17:11 <gamiee> Hi borneoa___ , in RISC-V DM during reset-assert, you can override executing C's assert_reset func. Although this function does quite a lot of stuff, and doing it manually in TCL involves kinda a lot of redundant code. (We have been talking abouth this some time ago). As a good alternative, there is reset-assert-post callback, which can be in most occasions enough for doing any post ndmreset stuff. I implemented my reset mechanism in it, instead of reset-

17:11 <gamiee> assert (and doing all the ndmreset and other stuff manually, without any hacks). Although there comes one problem.

17:11 <gamiee> In assert_rest func, there is part (riscv-013.c#L2384) where progbuf cache is cleaned up, although, if I use anything progbuf related (I couldn't avoid it), the cache is filled with stuff, and not cleared after reset (and this ofc causes troubles later).

17:11 <gamiee> My suggestion is to add progbuf cache clean on deassert_reset C function. Does this sounds reasonable? Thanks.

17:38 <borneoa___> gamiee: I'm not working on risc-v, I need to check the code

17:55 <gamiee> borneoa___ : okay thanks!

17:55 <gamiee> Moving it to deassert_reset shouldn't introduce any issues, eventually, we can have it on both places, but that seems redundant for me.