Thanks for you answer. It makes sense to your reply. We will work along
The problem is that the application is multi-thread and has about 300,000
It is not easy to find one problem of STACK.
Well, I would like to make a few questions, if you know have the answers.
1º- Do You know what is the default size of the stack in Gumstix?
2º- You know if you can change the size of the default stack?
3º- Is There a stack for each thread?
4º- Is there is some command of the compiler (GCC) to deploy a stack of
checks on each function?
We try with: -fstack-protector-all and -fstack-check
But have a problem with the libssp:
- warning: the 'gets' function is dangerous and should not be used.
(not important problem)
- This library isn´t in the Gumstix. :-(
Thanks for you help in advanced.
2008/11/2 Dave Hylands <dhylands@...>
> Hi Sir Graham,
> > We debuggin with the gdb in local, and when have the error, it shows:
> > Program received signal SIGSEGV, Segmentation fault.
> > 0x40150d64 in ?? ()
> > The address is odd (very high for our application), and if we try to
> > the stack only appears:
> > 0x40150d64 in ?? ()
> This typically means that the stack pointer itself has been corrupted.
> This can happen if you overwrite a variable on the stack.
> char foo;
> strcpy( foo, "This is a test" );
> will corrupt the stack beyond where the variable is declared. The way
> the stack is laid out, the return address and frame pointer are stored
> at a higher memory location than local variables (stack starts in high
> memory and grows down). When a function is called, parameters, the
> return address are pushed on the stack, and the called function may
> push what's known as a frame pointer. The debugger uses the frame
> pointer to figure out stack crawls.
> Not sure if this helps or not, but if you look at the address
> 0x40150d64 and do a byte swap (since the gumstix is little endian)
> 640d1540 then 640d looks likeit might be ascii 0x64 = 'd' and 0x0d =
> '\r' (or carriage return). 0x15 is a NAK and 0x40 is the @
> Hard to know if this is at all relevant, but I figured I'd point it out.
> Dave Hylands
> Shuswap, BC, Canada
> http://www.DaveHylands.com/ <http://www.davehylands.com/>
> This SF.Net email is sponsored by the Moblin Your Move Developer's
> Build the coolest Linux based applications with Moblin SDK & win great
> Grand prize is a trip for two to an Open Source event anywhere in the world
> gumstix-users mailing list