On Thu, May 10, 2012 at 8:13 PM, Hui Kang <[email protected]> wrote:
> Hi, I added some debug message as follows
>
> // ooo-pipe.cpp
> if unlikely (isclass(uop.opcode, OPCLASS_BRANCH)) {
> + if !uop.emo
> ptl_logfile << "uop eom is 0\n", uop.rip, flush;
> assert(uop.eom);
>
> Can you also add some more information in '
if (!uop.eom) {
ptl_logfile << "uop.eom is false. Dumping everything\n";
dump_all_info();
dump_bbcache_to_logfile();
}
This will dump all the CPU's internal state and decoded instruction cache
into logfile. If you can send that logfile, may be we can find the bug from
there.
- Avadh
I then started the VM and the simulation. When the VM crashes as before,
> the last line of log file is
> uop eom is 0 [0xffffffffa01235ce ss64 krn mfn 0]
> Could anyone figure out why the branch opcode's eom is not set? Please
> let me know if more information is needed.
>
> - Hui
>
>
> From: Hui Kang <[email protected]>
> Date: Thu, 10 May 2012 16:53:18 -0400
> To: Paul Rosenfeld <[email protected]>
>
> Cc: marss mailing list <[email protected]>
> Subject: Re: [marss86-devel] crash when simulating a fedora 14 OS
>
> Hi, Paul,
> No, it is fedora 14 x86_64 version.
>
> - Hui
>
> From: Paul Rosenfeld <[email protected]>
> Date: Thu, 10 May 2012 16:47:26 -0400
> To: Hui Kang <[email protected]>
> Cc: marss mailing list <[email protected]>
> Subject: Re: [marss86-devel] crash when simulating a fedora 14 OS
>
> Are you running a 32bit version of Fedora by chance?
>
> On Thu, May 10, 2012 at 4:26 PM, Hui Kang <[email protected]> wrote:
>
>> Hi,
>> I am using marssx86 to run a fedora 14 os. After I boot the machine and
>> start the simulation from the QEMU monitor, the VM exits with some error
>> message.
>>
>> Completed 321000 cycles, 152247 commits: 192571 Hz,
>> 226715
>> Completed 395000 cycles, 192288 commits: 368144 Hz,
>> 199200
>> Completed 509000 cycles, 219744 commits: 564596 Hz,
>> 135978
>> Completed 564000 cycles, 263019 commits: 274601 Hz,
>> 216061
>> Completed 607000 cycles, 309815 commits: 212271 Hz,
>> 231010 insns/sec: rip ffffffff811c2305qemu-system-x86_64:
>> ptlsim/build/core/ooo-core/ooo-pipe.cpp:2131: int
>> ooo::ReorderBufferEntry::commit(): Assertion `uop.eom' failed.
>>
>> I did a quick search online and find this maybe caused by an unsupported
>> opcode, but I am not sure about the cause.
>> Could anyone tell me how to debug this error message or find the
>> unsupported opcode? Thanks.
>>
>> - Hui
>>
>>
>> PS.
>> More information Š.
>> The code snippet is from ooo-pipe.cpp is
>>
>> if unlikely (isclass(uop.opcode, OPCLASS_BRANCH)) {
>> assert(uop.eom);
>>
>> If the uop is a branch class opcode and it should be set an eom. This
>> looks an exception for this case.
>> Any comment?
>>
>>
>>
>> _______________________________________________
>> http://www.marss86.org
>> Marss86-Devel mailing list
>> [email protected]
>> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>>
>
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
