On Wednesday 06 August 2008, jamal wrote:
quoted text
>
> Indeed there seems to be an issue here, but:
> Have you tried in your batch with a few "index"ed parameters to ipt with
> your fix? I think it will break with your changes. Of course that does
> not invalidate your concern. If you send me a two liner batch example
> which breaks for you, I will test it.
not sure about two lines
attaching example batch

quoted text
>
> > 2)Since target->t = fw_calloc(1, size); allocated memory in function
> > build_st, it have to be freed at the end, or in batch we will have memory
> > leak. TODO: Probably it must be freed in all "return -1" cases in
> > parse_ipt after build_st. About this i am not sure, up to Stephen.
> >
> > 3)lname and new_name was malloc'ed, but not freed
> >
> > Point 1 fix is critical, since without it m_ipt crashing iproute2 in
> > batch mode.
>
> I think #2 and #3 are leggit. Send two patches - one for #1 when you
> validate and the other for #2,#3 when you validate. BTW, for the
> equivalent of the latter - if you feel brave - theres a lot more of that
> type in the tc code.
#1 is validated 100%, because i was not able to run tc in batch mode for my 
tasks(2nd or 3rd ipt rule will crash), till i implement those patches. The 
batch i gave is short version, full version have around 11k lines. But it 
crashes usually max on 2rd or 3nd ipt.

Well, i can fix what i can sniff&smell in my daily job, and especially if i 
hit a wall because of some code :-) For me it is difficult to inspect the 
code, since i am very weak in C, but i will try.