whitequark changed the topic of #prjbureau to: Microchip (Atmel) ATF15xx CPLD reverse engineering · code https://github.com/whitequark/prjbureau · docs https://whitequark.github.io/prjbureau/ · logs https://libera.irclog.whitequark.org/prjbureau
astrobunny has joined #prjbureau
<astrobunny> hi hi
<astrobunny> what do you know, i get sleep and then i figure it out
<astrobunny> turns out i should look at the examples more
<astrobunny> using a plain assign statement causes yosys to generate some (rename id00001 "...") thing
<astrobunny> and the fitter just refuses to work with that
<whitequark> yes
<astrobunny> so im trying to walk through the steps to say make assign PIN43 = PIN42; by hand
<astrobunny> and i see that i blow the fuses M1_PAD UIM26, UIM26_P-40 on MC2, MC2 pt1_mux, MC2 xor_a_mux sum and dfast_mux
<astrobunny> but i dont get how the connection is made to drive o_mux/dfast_mux from the xor gate
<astrobunny> is it the d_mux fuse that muxes the xor to the dff and output?
<astrobunny> im so confuse
<astrobunny> i want a atf1502.v
<whitequark> haha
<whitequark> i tried to make that
<whitequark> it's hard
<whitequark> ask me again in 8 hours
<astrobunny> okey
pbsds has joined #prjbureau
<astrobunny> mmmm this good i guess the fitter has like some default settings for shorting the input to the output
<astrobunny> and i found it allocates output pins first MC32 backwards
<astrobunny> looks like it also allocates the non-mc pads for inputs first if there are no clocks or other things
<astrobunny> oh wow
<astrobunny> each PT is not just some inputs
<astrobunny> its 96 inputs
<whitequark> yeah
<astrobunny> wow ok i can see how this is flexible now
<astrobunny> i thought i was limited to jsut 5 and gates per macrocell
<astrobunny> yeah 5 and gates with up to 96 inputs each
<astrobunny> these chips are cool
<astrobunny> limited, but i can run 5v through them
<astrobunny> do you know if the arming switch should be turned off?
<astrobunny> i kept reading through the docs but it keeps dodging the question of if i could or could not prevent re-programming
<astrobunny> or maybe i keep getting to that section in sleepy mode
<whitequark> the arming switch is what makes the CPLD not toggle its outputs during programming
<whitequark> since the flash cells are directly connected to the fabric
<whitequark> there's no way to prevent re-programming iirc
<astrobunny> ah ok awesome
<whitequark> well you can turn off JTAG but that just needs 12 V
<astrobunny> Oh
<astrobunny> so its the same as the GAL chips
<astrobunny> you 12v them to reprogram
<astrobunny> oh my bad i missed that
<astrobunny> ok thats a good thing
<astrobunny> then i dont accidentally brick it
<astrobunny> i mean i can brick it in so many other ways but not this one
_whitelogger has joined #prjbureau