Re: RDONLY but for the good ( pledged ) guys

2018-07-18 Thread sven falempin
On Tue, Jun 26, 2018 at 7:05 PM Edgar Pettijohn III wrote: > > > > On 06/26/18 13:50, Raul Miller wrote: > > Personally, I can't totally figure out what this policy would be. > > > > My current best approximation is: there's a period of time when > > pkg_add and syspatch are running and that is a

Re: RDONLY but for the good ( pledged ) guys

2018-06-26 Thread Edgar Pettijohn III
On 06/26/18 13:50, Raul Miller wrote: Personally, I can't totally figure out what this policy would be. My current best approximation is: there's a period of time when pkg_add and syspatch are running and that is a time when writes are allowed, other than that, not. I could maybe rig up some

Re: RDONLY but for the good ( pledged ) guys

2018-06-26 Thread Raul Miller
Personally, I can't totally figure out what this policy would be. My current best approximation is: there's a period of time when pkg_add and syspatch are running and that is a time when writes are allowed, other than that, not. I could maybe rig up something more complicated using inherited cryp

RDONLY but for the good ( pledged ) guys

2018-06-26 Thread sven falempin
Hello, As a user i come across one use case where i m thinking : i do not want any program/exec to modify base or local base ( (/usr and /bin /bsd etc.. ) except syspatch and pkg_add -u. Please stop and tell if it does not make sense. I did look at pledge(2) and mount as pledge may force rdonl