Re: segmentation fault in sys_execve

[israel fdez <israel@seg.inf.cu>]

the best part is this: "I found this solution in the kernelnewbies Web site,
in the section code examples"
 :-[ 

mulix wrote:
> On Tue, Mar 05, 2002 at 10:49:42AM -0500, israel fdez wrote:
> > Hi Mulix, remember me? the one wih the sys_execve function problems?
> > I found the solution, the problem was that the eip register got lost 
> > when returnong thr regs structure to the original sys_execve function, 
> > the solution is not returning to the original function, and copy the 
> > source of this function to your my_sys_execve function.
> > Simple no?
> >     
>
> yes, pretty much. you should probably paste your solution to
> kernelnewbies, so other people know as well...
>
> > mulix wrote:
> >
> > > On Wed, Feb 27, 2002 at 03:43:30PM -0500, Israel Fernandez wrote:
> > >
> > > > Hi again... I'm changing the sys_call_table[__NR_excecve], in the
> > > > new function I do nothing, just return to "int sys_execve(struct
> > > > pt_regs regs)" doing this: return original_sys_execve(regs), where
> > > > regs is the pt_regs I receave in the new function, but when I excecute
> > > > any file I get a "Segmentation Fault" error message 
> > > > Any hint??
> > > >
> > > >           
> > > your problem likely stems from the fact that you're passing the
> > > original pt_regs by value on the stack instead of passing a pointer to
> > > it - in other words, sys_execve() relies on the location of the struct
> > > pt_regs on the stack.
> > >
> > >           
> >           
>
>