Hi, c18 uses #pragma to define the section for the following code. Do we have __attribute__(section) or something like that in SDCC?
If you can control the section for a function, you can play with the gplink linker script to place any "single function section" anywhere you want. Sebastien Le 30/08/2011 11:07, Kustaa Nyholm a écrit : > On 8/30/11 01:23, "Mario" <marbar...@gmail.com> wrote: > >> Just as described at section 3.6. >> I have managed to make it work partially, for example, I can decide >> the address in the RAM of variables only when they are globally >> defined. > How could you define the address of a local variable? Conceptually > it is in the stack and the address varies from invocation to > invocation. > > >> But I had to find out about this by myself, because the >> manual did not mention this. > Well in view of what I wrote to me that is obvious. > >> The way I had to do it: >> unsigned char __data myChar; > Isn't that what was in the manual? > >> But if I place the __data __at 0x30 part before the unsigned char >> part, it will work anyway. The manual does not mention about this >> either. > So? It is the nature of C that the order everything in definitions > is not fixed and does not always follow human logic. I don't think > the manual needs to spell out all the variations. I guess you could > write that as 'unsigned __at 0x30 char __data myChar'? > > >> I tried this example literally from the manual: >> __code __at (0x7ff0) char Id[5] = ²SDCC²; >> >> But when compiling it throws: >> error 9: FATAL Compiler Internal Error in file 'glue.c' line >> number '937' : !"Initialized char-arrays are not yet supported, assign >> at runtime instead." >> Contact Author with source code > This a limitation of the (current) compiler implementation, maybe could > have been mentioned in the manual, on the otherhand this maybe specific > to the backend/processor you are targeting and the error pretty > much spells out the situation and the remedy. Good enough for me. > > >> What I want to do now is to make a function start exactly at a certain >> address in code memory but I find no way how to do it. > Zafi showed you a workaround. Again this I do not see this as a > documentation issue, if the compiler does not support some non standard > feature I do not see that needs to be documented. There are endless > number of non standard features the compiler does not support. > > In a rare occasion this (placing a function at a fixed address) is > a legitime and reasonable requirement but no compiler (documentation) > (writer) can fore see all the corner cases. > >> I tried also to create a *var = "with-text", with its data located in >> RAM memory, but no luck, it is always stored somehow at code memory. > Makes sense to me as ANSI C defines string literals as unmodifiable. > Why would you want them to be in RAM (which typically is more scarce > than ROM in small processors and placing them in the RAM would > still require a copy in ROM)? > > br Kusti > > >> On Mon, Aug 29, 2011 at 3:10 AM, Kustaa Nyholm >> <kustaa.nyh...@planmeca.com> wrote: >>> On 8/29/11 00:00, "Mario" <marbar...@gmail.com> wrote: >>>> Is this a currently working feature? can someone provide me some piece >>>> of code where absolute addressing actually works? >>> I'm not sure what you re talking about with 'absolute addressing', >>> please show some code... >>> >>> br Kusti >>> >>> >>> >>> ------------------------------------------------------------------------- >>> ----- >>> EMC VNX: the world's simplest storage, starting under $10K >>> The only unified storage solution that offers unified management >>> Up to 160% more powerful than alternatives and 25% more efficient. >>> Guaranteed. http://p.sf.net/sfu/emc-vnx-dev2dev >>> _______________________________________________ >>> Sdcc-user mailing list >>> Sdcc-user@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/sdcc-user >>> >> -------------------------------------------------------------------------- >> ---- >> Special Offer -- Download ArcSight Logger for FREE! >> Finally, a world-class log management solution at an even better >> price-free! And you'll get a free "Love Thy Logs" t-shirt when you >> download Logger. Secure your free ArcSight Logger TODAY! >> http://p.sf.net/sfu/arcsisghtdev2dev >> _______________________________________________ >> Sdcc-user mailing list >> Sdcc-user@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/sdcc-user >> > > -- > Kustaa Nyholm > Research Manager, Software > Research and Technology Division > PLANMECA OY > Asentajankatu 6 > 00880 HELSINKI > FINLAND > > Please note our new telephone and fax numbers! > Tel: +358 20 7795 572 (direct) > Fax: +358 20 7795 676 > GSM: +358 40 580 5193 > e-mail: kustaa.nyh...@planmeca.com > > > > > > > ------------------------------------------------------------------------------ > Special Offer -- Download ArcSight Logger for FREE! > Finally, a world-class log management solution at an even better > price-free! And you'll get a free "Love Thy Logs" t-shirt when you > download Logger. Secure your free ArcSight Logger TODAY! > http://p.sf.net/sfu/arcsisghtdev2dev > _______________________________________________ > Sdcc-user mailing list > Sdcc-user@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/sdcc-user ------------------------------------------------------------------------------ Special Offer -- Download ArcSight Logger for FREE! Finally, a world-class log management solution at an even better price-free! And you'll get a free "Love Thy Logs" t-shirt when you download Logger. Secure your free ArcSight Logger TODAY! http://p.sf.net/sfu/arcsisghtdev2dev _______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
