Re: [PATCH 5.3 076/148] mmc: sdhci-of-esdhc: set DMA snooping based on DMA coherence

2019-10-10 Thread Greg Kroah-Hartman
On Thu, Oct 10, 2019 at 09:49:12AM +0100, Russell King - ARM Linux admin wrote: > Hi Greg, > > On 5th October, Christian Zigotzky reported a > problem with this on PowerPC (at a guess, it looks like there's a > PowerPC user of this where the DT does not mark the device as > dma-coherent, but the

Re: [PATCH 5.3 076/148] mmc: sdhci-of-esdhc: set DMA snooping based on DMA coherence

2019-10-10 Thread Russell King - ARM Linux admin
Hi Greg, On 5th October, Christian Zigotzky reported a problem with this on PowerPC (at a guess, it looks like there's a PowerPC user of this where the DT does not mark the device as dma-coherent, but the hardware requires it to be DMA coherent.) However, despite sending a reply to him within

[PATCH 5.3 076/148] mmc: sdhci-of-esdhc: set DMA snooping based on DMA coherence

2019-10-10 Thread Greg Kroah-Hartman
From: Russell King commit 121bd08b029e03404c451bb237729cdff76eafed upstream. We must not unconditionally set the DMA snoop bit; if the DMA API is assuming that the device is not DMA coherent, and the device snoops the CPU caches, the device can see stale cache lines brought in by speculative