Changes in faddr2line

56ac7bd2 - Revert "scripts/faddr2line: Check only two symbols when calculating symbol size"

Mon, 12 Aug 2024 23:01:20 +0000

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 symbolsize calculation. However, this can lead to an incorrect symbol sizecalculation when there are mapping symbols emitted by readelf.For instance, when feeding 'update_irq_load_avg+0x1c/0x1c4', faddr2linemight need to process 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_multiplierThe symbol size of 'update_irq_load_avg' should be calculated with theaddress of 'sched_pelt_multiplier', after skipping the mapping symbolsseen in between. However, the offending commit cuts the list short andfaddr2line incorrectly assumes 'update_irq_load_avg' is the last symbolin 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/0x1c4After 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:481Fixes: c02904f05ff8 ("scripts/faddr2line: Check only two symbols when calculating symbol size")Signed-off-by: Carlos Llamas Acked-by: Will Deacon Acked-by: Brian Johannesmeyer Signed-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

c02904f0 - scripts/faddr2line: Check only two symbols when calculating symbol size

Mon, 15 Apr 2024 14:55:38 +0000

scripts/faddr2line: Check only two symbols when calculating symbol sizeRather than looping through each symbol in a particular section tocalculate a symbol's size, grep for the symbol and its immediatesuccessor, and only use those two symbols.Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-8-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

406b5c12 - scripts/faddr2line: Remove call to addr2line from find_dir_prefix()

Mon, 15 Apr 2024 14:55:37 +0000

scripts/faddr2line: Remove call to addr2line from find_dir_prefix()Use the single long-running faddr2line process from find_dir_prefix().Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-7-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

e36b69e9 - scripts/faddr2line: Invoke addr2line as a single long-running process

Mon, 15 Apr 2024 14:55:36 +0000

scripts/faddr2line: Invoke addr2line as a single long-running processRather than invoking a separate addr2line process for each address, invokea single addr2line coprocess, and pass each address to its stdin. Previouswork [0] applied a similar change to perf, leading to a ~60x speed-up [1].If using an object file that is _not_ vmlinux, faddr2line passes a sectionname argument to addr2line. Because we do not know until runtime whichsection names will be passed to addr2line, we cannot apply this change tonon-vmlinux object files. Hence, it only applies to vmlinux.[0] commit be8ecc57f180 ("perf srcline: Use long-running addr2line perDSO")[1] Link:https://eighty-twenty.org/2021/09/09/perf-addr2line-speed-improvementSigned-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-6-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

5b280de4 - scripts/faddr2line: Pass --addresses argument to addr2line

Mon, 15 Apr 2024 14:55:35 +0000

scripts/faddr2line: Pass --addresses argument to addr2lineIn preparation for identifying an addr2line sentinel. See previous work[0], which applies a similar change to perf.[0] commit 8dc26b6f718a ("perf srcline: Make sentinel reading for binutilsaddr2line more robust")Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-5-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

2c809186 - scripts/faddr2line: Check vmlinux only once

Mon, 15 Apr 2024 14:55:34 +0000

scripts/faddr2line: Check vmlinux only onceRather than checking whether the object file is vmlinux for each invocationof __faddr2line, check it only once beforehand.Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-4-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

b8d9d949 - scripts/faddr2line: Combine three readelf calls into one

Mon, 15 Apr 2024 14:55:33 +0000

scripts/faddr2line: Combine three readelf calls into oneRather than calling readelf three separate times to collect three differenttypes of info, call it only once, and parse out the different types of infofrom its output.Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-3-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

39cf650d - scripts/faddr2line: Reduce number of readelf calls to three

Mon, 15 Apr 2024 14:55:32 +0000

scripts/faddr2line: Reduce number of readelf calls to threeRather than calling readelf several times for each invocation of__faddr2line, call readelf only three times at the beginning, and save itsresult for future use.Signed-off-by: Brian Johannesmeyer Link: https://lore.kernel.org/r/20240415145538.1938745-2-bjohannesmeyer@gmail.comSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

60fd39af - scripts/faddr2line: Skip over mapping symbols in output from readelf

Mon, 02 Oct 2023 16:57:49 +0000

scripts/faddr2line: Skip over mapping symbols in output from readelfMapping symbols emitted in the readelf output can confuse the'faddr2line' symbol size calculation, resulting in the erroneousrejection of valid offsets. This is especially prevalent when buildingan arm64 kernel with CONFIG_CFI_CLANG=y, where most functions areprefixed with a 32-bit data value in a '$d.n' section. For example:447538: ffff800080014b80 548 FUNC GLOBAL DEFAULT 2 do_one_initcall 104: ffff800080014c74 0 NOTYPE LOCAL DEFAULT 2 $x.73 106: ffff800080014d30 0 NOTYPE LOCAL DEFAULT 2 $x.75 111: ffff800080014da4 0 NOTYPE LOCAL DEFAULT 2 $d.78 112: ffff800080014da8 0 NOTYPE LOCAL DEFAULT 2 $x.79 36: ffff800080014de0 200 FUNC LOCAL DEFAULT 2 run_init_processAdding a warning to do_one_initcall() results in: | WARNING: CPU: 0 PID: 1 at init/main.c:1236 do_one_initcall+0xf4/0x260Which 'faddr2line' refuses to accept:$ ./scripts/faddr2line vmlinux do_one_initcall+0xf4/0x260skipping do_one_initcall address at 0xffff800080014c74 due to size mismatch (0x260 != 0x224)no match for do_one_initcall+0xf4/0x260Filter out these entries from readelf using a shell reimplementation ofis_mapping_symbol(), so that the size of a symbol is calculated as adelta to the next symbol present in ksymtab.Suggested-by: Masahiro Yamada Signed-off-by: Will Deacon Reviewed-by: Nick Desaulniers Link: https://lore.kernel.org/r/20231002165750.1661-4-will@kernel.orgSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

86bf86e1 - scripts/faddr2line: Use LLVM addr2line and readelf if LLVM=1

Mon, 02 Oct 2023 16:57:48 +0000

scripts/faddr2line: Use LLVM addr2line and readelf if LLVM=1GNU utilities cannot necessarily parse objects built by LLVM, which canresult in confusing errors when using 'faddr2line':$ CROSS_COMPILE=aarch64-linux-gnu- ./scripts/faddr2line vmlinux do_one_initcall+0xf4/0x260aarch64-linux-gnu-addr2line: vmlinux: unknown type [0x13] section `.relr.dyn'aarch64-linux-gnu-addr2line: DWARF error: invalid or unhandled FORM value: 0x25do_one_initcall+0xf4/0x260:aarch64-linux-gnu-addr2line: vmlinux: unknown type [0x13] section `.relr.dyn'aarch64-linux-gnu-addr2line: DWARF error: invalid or unhandled FORM value: 0x25$x.73 at main.c:?Although this can be worked around by setting CROSS_COMPILE to "llvm=-",it's cleaner to follow the same syntax as the top-level Makefile andaccept LLVM= as an indication to use the llvm- tools, optionallyspecifying their location or specific version number.Suggested-by: Masahiro Yamada Signed-off-by: Will Deacon Reviewed-by: Nick Desaulniers Link: https://lore.kernel.org/r/20231002165750.1661-3-will@kernel.orgSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

180af1a5 - scripts/faddr2line: Don't filter out non-function symbols from readelf

Mon, 02 Oct 2023 16:57:47 +0000

scripts/faddr2line: Don't filter out non-function symbols from readelfAs Josh points out in 20230724234734.zy67gm674vl3p3wv@treble:> Problem is, I think the kernel's symbol printing code prints the> nearest kallsyms symbol, and there are some valid non-FUNC code> symbols. For example, syscall_return_via_sysret.so we shouldn't be considering only 'FUNC'-type symbols in the outputfrom readelf.Drop the function symbol type filtering from the faddr2line outer loop.Suggested-by: Josh Poimboeuf Reviewed-by: Nick Desaulniers Link: https://lore.kernel.org/r/20230724234734.zy67gm674vl3p3wv@trebleSigned-off-by: Will Deacon Link: https://lore.kernel.org/r/20231002165750.1661-2-will@kernel.orgSigned-off-by: Josh Poimboeuf List of files: /linux-6.15/scripts/faddr2line

2d77de15 - scripts/faddr2line: Fix regression in name resolution on ppc64le

Tue, 27 Sep 2022 07:52:11 +0000

scripts/faddr2line: Fix regression in name resolution on ppc64leCommit 1d1a0e7c5100 ("scripts/faddr2line: Fix overlapping text sectionfailures") can cause faddr2line to fail on ppc64le on somedistributions, while it works fine on other distributions. The failurecan be attributed to differences in the readelf output. $ ./scripts/faddr2line vmlinux find_busiest_group+0x00 no match for find_busiest_group+0x00On ppc64le, readelf adds the localentry tag before the symbol name onsome distributions, and adds the localentry tag after the symbol name onother distributions. This problem has been discussed previously: https://lore.kernel.org/bpf/20191211160133.GB4580@calabresa/This problem can be overcome by filtering out the localentry tags in thereadelf output. Similar fixes are already present in the kernel by wayof the following commits: 1fd6cee127e2 ("libbpf: Fix VERSIONED_SYM_COUNT number parsing") aa915931ac3e ("libbpf: Fix readelf output parsing for Fedora")[jpoimboe: rework commit log]Fixes: 1d1a0e7c5100 ("scripts/faddr2line: Fix overlapping text section failures")Signed-off-by: Srikar Dronamraju Acked-by: Naveen N. Rao Reviewed-by: Thadeu Lima de Souza Cascardo Link: https://lore.kernel.org/r/20220927075211.897152-1-srikar@linux.vnet.ibm.comSigned-off-by: Josh Poimboeuf Signed-off-by: Peter Zijlstra List of files: /linux-6.15/scripts/faddr2line

a41a2e2e - scripts/faddr2line: Add CONFIG_DEBUG_INFO check

Thu, 21 Jul 2022 18:01:24 +0000

scripts/faddr2line: Add CONFIG_DEBUG_INFO checkOtherwise without DWARF it spits out gibberish and gives no indicationof what the problem is.Suggested-by: John Garry Signed-off-by: Josh Poimboeuf Signed-off-by: Ingo Molnar Tested-by: John Garry Link: https://lore.kernel.org/r/ffa7734c929445caa374bf9e68078300174f09b4.1658426357.git.jpoimboe@kernel.org List of files: /linux-6.15/scripts/faddr2line

b6a50688 - scripts/faddr2line: Fix vmlinux detection on arm64

Thu, 21 Jul 2022 18:01:23 +0000

scripts/faddr2line: Fix vmlinux detection on arm64Since commit dcea997beed6 ("faddr2line: Fix overlapping text sectionfailures, the sequel"), faddr2line is completely broken on arm64.For some reason, on arm64, the vmlinux ELF object file type is ET_DYNrather than ET_EXEC. Check for both when determining whether the objectis vmlinux.Modules and vmlinux.o have type ET_REL on all arches.Fixes: dcea997beed6 ("faddr2line: Fix overlapping text section failures, the sequel")Reported-by: John Garry Signed-off-by: Josh Poimboeuf Signed-off-by: Ingo Molnar Tested-by: John Garry Link: https://lore.kernel.org/r/dad1999737471b06d6188ce4cdb11329aa41682c.1658426357.git.jpoimboe@kernel.org List of files: /linux-6.15/scripts/faddr2line

dcea997b - faddr2line: Fix overlapping text section failures, the sequel

Thu, 02 Jun 2022 00:42:22 +0000

faddr2line: Fix overlapping text section failures, the sequelIf a function lives in a section other than .text, but .text also existsin the object, faddr2line may wrongly assume .text. This can result incomically wrong output. For example: $ scripts/faddr2line vmlinux.o enter_from_user_mode+0x1c enter_from_user_mode+0x1c/0x30: find_next_bit at /home/jpoimboe/git/linux/./include/linux/find.h:40 (inlined by) perf_clear_dirty_counters at /home/jpoimboe/git/linux/arch/x86/events/core.c:2504Fix it by passing the section name to addr2line, unless the object fileis vmlinux, in which case the symbol table uses absolute addresses.Fixes: 1d1a0e7c5100 ("scripts/faddr2line: Fix overlapping text section failures")Reported-by: Peter Zijlstra Signed-off-by: Josh Poimboeuf Link: https://lore.kernel.org/r/7d25bc1408bd3a750ac26e60d2f2815a5f4a8363.1654130536.git.jpoimboe@kernel.org List of files: /linux-6.15/scripts/faddr2line

1d1a0e7c - scripts/faddr2line: Fix overlapping text section failures

Thu, 12 May 2022 19:05:27 +0000

scripts/faddr2line: Fix overlapping text section failuresThere have been some recent reports of faddr2line failures: $ scripts/faddr2line sound/soundcore.ko sound_devnode+0x5/0x35 bad symbol size: base: 0x0000000000000000 end: 0x0000000000000000 $ ./scripts/faddr2line vmlinux.o enter_from_user_mode+0x24 bad symbol size: base: 0x0000000000005fe0 end: 0x0000000000005fe0The problem is that faddr2line is based on 'nm', which has a majorlimitation: it doesn't know how to distinguish between different textsections. So if an offset exists in multiple text sections in theobject, it may fail.Rewrite faddr2line to be section-aware, by basing it on readelf.Fixes: 67326666e2d4 ("scripts: add script for translating stack dump function offsets")Reported-by: Kaiwan N Billimoria Reported-by: Peter Zijlstra Signed-off-by: Josh Poimboeuf Link: https://lore.kernel.org/r/29ff99f86e3da965b6e46c1cc2d72ce6528c17c3.1652382321.git.jpoimboe@kernel.org List of files: /linux-6.15/scripts/faddr2line

f5f67cc0 - scripts/faddr2line: fix location of start_kernel in comment

Fri, 16 Nov 2018 23:08:22 +0000

scripts/faddr2line: fix location of start_kernel in commentFix a source file reference location to the correct path name.Link: http://lkml.kernel.org/r/1d50bd3d-178e-dcd8-779f-9711887440eb@infradead.orgSigned-off-by: Randy Dunlap Acked-by: Josh Poimboeuf Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds List of files: /linux-6.15/scripts/faddr2line

689135f0 - scripts/faddr2line: make the new code listing format optional

Mon, 04 Jun 2018 18:48:31 +0000

scripts/faddr2line: make the new code listing format optionalCommit 6870c0165feaa5 ("scripts/faddr2line: show the code context")radically altered the output format of the faddr2line tool. And whilethe new list output format might have merit it broke my vim usage andwas hard to read.Make the new format optional; using a '--list' argument and attempt tomake the output slightly easier to read by adding a little whitespace toseparate the different files and explicitly mark the line in question.Cc: Changbin Du Fixes: 6870c0165feaa5 ("scripts/faddr2line: show the code context")Signed-off-by: Peter Zijlstra (Intel) Signed-off-by: Josh Poimboeuf Signed-off-by: Linus Torvalds List of files: /linux-6.15/scripts/faddr2line

78eb0c63 - scripts/faddr2line: fix error when addr2line output contains discriminator

Fri, 11 May 2018 23:02:11 +0000

scripts/faddr2line: fix error when addr2line output contains discriminatorWhen addr2line output contains discriminator, the current awk scriptcannot parse it. This patch fixes it by extracting key words usingregex which is more reliable. $ scripts/faddr2line vmlinux tlb_flush_mmu_free+0x26 tlb_flush_mmu_free+0x26/0x50: tlb_flush_mmu_free at mm/memory.c:258 (discriminator 3) scripts/faddr2line: eval: line 173: unexpected EOF while looking for matching `)'Link: http://lkml.kernel.org/r/1525323379-25193-1-git-send-email-changbin.du@intel.comFixes: 6870c0165feaa5 ("scripts/faddr2line: show the code context")Signed-off-by: Changbin Du Cc: Thomas Gleixner Cc: Greg Kroah-Hartman Cc: Philippe Ombredanne Cc: NeilBrown Cc: Richard Weinberger Cc: Kate Stewart Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds List of files: /linux-6.15/scripts/faddr2line

6870c016 - scripts/faddr2line: show the code context

Thu, 05 Apr 2018 23:18:29 +0000

scripts/faddr2line: show the code contextInspired by gdb command 'list', show the code context of target lines.Here is a example:$ scripts/faddr2line vmlinux native_write_msr+0x6native_write_msr+0x6/0x20:arch_static_branch at arch/x86/include/asm/msr.h:105100 return EAX_EDX_VAL(val, low, high);101 }102103 static inline void notrace __wrmsr(unsigned int msr, u32 low, u32 high)104 {105 asm volatile("1: wrmsr\n"106 "2:\n"107 _ASM_EXTABLE_HANDLE(1b, 2b, ex_handler_wrmsr_unsafe)108 : : "c" (msr), "a"(low), "d" (high) : "memory");109 }110(inlined by) static_key_false at include/linux/jump_label.h:142137 #define JUMP_TYPE_LINKED 2UL138 #define JUMP_TYPE_MASK 3UL139140 static __always_inline bool static_key_false(struct static_key *key)141 {142 return arch_static_branch(key, false);143 }144145 static __always_inline bool static_key_true(struct static_key *key)146 {147 return !arch_static_branch(key, true);(inlined by) native_write_msr at arch/x86/include/asm/msr.h:150145 static inline void notrace146 native_write_msr(unsigned int msr, u32 low, u32 high)147 {148 __wrmsr(msr, low, high);149150 if (msr_tracepoint_active(__tracepoint_write_msr))151 do_trace_write_msr(msr, ((u64)high << 32 | low), 0);152 }153154 /* Can be uninlined because referenced by paravirt */155 static inline int notraceLink: http://lkml.kernel.org/r/1521444205-2259-1-git-send-email-changbin.du@intel.comSigned-off-by: Changbin Du Cc: Thomas Gleixner Cc: Philippe Ombredanne Cc: NeilBrown Cc: Richard Weinberger Cc: Kate Stewart Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds List of files: /linux-6.15/scripts/faddr2line