On Mar 18, 8:15am, al...@yandex.ru (Alexander Nasonov) wrote:
-- Subject: Re: KASSERT in exec_elf.c for DYN executable when p_align==0
| Christos Zoulas wrote:
| > In article <20180317225722.GA1538@neva>,
| > Alexander Nasonov wrote:
| > >Coverity (CID 1427746) complains a
Christos Zoulas wrote:
> In article <20180317225722.GA1538@neva>,
> Alexander Nasonov wrote:
> >Coverity (CID 1427746) complains about a division by zero when
> >align is 0 in all PT_LOAD headers.
> >...
> >I would be nice to perform sanity checks of tainted executable
> >instead of panicing.
>
In article <20180317225722.GA1538@neva>,
Alexander Nasonov wrote:
>Coverity (CID 1427746) complains about a division by zero when
>align is 0 in all PT_LOAD headers.
>
>I tried reproducing the problem but the code in question is inside
>'if (offset < epp->ep_vm_minaddr)' and it isn't easily repro
Alexander Nasonov wrote:
> Steps to reproduce (on amd64 compiled with MKPIE=yes):
>
> bvi -s 0x0e2 /bin/echo # change 20 to 00
> bvi -s 0x11a /bin/echo # change 20 to 00
>
> /bin/echo # boom!
>
> I would be nice to perform sanity checks of tainted executable
> instead of panicing.
Attached is a
Coverity (CID 1427746) complains about a division by zero when
align is 0 in all PT_LOAD headers.
I tried reproducing the problem but the code in question is inside
'if (offset < epp->ep_vm_minaddr)' and it isn't easily reproducable.
However, I hit KASSERT panic:
"(offset & (align - 1)) == 0" fi