> -----Original Message----- > From: David Howells <dhowells@xxxxxxxxxx> > Sent: Monday, 20 March 2023 14:13 > To: Bernard Metzler <BMT@xxxxxxxxxxxxxx> > Cc: David Howells <dhowells@xxxxxxxxxx>; Tom Talpey <tom@xxxxxxxxxx>; > linux-rdma@xxxxxxxxxxxxxxx > Subject: [EXTERNAL] Re: [RFC PATCH 08/28] siw: Inline do_tcp_sendpages() > > Bernard Metzler <BMT@xxxxxxxxxxxxxx> wrote: > > > > > Back in the days, I introduced that zcopy path for efficiency > > > > reasons - getting both better throughput and less CPU load. > > > > I looked at both WRITE and READ performance. Using > > > > do_tcp_sendpages() is currently limited to processing work > > > > which is not registered with local completion generation. > > > > Replying to a remote READ request is a typical case. Did > > > > you check with READ? > > > > > > Ah - you're talking about ksmbd there? I haven't tested the patch with > > > that. > > > > Did you test with both kernel ULPs and user level applications? > > Kernel "ULPs"? I was trying to refer to kernel applications or clients or upper layer protocols (ulp, like nfs). > > As far as cifs goes, I've tested the fs with large dd commands for the > moment, > but that's all. This post was more to find out how attached people were to > ->sendpage() and to see if anyone had any preferences on a couple of things > mentioned in the cover note. This isn't aimed at the next merge window. I like your patches to siw a lot, since it would significantly simplify the transmit code path. Thank you, Bernard. > > David
