Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-08-02 Thread Andy Lutomirski
On Thu, Aug 2, 2012 at 7:32 AM, Will Drewry wrote: > On Thu, Jul 26, 2012 at 10:41 AM, Andy Lutomirski wrote: >> On Tue, Jul 17, 2012 at 4:19 PM, Andy Lutomirski wrote: >>> Currently, if a tracer changes a syscall nr to __NR_future_enosys, >>> behavior will differ between kernels that know about

Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-08-02 Thread Will Drewry
On Thu, Jul 26, 2012 at 10:41 AM, Andy Lutomirski wrote: > On Tue, Jul 17, 2012 at 4:19 PM, Andy Lutomirski wrote: >> Currently, if a tracer changes a syscall nr to __NR_future_enosys, >> behavior will differ between kernels that know about >> __NR_future_enosys (and return -ENOSYS) and older ker

Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-07-26 Thread Andy Lutomirski
On Tue, Jul 17, 2012 at 4:19 PM, Andy Lutomirski wrote: > Currently, if a tracer changes a syscall nr to __NR_future_enosys, > behavior will differ between kernels that know about > __NR_future_enosys (and return -ENOSYS) and older kernels (which > return the value from pt_regs). This is silly; w

Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-07-18 Thread Andy Lutomirski
On Wed, Jul 18, 2012 at 11:35 AM, Will Drewry wrote: > On Tue, Jul 17, 2012 at 9:13 PM, Will Drewry wrote: >> On Tue, Jul 17, 2012 at 6:19 PM, Andy Lutomirski wrote: >>> Currently, if a tracer changes a syscall nr to __NR_future_enosys, >>> behavior will differ between kernels that know about >>

Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-07-18 Thread Will Drewry
On Tue, Jul 17, 2012 at 9:13 PM, Will Drewry wrote: > On Tue, Jul 17, 2012 at 6:19 PM, Andy Lutomirski wrote: >> Currently, if a tracer changes a syscall nr to __NR_future_enosys, >> behavior will differ between kernels that know about >> __NR_future_enosys (and return -ENOSYS) and older kernels

Re: [PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-07-17 Thread Will Drewry
On Tue, Jul 17, 2012 at 6:19 PM, Andy Lutomirski wrote: > Currently, if a tracer changes a syscall nr to __NR_future_enosys, > behavior will differ between kernels that know about > __NR_future_enosys (and return -ENOSYS) and older kernels (which > return the value from pt_regs). This is silly; w

[PATCH 3.5 2/2] seccomp: Future-proof against silly tracers

2012-07-17 Thread Andy Lutomirski
Currently, if a tracer changes a syscall nr to __NR_future_enosys, behavior will differ between kernels that know about __NR_future_enosys (and return -ENOSYS) and older kernels (which return the value from pt_regs). This is silly; we should just return -ENOSYS. This is unlikely to ever happen on