Thanks guys. I really don't want to have to install an additional toolchain if I can avoid it.
I just discovered the following messages which provide some seemingly relevant clues but doesn't really address collapsing INITIALIZER and INITIALIZED areas into one: https://sourceforge.net/p/sdcc/mailman/message/34781898/ I guess I need to dig into understanding the linker scripts and see if I can figure it out. Brad On Fri, Apr 3, 2020 at 6:11 AM Sergey Belyashov <sergey.belyas...@gmail.com> wrote: > > Hi, > SDCC uses very simple scheme for storing initialization data. It generates > two copies of variables, one copy is initialization data which is stored > after code section, and copied to variables area on startup. I think, it is > caused by very simple linkers supported. More complex linkers can do this job > instead of compiler, so result binary can be customized to generate data > section initialized, or separate initialization data. For example, GNU LD may > change LMA (load memory address, address after loading, before startup) of > data segment, so result will be same as produced by SDCC. But you may exclude > LMA changing, so VMA (virtual memory address, address where program expects > objects to be, "CPU memory") and LMA will be same - it is what you want. > > I have upgraded GNU Binutils to support SDCC. So you may generate assembler > for your code, then compile it by GNU AS and link by GNU LD (I suggest use > z80-unknown-elf toolchain). You may try to create your own liker script, > which stores INITIALIZER section to .data output section, and creates > INITIALIZED as separate section with same VMA as INITIALIZER but with > different LMA (I suggest put it out of address space). You need binutils > 2.34+. I suggest master branch of their GIT repo (there are some useful fixes > and upgrades). Build using --target=z80-unknown-elf as configure option. > Also, you may vote for: https://sourceforge.net/p/sdcc/feature-requests/666/ > > ... > .text 0x8000 : > { > __text_start = .; > *(_HOME) > *(_CABS) > *(_CODE) > *(_RODATA) > *(.text *.text.* .gnu.linkonce.t.*) /* it is required for proper section > attributes */ > __text_end = .; > } > .data : > { > __data_start = .; > *(_INITIALIZER) > *(.data *.data.* .gnu.linkonce.d.*) /* it is required for proper section > attributes */ > __data_end = .; > } > .bss : > { > __bss_start = .; > *(_DABS) > *(_DATA) > *(.bss *.bss.* .gnu.linkonce.b.*) /* it is required for proper section > attributes */ > __bss_end = .; > } > .data_init ADDR(__data_start) : AT(0x10000) /* LMA is out of address space */ > { > *(_INITIALIZED) > } > ... > > Sergey > > ср, 1 апр. 2020 г. в 11:01, Brad Robinson <brobin...@toptensoftware.com>: >> >> Hey All, >> >> I've got a Z80 project where the generated image will be loaded >> directly into RAM and therefore the initialization data doesn't need >> to be copied from the image to a specified RAM area. >> >> How can I configure SDCC to generate an image like this? A working >> example would be great. >> >> Brad >> >> >> _______________________________________________ >> Sdcc-user mailing list >> Sdcc-user@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/sdcc-user > > _______________________________________________ > Sdcc-user mailing list > Sdcc-user@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/sdcc-user _______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
