Hi Michael, The tls.pdf (ELF Handling for Thread Local Storage)
#define R_386_TLS_TPOFF 14 /* Negative offset in the static TLS Block */ #define R_386_TLS_TPOFF32 37 /* Offset in TLS Static Block */ However the Solaris uses on R_386_TLS_TPOFF and not R_386_TLS_TPOFF32 in the file elf_386.h. Regards Deepak Bhatia -----Original Message----- From: michael.s.walker at gmail.com [mailto:michael.s.wal...@gmail.com]on Behalf Of Michael Walker Sent: Thursday, February 22, 2007 10:21 PM To: Deepak Bhatia Cc: tools-linking at opensolaris.org Subject: Re: [tools-linking] relative relocation I'm not familiar with the ARM processor - so I can't say definitively one way or the other. But I would assume so - yes. Have TLS instruction sequences been defined for the ARM processor? That would answer your question I think. _Mike_ On 2/22/07, Deepak Bhatia <deepak.bhatia at iap-online.com> wrote: > Hi Michael, > > I see that you have written the TLS related stuff for Sparc and i386. Does > the ARM > has the same implications for various TLS models like General Dynamic, Local > Dynamic, > , Initial Exec and local-exec. > > Regards > > Deepak Bhatia > > -----Original Message----- > From: tools-linking-bounces at opensolaris.org > [mailto:tools-linking-bounces at opensolaris.org]On Behalf Of Michael Walker > Sent: Thursday, February 22, 2007 9:49 PM > To: bhushan verma > Cc: tools-linking at opensolaris.org > Subject: Re: [tools-linking] relative relocation > > > Here's a quick example: > > collage0021# elfdump -c data.so > collage0021# more data.c > static int local_var = 0; > int *global_data = &local_var; > collage0021# gcc -fpic -shared data.c -o data.so > collage0021# elfdump -r data.so | grep RELATIVE > R_386_RELATIVE 0x105cc .rel.data.rel. > R_386_RELATIVE 0x105d0 .rel.data.rel. > collage0021# elfdump -r data.so | grep RELATIVE > collage0021# nm -xv data.so | grep 105d0 > [63] |0x000105d0|0x00000004|OBJT |GLOB |0 |20 |global_data > collage0021# > > Notice that the RELATIVE relocation against 0x105d0 is against the > 'global_data' > symbol. So - it's going to adjust the value of '&local_data' (which > is already stored > there) with the offset of the library (data.so) when it's loaded into its > final > location in memory. > > _Mike_ > > > > On 2/21/07, bhushan verma <bhushan.verma at iap-online.net> wrote: > > Hi > > can anyone give me an c example which produces R_386_RELATIVE type > relocation > > Thanks > > > > > > This message posted from opensolaris.org > > _______________________________________________ > > tools-linking mailing list > > tools-linking at opensolaris.org > > > > > > > -- > Michael Walker > mwalker at fins.com > _______________________________________________ > tools-linking mailing list > tools-linking at opensolaris.org > > > -- Michael Walker mwalker at fins.com