02:34 <lvrp16> the SPICC DMA on GXL seems to be buggy for larger chunk sizes (> 1024 bytes broken down into two dma bursts), it keeps adding extra 1's in the data after one burst count after the 144th 64-bit word.

09:13 <narmstrong> yeah spicc is total garbage

18:03 <lvrp16> narmstrong: I got all the PIO modes from 1 bit to 64 bit working. I'm almost there with DMA for 8 to 64 bpw. However, I am not familiar with DMA mapping and there's some synchronization missing. About half of the transfers work when in DMA mode and half not working.

18:03 <lvrp16> I'm padding the data to 64-bit for DMA: https://github.com/libre-computer-project/libretech-linux/commit/aff2e443ab8da7f3532f2c617accf3d91ff4c218#diff-feaa520439b458b6c5ff6bf593c0ee324859210f365798a85d376c026313f003R237

18:03 <lvrp16> This also causes 64-bit DMA to fail as well so it has to be something with the DMA mapping.

18:04 <lvrp16> If you have time can you take a look?

18:05 <lvrp16> https://paste.debian.net/1373562/

18:12 <lvrp16> The transmit works fine on the scope. It's the RX portion that returns all 0 sometimes.

18:28 <f_> [ 30.079926@2]- btHAL->++++++sdio bt driver insmod start.++++++

18:28 <f_> wut

18:28 <f_> I thought W1 used uart

18:29 <f_> [ 1.716280@3]- Bluetooth: HCI UART driver ver 2.3

18:29 <f_> [ 1.716287@3]- Bluetooth: HCI UART protocol H4 registered

18:29 <f_> O.O

18:30 <f_> Ah it seems it's just probing and failing so I guess it's connected to UART

18:30 <f_> Needs more experimentation :P

18:31 <f_> (that's on the tvstick, reinstalled android there)

18:35 <lvrp16> lord help me the day I have to touch SDIO or BT.

18:49 <f_> :D

18:51 <lvrp16> *sigh* https://paste.debian.net/1373575/

18:59 <lvrp16> ahh, I'm shutting down the DMA too early before it empties the RX fifo.

19:00 <lvrp16> that's dumb...

19:01 <narmstrong> f_: it uses UART but you need to do the init via sdio

19:01 <narmstrong> W1 is a wild beast

19:01 <f_> narmstrong: great

19:02 <f_> Last time I checked I saw some W1 support for BT I think?

19:02 <f_> does it actually work? :P

19:06 <f_> Guess we'll find out!

19:06 <f_> after my kernel finishes compiling

19:07 <f_> What's the difference between W155S1 and W155S2?

19:07 <f_> (ahh, if only amlogic employees actually were active on irc)

23:19 <lvrp16> I have it on one of the pamphlets if I could find it.

23:20 <lvrp16> if memory serves me correctly, wifi 5 vs wifi 6.