[panda-users] Recording replays fine on one system, has infinite loop on another

[Brendan Dolan-Gavitt]

My guess is that this is caused by the recent addition of VAPIC recording
that was added in a pull request to make Windows XP guests record and
replay properly. Unfortunately since that bug fix adds a new event that
PANDA expects to be in the log old replays may now fail.

I general I’m not sure how to handle this situation properly. We definitely
want to maintain compatibility with old recordings, but we also need to be
able to fix bugs.

One option might be to use a versioned file format - but then the
corresponding code in PANDA would also need to have version checks added at
every place where events are read from the log. Maybe we could make this
easier through some macros though...

For now if you want to try to pin down the exact commit(s) via bisect that
would help us figure out how to handle it.

-Brendan

On Sun, May 27, 2018 at 7:40 AM Bridgey theGeek <bridgeythegeek at gmail.com>
wrote:

> Thanks for the reply! Glad to know it's not just me.
>
> I really don't want to have to re-record so I guess I need to copy my
> "old" PANDA across to the VPC.
>
> My recordings are only a matter of weeks old. Has their been a fundamental
> change in the recording format??
>
> Thanks,
> Adam
>
>
> On Sun, 27 May 2018, 12:27 Gabriele Viglianisi, <vigliag at gmail.com> wrote:
>
>> I have updated PANDA on my machine to the last version earlier today and
>> it
>> seems unable to replay old recordings (with the infinite loop error). I
>> think that's due to some recent changes.
>> Il giorno dom 27 mag 2018 alle ore 12:42 Bridgey theGeek <
>> bridgeythegeek at gmail.com> ha scritto:
>>
>> > Hi PANDAs,
>>
>> > I have a recording which when I replay on my local box (using the PANDA
>> Docker image) all works fine.
>>
>> > However, when I upload the same replay to my VPC in the cloud (Ubuntu
>> 16.04.4, not Docker), PANDA complains that an infinite loop has been
>> detected.
>>
>> > I've confirmed that both files (*-rr-nondet.log and *-rr-snp) have been
>> copied correctly by checking their MD5s.
>>
>> > Why might one system cause an infinite loop and the other be happy?
>>
>> > Thank you!
>> > Adam
>> > _______________________________________________
>> > 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
>
-- 
Brendan Dolan-Gavitt
Assistant Professor, Department of Computer Science and Engineering
NYU Tandon School of Engineering
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/panda-users/attachments/20180527/c21cbbe7/attachment.html