Build for RISC-V chips? #53
Description
Has anyone gotten this working, or tried, on any of the RISC-V based chips (C3, C6, H2, etc.) yet? It seems like new products will be RISC-V going forward, so something to think about.
I tried compiling as-is for my esp32c3 at first. Got an error that the -mlongcalls flag wasn't recognized, so I removed all the flags and copied all I could find in the IDF's gcc config for the esp32c3:
cc.flags << '-march=rv32imc_zicsr_zifencei'
cc.flags << '-std=gnu17'
cc.flags << ["-ffunction-sections", "-fdata-sections", "-gdwarf-4", "-ggdb"]
cc.flags = cc.flags.flatten.collect { |x| x.gsub('-MP', '') }In mruby_task I'm not even trying to load a file, just doing:
mrb_load_string(mrb, "puts \"mruby works!\"");It does not work. This is the output:
Guru Meditation Error: Core 0 panic'ed (Load access fault). Exception was unhandled.
Stack dump detected
Core 0 register dump:
MEPC : 0x40058eb6 RA : 0x4202820e SP : 0x3fc9ed60 GP : 0x3fc8d200
0x40058eb6: strlen in ROM
0x4202820e: mrb_load_string_cxt at components/mruby_component/mruby/mrbgems/mruby-compiler/core/parse.y:7003
(inlined by) mrb_load_string at components/mruby_component/mruby/mrbgems/mruby-compiler/core/parse.y:7009
TP : 0x3fc5afcc T0 : 0x4005890e T1 : 0x10000000 T2 : 0x000009fd
0x4005890e: memset in ROM
S0/FP : 0x3fc9ef20 S1 : 0x3c0a4a94 A0 : 0x00000000 A1 : 0xffffffff
A2 : 0x00000000 A3 : 0x7f7f7f7f A4 : 0x00000000 A5 : 0x00000004
A6 : 0x60023000 A7 : 0x0000000a S2 : 0x00000000 S3 : 0x00000000
S4 : 0x3fc9ef20 S5 : 0x00000000 S6 : 0x00000000 S7 : 0x00000000
S8 : 0x00000000 S9 : 0x00000000 S10 : 0x00000000 S11 : 0x00000000
T3 : 0x00000000 T4 : 0x0000000f T5 : 0x3fcb567c T6 : 0x0000000d
MSTATUS : 0x00001881 MTVEC : 0x40380001 MCAUSE : 0x00000005 MTVAL : 0x00000000
0x40380001: _vector_table at ??:?
MHARTID : 0x00000000
Backtrace:
0x40058eb6 in ?? ()
#0 0x40058eb6 in ?? ()
#1 0x4202820e in mrb_load_string_cxt (c=0x0, s=0x0, mrb=0x3c0a4a94) at mrbgems/mruby-compiler/core/parse.y:7003
#2 mrb_load_string (mrb=0x3c0a4a94, s=0x0) at mrbgems/mruby-compiler/core/parse.y:7009
#3 0x00000000 in ?? ()
Backtrace stopped: frame did not save the PC
ELF file SHA256: ec3ce0cb7
I think this might have something to do with RISC-V gcc not having an equivalent to Xtensa's -mlongcalls, but I'm out of my depth here. Any suggestions?