On Thu, 10 Jun 2021, Christoph Hellwig wrote: > On Wed, Jun 09, 2021 at 09:38:38AM -0400, Mikulas Patocka wrote: > > Hi > > > > This is the second version of the patch that avoids kcopyd. It sorts the > > requests according to sector number when writing to the slow device. But > > still, kcopyd is faster in my tests. > > So kcopyd is faster but you replace it by an open coded version? > Why? I thought that the open-coded version would perform better. This is just a test patch, I do not intend to push upstream. I tested it on IDE disk and SATA300 ssd. I am interested to see how it will perform on NVME ssd. Mikulas -- dm-devel mailing list dm-devel@xxxxxxxxxx https://listman.redhat.com/mailman/listinfo/dm-devel
