Conversation
Apaczer
left a comment
Apaczer
left a comment
There was a problem hiding this comment.
Great! This is much better (no weird scanlines on top-right), also performance differse little between tefix variants:
new-tefix=0 (223fps)
new-tefix=1 (209fps)
new-tefix=2 (204fps)
new-tefix=3 (202fps)
10% gain for 2&3 tefix compering to former code 👍
| lcdc_wr_cmd(0x45); | ||
| lcdc_rd_dat(); | ||
| lcdc_rd_dat(); | ||
| mycpu_clock = readl(iomm.ccm + PLL_CPU_CTRL_REG); |
There was a problem hiding this comment.
considering we remove PLL_CPU_CTRL_REG condition, do you think we should just leave tefix=2 (since it's more performant?) and dump tefix=3. AFAIR we had those two for either overclocking or underclocking CPUfreq.
There was a problem hiding this comment.
As far as I remember, some panels couldn't handle tefix 3 (or 2?) — for those, tefix 2( or 3?) with a reduced clock in the panel was used
There was a problem hiding this comment.
tefix=2 or 3 was running with 96MHz (0x91001303), but now we will use 108MHz (tefix=2) & 114MHz (tefix=3) from what I see. Will this be ok for those inferior panels?
There was a problem hiding this comment.
ah I believe 0x91001303 is 120MHz so much higher than current clocks
There was a problem hiding this comment.
Indeed tcon clock increased but panel clock was lowered
lcdc_wr_cmd(0xc6);
if (tefix == 3)
lcdc_wr_dat(0x04);
else if (tefix == 2)
lcdc_wr_dat(0x05);
2 participants
Improve vsync feature: