Bug 469878 - unhandled instruction bytes: 0x62 0xF1 0x7F 0x48 0x7F 0x44 0x24 0x1 0x83 0xE7
Summary: unhandled instruction bytes: 0x62 0xF1 0x7F 0x48 0x7F 0x44 0x24 0x1 0x83 0xE7
Status: RESOLVED DUPLICATE of bug 383010
Alias: None
Product: valgrind
Classification: Developer tools
Component: vex (show other bugs)
Version: 3.20.0
Platform: Other Linux
: NOR normal
Target Milestone: ---
Assignee: Julian Seward
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2023-05-17 01:31 UTC by tusooa
Modified: 2024-02-19 21:26 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments
cpuinfo (20.44 KB, text/plain)
2023-05-17 01:31 UTC, tusooa
Details

Note You need to log in before you can comment on or make changes to this bug.
Description tusooa 2023-05-17 01:31:12 UTC
Created attachment 159023 [details]
cpuinfo

SUMMARY
***
NOTE: If you are reporting a crash, please try to attach a backtrace with debug symbols.
See https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports
***


STEPS TO REPRODUCE
1. valgrind ls
2. 
3. 

OBSERVED RESULT

==8232== Memcheck, a memory error detector
==8232== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==8232== Using Valgrind-3.20.0 and LibVEX; rerun with -h for copyright info
==8232== Command: ls
==8232==
vex amd64->IR: unhandled instruction bytes: 0x62 0xF1 0x7F 0x48 0x7F 0x44 0x24 0x1 0x83 0xE7
vex amd64->IR:   REX=0 REX.W=0 REX.R=0 REX.X=0 REX.B=0
vex amd64->IR:   VEX=0 VEX.L=0 VEX.nVVVV=0x0 ESC=NONE
vex amd64->IR:   PFX.66=0 PFX.F2=0 PFX.F3=0
==8232== valgrind: Unrecognised instruction at address 0x401c126.
==8232==    at 0x401C126: _dl_start (rtld.c:566)
==8232==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
==8232== Your program just tried to execute an instruction that Valgrind
==8232== did not recognise.  There are two possible reasons for this.
==8232== 1. Your program has a bug and erroneously jumped to a non-code
==8232==    location.  If you are running Memcheck and you just saw a
==8232==    warning about a bad jump, it's probably your program's fault.
==8232== 2. The instruction is legitimate but Valgrind doesn't handle it,
==8232==    i.e. it's Valgrind's fault.  If you think this is the case or
==8232==    you are not sure, please let us know and we'll try to fix it.
==8232== Either way, Valgrind will now raise a SIGILL signal which will
==8232== probably kill your program.
==8232==
==8232== Process terminating with default action of signal 4 (SIGILL): dumping core
==8232==  Illegal opcode at address 0x401C126
==8232==    at 0x401C126: _dl_start (rtld.c:566)
==8232==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
==8232==
==8232== HEAP SUMMARY:
==8232==     in use at exit: 0 bytes in 0 blocks
==8232==   total heap usage: 0 allocs, 0 frees, 0 bytes allocated
==8232==
==8232== All heap blocks were freed -- no leaks are possible
==8232==
==8232== For lists of detected and suppressed errors, rerun with: -s
==8232== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)

EXPECTED RESULT
it should work

SOFTWARE/OS VERSIONS
Windows: 
macOS: 
Linux/KDE Plasma: 6.1.27
(available in About System)
KDE Plasma Version: 
KDE Frameworks Version: 
Qt Version: 

ADDITIONAL INFORMATION
/proc/cpuinfo is attached.
This is a Gentoo compiled with

*/* CPU_FLAGS_X86: aes avx avx2 avx512f avx512dq avx512cd avx512bw avx512vl avx512vbmi f16c fma3 mmx mmxext pclmul popcnt rdrand sha sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3

and glibc is compiled with
CFLAGS="${CFLAGS} -fno-builtin-strlen -ggdb3"
CXXFLAGS="${CXXFLAGS} -ggdb3"
FEATURES="${FEATURES} splitdebug compressdebug -nostrip installsources"
Comment 1 tusooa 2023-05-17 01:44:16 UTC
==9898== Memcheck, a memory error detector
==9898== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
==9898== Using Valgrind-3.20.0-5147d671e4-20221024 and LibVEX; rerun with -h for copyright info
==9898== Command: ls
==9898==
--9898-- Valgrind options:
--9898--    -v
--9898-- Contents of /proc/version:
--9898--   Linux version 6.1.27-gentoo-r1-x86_64 (REDACTED@REDACTED) (x86_64-pc-linux-gnu-gcc (Gentoo 12.2.1_p20230428-r1 p2) 12.2.1 20230428, GNU ld (Gentoo 2.39 p6) 2.39.0) #1 SMP PREEMPT_DYNAMIC Wed May 10 18:30:00 EDT 2023
--9898--
--9898-- Arch and hwcaps: AMD64, LittleEndian, amd64-cx16-lzcnt-rdtscp-sse3-ssse3-avx-avx2-bmi-f16c-rdrand-rdseed
--9898-- Page sizes: currently 4096, max supported 4096
--9898-- Valgrind library directory: /usr/libexec/valgrind
--9898-- Reading syms from /bin/ls
--9898--    object doesn't have a symbol table
--9898-- Reading syms from /lib64/ld-linux-x86-64.so.2
--9898--   Considering /usr/lib/debug/lib64/ld-linux-x86-64.so.2.debug ..
--9898--   .. CRC is valid
--9898-- Reading syms from /usr/libexec/valgrind/memcheck-amd64-linux
--9898--    object doesn't have a symbol table
--9898--    object doesn't have a dynamic symbol table
--9898-- Scheduler: using generic scheduler lock implementation.
--9898-- Reading suppressions file: /usr/libexec/valgrind/default.supp
==9898== embedded gdbserver: reading from /tmp/vgdb-pipe-from-vgdb-to-9898-by-user-on-???
==9898== embedded gdbserver: writing to   /tmp/vgdb-pipe-to-vgdb-from-9898-by-user-on-???
==9898== embedded gdbserver: shared mem   /tmp/vgdb-pipe-shared-mem-vgdb-9898-by-user-on-???
==9898==
==9898== TO CONTROL THIS PROCESS USING vgdb (which you probably
==9898== don't want to do, unless you know exactly what you're doing,
==9898== or are doing some strange experiment):
==9898==   /usr/libexec/valgrind/../../bin/vgdb --pid=9898 ...command...
==9898==
==9898== TO DEBUG THIS PROCESS USING GDB: start GDB like this
==9898==   /path/to/gdb ls
==9898== and then give GDB the following command
==9898==   target remote | /usr/libexec/valgrind/../../bin/vgdb --pid=9898
==9898== --pid is optional if only one valgrind process is running
==9898==
vex amd64->IR: unhandled instruction bytes: 0x62 0xF1 0x7F 0x48 0x7F 0x44 0x24 0x1 0x83 0xE7
vex amd64->IR:   REX=0 REX.W=0 REX.R=0 REX.X=0 REX.B=0
vex amd64->IR:   VEX=0 VEX.L=0 VEX.nVVVV=0x0 ESC=NONE
vex amd64->IR:   PFX.66=0 PFX.F2=0 PFX.F3=0
==9898== valgrind: Unrecognised instruction at address 0x401c126.
==9898==    at 0x401C126: _dl_start (rtld.c:566)
==9898==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
==9898== Your program just tried to execute an instruction that Valgrind
==9898== did not recognise.  There are two possible reasons for this.
==9898== 1. Your program has a bug and erroneously jumped to a non-code
==9898==    location.  If you are running Memcheck and you just saw a
==9898==    warning about a bad jump, it's probably your program's fault.
==9898== 2. The instruction is legitimate but Valgrind doesn't handle it,
==9898==    i.e. it's Valgrind's fault.  If you think this is the case or
==9898==    you are not sure, please let us know and we'll try to fix it.
==9898== Either way, Valgrind will now raise a SIGILL signal which will
==9898== probably kill your program.
==9898==
==9898== Process terminating with default action of signal 4 (SIGILL): dumping core
==9898==  Illegal opcode at address 0x401C126
==9898==    at 0x401C126: _dl_start (rtld.c:566)
==9898==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
==9898==
==9898== HEAP SUMMARY:
==9898==     in use at exit: 0 bytes in 0 blocks
==9898==   total heap usage: 0 allocs, 0 frees, 0 bytes allocated
==9898==
==9898== All heap blocks were freed -- no leaks are possible
==9898==
==9898== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
zsh: illegal hardware instruction  valgrind -v ls
Comment 2 Tom Hughes 2023-05-17 06:05:14 UTC
That's an AVX-512 instruction.

*** This bug has been marked as a duplicate of bug 383010 ***
Comment 3 Martin Kjær Jørgensen 2024-02-19 21:26:02 UTC
(In reply to tusooa from comment #1)
> ==9898== Memcheck, a memory error detector
> ==9898== Copyright (C) 2002-2022, and GNU GPL'd, by Julian Seward et al.
> ==9898== Using Valgrind-3.20.0-5147d671e4-20221024 and LibVEX; rerun with -h
> for copyright info
> ==9898== Command: ls
> ==9898==
> --9898-- Valgrind options:
> --9898--    -v
> --9898-- Contents of /proc/version:
> --9898--   Linux version 6.1.27-gentoo-r1-x86_64 (REDACTED@REDACTED)
> (x86_64-pc-linux-gnu-gcc (Gentoo 12.2.1_p20230428-r1 p2) 12.2.1 20230428,
> GNU ld (Gentoo 2.39 p6) 2.39.0) #1 SMP PREEMPT_DYNAMIC Wed May 10 18:30:00
> EDT 2023
> --9898--
> --9898-- Arch and hwcaps: AMD64, LittleEndian,
> amd64-cx16-lzcnt-rdtscp-sse3-ssse3-avx-avx2-bmi-f16c-rdrand-rdseed
> --9898-- Page sizes: currently 4096, max supported 4096
> --9898-- Valgrind library directory: /usr/libexec/valgrind
> --9898-- Reading syms from /bin/ls
> --9898--    object doesn't have a symbol table
> --9898-- Reading syms from /lib64/ld-linux-x86-64.so.2
> --9898--   Considering /usr/lib/debug/lib64/ld-linux-x86-64.so.2.debug ..
> --9898--   .. CRC is valid
> --9898-- Reading syms from /usr/libexec/valgrind/memcheck-amd64-linux
> --9898--    object doesn't have a symbol table
> --9898--    object doesn't have a dynamic symbol table
> --9898-- Scheduler: using generic scheduler lock implementation.
> --9898-- Reading suppressions file: /usr/libexec/valgrind/default.supp
> ==9898== embedded gdbserver: reading from
> /tmp/vgdb-pipe-from-vgdb-to-9898-by-user-on-???
> ==9898== embedded gdbserver: writing to  
> /tmp/vgdb-pipe-to-vgdb-from-9898-by-user-on-???
> ==9898== embedded gdbserver: shared mem  
> /tmp/vgdb-pipe-shared-mem-vgdb-9898-by-user-on-???
> ==9898==
> ==9898== TO CONTROL THIS PROCESS USING vgdb (which you probably
> ==9898== don't want to do, unless you know exactly what you're doing,
> ==9898== or are doing some strange experiment):
> ==9898==   /usr/libexec/valgrind/../../bin/vgdb --pid=9898 ...command...
> ==9898==
> ==9898== TO DEBUG THIS PROCESS USING GDB: start GDB like this
> ==9898==   /path/to/gdb ls
> ==9898== and then give GDB the following command
> ==9898==   target remote | /usr/libexec/valgrind/../../bin/vgdb --pid=9898
> ==9898== --pid is optional if only one valgrind process is running
> ==9898==
> vex amd64->IR: unhandled instruction bytes: 0x62 0xF1 0x7F 0x48 0x7F 0x44
> 0x24 0x1 0x83 0xE7
> vex amd64->IR:   REX=0 REX.W=0 REX.R=0 REX.X=0 REX.B=0
> vex amd64->IR:   VEX=0 VEX.L=0 VEX.nVVVV=0x0 ESC=NONE
> vex amd64->IR:   PFX.66=0 PFX.F2=0 PFX.F3=0
> ==9898== valgrind: Unrecognised instruction at address 0x401c126.
> ==9898==    at 0x401C126: _dl_start (rtld.c:566)
> ==9898==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
> ==9898== Your program just tried to execute an instruction that Valgrind
> ==9898== did not recognise.  There are two possible reasons for this.
> ==9898== 1. Your program has a bug and erroneously jumped to a non-code
> ==9898==    location.  If you are running Memcheck and you just saw a
> ==9898==    warning about a bad jump, it's probably your program's fault.
> ==9898== 2. The instruction is legitimate but Valgrind doesn't handle it,
> ==9898==    i.e. it's Valgrind's fault.  If you think this is the case or
> ==9898==    you are not sure, please let us know and we'll try to fix it.
> ==9898== Either way, Valgrind will now raise a SIGILL signal which will
> ==9898== probably kill your program.
> ==9898==
> ==9898== Process terminating with default action of signal 4 (SIGILL):
> dumping core
> ==9898==  Illegal opcode at address 0x401C126
> ==9898==    at 0x401C126: _dl_start (rtld.c:566)
> ==9898==    by 0x401B1D7: ??? (in /lib64/ld-linux-x86-64.so.2)
> ==9898==
> ==9898== HEAP SUMMARY:
> ==9898==     in use at exit: 0 bytes in 0 blocks
> ==9898==   total heap usage: 0 allocs, 0 frees, 0 bytes allocated
> ==9898==
> ==9898== All heap blocks were freed -- no leaks are possible
> ==9898==
> ==9898== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 0 from 0)
> zsh: illegal hardware instruction  valgrind -v ls

Did you work around your AXV512 issue?