Re: [PATCH v4 00/14] arm64: Optimise and update memcpy, user copy and string routines

2020-09-11 Thread Oli Swede
Hi Will and Catalin, Thank you both for having a look at this, the coupling between the usercopy routines and the fixup routines was my main concern as well and I understand the desire to take an approach that avoids this dependency when it comes to importing new copy routines. On 9/11/20 12

Re: [PATCH v4 00/14] arm64: Optimise and update memcpy, user copy and string routines

2020-09-11 Thread Catalin Marinas
On Mon, Sep 07, 2020 at 11:10:03AM +0100, Will Deacon wrote: > On Wed, Jul 01, 2020 at 09:12:49AM +0100, Oli Swede wrote: > > > Version 3 addressed this but I later found some issues with the fixup > > > correctness after further testing, and have partially re-written them > > > here, and addressed

Re: [PATCH v4 00/14] arm64: Optimise and update memcpy, user copy and string routines

2020-09-07 Thread Will Deacon
Hi Oli, Thanks for this. Just a few high-level comments below. On Wed, Jul 01, 2020 at 09:12:49AM +0100, Oli Swede wrote: > > Version 3 addressed this but I later found some issues with the fixup > > correctness after further testing, and have partially re-written them > > here, and addressed som

Re: [PATCH v4 00/14] arm64: Optimise and update memcpy, user copy and string routines

2020-07-01 Thread Oli Swede
Hi all, (Apologies if you have already received this follow-up, please ignore any previous duplicates; I am re-sending this with the lists in the CC as the formatting filters caught me out :) Thanks ] > Version 3 addressed this but I later found some issues with the fixup > correctness after

[PATCH v4 00/14] arm64: Optimise and update memcpy, user copy and string routines

2020-06-30 Thread Oliver Swede
Hi, This contains an update to the cortex-strings patchset: the correctness of the fixup routines are improved, with the aim being to return the exact number of remaining bytes for all copy sizes. To ensure they are exact - which the current fixups are not for some copy sizes and are off by a few