You should remove *everything* *from* line 86 *to the end* of the file crt.S
So remove the blocks : _itlb_miss_handler: [...] _dtlb_miss_handler: [...] _dtlb_fault_handler: [...] _privilege_exception_handler: [...] 2014/1/10 Salman Sheikh <salmanishe...@gmail.com> > I removed the following lines from unittests/crt0.S and also commented out > the lines 53, 84, 85 and 86 from the top level Makefile > > 85a86,95 > > _itlb_miss_handler: > > ori r25, r25, 128 > > rcsr r24, TLBVADDR > > ret > > nop > > nop > > nop > > nop > > nop > > And I get this: > > $ make unittest > make -C unittests check > make[1]: Entering directory > '/cygdrive/c/GMSAG/CPU_TESTING/lm32-master/test/unittests' > lm32-elf-gcc -x assembler -Wa,-I,. -c crt.S -o crt.o > crt.S: Assembler messages: > crt.S:88: Error: unrecognized keyword/register name `rcsr r24,TLBVADDR' > crt.S:98: Error: unrecognized keyword/register name `rcsr r24,TLBVADDR' > Makefile:103: recipe for target 'crt.o' failed > make[1]: *** [crt.o] Error 1 > make[1]: Leaving directory > '/cygdrive/c/GMSAG/CPU_TESTING/lm32-master/test/unittests' > Makefile:52: recipe for target 'unittest' failed > make: *** [unittest] Error 2 > > > > > On Fri, Jan 10, 2014 at 1:02 PM, Yann Sionneau <y...@minet.net> wrote: > >> You didn't say if removing the lines in crt.S and Makefile fix your issue. >> Does it? >> >> Le 2014-01-10 18:41, Salman Sheikh a écrit : >> >>> Now in that misoc/README you referenced, it mentioned installing >>> gcc...I am attempting to do so but I don't find a gcc-core and gcc-g++ >>> but just one large (105MB) gcc file. I tried running configure and it >>> complained about needing these before proceeding, do they need any >>> special target options or just plain vanilla compile.. >>> >>> configure: error: Building GCC requires GMP 4.2+, MPFR 2.4.0+ and MPC >>> 0.8.0+. >>> >>> On Fri, Jan 10, 2014 at 12:37 PM, Yann Sionneau <y...@minet.net> >>> wrote: >>> >>> Le 2014-01-10 18:34, Yann Sionneau a écrit : >>>> >>>> answer inline: >>>> >>>> Le 2014-01-10 18:20, Salman Sheikh a écrit : >>>> Okay, got a new binutils installed and the >>>> >>>> make sim_hello_world spits out "Hello World".. >>>> >>>> but the make sim_pipe1 doesn't seem to do anything on the screen >>>> (no >>>> output but no errors)... >>>> >>>> but the make unittest fails....I suspect (and dread) I might have >>>> to >>>> recompile the entire gcc for lm32...its still using the >>>> lm32-elf-gcc >>>> from Lattice... >>>> >>>> make unittest >>>> make -C unittests check >>>> make[1]: Entering directory >>>> '/cygdrive/c/GMSAG/CPU_TESTING/lm32-master/test/unittests' >>>> lm32-elf-gcc -x assembler -Wa,-I,. -c crt.S -o crt.o >>>> crt.S: Assembler messages: >>>> crt.S:88: Error: unrecognized keyword/register name `rcsr >>>> r24,TLBVADDR' >>>> crt.S:98: Error: unrecognized keyword/register name `rcsr >>>> r24,TLBVADDR' >>>> crt.S:108: Error: unrecognized keyword/register name `rcsr >>>> r24,TLBVADDR' >>>> Makefile:103: recipe for target 'crt.o' failed >>>> make[1]: *** [crt.o] Error 1 >>>> make[1]: Leaving directory >>>> '/cygdrive/c/GMSAG/CPU_TESTING/lm32-master/test/unittests' >>>> Makefile:52: recipe for target 'unittest' failed >>>> make: *** [unittest] Error 2 >>>> >>> >>> Your binutils does not support lm32 MMU, you should at least use >>> binutils 2.23.2 >>> OR you can use binutils from my github repository: >>> https://github.com/fallen/lm32-binutils-mmu [1] but I would strongly >>> >>> recommend using upstream binutils 2.23.2 or superior. >>> >>> OR if you don't want to use the MMU anyway, you can just remove the >>> MMU unit test from the list by removing the lines 53 84 85 86 in >>> https://github.com/m-labs/lm32/blob/master/test/unittests/Makefile >>> [2] >>> >>> >>> That would be the fastest and cleanest solution for you if you don't >>> care about unit testing the MMU :) >>> >>> I forgot to say to remove everything from line 86 in this file: >>> https://github.com/m-labs/lm32/blob/master/test/unittests/crt.S#L86 >>> [3] >>> >>> >>> On Fri, Jan 10, 2014 at 11:06 AM, Sébastien Bourdeauducq >>>> <sebastien.bourdeaud...@lekernel.net> wrote: >>>> >>>> On 01/10/2014 05:08 PM, Yann Sionneau wrote: >>>> issues with your lm32 objcopy not supporting the -O verilog >>>> backend >>>> >>>> Do those issues manifest themselves with a fresh binutils, or the >>>> outdated Lattice binaries? >>>> >>> _______________________________________________ >>> >>> Devel mailing list >>> Devel@lists.milkymist.org >>> https://ssl.serverraum.org/lists/listinfo/devel [4] >>> >>> >>> -- >>> Yann Sionneau >>> >>> -- >>> >>> Even a Smile is charity :) >>> - Prophet Muhammad >>> >>> >>> >>> Links: >>> ------ >>> [1] https://github.com/fallen/lm32-binutils-mmu >>> [2] https://github.com/m-labs/lm32/blob/master/test/unittests/Makefile >>> [3] https://github.com/m-labs/lm32/blob/master/test/unittests/crt.S#L86 >>> [4] https://ssl.serverraum.org/lists/listinfo/devel >>> >> >> -- >> Yann Sionneau >> > > > > -- > Even a Smile is charity :) > - Prophet Muhammad > > > > _______________________________________________ > Devel mailing list > Devel@lists.milkymist.org > https://ssl.serverraum.org/lists/listinfo/devel > > -- Yann Sionneau
_______________________________________________ Devel mailing list Devel@lists.milkymist.org https://ssl.serverraum.org/lists/listinfo/devel
