| Message ID | 20180327180929.8D1572BB7D@smtp.opengridcomputing.com (mailing list archive) |
|---|---|
| State | Accepted |
| Delegated to: | Jason Gunthorpe |
| Headers | show |
Doug/Jason, Seems like I got the authdates screwed up again. These show up in reverse order in patchwork. Would you like a respin? These two will enable me to finalize the associated rdma tool submission. Thanks, Steve. On 3/27/2018 10:38 AM, Steve Wise wrote: > Since the rdma_port_space enum is being passed between > user and kernel for user cm_id setup, we need it in a > UAPI header. So add it to rdma_user_cm.h. > > This also fixes the cm_id restrack changes which pass up the port space > value via the RDMA_NLDEV_ATTR_RES_PS attribute. > > Fixes: 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") > > Signed-off-by: Steve Wise <swise@opengridcomputing.com> > --- > include/rdma/rdma_cm.h | 8 +------- > include/uapi/rdma/rdma_user_cm.h | 7 +++++++ > 2 files changed, 8 insertions(+), 7 deletions(-) > > diff --git a/include/rdma/rdma_cm.h b/include/rdma/rdma_cm.h > index 7652efc..86652a0 100644 > --- a/include/rdma/rdma_cm.h > +++ b/include/rdma/rdma_cm.h > @@ -38,6 +38,7 @@ > #include <linux/in6.h> > #include <rdma/ib_addr.h> > #include <rdma/ib_sa.h> > +#include <uapi/rdma/rdma_user_cm.h> > > /* > * Upon receiving a device removal event, users must destroy the associated > @@ -64,13 +65,6 @@ enum rdma_cm_event_type { > > const char *__attribute_const__ rdma_event_msg(enum rdma_cm_event_type event); > > -enum rdma_port_space { > - RDMA_PS_IPOIB = 0x0002, > - RDMA_PS_IB = 0x013F, > - RDMA_PS_TCP = 0x0106, > - RDMA_PS_UDP = 0x0111, > -}; > - > #define RDMA_IB_IP_PS_MASK 0xFFFFFFFFFFFF0000ULL > #define RDMA_IB_IP_PS_TCP 0x0000000001060000ULL > #define RDMA_IB_IP_PS_UDP 0x0000000001110000ULL > diff --git a/include/uapi/rdma/rdma_user_cm.h b/include/uapi/rdma/rdma_user_cm.h > index c83ef00..851652d 100644 > --- a/include/uapi/rdma/rdma_user_cm.h > +++ b/include/uapi/rdma/rdma_user_cm.h > @@ -70,6 +70,13 @@ enum { > RDMA_USER_CM_CMD_JOIN_MCAST > }; > > +enum rdma_port_space { > + RDMA_PS_IPOIB = 0x0002, > + RDMA_PS_IB = 0x013F, > + RDMA_PS_TCP = 0x0106, > + RDMA_PS_UDP = 0x0111, > +}; > + > /* > * command ABI structures. > */ -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Mar 28, 2018 at 10:40:56AM -0500, Steve Wise wrote: > Doug/Jason, > > Seems like I got the authdates screwed up again. These show up in > reverse order in patchwork. Would you like a respin? I can manage.. > These two will enable me to finalize the associated rdma tool submission. I can't find a user API user of this though? Is netlink introducing the first user? I'm confused. I thought umca used this someplace. Maybe I just can't find it today. Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 3/28/2018 11:44 AM, Jason Gunthorpe wrote: > On Wed, Mar 28, 2018 at 10:40:56AM -0500, Steve Wise wrote: >> Doug/Jason, >> >> Seems like I got the authdates screwed up again. These show up in >> reverse order in patchwork. Would you like a respin? > I can manage.. > >> These two will enable me to finalize the associated rdma tool submission. > I can't find a user API user of this though? Is netlink introducing > the first user? I'm confused. I thought umca used this > someplace. Maybe I just can't find it today. Hrm... you're right. I don't see any code using rdma_ucm_create_id. I do see ucma_abi_create_id from rdma_cma_abi.h used in librdmacm:rdma_create_id2() called by rdma_create_id() . And ucma_abi_create_id has a ps field. So perhaps the correct header file is rdma_cma_abi.h? But I don't see that file in the kernel at all. /me confused. > Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 3/28/2018 11:57 AM, Steve Wise wrote: > > On 3/28/2018 11:44 AM, Jason Gunthorpe wrote: >> On Wed, Mar 28, 2018 at 10:40:56AM -0500, Steve Wise wrote: >>> Doug/Jason, >>> >>> Seems like I got the authdates screwed up again. These show up in >>> reverse order in patchwork. Would you like a respin? >> I can manage.. >> >>> These two will enable me to finalize the associated rdma tool submission. >> I can't find a user API user of this though? Is netlink introducing >> the first user? I'm confused. I thought umca used this >> someplace. Maybe I just can't find it today. > Hrm... you're right. I don't see any code using rdma_ucm_create_id. I > do see ucma_abi_create_id from rdma_cma_abi.h used in > librdmacm:rdma_create_id2() called by rdma_create_id() . And > ucma_abi_create_id has a ps field. So perhaps the correct header file > is rdma_cma_abi.h? But I don't see that file in the kernel at all. > > /me confused. The kernel side, in ucma.c, is using struct rdma_ucm_create_id to parse the u->k RDMA_USER_CM_CMD_CREATE_ID command. So its UAPI header is include/uapi/rdma/rdma_user_cm.h. The user side is using struct ucma_abi_create_id from rdma-core/librdmacm/rdma_cma_abi.h. Jason, how should I proceed? -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Tue, Mar 27, 2018 at 08:38:07AM -0700, Steve Wise wrote: > Since the rdma_port_space enum is being passed between > user and kernel for user cm_id setup, we need it in a > UAPI header. So add it to rdma_user_cm.h. > > This also fixes the cm_id restrack changes which pass up the port space > value via the RDMA_NLDEV_ATTR_RES_PS attribute. > > Fixes: 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") > Signed-off-by: Steve Wise <swise@opengridcomputing.com> > --- > include/rdma/rdma_cm.h | 8 +------- > include/uapi/rdma/rdma_user_cm.h | 7 +++++++ > 2 files changed, 8 insertions(+), 7 deletions(-) I applied this patch and one before after modifying it as discussed in the other thread.. Please check it.. Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
> > On Tue, Mar 27, 2018 at 08:38:07AM -0700, Steve Wise wrote: > > Since the rdma_port_space enum is being passed between > > user and kernel for user cm_id setup, we need it in a > > UAPI header. So add it to rdma_user_cm.h. > > > > This also fixes the cm_id restrack changes which pass up the port space > > value via the RDMA_NLDEV_ATTR_RES_PS attribute. > > > > Fixes: 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") > > Signed-off-by: Steve Wise <swise@opengridcomputing.com> > > --- > > include/rdma/rdma_cm.h | 8 +------- > > include/uapi/rdma/rdma_user_cm.h | 7 +++++++ > > 2 files changed, 8 insertions(+), 7 deletions(-) > > I applied this patch and one before after modifying it as discussed in > the other thread.. Please check it.. > > Jason Thanks Jason! This: @@ -82,7 +90,7 @@ struct rdma_ucm_cmd_hdr { struct rdma_ucm_create_id { __aligned_u64 uid; __aligned_u64 response; - __u16 ps; + __u16 ps; /* use enum rdma_port_space */ __u8 qp_type; __u8 reserved[5]; }; Should the comment be "use rdma_ucm_port_space? " I'll rebase my series and test it all tomorrow first thing. Steve. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Mar 28, 2018 at 06:40:22PM -0500, Steve Wise wrote: > > > > > On Tue, Mar 27, 2018 at 08:38:07AM -0700, Steve Wise wrote: > > > Since the rdma_port_space enum is being passed between > > > user and kernel for user cm_id setup, we need it in a > > > UAPI header. So add it to rdma_user_cm.h. > > > > > > This also fixes the cm_id restrack changes which pass up the port space > > > value via the RDMA_NLDEV_ATTR_RES_PS attribute. > > > > > > Fixes: 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") > > > Signed-off-by: Steve Wise <swise@opengridcomputing.com> > > > include/rdma/rdma_cm.h | 8 +------- > > > include/uapi/rdma/rdma_user_cm.h | 7 +++++++ > > > 2 files changed, 8 insertions(+), 7 deletions(-) > > > > I applied this patch and one before after modifying it as discussed in > > the other thread.. Please check it.. > > > > Jason > > Thanks Jason! This: > > @@ -82,7 +90,7 @@ struct rdma_ucm_cmd_hdr { > struct rdma_ucm_create_id { > __aligned_u64 uid; > __aligned_u64 response; > - __u16 ps; > + __u16 ps; /* use enum rdma_port_space */ > __u8 qp_type; > __u8 reserved[5]; > }; > > Should the comment be "use rdma_ucm_port_space? " done Jason -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
>> I applied this patch and one before after modifying it as discussed in >> the other thread.. Please check it.. >> >> Jason > Thanks Jason! This: > > @@ -82,7 +90,7 @@ struct rdma_ucm_cmd_hdr { > struct rdma_ucm_create_id { > __aligned_u64 uid; > __aligned_u64 response; > - __u16 ps; > + __u16 ps; /* use enum rdma_port_space */ > __u8 qp_type; > __u8 reserved[5]; > }; > > Should the comment be "use rdma_ucm_port_space? " > > I'll rebase my series and test it all tomorrow first thing. > > Steve. > Testing looks good. Steve. -- To unsubscribe from this list: send the line "unsubscribe linux-rdma" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/include/rdma/rdma_cm.h b/include/rdma/rdma_cm.h index 7652efc..86652a0 100644 --- a/include/rdma/rdma_cm.h +++ b/include/rdma/rdma_cm.h @@ -38,6 +38,7 @@ #include <linux/in6.h> #include <rdma/ib_addr.h> #include <rdma/ib_sa.h> +#include <uapi/rdma/rdma_user_cm.h> /* * Upon receiving a device removal event, users must destroy the associated @@ -64,13 +65,6 @@ enum rdma_cm_event_type { const char *__attribute_const__ rdma_event_msg(enum rdma_cm_event_type event); -enum rdma_port_space { - RDMA_PS_IPOIB = 0x0002, - RDMA_PS_IB = 0x013F, - RDMA_PS_TCP = 0x0106, - RDMA_PS_UDP = 0x0111, -}; - #define RDMA_IB_IP_PS_MASK 0xFFFFFFFFFFFF0000ULL #define RDMA_IB_IP_PS_TCP 0x0000000001060000ULL #define RDMA_IB_IP_PS_UDP 0x0000000001110000ULL diff --git a/include/uapi/rdma/rdma_user_cm.h b/include/uapi/rdma/rdma_user_cm.h index c83ef00..851652d 100644 --- a/include/uapi/rdma/rdma_user_cm.h +++ b/include/uapi/rdma/rdma_user_cm.h @@ -70,6 +70,13 @@ enum { RDMA_USER_CM_CMD_JOIN_MCAST }; +enum rdma_port_space { + RDMA_PS_IPOIB = 0x0002, + RDMA_PS_IB = 0x013F, + RDMA_PS_TCP = 0x0106, + RDMA_PS_UDP = 0x0111, +}; + /* * command ABI structures. */
Since the rdma_port_space enum is being passed between user and kernel for user cm_id setup, we need it in a UAPI header. So add it to rdma_user_cm.h. This also fixes the cm_id restrack changes which pass up the port space value via the RDMA_NLDEV_ATTR_RES_PS attribute. Fixes: 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") Signed-off-by: Steve Wise <swise@opengridcomputing.com> --- include/rdma/rdma_cm.h | 8 +------- include/uapi/rdma/rdma_user_cm.h | 7 +++++++ 2 files changed, 8 insertions(+), 7 deletions(-)