[panda-users] Replay error

[Brendan Dolan-Gavitt]

Just specifying -cpu core2duo shouldn't put QEMU use multiple cores
(you need to specify -smp for that). It's possible though that the
core2duo CPU is enabling some OS feature that uses hardware in a way
that breaks replay.

One other idea first though: how was the image created? If it was
converted over from a VMWare image, it could just be this bug:

https://github.com/moyix/panda/issues/18

If so, uninstalling VMWare Tools will fix it.

Is the analysis you're doing dependent on using a core2duo as opposed
to the default QEMU CPU?

On Thu, Feb 5, 2015 at 2:24 PM, Leek, Timothy - 0559 - MITLL
<tleek at ll.mit.edu> wrote:
> Certainly our record / replay isn’t multi-core.
>
> From: Manolis Stamatogiannakis <mstamat at gmail.com>
> Date: Thursday, February 5, 2015 at 2:22 PM
> To: "Leek, Timothy - 0559 - MITLL" <tleek at ll.mit.edu>
> Cc: "Downing, Evan P" <edowning3 at gatech.edu>, "panda-users at mit.edu"
> <panda-users at mit.edu>
> Subject: Re: [panda-users] Replay error
>
> I'll take a lucky shot: Is "-cpu core2duo" supported?
> IIRC, I've read somewhere that only 1 CPU is currently supported.
>
> Best,
> Manolis
>
> 2015-02-05 20:18 GMT+01:00 Leek, Timothy - 0559 - MITLL <tleek at ll.mit.edu>:
>>
>> Hi Evan.  Dos it replay with any plugins?  That is, if you remove the
>> –panda replaymovie part?
>>
>> Also, you shouldn’t need to specify the –hda windowsxp.img for replay.
>>
>>
>> From:  <Downing>, Evan P <edowning3 at gatech.edu>
>> Date:  Thursday, February 5, 2015 at 2:09 PM
>> To:  "panda-users at mit.edu" <panda-users at mit.edu>
>> Subject:  [panda-users] Replay error
>>
>>
>> Hey all,
>>
>>
>> I recorded a session with PANDA and now I'm attempting to replay it back
>> using the plugin 'replaymovie' (I've been able to get replay to work
>> successfully before with a different recording session, so I'm confused as
>> to why it won't work for this one).
>>
>>
>> I recorded it by using the following:
>>
>>     $ sudo ./panda/qemu/x86_64-softmmu/qemu-system-x86_64 -cpu core2duo
>> -hda windowsxp.img  -vga std -net
>> nic,macaddr=02:00:00:00:00:01,model=rtl8139 -net tap,script=no,ifname=vm0
>> -vnc :0 -snapshot -monitor stdio
>>
>>     (qemu) being_record evan_test
>>
>>     (qemu) end_record
>>
>>
>>     (No errors seen while recording.)
>>
>>
>>
>> I'm attempting to replay it:
>>
>>     $ sudo ./panda/qemu/x86_64-softmmu/qemu-system-x86_64 -cpu core2duo
>> -vga std -net nic,macaddr=02:00:00:00:00:01,model=rtl8139 -net
>> tap,script=no,ifname=vm0 -vnc :0 -replay evan_test -panda ‘replaymovie'
>>
>>
>> The error I receive:
>>
>>
>> adding
>> /home/evan/panda/qemu/x86_64-softmmu/panda_plugins/panda_replaymovie.so to
>> panda_plugin_files 0
>> loading
>> /home/evan/panda/qemu/x86_64-softmmu/panda_plugins/panda_replaymovie.so
>> Success
>> loading snapshot
>> ... done.
>>
>> Logging all cpu states
>>
>> [A bunch of stuff here. Processes 19% of instructions successfully.]
>>
>> RR rr_assertion `0' failed at /home/evan/panda/qemu/rr_log.c:1025
>> Current log point:
>> {guest_instr_count=7600686304 pc=0x806d2797, secondary=0x80d1c0ec}
>> Next log entry type: RR_INPUT_4
>> Current replay point:
>> {guest_instr_count=7600686299 pc=0x806d2797, secondary=0x80d1c0ec}
>>
>>
>>
>>
>>
>> In the source code for "rr_log.c" where the error occurred, it says "mz
>> we're trying to replay too early or we have the wrong kind of
>> rr_nondet_log
>>  entry. this is cause for failure."
>>
>> Can someone help me determine why this happened?
>>
>>
>> Thanks,
>>
>> Evan
>>
>>
>>
>> _______________________________________________
>> panda-users mailing list
>> panda-users at mit.edu
>> http://mailman.mit.edu/mailman/listinfo/panda-users
>
>
>
> _______________________________________________
> panda-users mailing list
> panda-users at mit.edu
> http://mailman.mit.edu/mailman/listinfo/panda-users
>