<xobs>
_florent_: ah, yes, that is an issue. The idea of them being in the same CSR region is that it wouldn't shift the CSR table at all, and would be a true drop-in replacement.
<xobs>
Complete with PHY.
<_florent_>
xobs: ok sorry, i didn't catch this requirement. https://github.com/enjoy-digital/litex/commit/23175190d859e6d8fd7303cc78c43cf0e3b51326 adds a simplified UARTCrossover module with both UARTs in the same module as you initially did. I have a preference for crossover vs loopback since for me a loopback would mean connecting uart.source to uart.sink.
<xobs>
Alright, if it ends up with the same API I'm happy with the change. I'll need to update wishbone-tool to use the new register name of `_xover` but that's fine.
<xobs>
Oh. That's what they were called before. Nevermind! Looks good to me.
<_florent_>
Yes it should be in fact similar to your first version. I also reduced tx_fifo_depth/rx_fifo_depth on the crossover UART to minimize resource usage. (we already have buffering on the main uart so no need to have another buffering on the crossover UART).
CarlFK has quit [Ping timeout: 245 seconds]
CarlFK has joined #litex
<keesj>
I find the self.comb += self.uart.connect(self.uart_xover) not very elegand
<keesj>
I find the self.comb += self.uart.connect(self.uart_xover) not very elegant
<_florent_>
somlo: yes i tried keeping the same default values
ambro718 has joined #litex
<somlo>
_florent_: I was just wondering about why decimal and not hex -- and screwed up the joke in the process with a typo of my own...
<_florent_>
somlo: i was also wondering what was the best, i could revert to hex
<somlo>
_florent_: weak vote in favor of hex from me, since it's slightly easier to follow given the surrounding context (after getting used to seeing hex everywhere, decimal just looks random) :)
<somlo>
but it can probably wait for the next time you actually have to modify things in that neighborhood, as far as I'm concerned :)
ambro718 has quit [Quit: Konversation terminated!]
<somlo>
_florent_, bunnie[m]: as of LiteX commit #f818755c, this should work with litesdcard: https://pastebin.com/s4unsyrz
<somlo>
I'm getting the same results as before (successful "sdclk 10" and "sdinit", hangs during "sdtest 8"), but the values read from the SDCORE_RESPONSE CSR are the same :)