[Ltrace-devel] ltrace: Try to fix MIPS arch (tested against git-fcf256c)

Sedat Dilek sedat.dilek at gmail.com
Tue Sep 4 06:33:56 UTC 2012


On Tue, Sep 4, 2012 at 3:17 AM, Petr Machata <pmachata at redhat.com> wrote:
> Sedat Dilek <sedat.dilek at gmail.com> writes:
>> root at fritz:/var/tmp# ./ltrace --config=./ltrace.conf -L -x main
>> ./a.out 2>&1 | tee ltrace-L-x-main.txt
>>
>> main(1, 0x7fe3c9f4, 0x7fe3c9fc, 0 <unfinished ...>
>> main(1, 0x7fe3c9f4, 0x7fe3c9fc, 0 <unfinished ...>
>> main(1, 0x7fe3c9f4, 0x7fe3c9fc, 0 <unfinished ...>
>
> This looks like a double enablement.  When we hit the breakpoint and try
> to disable it, we instead overwrite it with itself.  Hard to tell what's
> causing this.
>
> Really, it's clear that MIPS support is pretty broken.  I'll try to get
> around to fixing this later this year, but first I need to take care of
> Itanium, ARM, reimplementing -l, and supporting IFUNC symbols.
>
> (Itanium generally works, but needs some support for homogenous floating
> aggregates and passing structures in registers.  I'll take this
> opportunity to convert the backend to ARCH_HAVE_FETCH_ARG.)
>
> Which is to say that you are free to hack on it yourself, if you feel
> like it.  I won't get around to it anytime soon.
>

Thanks also for this analysis and an overview what needs to be done.
I would need your assistance and for sure more know-how :-).

Let me know when you move towards MIPS.
( See also my comments in my other email. )

Thanks,
- Sedat -

> Thank you,
> PM



More information about the Ltrace-devel mailing list