From mboxrd@z Thu Jan 1 00:00:00 1970 From: Leon Romanovsky Subject: Re: [PATCH] ib_core: Enable and expose force_mr module parameter Date: Tue, 20 Jun 2017 21:40:46 +0300 Message-ID: <20170620184046.GW17846@mtr-leonro.local> References: <20170619152351.2866.11046.stgit@klimt.1015granger.net> <20170620073236.GO17846@mtr-leonro.local> <20170620180348.GV17846@mtr-leonro.local> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha256; protocol="application/pgp-signature"; boundary="87MiR7gHvrw39A9h" Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-rdma-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Chuck Lever Cc: doug.ledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org, linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-rdma@vger.kernel.org --87MiR7gHvrw39A9h Content-Type: text/plain; charset=us-ascii Content-Disposition: inline On Tue, Jun 20, 2017 at 02:23:21PM -0400, Chuck Lever wrote: > > > On Jun 20, 2017, at 2:03 PM, Leon Romanovsky wrote: > > > > On Tue, Jun 20, 2017 at 11:43:56AM -0400, Chuck Lever wrote: > >> Hi Leon- > >> > >> > >>> On Jun 20, 2017, at 3:32 AM, Leon Romanovsky wrote: > >>> > >>> On Mon, Jun 19, 2017 at 11:26:40AM -0400, Chuck Lever wrote: > >>>> The fourth parameter of the module_param_named macro is a set of > >>>> file permissions. Passing 0 there means that module parameter is > >>>> not created and that adding "options ib_core force_mr=1" to a > >>>> modprobe.conf file has no effect. > >>>> > >>>> The default setting of rdma_rw_force_mr continues to be 0, or false. > >>>> > >>>> Fixes: a060b5629ab0 ("IB/core: generic RDMA READ/WRITE API") > >>>> Signed-off-by: Chuck Lever > >>>> --- > >>>> Hi Doug- > >>>> > >>>> This doesn't seem appropriate to go through Bruce's tree for 4.13. > >>>> > >>>> Last discussion didn't seem to conclude with full consensus. > >>>> Probably people don't care enough one way or another. But I'd like > >>>> to see this get fixed if there aren't strong objections. Would you > >>>> take it for 4.13? > >>> > >>> I care and believe that it should be removed from exposure to users. > >>> The variable force_mr should be leaved in the code for the debug, > >>> but module_param should be removed. > >> > >> I don't understand the technical grounds of your objection. > >> Why is > >> > >> options mlx4_core internal_err_reset=0 > >> > >> acceptable, but > >> > >> options ib_core force_mr=1 > >> > >> not? > > > > That mlx4 variable was exposed before I started to follow after > > Mellanox's submissions. > > > > The force_mr parameter is for debug of new ULP and/or conversion of > > existing ULP to RW API. This is not very common situation and it is > > for the developers and not for the users. > > No disagreement on that, but that's still not an argument > not to expose it. What is the risk, as you see it? It will be "marked" as UAPI and we won't be able to change it, which IMHO bad for the debug. It also invites users to set it and in perfect world, we should test our HCAs with this flag on and off, because curious user can set it. Thanks > > > > Thanks > > > >> > >> > >>> Thanks > >>> > >>> > >>>> > >>>> > >>>> drivers/infiniband/core/rw.c | 2 +- > >>>> 1 file changed, 1 insertion(+), 1 deletion(-) > >>>> > >>>> diff --git a/drivers/infiniband/core/rw.c b/drivers/infiniband/core/rw.c > >>>> index dbfd854..1cc8f07 100644 > >>>> --- a/drivers/infiniband/core/rw.c > >>>> +++ b/drivers/infiniband/core/rw.c > >>>> @@ -23,7 +23,7 @@ enum { > >>>> }; > >>>> > >>>> static bool rdma_rw_force_mr; > >>>> -module_param_named(force_mr, rdma_rw_force_mr, bool, 0); > >>>> +module_param_named(force_mr, rdma_rw_force_mr, bool, 0644); > >>>> MODULE_PARM_DESC(force_mr, "Force usage of MRs for RDMA READ/WRITE operations"); > >>>> > >>>> /* > >>>> > >>>> -- > >>>> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in > >>>> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > >>>> More majordomo info at http://vger.kernel.org/majordomo-info.html > >> > >> -- > >> Chuck Lever > > -- > Chuck Lever > > > --87MiR7gHvrw39A9h Content-Type: application/pgp-signature; name="signature.asc" -----BEGIN PGP SIGNATURE----- iQIzBAEBCAAdFiEEkhr/r4Op1/04yqaB5GN7iDZyWKcFAllJbC0ACgkQ5GN7iDZy WKcpRQ/9FS527pV9hE0ua75uRw/YqVOL25MvqvMC4kOvS+GjGifsQmftpDpc/AuG qEtW9IRpUzztND3T/bu+E3rMIfIVF+Jl4NC2bhY/Er7CN7AELeXTZV5unYbLvDL6 L45xOm7kGrTazCjLKWMyzv8OT7yEDDBawV7EaAsGBlcd2pwThRrqzyMIcZoYN2SM z6ul8UwvYhCPi5eGq+piE5Gqla1xmSY96LR6xDnD8i8Oednp5VsVW6sdoKOfxWai y+3Gl0OcoYxOHtpOeVu/aSh34uTEXcFMPDnjsMG4372MpYlw5fnWVIPtFsBPrdyq veefDrIxAT+dFwOeVW8lA9McA5xRSrMkBT/UllMSILImh3fEBbGT+FFrdOca7+sG TyTKzFCI5B0s0ner9HQk09AZBBqVhFhq4XzbmIxNa78tLuaFMTucBMSQ59RyFwnr 4qWOGrXvJx7lWjRUcTv4GbAfmOiNvc99dzk9LO9gJPCE8Prwvlxcp6mpBqjBeqD6 Vb9+S9yV3fLaqOl5xR4wTnQyiwE1oHb1fZOlYVHDQwmTj38cRFgwRxNitOy17Bzw 9agZ58lCdXkyfYspcHnQY2+b62w7NIvt3XncVz+eXmKydAKnWtk3lpfIOVHpKJ9K Pruk8ijhn5E8JU8fU94SuLHy2EwJPvQWCDVrI0FCORbOI7H/M/g= =jPGc -----END PGP SIGNATURE----- --87MiR7gHvrw39A9h-- -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html