On Fri, Apr 08, 2022 at 02:32:46PM +0900, Wangseok Lee wrote: > I would like to continue reviewing the pcie-designware-host.c patch below. > https://lore.kernel.org/all/20220328023009epcms2p309a5dfc2ff29d0a9945f65799963193c@epcms2p3/ The only new comment that I have is that the existing code also looks completely broken. Why did the DMA_ATTR_SKIP_CPU_SYNC magically appear in commit 07940c369a6b ("PCI: dwc: Fix MSI page leakage in suspend/resume") and what is is tƦying to do except for completely breaking non-coherent plaforms?
