Re: [GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread James Morris
On Fri, 29 Sep 2017, Jason A. Donenfeld wrote: > On Fri, Sep 29, 2017 at 12:33 AM, James Morris wrote: > > Generally speaking, we likely need to improve the amount of crypto review > > for kernel crypto users including keys (I'll post a note separately to > > ksummit-discuss).

Re: [GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread James Morris
On Fri, 29 Sep 2017, Jason A. Donenfeld wrote: > On Fri, Sep 29, 2017 at 12:33 AM, James Morris wrote: > > Generally speaking, we likely need to improve the amount of crypto review > > for kernel crypto users including keys (I'll post a note separately to > > ksummit-discuss). > > Indeed. > >

Re: [GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread Jason A. Donenfeld
On Fri, Sep 29, 2017 at 12:33 AM, James Morris wrote: > Generally speaking, we likely need to improve the amount of crypto review > for kernel crypto users including keys (I'll post a note separately to > ksummit-discuss). Indeed. I won't be at kernel summit, regrettably, but

Re: [GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread Jason A. Donenfeld
On Fri, Sep 29, 2017 at 12:33 AM, James Morris wrote: > Generally speaking, we likely need to improve the amount of crypto review > for kernel crypto users including keys (I'll post a note separately to > ksummit-discuss). Indeed. I won't be at kernel summit, regrettably, but I do intend to

[GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread James Morris
Hi Linus, Please pull these fixes for the keys subsystem. Notable here is a rewrite of big_key crypto by Jason Donenfeld to address some issues in the original code. >From Jason's commit log: "This started out as just replacing the use of crypto/rng with get_random_bytes_wait, so that we

[GIT PULL] security: keys fixes for v4.14-rc3

2017-09-28 Thread James Morris
Hi Linus, Please pull these fixes for the keys subsystem. Notable here is a rewrite of big_key crypto by Jason Donenfeld to address some issues in the original code. >From Jason's commit log: "This started out as just replacing the use of crypto/rng with get_random_bytes_wait, so that we