Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/10/2016 12:32 PM, Eric Anholt wrote: Stephen Warren writes: On 04/08/2016 12:20 PM, Eric Anholt wrote: Stephen Warren writes: On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy".

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/10/2016 12:32 PM, Eric Anholt wrote: Stephen Warren writes: On 04/08/2016 12:20 PM, Eric Anholt wrote: Stephen Warren writes: On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

Stephen Warren writes: > On 04/08/2016 12:20 PM, Eric Anholt wrote: >> Stephen Warren writes: >> >>> On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

Stephen Warren writes: > On 04/08/2016 12:20 PM, Eric Anholt wrote: >> Stephen Warren writes: >> >>> On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/08/2016 12:20 PM, Eric Anholt wrote: Stephen Warren writes: On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs.

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/08/2016 12:20 PM, Eric Anholt wrote: Stephen Warren writes: On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs. This is a less strong

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

Stephen Warren writes: > On 04/04/2016 09:44 PM, Eric Anholt wrote: >> dsb() requires an argument on arm64, so we needed to add "sy". >> Instead, take this opportunity to switch to the same smp_wmb() call >> that gic uses for its IPIs. This is a less strong barrier than

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

Stephen Warren writes: > On 04/04/2016 09:44 PM, Eric Anholt wrote: >> dsb() requires an argument on arm64, so we needed to add "sy". >> Instead, take this opportunity to switch to the same smp_wmb() call >> that gic uses for its IPIs. This is a less strong barrier than we >> were doing before

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs. This is a less strong barrier than we were doing before (dmb(ishst) compared to dsb(sy)), but

Re: [PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

On 04/04/2016 09:44 PM, Eric Anholt wrote: dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs. This is a less strong barrier than we were doing before (dmb(ishst) compared to dsb(sy)), but

[PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs. This is a less strong barrier than we were doing before (dmb(ishst) compared to dsb(sy)), but it seems to be the correct one.

[PATCH 4/4] irqchip: bcm2836: Use a more generic memory barrier call

dsb() requires an argument on arm64, so we needed to add "sy". Instead, take this opportunity to switch to the same smp_wmb() call that gic uses for its IPIs. This is a less strong barrier than we were doing before (dmb(ishst) compared to dsb(sy)), but it seems to be the correct one.