| Message ID | cc280176-7383-0600-9298-f6c981475245@rimuhosting.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | livepatch-build-tools: Detect special section group sizes | expand |
On 4/12/19 5:50 AM, Glenn Enright wrote: > A recent xsa livepatch failed to generate due to the following > message in create-diff-object.log ... > > /livepatch-build-tools/create-diff-object: ERROR: grant_table.o: > kpatch_regenerate_special_section: 1162: group size mismatch for section > .altinstructions > > This is similar to the issue reported and fixed in > https://github.com/dynup/kpatch/pull/528 which says ... > "Hard-coding the special section group sizes is unreliable. > Instead, determine them dynamically by finding the related > struct definitions in the DWARF metadata." > > Signed-off-by: Glenn Enright <glenn@rimuhosting.com> > --- > CC: Ross Lagerwall <ross.lagerwall@citrix.com> > CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > This patch resulted in a loadable livepatch. The alt section size in my > case was actually 14. Thanks for the patch! Reviewed-by: Ross Lagerwall <ross.lagerwall@citrix.com>
On Tue, Apr 16, 2019 at 11:05:49AM +0100, Ross Lagerwall wrote: > On 4/12/19 5:50 AM, Glenn Enright wrote: > > A recent xsa livepatch failed to generate due to the following > > message in create-diff-object.log ... > > > > /livepatch-build-tools/create-diff-object: ERROR: grant_table.o: > > kpatch_regenerate_special_section: 1162: group size mismatch for section > > .altinstructions > > > > This is similar to the issue reported and fixed in > > https://github.com/dynup/kpatch/pull/528 which says ... > > "Hard-coding the special section group sizes is unreliable. > > Instead, determine them dynamically by finding the related > > struct definitions in the DWARF metadata." > > > > Signed-off-by: Glenn Enright <glenn@rimuhosting.com> > > --- > > CC: Ross Lagerwall <ross.lagerwall@citrix.com> > > CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > > > > This patch resulted in a loadable livepatch. The alt section size in my > > case was actually 14. > Thanks for the patch! > > Reviewed-by: Ross Lagerwall <ross.lagerwall@citrix.com> Could you resend it with the Reviewed-by please and include me on the To list please?
diff --git a/create-diff-object.c b/create-diff-object.c index 82f777e..f9f4abf 100644 --- a/create-diff-object.c +++ b/create-diff-object.c @@ -962,9 +962,35 @@ static int bug_frames_0_group_size(struct kpatch_elf *kelf, int offset) { return static int bug_frames_1_group_size(struct kpatch_elf *kelf, int offset) { return 8; } static int bug_frames_2_group_size(struct kpatch_elf *kelf, int offset) { return 8; } static int bug_frames_3_group_size(struct kpatch_elf *kelf, int offset) { return 16; } -static int ex_table_group_size(struct kpatch_elf *kelf, int offset) { return 8; } -static int altinstructions_group_size(struct kpatch_elf *kelf, int offset) { return 12; } +int ex_table_group_size(struct kpatch_elf *kelf, int offset) +{ + static int size = 0; + char *str; + + if (!size) { + str = getenv("EX_STRUCT_SIZE"); + if (!str) + ERROR("EX_STRUCT_SIZE not set"); + size = atoi(str); + } + + return size; +} +int altinstructions_group_size(struct kpatch_elf *kelf, int offset) +{ + static int size = 0; + char *str; + + if (!size) { + str = getenv("ALT_STRUCT_SIZE"); + if (!str) + ERROR("ALT_STRUCT_SIZE not set"); + size = atoi(str); + } + + return size; +} /* * The rela groups in the .fixup section vary in size. The beginning of each * .fixup rela group is referenced by the .ex_table section. To find the size diff --git a/livepatch-build b/livepatch-build index c057fa1..6c3409c 100755 --- a/livepatch-build +++ b/livepatch-build @@ -304,6 +304,27 @@ if [ "${SKIP}" != "build" ]; then XEN_DEBUG="debug=$XEN_DEBUG" fi + echo "Reading special section data" + SPECIAL_VARS=$(readelf -wi "$XENSYMS" | + gawk --non-decimal-data ' + BEGIN { a = e = 0 } + a == 0 && /DW_AT_name.* alt_instr$/ {a = 1; next} + e == 0 && /DW_AT_name.* exception_table_entry$/ {e = 1; next} + a == 1 {printf("export ALT_STRUCT_SIZE=%d\n", $4); a = 2} + e == 1 {printf("export EX_STRUCT_SIZE=%d\n", $4); e = 2} + a == 2 && b == 2 && p == 2 && e == 2 {exit}') + + [[ -n $SPECIAL_VARS ]] && eval "$SPECIAL_VARS" + + if [[ -z $ALT_STRUCT_SIZE ]] || [[ -z $EX_STRUCT_SIZE ]]; then + die "can't find special struct size" + fi + for i in $ALT_STRUCT_SIZE $EX_STRUCT_SIZE; do + if [[ ! $i -gt 0 ]] || [[ ! $i -le 16 ]]; then + die "invalid special struct size $i" + fi + done + echo "Perform full initial build with ${CPUS} CPU(s)..." build_full
A recent xsa livepatch failed to generate due to the following message in create-diff-object.log ... /livepatch-build-tools/create-diff-object: ERROR: grant_table.o: kpatch_regenerate_special_section: 1162: group size mismatch for section .altinstructions This is similar to the issue reported and fixed in https://github.com/dynup/kpatch/pull/528 which says ... "Hard-coding the special section group sizes is unreliable. Instead, determine them dynamically by finding the related struct definitions in the DWARF metadata." Signed-off-by: Glenn Enright <glenn@rimuhosting.com> --- CC: Ross Lagerwall <ross.lagerwall@citrix.com> CC: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> This patch resulted in a loadable livepatch. The alt section size in my case was actually 14. --- create-diff-object.c | 30 ++++++++++++++++++++++++++++-- livepatch-build | 21 +++++++++++++++++++++ 2 files changed, 49 insertions(+), 2 deletions(-)