12:58 <Wizzup> freemangordon: when can we meet about mcclient?

12:58 <Wizzup> would this weekend work for you?

13:30 <parazyd> wfm btw

16:12 <freemangordon> Wizzup: rather no, I will have some health related actions the next 2-3 days, in a hospital

16:17 <freemangordon> however, starting today I have 10 days of holiday, so hopefully 3rd or 4th will be ok. But lets see how it will pan out the next few days

16:25 <parazyd> freemangordon: Ok. Hope all will be well

16:26 <parazyd> freemangordon: Then when you can, just let us know some time that works for you

16:29 <tmlind> yeah hopefully just a minor tune-up

16:30 <tmlind> uvos: no idea about the ioctl not working with current kernels, i only had to do it for kexecboot

16:31 <tmlind> uvos: maybe send an email to sre asking about what is needed for mainline kernel to update the lcd, i thought for console it is done automatically already

16:31 <tmlind> uvos: i mean i though framebuffer is updated automatically already

16:54 <freemangordon> parazyd: ok

17:17 <Wizzup> freemangordon: check

18:03 <uvos> freemangordon: i wish you good health & we eagerly await your return when you are ready :)

19:24 <uvos> parazyd, tmlind: i figured it out. so omapdrmfb dosent implment thos ioctls or any others that let me refesh the display as far as i can tell. the reason wht some fb applications work ist that every wirte() to /dev/fb0 refreshes the display. Some applications however mmap() the framebuffer and then use memcpy as this is faster. ofc here the kernel cant know that the buffer has been updated so it dosent refresh.

19:25 <uvos> so the mmap() is broken on d4/bionic

19:25 <uvos> you can wirte 1 byte to the fb to make it refresh

19:25 <uvos> but this is unsatisfying

19:25 <uvos> ill report to sre

19:26 <uvos> http://uvos.xyz/maserati/testfb.c

19:26 <uvos> ^^^^ test app that can use both methods and shows the above behavior

19:27 <parazyd> Thanks

19:28 <parazyd> Yeah that's why I suggested mmap in the first place as I thought it'd be better for performance

19:29 <parazyd> So this is something that can be fixed in the kerne?

19:29 <parazyd> *kernel

19:30 <uvos> kernel needs to support an ioctl like OMAP_FB_SYNC.

19:30 <uvos> in the drm - fbdev emulation layer

19:31 <parazyd> tmlind: ^

19:31 <parazyd> ok, makes sense

19:31 <uvos> in the mean time i can patch directfb to wirte one byte after memcpy

19:31 <parazyd> Can you do it in sdl or charge-mode app?

19:32 <uvos> sure

19:32 <uvos> yeah

19:32 <parazyd> Please do so, so I don't have to fork yet another package :)

19:32 <uvos> it will break on pixel (because i dont know what to write)

19:32 <uvos> but thats ok

19:32 <uvos> *one pixel

19:44 <uvos> also i hate the the guy who wrote driectfb with a pasion

19:44 <uvos> he used 5 space indentation and 80 col max

19:51 <uvos> yeah write() in charging_sdl makes it work fine

20:43 <duskull> Hello!

20:43 <uvos> hi duskull

20:45 <duskull> I'm going to try out leste on my n900 these days, any tips after installing it?

20:47 <uvos> not really following the instructions here should be enough: https://leste.maemo.org/Nokia_N900

20:47 <Entitlement> uvos - [ Nokia N900 - Maemo Leste Wiki ]

20:49 <uvos> parazyd: https://github.com/IMbackK/charge-mode

20:49 <Entitlement> uvos - [ GitHub - IMbackK/charge-mode ]

20:50 <uvos> parazyd: uses driectfb now, dosent slow down the boot process mutch at all anymore

20:50 <uvos> parazyd: i say ship it

20:52 <duskull> @uvos Thanks!