Skip to contentSkip to conversation

failed to compile glibc 2.1.2 - BFD_RELOC_16_PCREL_S2 problem

DISCUSSION Ulf Carlsson 5 participants 6 messages Apr 25, 2000 – Apr 28, 2000 2 messages not in the archive MIPS

lore
Share & tools
lore
#1 lore raw
In reply to a message not in the archive
Quoted 4 lines · Jun Sun

I am having this problem while I am trying to build glibc v2.1.2. Does anybody know about this problem? Do I need some MIPS patch for building this? Thanks.

You can't use glibc 2.1.2 on MIPS, try glibc 2.2 from the official cvs tree instead. Andreas Jaeger posted an announcement for glibc 2.2 on MIPS to this list last Friday with build instructions.

Quoted 2 lines

BFD_RELOC_16_PCREL_S2 relocation in this object file format make[2]: *** [/root/rpm/BUILD/glibc-obj/setjmp/setjmp.o] Error 1

This is because it's not compiling as PIC.

Ulf

#3 lore raw
In reply to a message not in the archive
Quoted 11 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line

Jun Sun writes:

Quoted 8 lines · Ulf Carlsson
Quoted 4 lines · Jun Sun

I am having this problem while I am trying to build glibc v2.1.2. Does anybody know about this problem? Do I need some MIPS patch for building this? Thanks.

Quoted 5 lines
Quoted 1 line

You can't use glibc 2.1.2 on MIPS,

Is there a specific reason why I can't use glibc 2.1.2? We have been using this version for other platforms (ppc, i386). It would be nice to stick with the same version for MIPS.

It's just that nobody did the work of porting glibc. A lot of has changed between glibc 2.0 and 2.1 (and also between 2.1 and 2.2). I've decided to port glibc 2.2 for MIPS. This involved a number of patches and but I don't have the time to backport everything - and I will not support such a backport at all.

All my glibc 2.2 patches have been added to the glibc archives and therefore will be in the official sources of 2.2.

Andreas

Signature · 3 lines
 Andreas Jaeger
  SuSE Labs [email protected]
   private [email protected]
#4 lore raw
Quoted 8 lines · Andreas Jaeger

It's just that nobody did the work of porting glibc. A lot of has changed between glibc 2.0 and 2.1 (and also between 2.1 and 2.2). I've decided to port glibc 2.2 for MIPS. This involved a number of patches and but I don't have the time to backport everything - and I will not support such a backport at all.

All my glibc 2.2 patches have been added to the glibc archives and therefore will be in the official sources of 2.2.

All in all I think that given the work that has been invested into making a solid 2.2 port 2.2 might soon be the better choice than 2.0 and source compatibility issues should make a port of 2.1 mostly a non-issue.

Ralf

#5 lore raw
Quoted 3 lines · Ralf Baechle

All in all I think that given the work that has been invested into making a solid 2.2 port 2.2 might soon be the better choice than 2.0 and source compatibility issues should make a port of 2.1 mostly a non-issue.

If my builds of 2.1 are any indication:

2.0 may still be the better choice on very space-constrained platforms until there are good options for reducing the size of 2.2. I mean, 2.2 is not huge, but on devices with 2M non-volatile store *total*, it's an obvious target.

I am not volunteering to work on this (yet), but I suspect that people who care intensely about size will have to start digging around in the link maps. H. J. Lu has done some of the work in this direction already.

On the other hand, maybe the really tiny devices will go away.

Jay

#6 lore raw
Quoted 1 line · I wrote:

2.0 may still be the better choice on very space-constrained platforms

until

Quoted 3 lines

there are good options for reducing the size of 2.2. I mean, 2.2 is not huge, but on devices with 2M non-volatile store *total*, it's an obvious target.

To which I should add: I am very grateful for the work on 2.2. It needs to be done. Many people will benefit from it, including me. Great thanks to Andreas!

The problems with code size are not Andreas's fault... :-)

Jay

#7 lore raw
In reply to Jay Carlson
Quoted 8 lines · Jay Carlson

2.0 may still be the better choice on very space-constrained platforms until there are good options for reducing the size of 2.2. I mean, 2.2 is not huge, but on devices with 2M non-volatile store *total*, it's an obvious target.

I am not volunteering to work on this (yet), but I suspect that people who care intensely about size will have to start digging around in the link maps. H. J. Lu has done some of the work in this direction already.

FYI, on PPC some people are working on kernel size optimization, ranging from automatically throwing away never used functions and data at link time to getting rid of all the kernel messages (I heard that's about 150 kB of textual data).

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- Linux/{m68k~Amiga,PPC~CHRP} -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds