Skip to contentSkip to conversation

Revert "scripts/faddr2line: Check only two symbols when calculating symbol size"

PATCH Applied Carlos Llamas 4 participants 6 messages Aug 12, 2024 – Oct 17, 2024 FADDR2LINE

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 Revert "scripts/faddr2line: Check only two symbols when calculating symbol size"

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 <[email protected]>
Cc John Stultz <[email protected]>
Cc Will Deacon <[email protected]>
---
 scripts/faddr2line | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
scripts/faddr2line +1 -1
index fe0cc45f03be..1fa6beef9f97 100755
@@ -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"
Signature · 1 line
2.46.0.76.ge559c4bf1a-goog
Acked-by on v1 1/1
#1 lore raw
Quoted 37 lines · Carlos Llamas

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
Quoted 7 lines
Cc Brian Johannesmeyer <[email protected]>
Cc John Stultz <[email protected]>
Cc Will Deacon <[email protected]>
---
 scripts/faddr2line | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

I'm assuming that Josh will pick this up.

Cheers,

Will

#2 lore raw
Quoted 47 lines · Will Deacon
Quoted 37 lines · Carlos Llamas

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
Quoted 7 lines
Cc Brian Johannesmeyer <[email protected]>
Cc John Stultz <[email protected]>
Cc Will Deacon <[email protected]>
---
 scripts/faddr2line | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 7 lines

I'm assuming that Josh will pick this up.

Cheers,

Will

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

#3 lore raw
Quoted 9 lines · Brian Johannesmeyer

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

#4 lore raw
In reply to Will Deacon
Quoted 46 lines · Will Deacon
Quoted 37 lines · Carlos Llamas

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
Quoted 7 lines
Cc Brian Johannesmeyer <[email protected]>
Cc John Stultz <[email protected]>
Cc Will Deacon <[email protected]>
---
 scripts/faddr2line | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
Quoted 3 lines

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

maintainer Applied
#5 lore raw
Quoted 7 lines · Carlos Llamas
Quoted 3 lines · Will Deacon

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!

Signature · 1 line
Josh