On Wed, 2015-10-28 at 13:23 +0000, Marco Cisternino wrote: > Then I give gdb > > continue > > After few seconds gdb says > Program received signal SIGTRAP, Trace/breakpoint trap. > 0x0000000006018417 in __libc_writev (fd=9, vector=0x6af6710, count=3) > at ../sysdeps/unix/sysv/linux/writev.c:50 > 50 ../sysdeps/unix/sysv/linux/writev.c: File o directory non > esistente. > where "non esistente" means "doesn't exist" > I don't really know what happens. > Could you help me understand, please?
--vgdb-error=0 tells valgrind to stop at the begin of execution of your program and has to wait for gdb. It also indicates that valgrind has to give control to gdb for each error encountered. You can also put breakpoints in your program, if you want. You then succesfully connected gdb to valgrind, and the continue command you gave has continued the execution (till the first error). After that, valgrind stopped with a SIGTRAP. Normally, this means that valgrind has detected an error, and is giving back the control to gdb, so that you can investigate the error. The error message is shown on the stderr of your valgrind command. In gdb, you can use the various normal gdb commands (e.g. examine the variables, ...). You can also use several valgrind specific 'monitor commands': see e.g. http://www.valgrind.org/docs/manual/mc-manual.html#mc-manual.monitor-commands and http://www.valgrind.org/docs/manual/manual-core-adv.html#manual-core-adv.valgrind-monitor-commands for more info about memcheck and valgrind monitor commands. Possibly but unlikely, the SIGTRAP could have other causes (e.g. a real trap instruction). So, very probably, the above shows that Valgrind detected an error in the usage of the writev syscall (probably writing uninitialised data). Philippe ------------------------------------------------------------------------------ _______________________________________________ Valgrind-users mailing list Valgrind-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/valgrind-users
