Re: [PATCH] arm64: Implement optimised IP checksum helpers

2016-05-12 Thread Luke Starrett
Hi Robin, On 5/12/2016 1:08 PM, Robin Murphy wrote: Hi Luke, On 12/05/16 16:34, Luke Starrett wrote: Hi Robin, I pulled this in to a userspace test app expecting that the __uint128_t type might cause GCC to emit 'ldp'. Seems like that was that your intent based on your commit note. Instead

Re: [PATCH] arm64: Implement optimised IP checksum helpers

2016-05-12 Thread Luke Starrett
Hi Robin, On 5/12/2016 1:08 PM, Robin Murphy wrote: Hi Luke, On 12/05/16 16:34, Luke Starrett wrote: Hi Robin, I pulled this in to a userspace test app expecting that the __uint128_t type might cause GCC to emit 'ldp'. Seems like that was that your intent based on your commit note. Instead

Re: [PATCH] arm64: Implement optimised IP checksum helpers

2016-05-12 Thread Luke Starrett
different way round from the generic version, so throw that into the mix too. Suggested-by: Luke Starrett <luke.starr...@broadcom.com> Signed-off-by: Robin Murphy <robin.mur...@arm.com> --- Having spent an evening poring over disassembler output, it turns out that there's really no

Re: [PATCH] arm64: Implement optimised IP checksum helpers

2016-05-12 Thread Luke Starrett
different way round from the generic version, so throw that into the mix too. Suggested-by: Luke Starrett Signed-off-by: Robin Murphy --- Having spent an evening poring over disassembler output, it turns out that there's really no need to drop into asm for this, hurrah! I've taken it for a spin

Re: [PATCH 1/1] arm64: Add ARM64 optimized IP checksum routine

2016-05-11 Thread Luke Starrett
On 5/11/2016 8:55 AM, Robin Murphy wrote: Hi Luke, On 10/05/16 19:27, Luke Starrett wrote: This change implements an optimized checksum for arm64, based loosely on the original arch/arm implementation. Load-pair is used for the initial 16B load, reducing the overall number of loads to two

Re: [PATCH 1/1] arm64: Add ARM64 optimized IP checksum routine

2016-05-11 Thread Luke Starrett
On 5/11/2016 8:55 AM, Robin Murphy wrote: Hi Luke, On 10/05/16 19:27, Luke Starrett wrote: This change implements an optimized checksum for arm64, based loosely on the original arch/arm implementation. Load-pair is used for the initial 16B load, reducing the overall number of loads to two

[PATCH 1/1] arm64: Add ARM64 optimized IP checksum routine

2016-05-10 Thread Luke Starrett
implementation. Changes were tested against LE and BE operation and for all possible mis-alignments. Signed-off-by: Luke Starrett <luke.starr...@broadcom.com> --- arch/arm64/include/asm/checksum.h | 57 +++ 1 file changed, 57 insertions(+) create mode 10064

[PATCH 1/1] arm64: Add ARM64 optimized IP checksum routine

2016-05-10 Thread Luke Starrett
implementation. Changes were tested against LE and BE operation and for all possible mis-alignments. Signed-off-by: Luke Starrett --- arch/arm64/include/asm/checksum.h | 57 +++ 1 file changed, 57 insertions(+) create mode 100644 arch/arm64/include/asm/checksum.h

[PATCH v2 1/1] arm64: dts: NS2 secondary core enablement via PSCI

2016-04-20 Thread Luke Starrett
Declare PSCI-1.0 node and enable CPU_ON method via PSCI. Spin-table memreserve has been removed as well as syscon based reset, as PSCI-1.0 expects reset implementation in firmware. Signed-off-by: Luke Starrett <luke.starr...@broadcom.com> --- Changes from v1: - No code changes, adding m

[PATCH v2 1/1] arm64: dts: NS2 secondary core enablement via PSCI

2016-04-20 Thread Luke Starrett
Declare PSCI-1.0 node and enable CPU_ON method via PSCI. Spin-table memreserve has been removed as well as syscon based reset, as PSCI-1.0 expects reset implementation in firmware. Signed-off-by: Luke Starrett --- Changes from v1: - No code changes, adding missing reviewers to CC list arch