Re: pointer corruption in exec_script.c

2017-05-10 Thread Peter J. Philipp
Yeah, thanks... I don't know what I was drinking yesterday it was only ice-tea, sorry for that noise. In retrospect I unearthed another hole in my own non-committed implementation and it will need to be rewritten to work ever. While I profited on that knowledge, you guys did not or only

Re: pointer corruption in exec_script.c

2017-05-09 Thread Jimmy Hess
On Tue, May 9, 2017 at 3:27 PM, Peter J. Philipp wrote: > Sorry the patch is bad because I fail to check what I want to free for NULL. > You probably caught it. Better let someone with lots of experience handle > this. At a cursory glance. Check the referrentI

Re: pointer corruption in exec_script.c

2017-05-09 Thread Ted Unangst
Peter J. Philipp wrote: > In my tinkering with the ELFSEC mechanism, I have noticed something > possibly troubling. In /sys/kern/exec_script.c shellname is a pointer > to cp which is a pointer to hdrstr which is a pointer to epp->ep_hdr... > When calling the intended set shellname variable,

Re: pointer corruption in exec_script.c

2017-05-09 Thread Peter J. Philipp
On Tue, May 09, 2017 at 10:05:28PM +0200, Peter J. Philipp wrote: > Hi, > > In my tinkering with the ELFSEC mechanism, I have noticed something > possibly troubling. In /sys/kern/exec_script.c shellname is a pointer > to cp which is a pointer to hdrstr which is a pointer to epp->ep_hdr... > When

pointer corruption in exec_script.c

2017-05-09 Thread Peter J. Philipp
Hi, In my tinkering with the ELFSEC mechanism, I have noticed something possibly troubling. In /sys/kern/exec_script.c shellname is a pointer to cp which is a pointer to hdrstr which is a pointer to epp->ep_hdr... When calling the intended set shellname variable, later, I get part of the ELF