Re: Minor patch to correct buffer overrun
Eli Zaretskii <[email protected]>
| Newsgroups | gmane.comp.gnu.make.windows |
|---|---|
| Message-ID | <[email protected]> |
> Date: Wed, 18 Aug 2010 08:31:31 -0400 > From: Chris Sutcliffe <[email protected]> > Cc: [email protected], [email protected] > > On 16 August 2010 16:53, Chris Sutcliffe wrote: > > Program received signal SIGSEGV, Segmentation fault. > > 0x00403a03 in find_directory (name=<value optimized out>) at > > ../make-3.82/dir.c:537 > > 537 Â Â Â Â Â Â Â Â Â else if (!strcmp(fs_type, "FAT")) > > (gdb) bt full > > #0 Â 0x00403a03 in find_directory (name=<value optimized out>) at > > ../make-3.82/dir.c:537 > > Â Â Â Â dc = Cannot access memory at address 0x63726dcf > > (gdb) > > As previously stated, I'm no GDB expert, but does the fact that GDB > indicates that 'dc' could not access memory mean that the issue is > with 'dc' (which is a malloc)? Not necessarily. In an optimized program, dc could be in a register or in some other strange place, and GDB could be unable to access the correct value. What does "info address dc" display at that spot?