| 11:52 | _rmk_ | shesselba: I tested out page flipping support with my drm driver.. it works, though the version I posted it was only half the frame rate |
| 11:53 | shesselba | _rmk_: page flipping means interlaced here? |
| 11:54 | shesselba | have you seen the tda998x sync generation patch? |
| 11:54 | _rmk_ | no, page flipping means being able to switch the displayed framebuffer on a vsync irq - needed for things like DRI |
| 11:55 | _rmk_ | or indeed video if you want to use the graphic framebuffer (which itself could be YUV) for that |
| 12:06 | shesselba | About the sync generation patch, Darren Etheridge reported some issues but wanted to check his (tilcdc) sync generation first |
| 12:07 | shesselba | I have been able to get every interlaced/progressive mode perfectly on different monitors |
| 12:07 | shesselba | and I confirmed that your driver generates sync perfectly.. |
| 12:07 | _rmk_ | great news |
| 12:08 | _rmk_ | if the cubox wasn't soo tiny I'd get my scope etc on it too |
| 12:09 | shesselba | yeah |
| 12:09 | shesselba | d2plug has it right on top. not easy but possible |
| 12:10 | shesselba | so, about the OF support patch for drm_i2c_slave_encoder, I wonder if they take it.. |
| 12:11 | _rmk_ | I've been up to my eyeballs in converting various bits over to the stuff they've talked about in the reviews |
| 12:12 | _rmk_ | dmabuf of course was quite horrid, and needs a custom extension in the kernel galcore |
| 12:12 | shesselba | any impact on performance, because that was what you were most concerned about? |
| 12:13 | _rmk_ | I've not done the video based side of it yet - that's where I think the problem may be |
| 12:14 | _rmk_ | I have converted the video playback side to a drm plane, and that seems to work quite well for the standard Xv protocol (iow, not the vmeta/bmm hack) |
| 12:14 | _rmk_ | it is more ioctls though |
| 12:14 | _rmk_ | ... if you have to create a new framebuffer each time |
| 12:15 | _rmk_ | it would be much better to do it in a similar way to the Intel XvMC stuff - have the gst plugin know about DRM, and prepare the framebuffer there, passing the framebuffer ID to the X server Xv backend |
| 12:16 | _rmk_ | Intel XvMC passes a drm buffer object ID rather than a framebuffer ID, but that's because of their overlay interface wanting a buffer object id |
| 22:14 | dexter93 | Hello people |
| 22:14 | dexter93 | rabeeh: any chance you're available atm? |
| 22:19 | dexter93 | That cubox is stubborn |
| 23:51 | dv__ | dexter93: whats going on? |
| 23:55 | dexter93 | Probably it's dead. Unbrick won't work |
| 23:58 | twitch153 | dexter93, it IS dead. The red LED isn't turning on at all. |