Re: [PATCH v2] general protection fault in sock_has_perm

2018-02-02 Thread Greg KH
On Thu, Feb 01, 2018 at 07:37:04AM -0800, Mark Salyzyn wrote: > In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE socket > flag") and all the associated infrastructure changes to take advantage > of a RCU grace period before freeing, there is a heightened > possibility that a security

Re: [PATCH v2] general protection fault in sock_has_perm

2018-02-01 Thread Mark Salyzyn
On 02/01/2018 09:02 AM, Stephen Smalley wrote: On Thu, 2018-02-01 at 08:20 -0800, Mark Salyzyn wrote: On 02/01/2018 08:00 AM, Paul Moore wrote: On Thu, Feb 1, 2018 at 10:37 AM, Mark Salyzyn wrote: In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE socket

Re: [PATCH v2] general protection fault in sock_has_perm

2018-02-01 Thread Greg KH
On Thu, Feb 01, 2018 at 08:20:13AM -0800, Mark Salyzyn wrote: > On 02/01/2018 08:00 AM, Paul Moore wrote: > > On Thu, Feb 1, 2018 at 10:37 AM, Mark Salyzyn wrote: > > > In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE socket > > > flag") and all the associated

Re: [PATCH v2] general protection fault in sock_has_perm

2018-02-01 Thread Stephen Smalley
On Thu, 2018-02-01 at 08:20 -0800, Mark Salyzyn wrote: > On 02/01/2018 08:00 AM, Paul Moore wrote: > > On Thu, Feb 1, 2018 at 10:37 AM, Mark Salyzyn > > wrote: > > > In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE > > > socket > > > flag") and all the

Re: [PATCH v2] general protection fault in sock_has_perm

2018-02-01 Thread Paul Moore
On Thu, Feb 1, 2018 at 11:20 AM, Mark Salyzyn wrote: > On 02/01/2018 08:00 AM, Paul Moore wrote: >> >> On Thu, Feb 1, 2018 at 10:37 AM, Mark Salyzyn wrote: >>> >>> In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE socket >>> flag") and all

[PATCH v2] general protection fault in sock_has_perm

2018-02-01 Thread Mark Salyzyn
In the absence of commit a4298e4522d6 ("net: add SOCK_RCU_FREE socket flag") and all the associated infrastructure changes to take advantage of a RCU grace period before freeing, there is a heightened possibility that a security check is performed while an ill-timed setsockopt call races in from