We just ran into this ourselves very recently.  We haven't tracked it down,
but our suspicion is that there's a bug in the default Ruby protocol
(MI_example) that is somehow triggered by the newer version of glibc, or
perhaps by the code generated by the newer version of gcc.

Please try another Ruby protocol (for example, X86_MESI_Two_Level).  That
solved the problem for us.  You really don't want to be using MI_example
anyway.

Steve


On Tue, Jun 17, 2014 at 2:56 AM, Severin Wischmann via gem5-users <
[email protected]> wrote:

> Hi all,
>
> I recompiled the "Hello World" test program, to ensure that future
> programs work properly and I encountered an issue, when running it with
> ruby in SE-mode.
>
> Running on only one core worked, but produced the output
> info: Increasing stack size by one page.
> warn: ignoring syscall access(0, 4893217, ...)
>
> When executing the command
>
> build/X86/gem5.debug configs/example/se.py --mesh-rows=1 --num-cpus 2
> --num-dir 2 --ruby --topology=Mesh --garnet-network=fixed -c
> tests/test-progs/hello/src/hello1,tests/test-progs/hello/src/hello1,tests/test-progs/hello/src/hello1,tests/test-progs/hello/src/hello1
>
> The same output is produced once per CPU and followed by
>
> panic: Tried to read unmapped address 0.
>  @ cycle 135796
> [invoke:build/X86/arch/x86/faults.cc, line 160]
> Memory Usage: 135796 KBytes
> Program aborted at cycle 59965000
> Aborted (core dumped)
>
> Dropping everything ruby related again produces an output with the
> warnings from before.
>
> gem5 is on the newest version,
> gcc: 4.8.2
> kernel (if relevant): 3.13.0-29
>
> Compilation command
>
>  gcc -static -o hello1 hello.c
>
> Is this a problem with my compiler? Or are some syscalls not supported
> anymore. I don't see how hello world can segfault...
>
> Any input is highly appreciated
>
> Best,
>
> Severin
>
> _______________________________________________
> gem5-users mailing list
> [email protected]
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users

Reply via email to