kernellabs.com/hg/~dheitmueller/v4l-dvb-em28xx-isoc for the > following: > > * Make em28xx s_ctrl not always return error > * Fix case where fields were not at the correct start location. > > This addresses two bugs found in the em28xx driver (one with video > rendering and one with the s_ctrl ioctl always returning an error). > > Thanks to Eugeniy Meshcheryakov for bringing the field rendering issue > to my attention. > Devin, There were one additional patch on your tree: patches/hg_14995_make_em28xx_s_ctrl_not_always_return_error.patch Changeset: 14995 From: Devin Heitmueller <dheitmueller@xxxxxxxxxxxxxx> Commiter: Devin Heitmueller <dheitmueller@xxxxxxxxxxxxxx> Tag: tip Date: Sun Jun 13 16:12:29 2010 -0400 Subject: Make em28xx s_ctrl not always return error The em28xx driver was always returning 1 for calls to s_ctrl, disregarding whatever value was set by the subdev. This resulted in errors being shown in application such as "v4l2-ctl --set-ctrl", even though the command actually succeeded. Priority: normal Signed-off-by: Devin Heitmueller <dheitmueller@xxxxxxxxxxxxxx> --- diff -upNr oldtree/drivers/media/video/em28xx/em28xx-video.c linux/drivers/media/video/em28xx/em28xx-video.c --- oldtree/drivers/media/video/em28xx/em28xx-video.c 2010-06-26 10:02:43.000000000 -0300 +++ linux/drivers/media/video/em28xx/em28xx-video.c 2010-06-26 10:02:39.000000000 -0300 @@ -1462,7 +1462,8 @@ static int vidioc_s_ctrl(struct file *fi /* It isn't an AC97 control. Sends it to the v4l2 dev interface */ if (rc == 1) { - v4l2_device_call_all(&dev->v4l2_dev, 0, core, s_ctrl, ctrl); + rc = v4l2_device_call_until_err(&dev->v4l2_dev, 0, core, + s_ctrl, ctrl); /* * In the case of non-AC97 volume controls, we still need This patch tries to fix a problem, but it is broken: Let's suppose that you have 2 i2c drivers that implements control, one for video (like tvp5150) and another for audio (like msp3400). That's the case of HVR-950, for example. As both drivers implement s_ctrl, the -ENOIOCTLCMD will not be returned by the drivers. Instead, if you try to set an audio CTRL to tvp5150, it will return -EINVAL. The same happens if you try to set a video CTRL to msp3400. If you use v4l2_device_call_until_err(), depending on the order that tvp5150 and msp3400 will be loaded, or video or audio CTRL's will always fail, due to -EINVAL. What we need here is something like v4l2_device_call_until_not_err(), e. g. something that will stop sending CTRL's if the return code is 0. > Devin > -- 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
