search

pulp-platform / pulpino

An open-source microcontroller system based on RISC-V
http://www.pulp-platform.org
Other
890 stars 297 forks source link

problem in running make vcompile #170

Closed thotaraghava closed 6 years ago

thotaraghava commented 6 years ago

Hi @FrancescoConti ,

I ran successfully cmake_configure.riscv.gcc.sh. Then as a next step, I am trying to run make vcompile Here, I am getting below errors. 

Scanning dependencies of target vcompile
    ./vcompile/build_rtl_sim.csh: Command not found
make[3]: *** [CMakeFiles/vcompile] Error 1
make[2]: *** [CMakeFiles/vcompile.dir/all] Error 2
make[1]: *** [CMakeFiles/vcompile.dir/rule] Error 2
make: *** [vcompile] Error 2.

Can you please help me here. I am doing stand alone PULPino simulation.

Thanks, Raghav

thotaraghava commented 6 years ago

As per the README.md file, > make vcompile compiles RTL. Is n't it? After running testcase by below command, where waveform dumps will be generated.

make helloworld.vsim

Thanks

WaleedOsama commented 6 years ago

put the path to folder of modelsim which contains vsim in PATH , then try it again , note that you run this command "make vcompile" , in same directory you ran the script in

thotaraghava commented 6 years ago

Yes, I am running this command in /sw/build/ path only. Could you elaborate, in which file, I need to mention the path of modelsim. Thanks,

FrancescoConti commented 6 years ago

The vsim command must be in your PATH environment variable . To append the ModelSim path to the PATH variable, you can use

export PATH=/path/to/vsim/bin:$PATH

assuming your shell is Bash.

thotaraghava commented 6 years ago

@FrancescoConti , I understand as below. I should run below command in linux terminal.

export PATH=/path/to/vsim/bin:$PATH Can you guide me, how to find /path/to/vsim/ here.

Thanks

thotaraghava commented 6 years ago

Hi @FrancescoConti ,

  We do not use Modelsim here, rather we use ncSim. Can we change script for this.
  Please guide me in this regard.

Thanks,

FrancescoConti commented 6 years ago

Why did you have us lose our time in answering questions on ModelSim if you don't use it in the end???

Anyways. We do not support NCSim in PULPino. We know it would work because other people have done it, but you have to do that on your own.

lfernand4 commented 6 years ago

thotaraghava, I want to use NCSim to, if you found how to do it, explain me.

thotaraghava commented 6 years ago

Hi @FrancescoConti and @gautschimi 

 My interest is riscv only (not zeroriscy), so I should run 'cmake_configure.riscv.gcc.sh' only right in /sw/build.

 I ran only this and when I am  trying to run 'make helloworld', I am getting below errors.

Building C object libs/sys_lib/CMakeFiles/sys.dir/src/exceptions.c.o cc1: error: unrecognized command line option "-fdiagnostics-color=always /sw/libs/sys_lib/src/exceptions.c:1: error: bad value (RV32I) for -march= switch /sw/libs/sys_lib/src/exceptions.c:1: error: bad value (RV32I) for -mtune= switch make[3]: [libs/sys_lib/CMakeFiles/sys.dir/src/exceptions.c.o] Error 1 make[2]: [libs/sys_lib/CMakeFiles/sys.dir/all] Error 2 make[1]: *** [apps/helloworld/CMakeFiles/helloworld.dir/rule] Error 2

Can you look into this.

Thanks

thotaraghava commented 6 years ago

Did I miss any settings in cmake_configure_riscy.gcc.sh.

Quick response would be highly appreciated Prior to running make helloworld, should I 'make vcompile'. My GCC version is as below.

gcc (GCC) 4.4.7 20120313 (Red Hat 4.4.7-11)

My aim is run one testcase (helloworld) and observe waveforms !!!!! Please guide me in detail.

Thanks in advance

quangdaovu commented 6 years ago

All the details you need is on the README file. So PLEASE read it carefully before make a question, the amount time that it takes to ask a question is enough for you to read it once.

This time, your problem lies on the very first part of README.

Requirements PULPino has the following requirements

riscv-toolchain, specifically you need riscv32-unknown-elf-gcc compiler and friends. There are two choices for this toolchain: Either using the official RISC-V toolchain supported by Berkeley or the custom RISC-V toolchain from ETH. The ETH versions supports all the ISA extensions that were incorporated into the RI5CY core as well as the reduced base instruction set for zero-riscy. Please make sure you are using the newlib version of the toolchain.

The toolchain can be found here: https://github.com/pulp-platform/ri5cy_gnu_toolchain

thotaraghava commented 6 years ago

thank you.

thotaraghava commented 6 years ago

One thing, I am missing in this platform. After installing ri5cy_gnu_toolchain, how this should be linked to pulpino. May be I am not understanding this part thoroughly.

Thanks,

quangdaovu commented 6 years ago

Yet again your binary folder of compiled toolchain must be on your PATH.

thotaraghava commented 6 years ago

Thanks for responding. In PATH means, how. Can you please elaborate. @FrancescoConti , could you please reopen this ticket.

Thanks,

FrancescoConti commented 6 years ago

http://lmgtfy.com/?q=what+is+PATH+on+Linux

quangdaovu commented 6 years ago

I suggest you take a simple course about Linux first. And as Francesco already answered 4 days ago:

export PATH=/path/to/built_toolchain/bin:$PATH

thotaraghava commented 6 years ago

Ok, ok. I understood. :)

thotaraghava commented 6 years ago

Hi all,

Is this should be in my PATH. /PATH/ri5cy_gnu_toolchain. I tried below paths also.

/PATH/ri5cy_gnu_toolchain/toolchain/ /PATH/ri5cy_gnu_toolchain./build

I am still getting below error. Building C object libs/sys_lib/CMakeFiles/sys.dir/src/exceptions.c.o cc1: error: unrecognized command line option "-fdiagnostics-color=always /sw/libs/sys_lib/src/exceptions.c:1: error: bad value (RV32I) for -march= switch /sw/libs/sys_lib/src/exceptions.c:1: error: bad value (RV32I) for -mtune= switch make[3]: [libs/sys_lib/CMakeFiles/sys.dir/src/exceptions.c.o] Error 1 make[2]: [libs/sys_lib/CMakeFiles/sys.dir/all] Error 2 make[1]: *** [apps/helloworld/CMakeFiles/helloworld.dir/rule] Error 2

Means, there could be issue in my toolchain building ?

Thanks

quangdaovu commented 6 years ago

Have you ever look at Francesco's previous answer about PATH? PATH is not a directory. Please study about it first.

Then, considering your ri5cy_gnu_toolchain directory is /somewhere/ri5cy_gnu_toolchain/, and of course you correctly built it, then you should include this on your PATH: /somewhere/ri5cy_gnu_toolchain/install/bin

I don't mean to be rude, but you really have to study before making question.

thotaraghava commented 6 years ago

I have installed ri5cy_gnu_toolchain in some other machine (server) and copied to my working server(which does not have internet) and running. Should I run "make" again here, to make sure installation proper. I tried to run but its getting halted with some error. Will this work as toolchain been installed in other server.

Thanks

thotaraghava commented 6 years ago

Even after setting /somewhere/ri5cy_gnu_toolchain/install/bin in my PATH, I am still getting the above errors. Any clue ? I am clueless here.

quangdaovu commented 6 years ago

Are you really put somewhere there??? Could you give me the output of echo $PATH

thotaraghava commented 6 years ago

path1:path2:/user/organization/USERS/raghav.thota/ri5cy_gnu_toolchain/install/bin

I used below command as mine cShell. setenv PATH $PATH\:/path/install/bin

quangdaovu commented 6 years ago

I have no idea how could your system works with that kind of PATH. I'm speechless.

FrancescoConti commented 6 years ago

@quangdaovu thanks!!!!!!!!!!

@thotaraghava , your setup is clearly completely different from any "normal" supported Linux system. I mean, not only you have a system where you cannot access the internet from Linux, but also a strongly customized Linux environment, very different from "normal" setups. This means that you are likely working in a big industrial organization and you have a system administrator. Ask for his help. It's impossible for us to help you solve problems for such an environment.

thotaraghava commented 6 years ago

Hi,

My cmake_configure.riscv.gcc.sh settings are as below.

TARGET_C_FLAGS="-O3 -g"

ratands TARGET_C_FLAGS="-O3 -m32 -g"

TARGET_C_FLAGS="-O2 -g -falign-functions=16 -funroll-all-loops"

RVC=0 USE_ZERO_RISCY=0 RISCY_RV32F=0 ZERO_RV32M=0 ZERO_RV32E=0 GCC_MARCH="IMXpulpv2" ARDUINO_LIB=1

I am interested in riscv core (4stage). Can you please confirm my above settings.

Thanks,

thotaraghava commented 6 years ago

How USE_RISCY variable in sw/CMakeLists.txt will be decided. I guess, somehow its taking 0. I hope you people can answer this.

thotaraghava commented 6 years ago

Any clue is highly appreciated :) Somehow it is taking as below. GCC_MARCH= RV32I

quangdaovu commented 6 years ago

For the first question, everything you need is listed in README section. Read it. The following part comes from sw's README.

CMake uses out-of-source builds which means you will need a separate build folder for the software, e.g. build mkdir build Then switch to the build folder and copy the cmake template configuration script there which resides in the sw folder. The name of template follows the following naming scheme: cmake-configure.{or1k/riscv}.{gcc/llvm}.sh Choose, copy, modify and then execute this script. It will setup the build environment for you.

So, cmake_configure.riscv.gcc.sh means it'll setup for RISCY core.

For the second question, According to cmake_configure.riscv.gcc.sh, the first part it sets the variables value and the last part of it calls cmake with these values. Note that non-included parameters will be 0. So, if you're using cmake_configure.riscv.gcc.sh then USE_RISCY will be 0.

I suggest again that you have to study. It's the third time already. If you have time to ask a question and wait for others to answer, why can you just give a little more time to learn? You can open cmake_configure.riscv.gcc.sh to copy its content (in first question), why didn't you read all the file? Why didn't you try to google "HOW CMAKE WORKS"? To be honest, I've lost all of my willing to help. Good luck on your proj.

thotaraghava commented 6 years ago

thanks @quangdaovu

thotaraghava commented 6 years ago

flash_stim.slm I2_ram.slm I2_ram_cut0_hi.slm I2_ram_cut0_lo.slm I2_stim.slm spi_stim.txt tcdm_bank0.slm.

Which files of these are considered as simulators for simulation

Regards, Raghav

thotaraghava commented 6 years ago

Hello @FrancescoConti ,

I may need information about this. I would like to pick the concerned file and put in the test bench of my system Tahnks

thotaraghava commented 6 years ago

Hello anybody,

I am clueless on this. Can someone break the silence. Thanks in advance,

FrancescoConti commented 6 years ago

No one of that files is a ``simulator'' and if you ask this, it means that you have no idea of what a simulator is. ModelSim is a simulator, QuestaSim is a simulator, Cadence Incisive is a simulator, Synopsys VCS is a simulator. Those files host stimuli generated by the SDK when you compile and run your program, as stated in the README.

FrancescoConti commented 6 years ago

@thotaraghava it is clear to me you are proceeding blindly without much idea of what you are doing. I really suggest you do your own personal experiments and study, study, study. PULPino is not a particularly complicated system, but it seems it is well beyond your current knowledge of how to handle an RTL platform.

As a person who has followed several student projects, some of them good others less, let me give you one piece of advice: if you don't spend a few days on a problem completely on your own, you won't learn. I personally learned that the hard way, and I think this is true for 99% of engineers in the world (at least those in this field). In real engineering problems are hard and most of the time if you're the person on a problem, it means that you are the most competent and skilled person that there is for that specific problem - so there's literally nobody that can solve the thing for you. Take these as occasions to grow your problem-solving skills.

thotaraghava commented 6 years ago

absolutely brilliant message.

thotaraghava commented 6 years ago

My intention was to ask like this

Which files of these are considered as s"t"imulators for simulation (this sentence is from my previous post) Typing mistake ("t" was missed). But its good that I got confirmed and got some good message.

Thanks

thotaraghava commented 6 years ago

I tried my best to find out which file of the below files should be used as stimuli.

flash_stim.slm I2_ram.slm I2_ram_cut0_hi.slm I2_ram_cut0_lo.slm I2_stim.slm spi_stim.txt tcdm_bank0.slm

Not able to find out. Any clue on where can I get.

Regards,

gautschimi commented 6 years ago

there are different methods on booting in pulpino.

either you preload all memories (fast) or you load the memory content over spi (slow)

option 1: preload: l2_ram.slm = content of the instruction memory tcdm_bank0.slm = content of the data memory

option 2: load over spi: spi_stim.txt is used

you will see in tb.sv and its packages which stimuli is used under what conditions

On 02/23/2018 11:34 AM, thotaraghava wrote:

I tried my best to find out which file of the below files should be used as stimuli.

flash_stim.slm I2_ram.slm I2_ram_cut0_hi.slm I2_ram_cut0_lo.slm I2_stim.slm spi_stim.txt tcdm_bank0.slm

Not able to find out. Any clue on where can I get.

Regards,

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/pulp-platform/pulpino/issues/170#issuecomment-367972681, or mute the thread https://github.com/notifications/unsubscribe-auth/ARsRnb9tPnOP4mFoO2O0ypiJX8uMCYzOks5tXpRIgaJpZM4SGnCb.

thotaraghava commented 6 years ago

@gautschimi , Excellent !!!.

Let me work on myself .

Thanks,

thotaraghava commented 6 years ago

Hi @gautschimi ,

  For "helloworld" testcase, I generated slm files .
  Interestingly, I could not see any data in l2_ram.slm, but as expected, data exists in tcdm_bank0.slm.
  I came to know that, I should input scatter file during compilation which I did not.
  Was this the reason for the situation.
  And if I have to mention scatter file, where I need to.

  PS: This is my first processor related project.

Thanks

quangdaovu commented 6 years ago

you will see in tb.sv and its packages which stimuli is used under what conditions

I doubt that you really have no intent to read source code.

gautschimi already mentioned about tb.sv and its package, and the method that tb used to preload memory. You should have read and trace it back.

https://github.com/pulp-platform/pulpino/blob/0cdfddd93531acb5291d8ed4a449a5173a02826f/tb/tb.sv#L246-L250

If PRELOAD then execute mem_preload(), which lies on tb_mem_pkg.sv. And from there you will be able to find this:

https://github.com/pulp-platform/pulpino/blob/05f0dbeeb8823e215a8fc526145af24ad18417ab/tb/tb_mem_pkg.sv#L38-L48

The files that you need are slm_files/l2_stim.slm for instruction mem and slm_files/tcdm_bank0.slm for data mem.

thotaraghava commented 6 years ago

HI all,

I am able to generate both instr and data files.
In my system, we have only one ROM in which both instruction and data should sit.
So, I put instruction data on top of the ROM and data just below that. 
What could be the value of "boot_addr_i" of core_region module to drive in this case.

Please correct me if something is wrong in this procedure.

Regards,

thotaraghava commented 6 years ago

Should it be 0x0000_0000.

I am not using "peripherals" block, as we dont have such peripherals in our system. So, I am trying to drive.

At present, I could not see any transactions with riscv_core.