Felix Fietkau wrote:
> Hinko Kočevar wrote:
>>> I've confirmed that the 610-cris_target.patch actually applied to 2.18
>>> and 2.17... weird.
>>>
>> Success with binutils 2.17 and the attached patch.
>>
>> I haven't compared your patch and the one attached here, but I think
>> they should be pretty
Hinko Kočevar wrote:
>> I've confirmed that the 610-cris_target.patch actually applied to 2.18
>> and 2.17... weird.
>>
>
> Success with binutils 2.17 and the attached patch.
>
> I haven't compared your patch and the one attached here, but I think
> they should be pretty musch the same - but I c
Hinko Kočevar wrote:
> Next I've tried out binutils 2.17 - the result is the same!
> ...
> LD vmlinux
> SYSMAP System.map
> OBJCOPY arch/cris/boot/Image
> Kernel: arch/cris/boot/Image is ready
> GZIParch/cris/boot/compressed/piggy.gz
> LD arch/cris/boot/compressed/decompr
Felix Fietkau wrote:
> Hinko Kočevar wrote:
>> Felix Fietkau wrote:
>>
>>> I've found a way to remove the need for a separate cris-elf toolchain
>>> altogether. The way it is in svn right now, head.S and misc.c can be
>>> compiled with the regular toolchain, and we tested it on a real system.
>>
Felix Fietkau wrote:
> I patched binutils 2.17 to support the criself emulation.
> The change is in toolchain/binutils/patches/2.17/610-cris_target.patch
> I wonder why it didn't apply on your system - we did not use the Axis
> toolchain at all during our test.
I know why - I've been using 2.18 :
Hinko Kočevar wrote:
> Felix Fietkau wrote:
>
>>
>> I've found a way to remove the need for a separate cris-elf toolchain
>> altogether. The way it is in svn right now, head.S and misc.c can be
>> compiled with the regular toolchain, and we tested it on a real system.
>
> Sounds great.
> I've
Felix Fietkau wrote:
>
> I've found a way to remove the need for a separate cris-elf toolchain
> altogether. The way it is in svn right now, head.S and misc.c can be
> compiled with the regular toolchain, and we tested it on a real system.
>
Sounds great.
I've checked out fresh copy of SVN an
Hinko Kočevar wrote:
> great job Imre!!! It boots now that I've added '-O binary -R .note -R
> .note.gnu.build-id -R .comment -S' to the objcopy line in
> arch/cris/arch-v10/boot/Makefile!!!
>
> Using above tools in uClibc toolchain and generated cris-foo-elf seems to
> work
> now too, since I
nice1 if they work for me tonight i will merge them with your usb driver
update
Hinko Kočevar wrote:
> Imre Kaloz wrote:
>> On Fri, 30 May 2008 13:46:53 +0200, Hinko Kočevar
>> <[EMAIL PROTECTED]> wrote:
>>
>>> It does not matter - the objcopy creates 3GB file in both cases.
>>> [EMAIL PROTECTE
Imre Kaloz wrote:
> On Fri, 30 May 2008 13:46:53 +0200, Hinko Kočevar
> <[EMAIL PROTECTED]> wrote:
>
>> It does not matter - the objcopy creates 3GB file in both cases.
>> [EMAIL PROTECTED] ~/work/ws/linux-2.6-git $
>> /work/stage/openwrt-trunk/staging_dir/toolchain-cris_gcc4.2.3/bin/cris-linux-
On Fri, 30 May 2008 13:46:53 +0200, Hinko Kočevar
<[EMAIL PROTECTED]> wrote:
> It does not matter - the objcopy creates 3GB file in both cases.
> [EMAIL PROTECTED] ~/work/ws/linux-2.6-git $
> /work/stage/openwrt-trunk/staging_dir/toolchain-cris_gcc4.2.3/bin/cris-linux-uclibc-objcopy
> -O binary
John Crispin wrote:
>>> But I have another problem when compiling kernel with uvlibc generated
>>> toochain
>>> - the resulting arch/cris/boot/Image is 3GB large!? arch/cris/boot/zImage
>>> is
>>> later trimmed down to 'just' 4MB but still too big!!
>
> have you got debugging turned on in the
>>
>> But I have another problem when compiling kernel with uvlibc generated
>> toochain
>> - the resulting arch/cris/boot/Image is 3GB large!? arch/cris/boot/zImage is
>> later trimmed down to 'just' 4MB but still too big!!
>
have you got debugging turned on in the kconfig ?
Hinko Kočevar wrote:
> John Crispin wrote:
>
>> yes, the cris-elf seems the problem. if we have a cris-elf that can
>> translate a head.S in to a working bin, then i can get the rest running
>> in no time. but until then you need the axis toolchain installed so we
>> can use their gcc_cris -mel
John Crispin wrote:
>
> yes, the cris-elf seems the problem. if we have a cris-elf that can
> translate a head.S in to a working bin, then i can get the rest running
> in no time. but until then you need the axis toolchain installed so we
> can use their gcc_cris -melf
Great.
But I have anot
Hinko Kočevar wrote:
> John Crispin wrote:
>>> What is the etrax status in openwrt? Does it use generated uclibc toolchain
>>> or
>>> vendor (Axis) provided glibc toochain?
>> i can run the generated images on a foxboard, only draw back is, that
>> the head.S in kernel needs to be compiled with
John Crispin wrote:
>> What is the etrax status in openwrt? Does it use generated uclibc toolchain
>> or
>> vendor (Axis) provided glibc toochain?
>
> i can run the generated images on a foxboard, only draw back is, that
> the head.S in kernel needs to be compiled with the axis toolchain.
> us
> What is the etrax status in openwrt? Does it use generated uclibc toolchain
> or
> vendor (Axis) provided glibc toochain?
i can run the generated images on a foxboard, only draw back is, that
the head.S in kernel needs to be compiled with the axis toolchain.
using the stage3 elf compiler, i
John Crispin wrote:
> i will try it out the following days
Have you had the chance to test the cris compiler and kernel?
I had partial success with the uclibc toolchain that openwrt produced.
The toolchain is based on uClibc 0.29.9 and gcc 4.2.3. It manages to build the
packages but, running the
i will try it out the following days
Hinko Koc(evar wrote:
> Hi,
>
> Here are two patches that enable building of cris-foo-elf and
> cris-uclibc-linux tools inside one cris toolchain. This is needed
> because kernel head.o needs ELF based gcc to link the head.S (haven't
> found any references o
Hinko Kočevar wrote:
> [EMAIL PROTECTED] wrote:
>> nice to see someone actually use it. The build works succesfully on
>> my machine. the cris-axis-elf thin is what i was thinking of aswell.
>
> So uclibc+gcc+binutils toolchain that gets built works (is usable)?
> I'm looking for buildroot solut
[EMAIL PROTECTED] wrote:
nice to see someone actually use it. The build works succesfully on my
machine. the cris-axis-elf thin is what i was thinking of aswell.
So uclibc+gcc+binutils toolchain that gets built works (is usable)?
I'm looking for buildroot solution for our target board and came
nice to see someone actually use it. The build works succesfully on my
machine. the cris-axis-elf thin is what i was thinking of aswell.
i am on the move atm. but will contact you again about this
john
Quoting Hinko Ko?evar <[EMAIL PROTECTED]>:
> Hinko Ko?evar wrote:
>
>>
>> Next problem eme
Hinko Kočevar wrote:
>
> Next problem emerges when using $(CROSS_COMPILE)gcc instead of cris-gcc -
> toolchain that buildroot creates is cris-linux-uclibc and does not recognize
> -melf switch when linking head.o. Using $(CROSS_COMPILE)gcc fails with
> assembler
> errors Error: Illegal operan
24 matches
Mail list logo
