Re: [GSOC] capsicum week6 report

[Loganaden Velvindron <loganaden@xxxxxxxxx>]

--047d7b5d57c816901904e2a13241
Content-Type: text/plain; charset=ISO-8859-1

Awesome man :-)

Keep it up !


On Mon, Jul 29, 2013 at 10:29 AM, Joris Giovannangeli <
joris@giovannangeli.fr> wrote:

> Hi,
>
> this week I've been stabilizing my work a bit more. The implementation
> is still rough but "functionnal". I've finished the capability mode by
> modifying the sycall path and the syscall table to check if a syscall is
> allowed in capability mode and return ECAPMODE if not. I've converted
> most of the holdfp calls to check capability rights, but there is still
> some calls I must read more carefully to find which rights they need.
> I've made the nlookup path return ECAPMODE when doing a non-relative
> lookup. For now, the code is not perfect because i can leak information
> from outside the sandbox. Indeed, if an early error is returned, I don't
> check if the current namecache entry is still in sandbox and return the
> error (EEXIST, ENOENT, etc). This is a bug. I've been trying to modify
> this behaviour, but for now, it's completely broken.
> To test the work, i've ported the capsicum-enabled bzip2 a student has
> done for FreeBSD during gsoc :
> https://svnweb.freebsd.org/socsvn/soc2013/dpl/head/ . I've run it on a
> vkernel, and it can compress and decompress files w/o troubles.
>
> Joris
>
>


-- 
This message is strictly personal and the opinions expressed do not
represent those of my employers, either past or present.

--047d7b5d57c816901904e2a13241
Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Awesome man :-)<div><br></div><div>Keep it up !</div></div=
><div class=3D"gmail_extra"><br><br><div class=3D"gmail_quote">On Mon, Jul =
29, 2013 at 10:29 AM, Joris Giovannangeli <span dir=3D"ltr">&lt;<a href=3D"=
mailto:joris@giovannangeli.fr"; target=3D"_blank">joris@giovannangeli.fr</a>=
&gt;</span> wrote:<br>
<blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1p=
x #ccc solid;padding-left:1ex">Hi,<br>
<br>
this week I&#39;ve been stabilizing my work a bit more. The implementation<=
br>
is still rough but &quot;functionnal&quot;. I&#39;ve finished the capabilit=
y mode by<br>
modifying the sycall path and the syscall table to check if a syscall is<br=
>
allowed in capability mode and return ECAPMODE if not. I&#39;ve converted<b=
r>
most of the holdfp calls to check capability rights, but there is still<br>
some calls I must read more carefully to find which rights they need.<br>
I&#39;ve made the nlookup path return ECAPMODE when doing a non-relative<br=
>
lookup. For now, the code is not perfect because i can leak information<br>
from outside the sandbox. Indeed, if an early error is returned, I don&#39;=
t<br>
check if the current namecache entry is still in sandbox and return the<br>
error (EEXIST, ENOENT, etc). This is a bug. I&#39;ve been trying to modify<=
br>
this behaviour, but for now, it&#39;s completely broken.<br>
To test the work, i&#39;ve ported the capsicum-enabled bzip2 a student has<=
br>
done for FreeBSD during gsoc :<br>
<a href=3D"https://svnweb.freebsd.org/socsvn/soc2013/dpl/head/"; target=3D"_=
blank">https://svnweb.freebsd.org/socsvn/soc2013/dpl/head/</a> . I&#39;ve r=
un it on a<br>
vkernel, and it can compress and decompress files w/o troubles.<br>
<span class=3D"HOEnZb"><font color=3D"#888888"><br>
Joris<br>
<br>
</font></span></blockquote></div><br><br clear=3D"all"><div><br></div>-- <b=
r><div dir=3D"ltr"><div style=3D"text-align:left">This message is strictly =
personal and the opinions expressed do not represent those of my employers,=
 either past or present.</div>
<br><br>=A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
 =A0 =A0 =A0 =A0 =A0 <br><br></div>
</div>

--047d7b5d57c816901904e2a13241--