| 09:18 | mk01 | jnettlet: please once you come back can you check what lib ver you have got ? I synced the repos, but got again p4.5 only (dated May, but cksum is the same) |
| 09:49 | jnettlet | mk01, p4.5 but it should be dated June with a different md5 sum. I downloaded the version from fsl to verify the April and June packages are different |
| 09:51 | mk01 | 8314408acb6b3bc58fcbbb8a0f48b54b imx-gpu-viv-5.0.11.p4.5-hfp.bin |
| 09:52 | jnettlet | I am still out and cant check right now. |
| 09:52 | mk01 | are you on default (regarding repo version or next?) |
| 09:53 | jnettlet | Neither right now. I do development on debian or fedora. Yocto is too slow for that. |
| 09:53 | jnettlet | I will start the integration testing next. |
| 10:09 | mk01 | btw: I implemented refclock (for xbmc's sync to display layer) and fractional modes (1000/1001) |
| 10:09 | mk01 | maybe would be interesting for rabeeh ? |
| 13:19 | jnettlet | mk01, sure those would be great. We had the frational modes, but it throws audio timing off about half an hour into a movie |
| 13:34 | mk01 | I have seen rabeehs patch - basically the same, only he forgot to set new aviinfo frames for frame composer and stuff. |
| 13:35 | mk01 | do you care in any way of mxc_dcic? |
| 13:35 | mk01 | (driver)? |
| 13:36 | mk01 | I used it for the refclock - by adding a variant of it's run when it is not stopping irqs and didn't wait for client to request results. |
| 13:37 | mk01 | this then provides vsync signal |
| 13:37 | jnettlet | sure I will take a look. |
| 13:38 | mk01 | currently I directly hacked the driver, but If you want very nice commit, I could create separate "driver" which would use like 25% of the original code |
| 13:39 | mk01 | ok, so you don't mind direct edit of it |
| 13:39 | jnettlet | I don't think so. just post what you have for now and I will take a look. We figure out what works best. |
| 14:00 | mk01 | refclock is PR at your new repo |
| 14:03 | jnettlet | Okay great. I will take a look after lunch |
| 14:07 | mk01 | with fract modes there are two parts, the code itself and then the question how to get those modes into xbmc. And you will not like much how I did it - I updated fbsysfs.c to handle the fract modes as extra modes. the mode has same Refresh as it's original mode, but I;m changing the S/D/V character to lowercase |
| 14:09 | mk01 | then XBMC just checks upper/lower first character and in case of lower, it recalculates Refresh * (1000/1001) |