Re: [PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-17 Thread Michael Ellerman
Christoph Hellwig writes: > diff --git a/arch/powerpc/kernel/dma.c b/arch/powerpc/kernel/dma.c > index dbfc7056d7df..d442d23e182b 100644 > --- a/arch/powerpc/kernel/dma.c > +++ b/arch/powerpc/kernel/dma.c > @@ -247,6 +252,8 @@ static inline void dma_nommu_unmap_page(struct device > *dev, >

Re: [PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-17 Thread Christoph Hellwig
On Mon, Dec 17, 2018 at 08:39:17AM +0100, Christophe Leroy wrote: > I can help you with powerpc 8xx actually. Below is a patch that implements the proper scheme on top of the series in this thread. Compile tested with tqm8xx_defconfig and tqm8xx_defconfig + CONFIG_HIGHMEM only. diff --git a/arch

Re: [PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-16 Thread Christophe Leroy
Le 17/12/2018 à 08:34, Christoph Hellwig a écrit : On Mon, Dec 17, 2018 at 07:41:54AM +0100, Christophe Leroy wrote: Le 16/12/2018 à 18:19, Christoph Hellwig a écrit : The unmap methods need to transfer memory ownership back from the device to the cpu by identical means as dma_sync_*_to_cp

Re: [PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-16 Thread Christoph Hellwig
On Mon, Dec 17, 2018 at 07:41:54AM +0100, Christophe Leroy wrote: > > > Le 16/12/2018 à 18:19, Christoph Hellwig a écrit : >> The unmap methods need to transfer memory ownership back from the device >> to the cpu by identical means as dma_sync_*_to_cpu. I'm not sure powerpc >> needs to do any work

Re: [PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-16 Thread Christophe Leroy
Le 16/12/2018 à 18:19, Christoph Hellwig a écrit : The unmap methods need to transfer memory ownership back from the device to the cpu by identical means as dma_sync_*_to_cpu. I'm not sure powerpc needs to do any work in this transfer direction, but given that it does invalidate the caches in

[PATCH 2/8] powerpc/dma: properly wire up the unmap_page and unmap_sg methods

2018-12-16 Thread Christoph Hellwig
The unmap methods need to transfer memory ownership back from the device to the cpu by identical means as dma_sync_*_to_cpu. I'm not sure powerpc needs to do any work in this transfer direction, but given that it does invalidate the caches in dma_sync_*_to_cpu already we should make sure we also d