kernel_optimize_test/arch/powerpc/tools/head_check.sh
Nicholas Piggin c494adefef powerpc/64: Tool to check head sections location sanity
Use a tool to check that the location of "fixed sections" are where
we expected them to be, which catches cases the linker script can't
(stubs being added to start of .text section), and which ends up
being neater.

Sample output:

  ERROR: start_text address is c000000000008100, should be c000000000008000
  ERROR: see comments in arch/powerpc/tools/head_check.sh

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
[mpe: Fold in fix from Nick for 4.6 era toolchains]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
2017-05-30 14:59:51 +10:00

79 lines
2.9 KiB
Bash

# Copyright © 2016 IBM Corporation
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version
# 2 of the License, or (at your option) any later version.
# This script checks the head of a vmlinux for linker stubs that
# break our placement of fixed-location code for 64-bit.
# based on relocs_check.pl
# Copyright © 2009 IBM Corporation
# NOTE!
#
# If the build dies here, it's likely code in head_64.S/exception-64*.S or
# nearby, is branching to labels it can't reach directly, which results in the
# linker inserting branch stubs. This can move code around in ways that break
# the fixed section calculations (head-64.h). To debug this, disassemble the
# vmlinux and look for branch stubs (long_branch, plt_branch, etc.) in the
# fixed section region (0 - 0x8000ish). Check what code is calling those stubs,
# and perhaps change so a direct branch can reach.
#
# A ".linker_stub_catch" section is used to catch some stubs generated by
# early .text code, which tend to get placed at the start of the section.
# If there are too many such stubs, they can overflow this section. Expanding
# it may help (or reducing the number of stub branches).
#
# Linker stubs use the TOC pointer, so even if fixed section code could
# tolerate them being inserted into head code, they can't be allowed in low
# level entry code (boot, interrupt vectors, etc) until r2 is set up. This
# could cause the kernel to die in early boot.
# Turn this on if you want more debug output:
# set -x
if [ $# -lt 2 ]; then
echo "$0 [path to nm] [path to vmlinux]" 1>&2
exit 1
fi
# Have Kbuild supply the path to nm so we handle cross compilation.
nm="$1"
vmlinux="$2"
# gcc-4.6-era toolchain make _stext an A (absolute) symbol rather than T
$nm "$vmlinux" | grep -e " [TA] _stext$" -e " t start_first_256B$" -e " a text_start$" -e " t start_text$" -m4 > .tmp_symbols.txt
vma=$(cat .tmp_symbols.txt | grep -e " [TA] _stext$" | cut -d' ' -f1)
expected_start_head_addr=$vma
start_head_addr=$(cat .tmp_symbols.txt | grep " t start_first_256B$" | cut -d' ' -f1)
if [ "$start_head_addr" != "$expected_start_head_addr" ]; then
echo "ERROR: head code starts at $start_head_addr, should be $expected_start_head_addr"
echo "ERROR: try to enable LD_HEAD_STUB_CATCH config option"
echo "ERROR: see comments in arch/powerpc/tools/head_check.sh"
exit 1
fi
top_vma=$(echo $vma | cut -d'0' -f1)
expected_start_text_addr=$(cat .tmp_symbols.txt | grep " a text_start$" | cut -d' ' -f1 | sed "s/^0/$top_vma/")
start_text_addr=$(cat .tmp_symbols.txt | grep " t start_text$" | cut -d' ' -f1)
if [ "$start_text_addr" != "$expected_start_text_addr" ]; then
echo "ERROR: start_text address is $start_text_addr, should be $expected_start_text_addr"
echo "ERROR: try to enable LD_HEAD_STUB_CATCH config option"
echo "ERROR: see comments in arch/powerpc/tools/head_check.sh"
exit 1
fi
rm -f .tmp_symbols.txt