Re: [PATCH] capabilities: Ambient capability set V1

On Thu, Feb 26, 2015 at 11:38:01AM -0800, Andy Lutomirski wrote:

On Thu, Feb 26, 2015 at 11:32 AM, Serge E. Hallyn [off-list ref] wrote:

quoted

On Thu, Feb 26, 2015 at 12:28:30PM -0600, Christoph Lameter wrote:

quoted

On Thu, 26 Feb 2015, Serge E. Hallyn wrote:

quoted

quoted

Well doing that breaks su.

Don't what exactly?  You're saying that doing

pI' = pI
pA' = pA  (pA is ambient)
pP' = (X & fP) | (pI & (fI | pA))
pE' = pP' & (fE | pA)

stopped su from having CAP_SETGID while

pI' = pI
pA' = pA  (pA is ambient)
pP' = (X & fP) | (pI & (fI | pA))
pE' = pP' & fE

worked?  I'll hope you're saying both "fail", in which case

fE does not exist in cpu_vfs_cap_data. We only get fI and fP. Why in the
world does setcap allow setting fE?

It sets a bit in the magic_etc.  So fE is either all on or all off.

quoted

V1 does not use fE. In my newer attempt, I seemed to have worked
with zeroed field that I assumed was filled in.

I will just do

pE' = pP'

Ok?

Andrew Morgan was against that.  What if we changed

pE' = pP' & (fE | pA)

to

        if (pA)
                pE' = pP' & fE
        else
                pE' = pP'

Is this backwards?

D'oh, yes.

Also, on further bikeshedding consideration, I think I like the name
"persistent" much better than "ambient".  Alas, "persistent" starts
with a P.

--Andy