From: Leon Romanovsky Subject: Re: [PATCH mlx5-next 3/4] net/mlx5: Decrease default mr cache size Date: Wed, 27 Mar 2019 15:36:07 +0200 Message-ID: <20190327133606.GY22899@mtr-leonro.mtl.com> References: <20190319092439.10701-1-leon@kernel.org> <20190319092439.10701-4-leon@kernel.org> <20190327114101.GT22899@mtr-leonro.mtl.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="K0M4iRgkWWXhiQ9N" Return-path: Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org To: Or Gerlitz Cc: Doug Ledford , Jason Gunthorpe , RDMA mailing list , Artemy Kovalyov , Guy Levi , Haggai Eran , Jerome Glisse , Moni Shoua , Saeed Mahameed , linux-netdev List-Id: linux-rdma@vger.kernel.org --K0M4iRgkWWXhiQ9N Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Wed, Mar 27, 2019 at 01:58:17PM +0200, Or Gerlitz wrote: > On Wed, Mar 27, 2019 at 1:41 PM Leon Romanovsky wrote: > > On Wed, Mar 27, 2019 at 12:07:54PM +0200, Or Gerlitz wrote: > > > On Tue, Mar 19, 2019 at 11:25 AM Leon Romanovsky wrote: > > > > From: Artemy Kovalyov > > > > > > > > Delete initialization of high order entries in mr cache to decrease initial > > > > memory footprint. When required, the administrator can populate the > > > > entries with memory keys via the /sys interface. > > > > > > Please add here: > > > > > > This approach is very helpful to reduce the per HW function memory > > > footprint in environments such as VMs. Before the patch we see > > > consumption of 0.9GB per function and after the patch about 0.1GB > > > > > > > Signed-off-by: Artemy Kovalyov > > > > Signed-off-by: Moni Shoua > > > > Signed-off-by: Leon Romanovsky > > > > > > Reported-by: Shalom Toledo > > > Acked-by: Or Gerlitz > > > > > > Lets push it into stable kernels, a Fixes tag here will cause that > > > to happen more easily, so please consider that. > > > I'll add it at the "apply" stage. > > Just to make sure, by "add it" you mean the signatures, the text and > the Fixes that? Yes, of course. > > Also, is there any reason not to eliminate the mr cache pre-population > all together? AFAIK, pre-populated cache helps MPI application to start faster and with out-of-box experience. Nobody seems to care enough to challenge this internal assumption. > > Currently we consume 100-200MB per function after the patch which is > also problematic for some > environments. --K0M4iRgkWWXhiQ9N Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIcBAEBAgAGBQJcm3xGAAoJEORje4g2clinLcsP/0YcbbLKj6k71w6Hv7c8U0Rj JdwBtuSBJXUo/VCPUkO4U1uCZRROdx+YfkMAmfYHWxBaKzw8OAQksc3m3xt/FqPu vmOwPfabpP4mP769p86tZz1nTBlgTBoQKOPoCm+i1U+0TTV79b5ykaIuuRReLEgK 7PVGyi8cdsR5B2lNyO4fStoss4ZRm8n3nN9qyZAO/aAwKlrvTltbKAMnjmOxdyw1 PQ1eX308E3fjUm6YrnbywALqBjko8Q/5YDt777TUoyS66A0d16jAjikoGfeVw/dA djTg4Ge9u/VfSiUkgVy3ZXw8pRx1zX9ZBAmEa+L+1DLvc8KJKXohxqi/ziBM2X41 AEELhttresXkKIOygJ2CZ1aTooyDVzxGC9jZoTQUsDbvSa8xXwIaeEOhg7XQ9Zy3 FVpMFKTU9SNWxuunAKvRaoHIiEGSlDLZqgHn8ItPSBMRXT+3lmI8epdF9//fi9w3 VrZTFvrZILptU9GhlygXMYredgTNpnfXrIx4by3NCgBX294LlZfgw5Rgn+sE2DP7 HxG/kGLOh8kG89U0qWuq+ApnjWcQZTozHA/sZ+5haGpye24Jjq9sPdyUlrzNyntM KG64jR2j58bHNy5tKRMW2AuGxJUH7dhRcg2B8nY0QcFkD/aHScFaFHn+hTEUNYqN D2OnF+jqazIH8i5k2VsV =QJYi -----END PGP SIGNATURE----- --K0M4iRgkWWXhiQ9N--