scripts/faddr2line | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
This reverts commit c02904f05ff805d6c0631634d5751ebd338f75ec.
Such commit assumed that only two symbols are relevant for the symbol
size calculation. However, this can lead to an incorrect symbol size
calculation when there are mapping symbols emitted by readelf.
For instance, when feeding 'update_irq_load_avg+0x1c/0x1c4', faddr2line
might need to processes the following readelf lines:
784284: ffffffc0081cca30 428 FUNC GLOBAL DEFAULT 2 update_irq_load_avg
87319: ffffffc0081ccb0c 0 NOTYPE LOCAL DEFAULT 2 $x.62522
87321: ffffffc0081ccbdc 0 NOTYPE LOCAL DEFAULT 2 $x.62524
87323: ffffffc0081ccbe0 0 NOTYPE LOCAL DEFAULT 2 $x.62526
87325: ffffffc0081ccbe4 0 NOTYPE LOCAL DEFAULT 2 $x.62528
87327: ffffffc0081ccbe8 0 NOTYPE LOCAL DEFAULT 2 $x.62530
87329: ffffffc0081ccbec 0 NOTYPE LOCAL DEFAULT 2 $x.62532
87331: ffffffc0081ccbf0 0 NOTYPE LOCAL DEFAULT 2 $x.62534
87332: ffffffc0081ccbf4 0 NOTYPE LOCAL DEFAULT 2 $x.62535
783403: ffffffc0081ccbf4 424 FUNC GLOBAL DEFAULT 2 sched_pelt_multiplier
The symbol size of 'update_irq_load_avg' should be calculated with the
address of 'sched_pelt_multiplier', after skipping the mapping symbols
seen in between. However, the offending commit cuts the list short and
faddr2line incorrectly assumes 'update_irq_load_avg' is the last symbol
in the section, resulting in:
$ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4
skipping update_irq_load_avg address at 0xffffffc0081cca4c due to size mismatch (0x1c4 != 0x3ff9a59988)
no match for update_irq_load_avg+0x1c/0x1c4
After reverting the commit the issue is resolved:
$ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4
update_irq_load_avg+0x1c/0x1c4:
cpu_of at kernel/sched/sched.h:1109
(inlined by) update_irq_load_avg at kernel/sched/pelt.c:481
Cc: Brian Johannesmeyer <bjohannesmeyer@gmail.com>
Cc: John Stultz <jstultz@google.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Carlos Llamas <cmllamas@google.com>
---
scripts/faddr2line | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/scripts/faddr2line b/scripts/faddr2line
index fe0cc45f03be..1fa6beef9f97 100755
--- a/scripts/faddr2line
+++ b/scripts/faddr2line
@@ -252,7 +252,7 @@ __faddr2line() {
found=2
break
fi
- done < <(echo "${ELF_SYMS}" | sed 's/\[.*\]//' | ${AWK} -v sec=$sym_sec '$7 == sec' | sort --key=2 | ${GREP} -A1 --no-group-separator " ${sym_name}$")
+ done < <(echo "${ELF_SYMS}" | sed 's/\[.*\]//' | ${AWK} -v sec=$sym_sec '$7 == sec' | sort --key=2)
if [[ $found = 0 ]]; then
warn "can't find symbol: sym_name: $sym_name sym_sec: $sym_sec sym_addr: $sym_addr sym_elf_size: $sym_elf_size"
--
2.46.0.76.ge559c4bf1a-googOn Mon, Aug 12, 2024 at 11:01:20PM +0000, Carlos Llamas wrote: > This reverts commit c02904f05ff805d6c0631634d5751ebd338f75ec. > > Such commit assumed that only two symbols are relevant for the symbol > size calculation. However, this can lead to an incorrect symbol size > calculation when there are mapping symbols emitted by readelf. > > For instance, when feeding 'update_irq_load_avg+0x1c/0x1c4', faddr2line > might need to processes the following readelf lines: > > 784284: ffffffc0081cca30 428 FUNC GLOBAL DEFAULT 2 update_irq_load_avg > 87319: ffffffc0081ccb0c 0 NOTYPE LOCAL DEFAULT 2 $x.62522 > 87321: ffffffc0081ccbdc 0 NOTYPE LOCAL DEFAULT 2 $x.62524 > 87323: ffffffc0081ccbe0 0 NOTYPE LOCAL DEFAULT 2 $x.62526 > 87325: ffffffc0081ccbe4 0 NOTYPE LOCAL DEFAULT 2 $x.62528 > 87327: ffffffc0081ccbe8 0 NOTYPE LOCAL DEFAULT 2 $x.62530 > 87329: ffffffc0081ccbec 0 NOTYPE LOCAL DEFAULT 2 $x.62532 > 87331: ffffffc0081ccbf0 0 NOTYPE LOCAL DEFAULT 2 $x.62534 > 87332: ffffffc0081ccbf4 0 NOTYPE LOCAL DEFAULT 2 $x.62535 > 783403: ffffffc0081ccbf4 424 FUNC GLOBAL DEFAULT 2 sched_pelt_multiplier > > The symbol size of 'update_irq_load_avg' should be calculated with the > address of 'sched_pelt_multiplier', after skipping the mapping symbols > seen in between. However, the offending commit cuts the list short and > faddr2line incorrectly assumes 'update_irq_load_avg' is the last symbol > in the section, resulting in: > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > skipping update_irq_load_avg address at 0xffffffc0081cca4c due to size mismatch (0x1c4 != 0x3ff9a59988) > no match for update_irq_load_avg+0x1c/0x1c4 > > After reverting the commit the issue is resolved: > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > update_irq_load_avg+0x1c/0x1c4: > cpu_of at kernel/sched/sched.h:1109 > (inlined by) update_irq_load_avg at kernel/sched/pelt.c:481 > > Cc: Brian Johannesmeyer <bjohannesmeyer@gmail.com> > Cc: John Stultz <jstultz@google.com> > Cc: Will Deacon <will@kernel.org> > Signed-off-by: Carlos Llamas <cmllamas@google.com> > --- > scripts/faddr2line | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) Acked-by: Will Deacon <will@kernel.org> I'm assuming that Josh will pick this up. Cheers, Will
On Fri, Aug 16, 2024 at 01:50:46PM +0100, Will Deacon wrote: > On Mon, Aug 12, 2024 at 11:01:20PM +0000, Carlos Llamas wrote: > > This reverts commit c02904f05ff805d6c0631634d5751ebd338f75ec. > > > > Such commit assumed that only two symbols are relevant for the symbol > > size calculation. However, this can lead to an incorrect symbol size > > calculation when there are mapping symbols emitted by readelf. > > > > For instance, when feeding 'update_irq_load_avg+0x1c/0x1c4', faddr2line > > might need to processes the following readelf lines: > > > > 784284: ffffffc0081cca30 428 FUNC GLOBAL DEFAULT 2 update_irq_load_avg > > 87319: ffffffc0081ccb0c 0 NOTYPE LOCAL DEFAULT 2 $x.62522 > > 87321: ffffffc0081ccbdc 0 NOTYPE LOCAL DEFAULT 2 $x.62524 > > 87323: ffffffc0081ccbe0 0 NOTYPE LOCAL DEFAULT 2 $x.62526 > > 87325: ffffffc0081ccbe4 0 NOTYPE LOCAL DEFAULT 2 $x.62528 > > 87327: ffffffc0081ccbe8 0 NOTYPE LOCAL DEFAULT 2 $x.62530 > > 87329: ffffffc0081ccbec 0 NOTYPE LOCAL DEFAULT 2 $x.62532 > > 87331: ffffffc0081ccbf0 0 NOTYPE LOCAL DEFAULT 2 $x.62534 > > 87332: ffffffc0081ccbf4 0 NOTYPE LOCAL DEFAULT 2 $x.62535 > > 783403: ffffffc0081ccbf4 424 FUNC GLOBAL DEFAULT 2 sched_pelt_multiplier > > > > The symbol size of 'update_irq_load_avg' should be calculated with the > > address of 'sched_pelt_multiplier', after skipping the mapping symbols > > seen in between. However, the offending commit cuts the list short and > > faddr2line incorrectly assumes 'update_irq_load_avg' is the last symbol > > in the section, resulting in: > > > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > > skipping update_irq_load_avg address at 0xffffffc0081cca4c due to size mismatch (0x1c4 != 0x3ff9a59988) > > no match for update_irq_load_avg+0x1c/0x1c4 > > > > After reverting the commit the issue is resolved: > > > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > > update_irq_load_avg+0x1c/0x1c4: > > cpu_of at kernel/sched/sched.h:1109 > > (inlined by) update_irq_load_avg at kernel/sched/pelt.c:481 > > > > Cc: Brian Johannesmeyer <bjohannesmeyer@gmail.com> > > Cc: John Stultz <jstultz@google.com> > > Cc: Will Deacon <will@kernel.org> > > Signed-off-by: Carlos Llamas <cmllamas@google.com> > > --- > > scripts/faddr2line | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > Acked-by: Will Deacon <will@kernel.org> > > I'm assuming that Josh will pick this up. I don't think this was picked up yet, at least I don't see it in tip. Or maybe I looked in the wrong place? Cheers, Carlos Llamas
On Thu, Oct 17, 2024 at 12:12:16AM +0000, Carlos Llamas wrote: > On Fri, Aug 16, 2024 at 01:50:46PM +0100, Will Deacon wrote: > > Acked-by: Will Deacon <will@kernel.org> > > > > I'm assuming that Josh will pick this up. > > I don't think this was picked up yet, at least I don't see it in tip. Or > maybe I looked in the wrong place? Queued, thanks! -- Josh
On Fri, Aug 16, 2024 at 5:50 AM Will Deacon <will@kernel.org> wrote: > > On Mon, Aug 12, 2024 at 11:01:20PM +0000, Carlos Llamas wrote: > > This reverts commit c02904f05ff805d6c0631634d5751ebd338f75ec. > > > > Such commit assumed that only two symbols are relevant for the symbol > > size calculation. However, this can lead to an incorrect symbol size > > calculation when there are mapping symbols emitted by readelf. > > > > For instance, when feeding 'update_irq_load_avg+0x1c/0x1c4', faddr2line > > might need to processes the following readelf lines: > > > > 784284: ffffffc0081cca30 428 FUNC GLOBAL DEFAULT 2 update_irq_load_avg > > 87319: ffffffc0081ccb0c 0 NOTYPE LOCAL DEFAULT 2 $x.62522 > > 87321: ffffffc0081ccbdc 0 NOTYPE LOCAL DEFAULT 2 $x.62524 > > 87323: ffffffc0081ccbe0 0 NOTYPE LOCAL DEFAULT 2 $x.62526 > > 87325: ffffffc0081ccbe4 0 NOTYPE LOCAL DEFAULT 2 $x.62528 > > 87327: ffffffc0081ccbe8 0 NOTYPE LOCAL DEFAULT 2 $x.62530 > > 87329: ffffffc0081ccbec 0 NOTYPE LOCAL DEFAULT 2 $x.62532 > > 87331: ffffffc0081ccbf0 0 NOTYPE LOCAL DEFAULT 2 $x.62534 > > 87332: ffffffc0081ccbf4 0 NOTYPE LOCAL DEFAULT 2 $x.62535 > > 783403: ffffffc0081ccbf4 424 FUNC GLOBAL DEFAULT 2 sched_pelt_multiplier > > > > The symbol size of 'update_irq_load_avg' should be calculated with the > > address of 'sched_pelt_multiplier', after skipping the mapping symbols > > seen in between. However, the offending commit cuts the list short and > > faddr2line incorrectly assumes 'update_irq_load_avg' is the last symbol > > in the section, resulting in: > > > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > > skipping update_irq_load_avg address at 0xffffffc0081cca4c due to size mismatch (0x1c4 != 0x3ff9a59988) > > no match for update_irq_load_avg+0x1c/0x1c4 > > > > After reverting the commit the issue is resolved: > > > > $ scripts/faddr2line vmlinux update_irq_load_avg+0x1c/0x1c4 > > update_irq_load_avg+0x1c/0x1c4: > > cpu_of at kernel/sched/sched.h:1109 > > (inlined by) update_irq_load_avg at kernel/sched/pelt.c:481 > > > > Cc: Brian Johannesmeyer <bjohannesmeyer@gmail.com> > > Cc: John Stultz <jstultz@google.com> > > Cc: Will Deacon <will@kernel.org> > > Signed-off-by: Carlos Llamas <cmllamas@google.com> > > --- > > scripts/faddr2line | 2 +- > > 1 file changed, 1 insertion(+), 1 deletion(-) > > Acked-by: Will Deacon <will@kernel.org> > > I'm assuming that Josh will pick this up. > > Cheers, > > Will Acked-by: Brian Johannesmeyer <bjohannesmeyer@gmail.com> Good catch. Thanks for the fix. Just curious: What compiler did you use to build the kernel, where you hit this error with 'update_irq_load_avg'? I was unable to reproduce it with LLVM 14. Thanks! -Brian -Brian
On Tue, Aug 20, 2024 at 09:38:46AM -0700, Brian Johannesmeyer wrote: > > Acked-by: Brian Johannesmeyer <bjohannesmeyer@gmail.com> > > Good catch. Thanks for the fix. > > Just curious: What compiler did you use to build the kernel, where you > hit this error with 'update_irq_load_avg'? I was unable to reproduce > it with LLVM 14. Thanks! > You should be able to reproduce this issue as long as you hit a NOTYPE symbol as the second entry. Using aarch64 and CFI_CLANG will make this scenario more likely to happen. I believe I used the following: $ make LLVM=1 ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- defconfig $ ./scripts/config -e CFI_CLANG ... Cheers, Carlos Llamas
© 2016 - 2026 Red Hat, Inc.