Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-07-02 Thread Kees Cook
On Fri, May 10, 2019 at 01:43:36PM -0700, Kees Cook wrote: > This feature continues to cause more problems than it solves[1]. Its > intention was to check the bounds of page-allocator allocations by using > __GFP_COMP, for which we would need to find all missing __GFP_COMP > markings. This work

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-06-10 Thread Kees Cook
On Mon, Jun 10, 2019 at 03:30:55PM -0700, Eric Biggers wrote: > Any progress on this patch? I have no had time yet; sorry. If anyone else would like to take a stab at it, I'd appreciate it. :) -- Kees Cook

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-06-10 Thread Eric Biggers
On Mon, May 13, 2019 at 02:32:43PM -0700, Kees Cook wrote: > On Sat, May 11, 2019 at 09:11:42PM -0700, Matthew Wilcox wrote: > > On Sat, May 11, 2019 at 05:03:08PM -0700, Kees Cook wrote: > > > On Fri, May 10, 2019 at 08:41:43PM -0400, Laura Abbott wrote: > > > > On 5/10/19 3:43 PM, Kees Cook

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-05-13 Thread Kees Cook
On Sat, May 11, 2019 at 09:11:42PM -0700, Matthew Wilcox wrote: > On Sat, May 11, 2019 at 05:03:08PM -0700, Kees Cook wrote: > > On Fri, May 10, 2019 at 08:41:43PM -0400, Laura Abbott wrote: > > > On 5/10/19 3:43 PM, Kees Cook wrote: > > > > This feature continues to cause more problems than it

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-05-11 Thread Matthew Wilcox
On Sat, May 11, 2019 at 05:03:08PM -0700, Kees Cook wrote: > On Fri, May 10, 2019 at 08:41:43PM -0400, Laura Abbott wrote: > > On 5/10/19 3:43 PM, Kees Cook wrote: > > > This feature continues to cause more problems than it solves[1]. Its > > > intention was to check the bounds of page-allocator

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-05-11 Thread Kees Cook
On Fri, May 10, 2019 at 08:41:43PM -0400, Laura Abbott wrote: > On 5/10/19 3:43 PM, Kees Cook wrote: > > This feature continues to cause more problems than it solves[1]. Its > > intention was to check the bounds of page-allocator allocations by using > > __GFP_COMP, for which we would need to find

Re: [PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-05-10 Thread Laura Abbott
On 5/10/19 3:43 PM, Kees Cook wrote: This feature continues to cause more problems than it solves[1]. Its intention was to check the bounds of page-allocator allocations by using __GFP_COMP, for which we would need to find all missing __GFP_COMP markings. This work has been on hold and there is

[PATCH] usercopy: Remove HARDENED_USERCOPY_PAGESPAN

2019-05-10 Thread Kees Cook
This feature continues to cause more problems than it solves[1]. Its intention was to check the bounds of page-allocator allocations by using __GFP_COMP, for which we would need to find all missing __GFP_COMP markings. This work has been on hold and there is an argument[2] that such markings are