I ran into a similar problem a while back when building M5 on 32-bit Cygwin.
Maybe this explains the early crash here:
It looked like some updates to the flags field defined on StaticInstBase were
being truncated. I was seeing IsLastMicroop being lost, leading to
assert-fails (m5.opt) / crashes (m5.fast).
decoder.cc was generating uint64_t initializers for the flags and these were
being passed intact to the X86MicroopBase constructor in microop.hh. The
constructor then initialized the flags bitset by using:
flags |= setFlags;
(where flags is a std::bitset, and setFlags is a uint64_t).
I think the implementation of std::bitset I had with Cygwin was only taking the
bottom 32-bits, rather than splitting a uint64_t over two 32-bit words.
I meant to look into this further, and see whether this was a bug in
std::bitset, or in M5's use of it... but I was travelling, switched to using
Virtual PC, and unfortunately forgot to get back to the problem.
--Tim
-----Original Message-----
From: gem5-users-boun...@m5sim.org [mailto:gem5-users-boun...@m5sim.org] On
Behalf Of Gabe Black
Sent: 01 June 2011 05:27
To: gem5-users@m5sim.org
Subject: Re: [gem5-users] Aborted at X86 FS SE Mode (Solved)
Please try again with m5.opt. m5.fast has all the asserts and some other
debugging support turned off to get slightly better performance, and there may
be an earlier error that's not being caught. I always use m5.opt, and I'd
recommend you use it until you're really confident no asserts would have gone
off had they been turned on. If it still fails in this same way, please send me
your binary.
Gabe
On 05/31/11 19:24, Yu Licheng wrote:
> Hi, Gabe
>
> The application source are compiled with:
> gcc --static hello.c -ohello
> I also tried x64, and got the same 'Invalid microop' error when start m5:
> gcc -m64 --static hello.c -ohello
>
> The command line for compiling m5:
> scons -j2 build/X86_SE/m5.fast
>
> The command line for running m5:
> build/X86_SE/m5.fast configs/example/se.py -c hello
>
> m5 source:
> changeset: 8325:76095b05f4da
> user: Steve Reinhardt
> last modified: Mon May 23 14:29:23 2011 -0700
>
> gcc version: 4.4.3 (ubuntu 10.04 x86(32bit) with x86_64 library
> installed)
>
> Thanks for response.
>
> Yu Licheng
>
> -----原始邮件----- From: Gabriel Michael Black
> Sent: Wednesday, June 01, 2011 9:58 AM
> To: gem5-users@m5sim.org
> Subject: Re: [gem5-users] Aborted at X86 FS SE Mode (Solved)
>
> Lets start with SE mode. How are you compiling your binary? What is
> your command line? Also, what version of the gem5 source are you using?
>
> Gabe
>
> Quoting Yu Licheng <gancuim...@hotmail.com>:
>
>> Hi Gabe,
>> I compiled the kernel with ARCH=x86_64, and got the error:
>> panic: Invalid microop!
>> @ cycle 1000
>>
>> In SE mode, whenever the binary file is compiled in 32bit or 64bit, I
>> always get:
>> panic: Invalid microop!
>> @ cycle 500
>>
>>
>> Yu Licheng
>>
>> -----原始邮件----- From: Gabe Black
>> Sent: Wednesday, June 01, 2011 1:26 AM
>> To: gem5-users@m5sim.org
>> Subject: Re: [gem5-users] Aborted at X86 FS SE Mode (Solved)
>>
>> As the message says, it doesn't support 32 bit kernels. If you want
>> to use X86 FS you have to use a 64 bit kernel.
>>
>> Gabe
>>
>> On 05/31/11 08:38, Yu Licheng wrote:
>>> Hello
>>> I got following error in x86 SE mode, always aborted at cycle 500
>>> with 'Invalid microop!':
>>>
>>> **** REAL SIMULATION ****
>>> info: Entering event queue @ 0. Starting simulation...
>>> panic: Invalid microop!
>>> @ cycle 500
>>> [invoke:build/X86_SE/arch/generic/debugfaults.hh, line 93] Memory
>>> Usage: 548060 KBytes For more information see:
>>> http://www.m5sim.org/panic/362dcab5
>>> Program aborted at cycle 500
>>>
>>> Did you mean your problems are due to the version of tools?
>>> Mine is as follows:
>>> gcc-4.4.3
>>> python-2.6.5
>>> scons-v1.2.0.d20100117.r4629
>>> swig-1.3.40
>>> m4-1.4.13
>>>
>>> On the other hand, in FS mode I got 'fatal: Loading a 32 bit x86
>>> kernel is not supported.', so I'm wondering if m5 already supports
>>> the FS mode of x86.
>>> The changeset of m5 is 8325:76095b05f4da.
>>>
>>> Yu Licheng
>>>
>>> -----原始邮件----- From: 冠男陳
>>> Sent: Tuesday, May 31, 2011 5:10 PM
>>> To: m5-us...@m5sim.org
>>> Subject: [gem5-users] Aborted at X86 FS SE Mode (Solved)
>>>
>>> Hi M5,
>>> Few days ago, I got errors when running X86 FS and SE:
>>> 1. m5.debug: build/X86_FS/arch/x86/insts/macroop.hh:78: virtual
>>> StaticInstPtr X86ISA::MacroopBase::fetchMicroop(MicroPC): Assertion
>>> `microPC < numMicroops' failed.
>>> 2. Segmentation fault
>>>
>>> Today, I use another server and again go through the steps mentioned
>>> in the following mails.
>>> http://www.mail-archive.com/m5-users@m5sim.org/msg04960.html
>>> And it seems fine now!
>>>
>>> Tool list of two servers:
>>> Server A (got errors),
>>> gcc-4.2.4
>>> python-2.5.4
>>> scons-v1.2.0.r3842
>>> swig-1.3.36
>>> m4-1.4.10
>>>
>>> Server B (fine),
>>> gcc-4.4.1
>>> python-2.6.4
>>> scons-v1.2.0.r3842
>>> swig-1.3.36
>>> m4-1.4.13
>>>
>>> Thanks for everybody's help.
>>> Mark Chen
>>> _______________________________________________
>>> gem5-users mailing list
>>> gem5-users@m5sim.org
>>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>> _______________________________________________
>>> gem5-users mailing list
>>> gem5-users@m5sim.org
>>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>> _______________________________________________
>> gem5-users mailing list
>> gem5-users@m5sim.org
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>> _______________________________________________
>> gem5-users mailing list
>> gem5-users@m5sim.org
>> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
>>
>
>
> _______________________________________________
> gem5-users mailing list
> gem5-users@m5sim.org
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
> _______________________________________________
> gem5-users mailing list
> gem5-users@m5sim.org
> http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
gem5-users mailing list
gem5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
_______________________________________________
gem5-users mailing list
gem5-users@m5sim.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
