Re: [Bug-xorriso] consultation about libburn

[Thomas Schmitt]

Hi,

address@hidden wrote:
> cdrskin: FATAL : Failed to transfer command to drive
> cdrskin: ( Most recent system error: 22 'Invalid argument' 
> cdrskin: malloc.c:3096: sYSMALLOc: Assertion `(old_top ==
> ...
> UNIX-SIGNAL: SIGABRT errno= 22.

This looks like the management of dynamic memory went bad.
Either because cdrskin and libburn did something wrong, or
because the system equipment has a flaw.

The error number 22 'Invalid argument' may be a remnant from
earlier normal operation or may have been set by the error
catching code in malloc.c (which is part of libc, i guess).


> I found the problem occurs at ioctl(d->fd, SG_IO, &s).

Regrettably the occasion of failure is often not the one which
causes such an error. It may well be that some previous memory
operation initiated the problem and that it needed some time
to become obvious.


> but if i use linux 3.13.0-24-generic on pc,there is no problem at
> all.

This points somewhat to the operating system software as culprit.
But it cannot be outruled that my stuff provides the trigger in
both cases, and that one system version can stand it whereas the
other cannot.


> looking for your help,thanks.

The program valgrind should be able to point us to the first
bad memory usage in the course of the program. If the cause is
in libburn or cdrskin, then there is hope we find it by valgrind.

Please repeat your cdrskin run on the problemeatic system with
a few additional words wrapped around the command line:

  valgrind --leak-check=full  cdrskin ...your.arguments... \
    2>&1 | tee -i /tmp/vagrind_cdrskin.log

The possibly lengthy messages of valgrind and cdrskin will
not only flood your terminal but will aslo be recorded in the
text file

  /tmp/vagrind_cdrskin.log

which you please send to this list (or to me in private, if you
do not want to publish the messages). 


Have a nice day :)

Thomas