<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style>

   ***: shik has quit IRC (Ping timeout: 245 seconds)
   hverkuil: <u>padovan</u>: thanks for letting me know.
   devinheitmueller: <u>hverkuil</u>: Are you around?
   hverkuil: <u>devinheitmueller</u>: yes, I am.
   devinheitmueller: Any chance you saw my discussion with Daniel on the list regarding the cx88 issue?
   <br> Now that VB2 has an explicit start/stop streaming callback it seems relatively safe to not start the DMA engine during buffer_queue(), but I wanted to get your opinion given you did the original VB2 conversion.
   hverkuil: I saw it, but did not read it in detail.
   <br> Correct, you shouldn't start the DMA in buffer_queue with vb2.
   <br> That's done in start_streaming.
   devinheitmueller: Right.  That’s what I assumed.  Wasn’t sure if you did it intentionally though to address some edge case for cx88.
   <br> If not, then I’ll just yank those lines from both the VBI and video buffer_queue callbacks and submit a patch.
   hverkuil: And if there are requirements for a minimum number of buffers before you can start the DMA, then you should set min_buffers_needed.
   devinheitmueller: Agreed.
   <br> In the case of cx88 it starts DMA on the first buffer, so I’m assuming there is no such requirement.
   hverkuil: which cx88 source does this?
   devinheitmueller: cx88-vbi.c, cx88-video.c.
   <br> https://git.linuxtv.org/media_tree.git/tree/drivers/media/pci/cx88/cx88-vbi.c#n181
   hverkuil: only in vbi.c, video.c doesn't start dma from buffer_queue.
   devinheitmueller: Oh, looks like it only does it in VBI.
   <br> Looks like it was already fixed for video at some point (the version from the original VB2 conversion did it in both).
   hverkuil: in commit 389208e1173e097590856ed24a505551510f78d4
   devinheitmueller: Yeah, I just found that myself.
   <br> Ok, so just need to fix it for VBI and we should be good to go.
   hverkuil: "It took me a whole friggin' day to track this down, and in the end it was just luck that my eye suddenly triggered on that line." :-)
   devinheitmueller: Yeah, I see that now.  If it’s any consolation, it took me more than a day.  :-)
   hverkuil: Hilarious :-) Well, not for you, but still...
   devinheitmueller: Ok, will send a patch to the ML.  Thanks for your input.
   hverkuil: Re the request API:
   <br> I've plowed my way through all the comments and where I agreed I've applied the fix/change.
   <br> There are a few vim2m patches remaining that I need to take a closer look at (the hang described by Paul).
   <br> s/patches/comments/
   <br> I probably won't be doing any work on this tomorrow, but hope to resume on Wednesday. If anyone is interested: the reqv12 branch contains all my changes.
   <br> This will be rebased and cleaned up, of course.
   <br> If I missed any comments, then ping me. I checked all messages starting with rfcv11, so I think I am good.
   <br> After I've rebased/cleaned up reqv12 I'll continue improving the compliance tests.
   <br> I don't have any public API changes planned, so that remains stable.
   ***: mmattice has quit IRC (Read error: Connection reset by peer)