Hi.

I just got a spare Nexus 5 for experimenting, and I'm wondering if anyone tried to run Plasma Mobile or any other proper glibc Linux on it without libhybris and Android blobs. I red recently, that Linux 4.9 added some support for Nexus 5 upstream.

From the blobs used by Android, I found this list: https://developers.google.com/android/d ... hammerhead

Freedreno should address graphics, and supposedly brcmfmac should work for WiFi?

See:
* https://stackoverflow.com/questions/302 ... or-bcm4339
* http://lxr.free-electrons.com/source/dr ... c/common.c

Are there some open drivers for managing LTE modem, camera, touchscreen and the rest of the specific hardware?

Are any Plasma Mobile developers around this forum sometimes? Or is there some IRC channel where I can find them?

Hello,

Main developer of Plasma Mobile here, about upstream kernel support in Nexus 5 (hammerhead) it was added by me in 4.9 release and as of now just UART and basic board bring-up is supported. Which is not useful to run full fledged system.

Thanks

bshah wrote:Main developer of Plasma Mobile here, about upstream kernel support in Nexus 5 (hammerhead) it was added by me in 4.9 release and as of now just UART and basic board bring-up is supported. Which is not useful to run full fledged system.

Thanks for that! So most of the blobs like for the modem, camera and etc. might have no open replacement still? Do you know if anyone is trying to reverse engineer those?

The parts that have been mainlined can be seen on this wiki page: http://elinux.org/Qualcomm_SOC_Mainlining_Project#IP_blocks_on_the_8974
To use the display, a panel driver needs to be written, see the guide here: https://github.com/freedreno/freedreno/wiki/DSI-Panel-Driver-Porting
I don't know if it's possible to use touch input or how this could be achieved.

The hardest part will be the modem, I guess. But maybe parts of the osmocon-project could be reused to write a free modem driver for Qualcom SoCs a well. See https://osmocom.org/projects/quectel-modems/wiki

But I'm not a kernel hacker, so I don't know how hard all this is in reality. If it was easy it would have been done already, I guess