The vast majority of optimization in gcc is independent of the target, happening either at near-source level with tree optimizations, or in the abstract back end with RTL optimizations which include machine-dependent peephole optimizations. It's more likely that, as Grant suggests, there's a user-level issue which is allowing the newer versions of gcc to more aggressively optimize out code. However, anytime I see IPv6 I think 64-bit, and there are a variety of bugs in mspgcc4 related to shift and bitwise operations on 64-bit values. Those should all be fixed in mspgcc-20110612, which is what "uniarch" has now turned into now that it's the sole supported MSPGCC.
I recommend moving to mspgcc-20110612 and seeing if the problem persists. If it does, then it's worth digging into why since we'll be able to fix it once the problem can be reproduced outside Contiki. Peter On Wed, Jul 13, 2011 at 11:29 AM, Ricky Wong Yung Fei <yfwongri...@gmail.com> wrote: > Hey guys, > > I'm trying to port Contiki OS for my custom fabricated board consisting of a > TI MSP430F2618 MCU and a TI CC2520 radio IC. Thus far, I have managed to > port all the neccesary drivers to get my project going. Most of the example > programs are working. However, when testing the unicast example application > with compiler optimization (-O1 and above), both nodes wouldn't be able to > receive anything. Otherwise, it worked flawlessly. Unoptimized code is not > an option for me because when I went on to enable uIPv6, the linking process > fails due to the text section overlapping. I know I could actually compare > the assembly outputs of the both OS image, but before I do so, I would like > to confirm the likelihood of this being a bug as I'm not very good with gcc > internals e.g. when programs work in O0 but not O1 and above, is it normally > a bug with the optimizer? I heard that mspgcc4 is no longer maintained and > attention are now being focused on the uniarch version of mspgcc instead. I > was wondering if there are any changes in the optimizer between the latest > snapshot of mspgcc4 and the latest uniarch mspgcc? If so, I would certainly > just jump on to the uniarch mspgcc bandwagon to give it a shot. What do you > say? Any suggestions are welcomed. > > Thanks in advance. > > Regards, > Ricky Wong Yung Fei. > > ------------------------------------------------------------------------------ > AppSumo Presents a FREE Video for the SourceForge Community by Eric > Ries, the creator of the Lean Startup Methodology on "Lean Startup > Secrets Revealed." This video shows you how to validate your ideas, > optimize your ideas and identify your business strategy. > http://p.sf.net/sfu/appsumosfdev2dev > _______________________________________________ > Mspgcc-users mailing list > Mspgcc-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/mspgcc-users > > ------------------------------------------------------------------------------ AppSumo Presents a FREE Video for the SourceForge Community by Eric Ries, the creator of the Lean Startup Methodology on "Lean Startup Secrets Revealed." This video shows you how to validate your ideas, optimize your ideas and identify your business strategy. http://p.sf.net/sfu/appsumosfdev2dev _______________________________________________ Mspgcc-users mailing list Mspgcc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mspgcc-users
