On Mon, Feb 25, 2013 at 8:51 PM, Roland Mainz <[email protected]> wrote: > On Fri, Feb 22, 2013 at 10:14 PM, Glenn Fowler <[email protected]> wrote: >> >> On Thu, 21 Feb 2013 04:19:00 +0100 Roland Mainz wrote: >>> On Mon, Feb 18, 2013 at 5:51 PM, Glenn Fowler <[email protected]> wrote: >>> > >>> > 2013-02-14 alpha posted to www.research.att.com/download/alpha/ >>> > INIT.2013-02-14.tgz >>> > ast-ksh.2013-02-14.tgz >>> > this is a work in progress >>> > there are known problems on some architectures >> >>> On SuSE 12.2 Linux/AMD64/64bit I get a couple of valgrind hits in the >>> spawnvex subsystem: >>> -- snip -- >>> $ valgrind --track-origins=yes --read-var-info=yes ~/bin/ksh -c 'ls -l ; >>> true' >>> ==19925== Memcheck, a memory error detector >>> ==19925== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al. >>> ==19925== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info >>> ==19925== Command: /home/test001/bin/ksh -c ls\ -l\ ;\ true >>> ==19925== >>> [snip] >>> ==19925== >>> ==19925== Conditional jump or move depends on uninitialised value(s) >>> ==19925== at 0x4DC865: spawnvex (spawnvex.c:1100) >>> ==19925== by 0x4653D3: _spawnveg (path.c:151) >>> ==19925== by 0x468973: path_spawn (path.c:1228) >>> ==19925== by 0x480B4A: sh_ntfork (xec.c:3824) >>> ==19925== by 0x479F15: sh_exec (xec.c:1675) >>> ==19925== by 0x47BC7A: sh_exec (xec.c:2200) >>> ==19925== by 0x40F0D5: exfile (main.c:588) >>> ==19925== by 0x40E29A: sh_main (main.c:360) >>> ==19925== by 0x40D450: main (pmain.c:45) >>> ==19925== Uninitialised value was created by a stack allocation >>> ==19925== at 0x4DC6C2: spawnvex (spawnvex.c:793) > [snip] >> >> roland, I'm having line sync probs between my src and the valgrind lines >> can you identify the varialble that is claimed to be uninitialized >> thanks > > valgrind hits and line numbers for ast-ksh.2013-02-22 look like this: > -- snip -- > $ valgrind --track-origins=yes --read-var-info=yes > ./arch/linux.i386-64/bin/ksh -c 'ls -l ; true' > ==2409== Memcheck, a memory error detector > ==2409== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al. > ==2409== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info > ==2409== Command: ./arch/linux.i386-64/bin/ksh -c ls\ -l\ ;\ true > ==2409== > [snip] > ==2409== > ==2409== Conditional jump or move depends on uninitialised value(s) > ==2409== at 0x4DC851: spawnvex (spawnvex.c:1100) > ==2409== by 0x4653BF: _spawnveg (path.c:151) > ==2409== by 0x46895F: path_spawn (path.c:1228) > ==2409== by 0x480B36: sh_ntfork (xec.c:3824) > ==2409== by 0x479F01: sh_exec (xec.c:1675) > ==2409== by 0x47BC66: sh_exec (xec.c:2200) > ==2409== by 0x40F0D5: exfile (main.c:588) > ==2409== by 0x40E29A: sh_main (main.c:360) > ==2409== by 0x40D450: main (pmain.c:45) > ==2409== Uninitialised value was created by a stack allocation > ==2409== at 0x4DC6AE: spawnvex (spawnvex.c:793) > ==2409== > ==2409== Conditional jump or move depends on uninitialised value(s) > ==2409== at 0x4DCC77: spawnvex (spawnvex.c:1193) > ==2409== by 0x4653BF: _spawnveg (path.c:151) > ==2409== by 0x46895F: path_spawn (path.c:1228) > ==2409== by 0x480B36: sh_ntfork (xec.c:3824) > ==2409== by 0x479F01: sh_exec (xec.c:1675) > ==2409== by 0x47BC66: sh_exec (xec.c:2200) > ==2409== by 0x40F0D5: exfile (main.c:588) > ==2409== by 0x40E29A: sh_main (main.c:360) > ==2409== by 0x40D450: main (pmain.c:45) > ==2409== Uninitialised value was created by a stack allocation > ==2409== at 0x4DC6AE: spawnvex (spawnvex.c:793) > ==2409== > ==2410== Conditional jump or move depends on uninitialised value(s) > ==2410== at 0x4DC8AB: spawnvex (spawnvex.c:1118) > ==2410== by 0x4653BF: _spawnveg (path.c:151) > ==2410== by 0x46895F: path_spawn (path.c:1228) > ==2410== by 0x480B36: sh_ntfork (xec.c:3824) > ==2410== by 0x479F01: sh_exec (xec.c:1675) > ==2410== by 0x47BC66: sh_exec (xec.c:2200) > ==2410== by 0x40F0D5: exfile (main.c:588) > ==2410== by 0x40E29A: sh_main (main.c:360) > ==2410== by 0x40D450: main (pmain.c:45) > ==2410== Uninitialised value was created by a stack allocation > ==2410== at 0x4DC6AE: spawnvex (spawnvex.c:793) > ==2410== > total 1536 > drwxr-xr-x 3 test001 users 4096 Feb 24 23:06 arch > drwxr-xr-x 2 test001 users 4096 Feb 24 23:07 bin > -rw-r--r-- 1 test001 users 1537649 Feb 25 02:29 buildlog.log > drwxr-xr-x 3 test001 users 4096 Feb 24 23:04 lib > -rw-r--r-- 1 test001 users 854 Feb 23 08:20 README > drwxr-xr-x 4 test001 users 4096 Feb 24 23:04 src > drwxr-xr-x 3 test001 users 4096 Feb 25 05:29 tmp > -rw-r--r-- 1 test001 users 3851 Feb 24 23:06 tmp_gnulinux_builtin_header.h > -- snip -- > > First hit is: > -- snip -- > ==2461== Conditional jump or move depends on uninitialised value(s) > ==2461== at 0x4DC851: spawnvex (spawnvex.c:1100) > ==2461== by 0x4653BF: _spawnveg (path.c:151) > ==2461== by 0x46895F: path_spawn (path.c:1228) > ==2461== by 0x480B36: sh_ntfork (xec.c:3824) > ==2461== by 0x479F01: sh_exec (xec.c:1675) > ==2461== by 0x47BC66: sh_exec (xec.c:2200) > ==2461== by 0x40F0D5: exfile (main.c:588) > ==2461== by 0x40E29A: sh_main (main.c:360) > ==2461== by 0x40D450: main (pmain.c:45) > ==2461== Uninitialised value was created by a stack allocation > ==2461== at 0x4DC6AE: spawnvex (spawnvex.c:793) > -- snip -- > ... which seems to be the |flags| variable in line 1100: > -- snip -- > 1097 fcntl(msg[1], F_SETFD, FD_CLOEXEC); > 1098 } > 1099 #endif > 1100 if (!(flags & SPAWN_FOREGROUND)) > 1101 sigcritical(SIG_REG_EXEC|SIG_REG_PROC); > 1102 #if _lib_vfork > 1103 #if _lib_fork > 1104 if (nx.flags & SPAWN_VFORK) > -- snip -- > > > 2nd hit is: > -- snip -- > ==2461== Conditional jump or move depends on uninitialised value(s) > ==2461== at 0x4DC851: spawnvex (spawnvex.c:1100) > ==2461== by 0x4653BF: _spawnveg (path.c:151) > ==2461== by 0x46895F: path_spawn (path.c:1228) > ==2461== by 0x480B36: sh_ntfork (xec.c:3824) > ==2461== by 0x479F01: sh_exec (xec.c:1675) > ==2461== by 0x47BC66: sh_exec (xec.c:2200) > ==2461== by 0x40F0D5: exfile (main.c:588) > ==2461== by 0x40E29A: sh_main (main.c:360) > ==2461== by 0x40D450: main (pmain.c:45) > ==2461== Uninitialised value was created by a stack allocation > ==2461== at 0x4DC6AE: spawnvex (spawnvex.c:793) > -- snip -- > ... again it seems to be the |flags| variable in line 1118: > -- snip -- > 1117 { > 1118 if (!(flags & SPAWN_FOREGROUND)) > 1119 sigcritical(0); > 1120 if (vex && (n = spawnvex_apply(vex, 0, > SPAWN_FRAME|SPAWN_NOCALL))) > 1121 errno = n; > -- snip -- > > > 3rd (and last) hit is this one: > -- snip -- > ==2461== Conditional jump or move depends on uninitialised value(s) > ==2461== at 0x4DCC77: spawnvex (spawnvex.c:1193) > ==2461== by 0x4653BF: _spawnveg (path.c:151) > ==2461== by 0x46895F: path_spawn (path.c:1228) > ==2461== by 0x480B36: sh_ntfork (xec.c:3824) > ==2461== by 0x479F01: sh_exec (xec.c:1675) > ==2461== by 0x47BC66: sh_exec (xec.c:2200) > ==2461== by 0x40F0D5: exfile (main.c:588) > ==2461== by 0x40E29A: sh_main (main.c:360) > ==2461== by 0x40D450: main (pmain.c:45) > ==2461== Uninitialised value was created by a stack allocation > ==2461== at 0x4DC6AE: spawnvex (spawnvex.c:793) > ==2461== > -- snip -- > ... which relates to this code in line 1193: > -- snip -- > 1192 } > 1193 if (!(flags & SPAWN_FOREGROUND)) > 1194 sigcritical(0); > 1195 if (pid != -1 && vex) > -- snip -- > > ... short summary: It's the |flags| variable which is uninitalised in > some usage cases. > > The frustrating part is that the --track-origins=yes option of > valgrind points to line 793: > -- snip -- > 791 pid_t > 792 spawnvex(const char* path, char* const argv[], char* const > envv[], Spawnvex_t* vex) > 793 { > 794 > -- snip -- > This is the beginning of the function, IMO useless and frustrating > because the binary was explicitly created with -g -ggdb and optimiser > disabled.
ping! ... is there a patch for the valgrind hits yet (well, I can "silence" the warnings via initalising the |flags| variable myself... but I'm not 100% sure whether this is the correct "fix") ? ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) [email protected] \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 3992797 (;O/ \/ \O;) _______________________________________________ ast-developers mailing list [email protected] http://lists.research.att.com/mailman/listinfo/ast-developers
