<kristianpaul>
wpwrak: from the current understanding of softusb core, do you think it is more intented for ie, 'migrate' linux drivers for some usb devices than actually implementing the whole usb stack?
<kristianpaul>
i ask this in the ignorance of the usb internals on Linux as well, so i relly on your comments
<kristianpaul>
relay*
<wolfspraul>
rely
<kristianpaul>
he
<kristianpaul>
Thanks !
azonenberg has joined #milkymist
<wpwrak>
kristianpaul: at the moment, it's all about implementing the stack on the navre. the FPGA core could be used for a more general architecture, though
<wpwrak>
(the FPGA core is very simple and thus quite generic. it's the navre who connects the dots)
<mwalle>
lekernel: the navre io bus is 6 bit wide isnt it?
<mwalle>
for some reason rio8(0x11) doesn't work, mhh
mumptai has joined #milkymist
<lekernel>
well, it should. what happens?
<mwalle>
lekernel: reads some garbage. i extended the switch statement in softusb_sie.v, with another (lower) address, it works
<mwalle>
lekernel: never mind, i'll try to simulate it tomorrow
<lekernel>
aha. maybe you've found a bug.
<jpbonn>
can anyone tell me why I get different values when reading 0x60001074 and 0xe0001074? Doesn't 0x60001074 shadow 0xe0001074? This is using gdb on qemu.
<lekernel>
0xe... is non cacheable (wrt L1) while 0x6... is - but afaik it shouldn't make a difference when using qemu
<lekernel>
though I'm not sure if qemu implement CSRs correctly at 0x6..., since they should always use non cacheable access :)