tech-toolchain archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Programmatically using NetBSD Version in ELF binaries



While learning how to write assembly code on NetBSD, I decided to extend
readelf.c (from GNU binutils) to show the NT_VERSION value in an ELF
header, which is the NetBSD version in the format of the
__NetBSD_Version__ macro from sys/param.h.  The value comes out as a 32-bit
integer: for example 601000500 for 6.1.5.  How do I handle that value
in a C program?  If I follow the logic of the __NetBSD_Prereq__ macro,
only the first digit of each field (MMmmrrpp major/minor/0/patch level) can
be used so major version 6 is represented as 60.  What if it goes to 10 or
higher?  While researching the source code for this issue, I found out that
the file utility in src/external/bsd/file/dist/src/readelf.c handles it the old
way still:

	 * M = major version
	 * m = minor version
	 * r = release ["",A-Z,Z[A-Z] but numeric]
	 * p = patchlevel

...which I understand was depreciated in 2004, so if I fake the value 99 for
minor version (to indicate current) in the assembler section note, the file
command shows "...for NetBSD 6.9ZZZL...".  If I put in a major version 10
in the assembly source, then my current binutils readelf.c changes (based on
the prereq macro's logic) shows version 1 instad of 10.



Home | Main Index | Thread Index | Old Index