Conversation

Notices

1. minute (mntmn@mastodon.social)'s status on Thursday, 18-Jul-2024 19:59:38 JST minute

   someone who's good at eMMC help me budget this... i wasted many hours yesterday on trying to figure out why eMMC doesn't work reliably in uboot on A311D. i enabled mmc tracing now and it looks like some initial commands work but it starts falling apart at CMD 6. often it does work, and the ROM and linux can always talk to the eMMC without problems. it's just something weird about uboot (driver is meson_gx_mmc.c). gist: https://gist.github.com/mntmn/56eb9981bab7a0336fd1de9b3b99350d

   • minute (mntmn@mastodon.social)'s status on Thursday, 18-Jul-2024 20:48:12 JST minute

     in reply to

     • patrislav #RIPNatenom

     @patrislav yeahl, the kernel does HS200 here too but the uboot driver doesn't seem to support that yet

   • patrislav #RIPNatenom (patrislav@chaos.social)'s status on Thursday, 18-Jul-2024 20:48:16 JST patrislav #RIPNatenom

     in reply to

     @mntmn if linux can talk to the eMMC and u-boot can't, could it also be worth to check the kernel log for in what mode it addresses the chip?

   • patrislav #RIPNatenom (patrislav@chaos.social)'s status on Thursday, 18-Jul-2024 20:48:17 JST patrislav #RIPNatenom

     in reply to

     @mntmn I had a similar issue when bringing up a board with a SD 3.0 interface; I don't remember the exact reason, I think it had to do with the protocol upgrade to 3.0 after the autonegotiation (something about 3.0 was not implemented correctly on the H/W or SoC config, maybe some voltage or enable signal missing)
     eMMC is simpler than SD due to less circuitry in between, so there should be less stuff that can go wrong...

     one debug step could be to disable the faster I/F versions and speeds?

   • minute (mntmn@mastodon.social)'s status on Thursday, 18-Jul-2024 22:06:38 JST minute

     in reply to

     • dok

     @dok hmm, this would make sense

   • dok (dok@rivals.space)'s status on Thursday, 18-Jul-2024 22:06:39 JST dok

     in reply to

     @mntmn maybe this could be caused by pins not being configured properly (pull-up & drive strenght) ?