ARM EABI buildroot
Daniel Smolik
marvin at mydatex.cz
Thu May 31 20:19:28 UTC 2007
Tzungder Lin napsal(a):
> Hi, I think you could try TARGET_CPU_arm7tdmi
> FYR
Thanks again but still nothing. May be kernel is now broken or my CPU is
not supported.
Dan
>
> On 5/31/07, Daniel Smolik <marvin at mydatex.cz> wrote:
>> Tzungder Lin napsal(a):
>> > If you use buildroot to create your toolchain
>> > It seems config gcc default to ARMv5T architecture
>> > Maybe that's why you got illegal instruction exceptions...
>> > I think you could try to config gcc default to ARMv4T by hand.
>> > (gcc/config/arm/linux-eabi.h)
>> > I did this to make my ARMv4 CPU support the EABI interface
>> > (though without thumb and normal mode interworking)
>> Thanks for advice. In linux-eabi.h I found this:
>>
>> #define SUBTARGET_CPU_DEFAULT TARGET_CPU_arm10tdmi
>> Which target can I use for my board ?
>>
>> Thanks a lot
>> Dan
>>
>>
>>
>>
>> > FYR
>> >
>> > On 5/30/07, Tzungder Lin <tzungder at gmail.com> wrote:
>> >> Hi,
>> >>
>> >> FYR
>> >> "The EABI includes thumb interworking, which means that 16-bit Thumb
>> >> and normal 32-bit ARM instructions can be mixed at function-level
>> >> granularity. With current gcc this requires at least an armv4t core,
>> >> because it uses the BX instruction which does not exist in armv4 or
>> >> earlier." --- from debian-eabi-arm
>> >>
>> >> I think it's ok.
>> >>
>> >>
>> >> On 5/30/07, Daniel Smolik <marvin at mydatex.cz> wrote:
>> >> > Tzungder Lin napsal(a):
>> >> > > Hi, what's your ARM version?
>> >> > > EABI will use BX, BLX instructions which only be supported in
>> ARMv4t
>> >> > > and the later versions.
>> >> > > FYR
>> >> > Hi,
>> >> > this is my /proc/cpuinfo
>> >> >
>> >> > Processor : ARM920T rev 0 (v4l)
>> >> >
>> >> > BogoMIPS : 49.86
>> >> >
>> >> > Features : swp half thumb
>> >> >
>> >> > CPU implementer : 0x41
>> >> >
>> >> > CPU architecture: 4T
>> >> >
>> >> > CPU variant : 0x1
>> >> >
>> >> > CPU part : 0x920
>> >> >
>> >> > CPU revision : 0
>> >> >
>> >> > Cache type : write-back
>> >> >
>> >> > Cache clean : cp15 c7 ops
>> >> >
>> >> > Cache lockdown : format A
>> >> >
>> >> > Cache format : Harvard
>> >> >
>> >> > I size : 16384
>> >> >
>> >> > I assoc : 64
>> >> >
>> >> > I line length : 32
>> >> >
>> >> > I sets : 8
>> >> >
>> >> > D size : 16384
>> >> >
>> >> > D assoc : 64
>> >> >
>> >> > D line length : 32
>> >> >
>> >> > D sets : 8
>> >> >
>> >> > Is it OK for EABI ?
>> >> >
>> >> > Dan
>> >> >
>> >> >
>> >> >
>> >> >
>> >> >
>> >> > >
>> >> > > On 5/29/07, Daniel Smolik <marvin at mydatex.cz> wrote:
>> >> > >> Fernando Rodriguez napsal(a):
>> >> > >> > Have you got EABI support built in the kernel?
>> >> > >> >
>> >> > >> > Fernando
>> >> > >> >
>> >> > >> > Daniel Smolik wrote:
>> >> > >> >> Hi,
>> >> > >> >> have anybody working uClibc with 2.6 kernel on ARM with EABI?
>> >> > >> >> I compile buildroot with EABI support but when kenel boots I
>> >> only see
>> >> > >> >> Freeing unused kernel memory and it's all. When I compile
>> >> buildroot
>> >> > >> and
>> >> > >> >> uClibc with OABI (and apply my ugly hack to _NR_syscall) all
>> >> works.
>> >> > >> >> I compile static version of busybox and boot to EABI root all
>> >> command
>> >> > >> >> crash with Illegal instruction.
>> >> > >> >>
>> >> > >> >>
>> >> > >> >> Dan
>> >> > >> >>
>> >> > >>
>> >> > >> Yes I have, I try disable/enable Maverick Crunch support and
>> >> still the
>> >> > >> same.
>> >> > >>
>> >> > >> Dan
>> >> > >> _______________________________________________
>> >> > >> uClibc mailing list
>> >> > >> uClibc at uclibc.org
>> >> > >> http://busybox.net/cgi-bin/mailman/listinfo/uclibc
>> >> > >>
>> >> >
>> >> >
>> >> > --
>> >> > Mydatex s r.o.
>> >> > http://www.mydatex.cz
>> >> > email: smolik at mydatex.cz
>> >> > mob: 604200362
>> >> > tel: 226210085
>> >> >
>> >>
>>
>>
>> --
>> Mydatex s r.o.
>> http://www.mydatex.cz
>> email: smolik at mydatex.cz
>> mob: 604200362
>> tel: 226210085
>>
--
Mydatex s r.o.
http://www.mydatex.cz
email: smolik at mydatex.cz
mob: 604200362
tel: 226210085
More information about the uClibc
mailing list