On 2019/06/27 17:25, Christoph Hellwig wrote: > On Thu, Jun 27, 2019 at 08:17:08AM +0000, Damien Le Moal wrote: >>> Are your sure that invalidate[|flush]_kernel_vmap_range is needed for >>> bio_copy_kernel? The vmalloc buffer isn't involved in IO, and only >>> accessed by CPU. >> >> Not sure at all. I am a little out of my league here. >> Christoph mentioned that this is necessary. >> >> Christoph, can you confirm ? > > Ming is right. Sorry for misleading you. No problem. Sending a V5 with cleanups. Thanks ! -- Damien Le Moal Western Digital Research
