| Message ID | 20170626183329.GA11716@ctung-MOBL3.amr.corp.intel.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
On Mon, Jun 26, 2017 at 01:33:29PM -0500, Chien Tin Tung wrote: > On Mon, Jun 26, 2017 at 09:13:24PM +0300, Leon Romanovsky wrote: > > On Mon, Jun 26, 2017 at 12:55:05PM -0500, Chien Tin Tung wrote: > > > On Tue, Jun 13, 2017 at 03:59:23AM -0500, Leon Romanovsky wrote: > > > > From: Leon Romanovsky <leonro@mellanox.com> > > > > > > > > RDMA netlink has complicated infrastructure to add and remove netlink > > > > clients to NETLINK_RDMA family. This complicates the code and not in > > > > use because not many clients are available (3 clients) and most of them > > > > (2 clients) are statically compiled together with netlink.c. > > > > > > > > The following patch refactors RDMA netlink and opens door for the future > > > > patches which will be able to get rid of a lot of dead iwcm* code. > > > > > > > > Signed-off-by: Leon Romanovsky <leonro@mellanox.com> > > > > Reviewed-by: Chien Tin Tung <chien.tin.tung@intel.com> > > > > --- > > > > drivers/infiniband/core/cma.c | 6 +- > > > > drivers/infiniband/core/device.c | 41 +++------ > > > > drivers/infiniband/core/iwcm.c | 10 +-- > > > > drivers/infiniband/core/netlink.c | 185 ++++++++++++++++++-------------------- > > > > include/rdma/rdma_netlink.h | 17 ++-- > > > > 5 files changed, 110 insertions(+), 149 deletions(-) > > > > > > > diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h > > > > index 585266144329..6735dcf5d2a3 100644 > > > > --- a/include/rdma/rdma_netlink.h > > > > +++ b/include/rdma/rdma_netlink.h > > > > @@ -10,27 +10,22 @@ struct ibnl_client_cbs { > > > > struct module *module; > > > > }; > > > > > > > > -int ibnl_init(void); > > > > -void ibnl_cleanup(void); > > > > +int rdma_nl_init(void); > > > > +void rdma_nl_exit(void); > > > > > > This hunk is in direct conflict with 233c1955 which you removed > > > ibnl_init and ibnl_cleanup. Are you putting back the init and > > > exit or do you want to leave them out? > > > > We have different git hashes and I don't know about which commit you are > > referring. > > There is only one true upstream, go look at Linus' git. > > > > > This commit simply renames exit/init functions in general header file. > > The init/exit is still needed, because it is called from device.c. > > > > The git log doesn't show me anything suspicious too. > > ➜ linux-rdma git:(topic/rdma-netlink-v1) git log -p include/rdma/rdma_netlink.h > > Your git and branch is out of sync. This is the relevant part of the patch. > > commit 233c1955835bd8649003be9bb3d8e79788b08be1 > Author: Leon Romanovsky <leonro@mellanox.com> > Date: Sun May 14 15:49:57 2017 +0300 > > RDMA/netlink: Reduce exposure of RDMA netlink functions > > RDMA netlink is part of ib_core, hence ibnl_chk_listeners(), > ibnl_init() and ibnl_cleanup() don't need to be published > in public header file. > > Let's remove EXPORT_SYMBOL from ibnl_chk_listeners() and move all these > functions to private header file. > > CC: Yuval Shaia <yuval.shaia@oracle.com> > Signed-off-by: Leon Romanovsky <leonro@mellanox.com> > Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com> > Signed-off-by: Doug Ledford <dledford@redhat.com> > > diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h > index 5852661..348c102 100644 > --- a/include/rdma/rdma_netlink.h > +++ b/include/rdma/rdma_netlink.h > @@ -10,9 +10,6 @@ struct ibnl_client_cbs { > struct module *module; > }; > > -int ibnl_init(void); > -void ibnl_cleanup(void); > - Thanks, It is rebase error, those functions needs to be in core_priv.h. I'll resend. > > > Chien
On Mon, Jun 26, 2017 at 10:11:54PM +0300, Leon Romanovsky wrote: > On Mon, Jun 26, 2017 at 01:33:29PM -0500, Chien Tin Tung wrote: > > On Mon, Jun 26, 2017 at 09:13:24PM +0300, Leon Romanovsky wrote: > > > On Mon, Jun 26, 2017 at 12:55:05PM -0500, Chien Tin Tung wrote: > > > > On Tue, Jun 13, 2017 at 03:59:23AM -0500, Leon Romanovsky wrote: > > > > > From: Leon Romanovsky <leonro@mellanox.com> > > > > > > > > > > RDMA netlink has complicated infrastructure to add and remove netlink > > > > > clients to NETLINK_RDMA family. This complicates the code and not in > > > > > use because not many clients are available (3 clients) and most of them > > > > > (2 clients) are statically compiled together with netlink.c. > > > > > > > > > > The following patch refactors RDMA netlink and opens door for the future > > > > > patches which will be able to get rid of a lot of dead iwcm* code. > > > > > > > > > > Signed-off-by: Leon Romanovsky <leonro@mellanox.com> > > > > > Reviewed-by: Chien Tin Tung <chien.tin.tung@intel.com> > > > > > --- > > > > > drivers/infiniband/core/cma.c | 6 +- > > > > > drivers/infiniband/core/device.c | 41 +++------ > > > > > drivers/infiniband/core/iwcm.c | 10 +-- > > > > > drivers/infiniband/core/netlink.c | 185 ++++++++++++++++++-------------------- > > > > > include/rdma/rdma_netlink.h | 17 ++-- > > > > > 5 files changed, 110 insertions(+), 149 deletions(-) > > > > > > > > > diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h > > > > > index 585266144329..6735dcf5d2a3 100644 > > > > > --- a/include/rdma/rdma_netlink.h > > > > > +++ b/include/rdma/rdma_netlink.h > > > > > @@ -10,27 +10,22 @@ struct ibnl_client_cbs { > > > > > struct module *module; > > > > > }; > > > > > > > > > > -int ibnl_init(void); > > > > > -void ibnl_cleanup(void); > > > > > +int rdma_nl_init(void); > > > > > +void rdma_nl_exit(void); > > > > > > > > This hunk is in direct conflict with 233c1955 which you removed > > > > ibnl_init and ibnl_cleanup. Are you putting back the init and > > > > exit or do you want to leave them out? > > > > > > We have different git hashes and I don't know about which commit you are > > > referring. > > > > There is only one true upstream, go look at Linus' git. > > > > > > > > This commit simply renames exit/init functions in general header file. > > > The init/exit is still needed, because it is called from device.c. > > > > > > The git log doesn't show me anything suspicious too. > > > ➜ linux-rdma git:(topic/rdma-netlink-v1) git log -p include/rdma/rdma_netlink.h > > > > Your git and branch is out of sync. This is the relevant part of the patch. > > > > commit 233c1955835bd8649003be9bb3d8e79788b08be1 > > Author: Leon Romanovsky <leonro@mellanox.com> > > Date: Sun May 14 15:49:57 2017 +0300 > > > > RDMA/netlink: Reduce exposure of RDMA netlink functions > > > > RDMA netlink is part of ib_core, hence ibnl_chk_listeners(), > > ibnl_init() and ibnl_cleanup() don't need to be published > > in public header file. > > > > Let's remove EXPORT_SYMBOL from ibnl_chk_listeners() and move all these > > functions to private header file. > > > > CC: Yuval Shaia <yuval.shaia@oracle.com> > > Signed-off-by: Leon Romanovsky <leonro@mellanox.com> > > Reviewed-by: Yuval Shaia <yuval.shaia@oracle.com> > > Signed-off-by: Doug Ledford <dledford@redhat.com> > > > > diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h > > index 5852661..348c102 100644 > > --- a/include/rdma/rdma_netlink.h > > +++ b/include/rdma/rdma_netlink.h > > @@ -10,9 +10,6 @@ struct ibnl_client_cbs { > > struct module *module; > > }; > > > > -int ibnl_init(void); > > -void ibnl_cleanup(void); > > - > > Thanks, > > It is rebase error, those functions needs to be in core_priv.h. I'll resend. I found the reason, this patch set is based on -rc2, rebasing on top of -rc4 solved the issue. To be on the safe side, I created new topic/rdma-netlink-v2 which is rebased version of topic/rdma-netlink-v1. Thanks for pointing it. > > > > > > > Chien
diff --git a/include/rdma/rdma_netlink.h b/include/rdma/rdma_netlink.h index 5852661..348c102 100644 --- a/include/rdma/rdma_netlink.h +++ b/include/rdma/rdma_netlink.h @@ -10,9 +10,6 @@ struct ibnl_client_cbs { struct module *module; }; -int ibnl_init(void); -void ibnl_cleanup(void); - Chien