LFS vs 64bit arches

Bernhard Fischer rep.dot.nop at gmail.com
Sat Jan 27 13:03:44 UTC 2007


On Fri, Jan 26, 2007 at 07:09:04PM +0100, Denis Vlasenko wrote:
>On Friday 26 January 2007 12:07, Mike Frysinger wrote:
>> after seeing the readahead depending on LFS, i have to wonder without looking 
>> at any code ... are we handling LFS properly for 64bit arches ?  which is to 
>> say, it doesnt really make sense to have LFS disabled for 64bit arches ...
>
>I would be able to try answering such questions as soon
>as I will be able to run 64-bit busybox compiled against uclibc.
>
>Currently I cannot do it because uclibc seems to suffer
>from stdio bugs. At least I definitely see them,
>but uclibc mailing list is totally deaf to my mails...

For the weekend, i planned to add a BR2_PREFER_IMA to buildroot and will
then redo the patch for uClibc's GETC/PUTC in order not to pollute the
(visible) namespace. After that i can try to read and/or reproduce the
problem you were seeing with stdio (offhand i don't remember what it was
about nor if i read it in the first place, fwiw).

Then there is also whirm's soekris update that sounds like i should
generalize the so called generic x86 support i added for my own
convenience to be more generic. That'd allow for using the linux26
kernel for non-i386 and non-soekris devices with just an updated
platform/board dependant linux.config.

I also like to try out the patch for bug #881 while i'm building
buildroot anyway.

There are also some shrinkages for busybox and uClibc that i ment to
send out, we'll see if i find enough time to take care of those this
weekend..

And there's also Ulf's atmel patches which will have to wait for the
soekris/ia32 addition and cleanup to be finished first, fwiw.



More information about the busybox mailing list