rtl for avr and sh

Joel Sherrill Joel.Sherrill at OARcorp.com
Wed Jun 19 02:07:24 UTC 2013


Avr is a tiny architecture. Largest CPU only had 64k address space for sixteen bit words. Also the compiler has bugs which avr provided tools have many patches. It really is not appropriate target for RTL.

Shsim may be a different matter. Look at the help from the sh-rtens4.11-run command and see if there is a memory size option. If there is we can use that argument, make the Sim script pass it and fix the linkcmds. Otherwise we need to look at the Sim source in gdb and check into increasing it. I did that with sis a few years ago.


Peng Fan <van.freenix at gmail.com> wrote:



hi,
I do not have a stable internet connection now and  was often dropped on the IRC, so I asked here.

1.
I am tring to make rtl run on avrtest and shsim bsp, but I found that they have small ram with avr 64KB and shsim 256KB. I am not sure whether it is needed to do this or not.
2.
 I compiled rtems for avr, but the compilation failed.
The configure command is :
../rtems/configure --prefix=/opt/rtems-4.11 --target=avr-rtems4.11 --enable-tests=samples
The following is the error message.
../../../../../../temp-rtems/c/src/../../cpukit/libblock/src/flashdisk.c: In function 'rtems_fdisk_recycle_segment':
../../../../../../temp-rtems/c/src/../../cpukit/libblock/src/flashdisk.c:1392:1: error: unable to find a register to spill in class 'POINTER_REGS'
 }
 ^
../../../../../../temp-rtems/c/src/../../cpukit/libblock/src/flashdisk.c:1392:1: error: this is the insn:
(insn 382 381 383 19 (set (reg:HI 26 r26)
        (reg/v/f:HI 69 [ dpd ])) ../../../../../../temp-rtems/c/src/../../cpukit/libblock/src/flashdisk.c:1321 82 {*movhi}
     (nil))
../../../../../../temp-rtems/c/src/../../cpukit/libblock/src/flashdisk.c:1392: confused by earlier errors, bailing out
gmake[5]: *** [src/flashdisk.o] Error 1
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20130618/b147e966/attachment.html>


More information about the devel mailing list