Re: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-19 Thread Herbert Xu
On Mon, Mar 19, 2018 at 11:31:24PM +0800, Ard Biesheuvel wrote: > > Apologies if this wasn't clear, but there are some cross dependencies > with the arm64 tree, which receives non-trivial modifications in > patches 10 and 11, which are subsequently depended upon by patches 12 > - 23. > > Without

Re: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-19 Thread Ard Biesheuvel
On 16 March 2018 at 23:57, Herbert Xu wrote: > On Sat, Mar 10, 2018 at 03:21:45PM +, Ard Biesheuvel wrote: >> As reported by Sebastian, the way the arm64 NEON crypto code currently >> keeps kernel mode NEON enabled across calls into skcipher_walk_xxx() is >>

Re: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-16 Thread Herbert Xu
On Sat, Mar 10, 2018 at 03:21:45PM +, Ard Biesheuvel wrote: > As reported by Sebastian, the way the arm64 NEON crypto code currently > keeps kernel mode NEON enabled across calls into skcipher_walk_xxx() is > causing problems with RT builds, given that the skcipher walk API may > allocate and

Re: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-11 Thread Ard Biesheuvel
us...@vger.kernel.org; Peter Zijlstra <pet...@infradead.org>; Catalin >> Marinas <catalin.mari...@arm.com>; Will Deacon >> <will.dea...@arm.com>; Steven Rostedt <rost...@goodmis.org>; Thomas >> Gleixner <t...@linutronix.de> >> Subject: [PATCH

RE: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-10 Thread Vakul Garg
lt;rost...@goodmis.org>; Thomas > Gleixner <t...@linutronix.de> > Subject: [PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT > > As reported by Sebastian, the way the arm64 NEON crypto code currently > keeps kernel mode NEON enabled across calls into skcipher_w

[PATCH v5 00/23] crypto: arm64 - play nice with CONFIG_PREEMPT

2018-03-10 Thread Ard Biesheuvel
As reported by Sebastian, the way the arm64 NEON crypto code currently keeps kernel mode NEON enabled across calls into skcipher_walk_xxx() is causing problems with RT builds, given that the skcipher walk API may allocate and free temporary buffers it uses to present the input and output arrays to