On Mon, 19 Jan 2015, Ben Hutchings wrote: > On Sun, 2015-01-18 at 22:23 +0100, Guennadi Liakhovetski wrote: > > On Thu, 18 Dec 2014, Ben Hutchings wrote: > > > > > From: William Towle <william.to...@codethink.co.uk> > > > > > > Move the buffer state test in the .buf_cleanup handler into > > > .stop_streaming so that a) the vb2_queue API is not subverted, and > > > b) tracking of active-state buffers via priv->queue_buf[] is handled > > > as early as is possible > > > > Huh... Sorry, patches 1, 2, 3, and 5 of this series look like a strange > > way to get from the present state to the destination. They all have to be > > merged IMHO. > [...] > > Well, I thought that too. Will's submission from last week has that > change: > http://thread.gmane.org/gmane.linux.drivers.video-input-infrastructure/87009
Hm... interesting, why I didn't get those mails in my INBOX, although I do see myself in CC... Only got them from the list, and no, I don't have "avoid copies" enabled. Anyway, yes, that looks better! But I would still consider keeping buffers on the list in .buf_clean(), in which case you can remove it. And walk the list instead of the VB2 internal buffer array, as others have pointed out. Thanks Guennadi -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
