Date: Fri, 23 Feb 2001 16:00:04 -0800 From: Marcel Moolenaar <marcel@cup.hp.com> To: John Baldwin <jhb@FreeBSD.ORG> Cc: arch@FreeBSD.ORG Subject: Re: ELF, OS and ABI [was: Re: sysctl kern.fallback_elf_brand] Message-ID: <3A96F984.7233C733@cup.hp.com> References: <XFMail.010223143316.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
John Baldwin wrote: > > > The use of the section seems broken to me. It's the first time I've > > heard of it (so I won't really know how it's used :-). AFAICT, ABI > > information is written in e_ident[EI_OSABI] and annotated with flags in > > e_flags. That should be enough for our ELF loader. > > I talked with O`Brien about this today. Many other people view this field as > being used when you _extend_ the ELF specification itself. Not as a mechanism > for running a plain ELF binary and marking which set of system calls, etc. it > assumes. I'm not sure that's a valid view, because ELF was originally defined as *part* of an ABI. Consequently, it does not *define* an ABI. With the adoption of ELF by different ABI providers, the lack of ABI information in ELF was apparent. It seems to me that the addition of EI_OSABI is to address this shortcoming in the ELF specification and further allows ABI specific flags to be used in the ELF file. The ELF modification/extension architects which fields and which bits in fields you can use for OS or machine specific information. Do any of you have pointers to ELF specifications that document the ABI section. The ELF specifications I use can be found: http://developer.intel.com/vtune/tis.htm http://developer.intel.com/design/ia-64/downloads/245370.htm -- Marcel Moolenaar mail: marcel@cup.hp.com / marcel@FreeBSD.org tel: (408) 447-4222 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3A96F984.7233C733>