Re: CVS commit: src/sys/lib/libsa

[Valery Ushakov <uwe%stderr.spb.ru@localhost>]

On Thu, Feb 23, 2017 at 12:13:05 +0000, NONAKA Kimihiro wrote:

> Module Name:	src
> Committed By:	nonaka
> Date:		Thu Feb 23 12:13:05 UTC 2017
> 
> Modified Files:
> 	src/sys/lib/libsa: loadfile_elf32.c
> 
> Log Message:
> fix CTF section symbol size was not counted with COUNT_KERNEL.
> 
> 
> To generate a diff of this commit:
> cvs rdiff -u -r1.34 -r1.35 src/sys/lib/libsa/loadfile_elf32.c

This change looks suspicious to me.

Reading shstrtab into the elf image we construct for the kernel
symbols is still protected by LOAD_SYM check, but the second part is
no longer protected, so maxp is advanced unconditionally and sh_offset
is also always updated.

What is really intended here?  Do we need shstr set so that we can
test for ".SUNW_ctf" section name later?

Thanks.

-uwe