On Sun, 2021-02-14 at 14:13 +0000, Daire Byrne wrote: > > ----- On 13 Feb, 2021, at 23:30, Chuck Lever > chuck.lever@xxxxxxxxxx wrote: > > > > I don't have a performance system to measure the improvement > > > accurately. > > > > Then let's have Daire try it out, if possible. > > I'm happy to test it out on one of our 2 x 40G NFS servers with 100 x > 1G clients (but it's trickier to patch the clients too atm). > > Just so I'm clear, this is in addition to Chuck's "Handle TCP socket > sends with kernel_sendpage() again" patch from bz #209439 (which I > think is now in 5.11 rc)? Or you want to see what this patch looks > like on it's own without that (e.g. v5.10)? > This patch is independent of which socket mechanism we're using: it should work with sendmsg(), sendpage(), and even writev(). So feel free to test whichever combination makes most sense to you. -- Trond Myklebust Linux NFS client maintainer, Hammerspace trond.myklebust@xxxxxxxxxxxxxxx
