find-provides.ksyms, find-requires.ksyms: rewrite indirect CRC parsing

Linux commit v5.19-rc1~139^2~2 ("kbuild: link symbol CRCs at final link,
removing CONFIG_MODULE_REL_CRCS") has broken the assumption aobut
the placement of non-absolute CRC symbols in .rodata (and also affects
the architectures that used to have these symbols as absolute);  rewrite
the parsing by utilising "objdump -t" output to figure out
the section(s) where the __crc_* symbols are stored and process
it instead of the hard-coded ".rodata" section.  The change also speeds
up the processing a bit, around 33% on synthetic tests:

    $ time sh -c 'find ./lib/modules/5.14.0-258.el9.ppc64le -name "*.ko.xz" | find-provides.ksyms.old > /dev/null 2> /dev/null'
    sh -c   10.36s user 5.58s system 137% cpu 11.613 total
    $ time sh -c 'find ./lib/modules/5.14.0-258.el9.ppc64le -name "*.ko.xz" | find-provides.ksyms.new > /dev/null 2> /dev/null'
    sh -c   7.82s user 4.59s system 142% cpu 8.686 total
    $ time sh -c 'find ./lib/modules/5.14.0-258.el9.ppc64le -name "*.ko.xz" -exec sh -c "echo {} | find-provides.ksyms.old" \; > /dev/null 2> /dev/null'
    sh -c   11.85s user 6.76s system 129% cpu 14.318 total
    $ time sh -c 'find ./lib/modules/5.14.0-258.el9.ppc64le -name "*.ko.xz" -exec sh -c "echo {} | find-provides.ksyms.new" \; > /dev/null 2> /dev/null'
    sh -c   8.91s user 5.51s system 135% cpu 10.647 total

* find-provides.ksyms: Process "objdump -t" output to get the list
of sections where __crc_* symbol contents are placed;  retrieve
each one with "readelf -R" and supply it to an awk script that cuts
the required part of it for each __crc_* symbol in that section.
* find-requires.ksyms (all_provides): Likewise.

Signed-off-by: Eugene Syromiatnikov <esyr@redhat.com>
Resolves: #2135047
This commit is contained in:
Eugene Syromiatnikov 2023-02-05 01:26:01 +01:00
parent 297a5b711f
commit cd7e9e8a2f
2 changed files with 39 additions and 33 deletions

View File

@ -42,23 +42,26 @@ for module in $(grep -E '/lib/modules/.+\.ko(\.gz|\.bz2|\.xz|\.zst)?$') "$@"; do
'match($0, /^([0-9a-f]+) A __crc_(.+)/, a) { printf("%s(%s) = 0x%08x\n", dep_pfx, a[2], strtonum("0x" a[1])) }' \
| sort -u
else
ELFRODATA=$(readelf -R .rodata $module | awk '/0x/{printf $2$3$4$5}')
objdump -t "$module" \
| sed -n 's/^[0-9a-f][0-9a-f]* g...... \(.*\) [0-9a-f][0-9a-f]* __crc_.*$/\1/p' \
| sort -u \
| while read sectname; do
[ -n "$sectname" ] || continue
ELFSECTDATA=$(readelf -R "$sectname" "$module" | awk '/0x/{printf $2$3$4$5}')
if [[ -n $(readelf -h $module | grep "little endian") ]]; then
RODATA=$(echo $ELFRODATA | sed 's/\(..\)\(..\)\(..\)\(..\)/\4\3\2\1/g')
SECTDATA=$(echo $ELFSECTDATA | sed 's/\(..\)\(..\)\(..\)\(..\)/\4\3\2\1/g')
else
RODATA=$ELFRODATA
SECTDATA=$ELFSECTDATA
fi
# Commit binutils-2_33~1385[1] has changed (and binutils-2_35~1768[2]
# has not reverted it) the calculated type for symbols in read-write
# .rodata section from 'R' to 'D', since, apparently, many kernel
# modules have it indeed read-write.
#
# [1] https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=a288c270991d
# [2] https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=49d9fd42acef
for sym in $(nm $module | sed -r -ne 's:^0*([0-9a-f]+) [DR] __crc_(.+):0x\1 \2:p'); do
echo $sym $RODATA
objdump -t "$module" \
| awk \
-v 'dep_pfx='"$dep_pfx" \
-v 'sectdata='"$SECTDATA" \
--non-decimal-data \
'match($0, /^([0-9a-f]+) g...... .* [0-9a-f]+ __crc_(.*)$/, a) { printf("%s(%s) = 0x%08s\n", dep_pfx, a[2], substr(sectdata, (strtonum("0x" a[1]) * 2) + 1, 8)) }'
done \
| awk --non-decimal-data '{printf("'"${dep_pfx}"'(%s) = 0x%08s\n", $2, substr($3,($1*2)+1,8))}' \
| sort -u
fi

View File

@ -42,23 +42,26 @@ all_provides() {
--non-decimal-data \
'match($0, /^([0-9a-f]+) A __crc_(.+)/, a) { printf("%s(%s) = 0x%08x\n", dep_pfx, a[2], strtonum("0x" a[1])) }'
else
ELFRODATA=$(readelf -R .rodata "$module" | awk '/0x/{printf $2$3$4$5}')
objdump -t "$module" \
| sed -n 's/^[0-9a-f][0-9a-f]* g...... \(.*\) [0-9a-f][0-9a-f]* __crc_.*$/\1/p' \
| sort -u \
| while read sectname; do
[ -n "$sectname" ] || continue
ELFSECTDATA=$(readelf -R .rodata "$module" | awk '/0x/{printf $2$3$4$5}')
if [[ -n $(readelf -h "$module" | grep "little endian") ]]; then
RODATA=$(echo $ELFRODATA | sed 's/\(..\)\(..\)\(..\)\(..\)/\4\3\2\1/g')
SECTDATA=$(echo $ELFSECTDATA | sed 's/\(..\)\(..\)\(..\)\(..\)/\4\3\2\1/g')
else
RODATA=$ELFRODATA
SECTDATA=$ELFSECTDATA
fi
# Commit binutils-2_33~1385[1] has changed (and binutils-2_35~1768[2]
# has not reverted it) the calculated type for symbols in read-write
# .rodata section from 'R' to 'D', since, apparently, many kernel
# modules have it indeed read-write.
#
# [1] https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=a288c270991d
# [2] https://sourceware.org/git/?p=binutils-gdb.git;a=commitdiff;h=49d9fd42acef
for sym in $(nm "$module" | sed -r -ne 's:^0*([0-9a-f]+) [DR] __crc_(.+):0x\1 \2:p'); do
echo $sym $RODATA
done \
| awk --non-decimal-data '{printf("%s:0x%08s\n", $2, substr($3,($1*2)+1,8))}'
objdump -t "$module" \
| awk \
-v 'dep_pfx='"$dep_pfx" \
-v 'sectdata='"$SECTDATA" \
--non-decimal-data \
'match($0, /^([0-9a-f]+) g...... .* [0-9a-f]+ __crc_(.*)$/, a) { printf("%s(%s) = 0x%08s\n", dep_pfx, a[2], substr(sectdata, (strtonum("0x" a[1]) * 2) + 1, 8)) }'
done
fi
[ -z "$tmpfile" ] || rm -f -- "$tmpfile"