Disk Image and Binaries

[DAVESH SHINGARI]

Hi

For running x86, from where should the binaries and disk image be downloaded. I am trying the following and getting the errors:

·       gem5-gpu/gem5/full_system_images# wget http://www.m5sim.org/dist/current/x86/x86-system.tar.bz2

·     /gem5-gpu/gem5/full_system_images# tar -xf x86-system.tar.bz2   >> Binaries  (x86_64-vmlinux-2.6.22.9  x86_64-vmlinux-2.6.22.9.smp), Disk (linux-x86.img)

     Export M5_PATH and export LINUX_IMAGE=$M5_PATH/disks/linux-x86.img

·        /gem5-gpu/gem5/full_system_images# wget www.m5sim.org/dist/current/m5_system_2.0b3.tar.bz2

·        /gem5-gpu/gem5/full_system_images# tar -xf m5_system_2.0b3.tar.bz2

      /gem5-gpu/gem5/full_system_images# cp ./m5_system_2.0b3/disks/linux-bigswap2.img ./disks/

      /gem5-gpu/gem5/full_system_images# wget http://www.cs.utexas.edu/~parsec_m5/x86_64-vmlinux-2.6.28.4-smp

      /gem5-gpu/gem5/full_system_images# mv x86_64-vmlinux-2.6.28.4-smp ./binaries/

       Finally the directory structure has following:

├── binaries

│   ├── console

│   ├── ts_osfpal

│   ├── vmlinux

│   ├── x86_64-vmlinux-2.6.22.9

│   ├── x86_64-vmlinux-2.6.22.9.smp

│   └── x86_64-vmlinux-2.6.28.4-smp

├── disks

│   ├── linux-bigswap2.img

│   ├── linux-latest.img

│   ├── linux-x86.img

│   ├── temp

│   └── x86root-parsec.img

       And the LINUX_IMAGE points to x86root-parsec.img

      But when I run 

      ./build/VI_hammer/gem5.opt ../gem5-gpu/configs/fs_fusion.py --kernel=/home/dshingar/simulator_x86/gem5-gpu/gem5/full_system_images/binaries/x86_64-vmlinux-2.6.28.4-smp

      After some time of running I get message like "“Kernel panic - not syncing: Attempted to kill the idle task!” And it get struck at that point.

     So please let me know the disk image and binaries to use and command line to use. With 22.9 kernel it work fines but doesnt support running of programs with error as 'kernel too old'. So thats why using 28.4 binary.

[Joel Hestness]

Hi Davesh,

  It's hard to say what could be going on here. Kernel panics are generally pretty tricky to debug. If you'd like to try to debug this yourself, I'd recommend dropping an Exec trace for the last couple million ticks before the panic using --debug-flag=Exec --debug-start=<trace start tick>. Given that trace, you should be able to find what portion of the kernel was being executed, and you can look up the code to figure out the cause (e.g. here: http://lxr.free-electrons.com/ident?v=2.6.28;i=do_page_fault).

  Problems with gem5 full-system mode, such as this one, are more likely to get answered on the gem5 email lists, though kernel panics tend to be highly varied and generally do not receive answers as a result. That said, I've run into problems booting Linux 2.6.28.4 with recent versions of gem5. Details on that kernel panic are here: http://permalink.gmane.org/gmane.comp.emulators.m5.devel/24965

  Joel

--

  Joel Hestness
  PhD Candidate, Computer Architecture
  Dept. of Computer Science, University of Wisconsin - Madison
  http://pages.cs.wisc.edu/~hestness/

[DAVESH SHINGARI]

Hi Joel

Thanks. I also observed the same. I used linux 2.6.28.4 image in December on my previous Gem5-GPU setup (based on hg update -r 10237) and it worked fine.

On the new version ARM linux image is working perfectly. So I will be using that for now.

And one silly doubt: "How to post on Gem5". I mailed at m5-u...@m5sim.org and gem5-...@gem5.org. But I couldn't see my question. Am I missing something?

[zzh...@gmail.com]

Hi，DAVESH SHINGARI，I have encountered the same problem with you. Did you solve the problem later? I need your help.

在 2015年4月1日星期三 UTC+8下午5:28:25，DAVESH SHINGARI写道：

[Ricks_gtech]

Hi,
I faced the same problem. This was solved using the following change:

/src/arch/x86/cpuid.cc Fri Jul 22 00:42:48 2016 -0400

                 result = CpuidResult(0x00020f51, 0x00000805,

-                                     0xe7dbfbff, 0x04000209);

+                                     0xe7dbfbff, 0x00000209);