| Message ID | 1510119893-24676-1-git-send-email-devesh.sharma@broadcom.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
> -----Original Message----- > From: Devesh Sharma [mailto:devesh.sharma@broadcom.com] > Sent: Tuesday, November 07, 2017 11:45 PM > To: dledford@redhat.com > Cc: linux-rdma@vger.kernel.org; Parav Pandit <parav@mellanox.com>; > leon@kernel.org; Devesh Sharma <devesh.sharma@broadcom.com> > Subject: [PATCH] RDMA/bnxt_re: report vlan_id and sl in qp1 recv completion > > In a real RoCE v2 network it is possible to have two Sections of network have > same IP hence same gid. However those may have different vlans. During > connection resolution it is important to report the actual vlan on which the MAD > packet was received instead of relying on other means to resolve vlan-id. > ib_find_gid_index should not be used to resolve the vlan-id using sgid of the > local system where the packet was received. > > Our device has the capability to report the actual VLAN-ID in the GSI qp > completions. Since we have the capability our driver should move away from > resolving the vlan-id with the help of SGID at the destination port. > > Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com> > Reported-by: Parav Pandit <parav@mellanox.com> > --- > drivers/infiniband/hw/bnxt_re/ib_verbs.c | 33 > ++++++++++++++++++++++++++++++++ > drivers/infiniband/hw/bnxt_re/qplib_fp.c | 1 + > 2 files changed, 34 insertions(+) > > diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c > b/drivers/infiniband/hw/bnxt_re/ib_verbs.c > index ebcdfb4..558d786 100644 > --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c > +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c > @@ -2771,6 +2771,32 @@ static void bnxt_re_process_res_rawqp1_wc(struct > ib_wc *wc, > wc->wc_flags |= IB_WC_GRH; > } > > +static bool bnxt_re_is_vlan_pkt(struct bnxt_qplib_cqe *orig_cqe, > + u16 *vid, u8 *sl) > +{ > + u32 metadata; > + u16 tpid; > + bool ret = false; > + Large number of places reverse Christmas tree reverse order is not followed, but let's try to follow for newly added code like below. bool ret = false; u32 metadata; u16 tpid; > + metadata = orig_cqe->raweth_qp1_metadata; > + if (orig_cqe->raweth_qp1_flags2 & > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_FLAGS2_META_FORMAT_VLAN) > { > + tpid = ((metadata & > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_MASK) >> > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_SFT); > + if (tpid == ETH_P_8021Q) { > + *vid = metadata & > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_VID_MASK; > + *sl = (metadata & > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_MASK) >> > + > CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_SFT; > + ret = true; > + } > + } > + > + return ret; > +} > + > static void bnxt_re_process_res_rc_wc(struct ib_wc *wc, > struct bnxt_qplib_cqe *cqe) > { > @@ -2796,6 +2822,8 @@ static void > bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, > struct bnxt_qplib_cqe *orig_cqe = NULL; > struct bnxt_re_sqp_entries *sqp_entry = NULL; > int nw_type; > + u16 vlan_id; > + u8 sl; > > tbl_idx = cqe->wr_id; > > @@ -2810,6 +2838,11 @@ static void > bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, > wc->ex.imm_data = orig_cqe->immdata; > wc->src_qp = orig_cqe->src_qp; > memcpy(wc->smac, orig_cqe->smac, ETH_ALEN); > + if (bnxt_re_is_vlan_pkt(orig_cqe, &vlan_id, &sl)) { > + wc->vlan_id = vlan_id; > + wc->sl = sl; > + wc->wc_flags |= IB_WC_WITH_VLAN; > + } > wc->port_num = 1; > wc->vendor_err = orig_cqe->status; > > diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c > b/drivers/infiniband/hw/bnxt_re/qplib_fp.c > index a82044d..8962621 100644 > --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c > +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c > @@ -2236,6 +2236,7 @@ static int > bnxt_qplib_cq_process_res_raweth_qp1(struct bnxt_qplib_cq *cq, > > cqe->raweth_qp1_flags = le16_to_cpu(hwcqe->raweth_qp1_flags); > cqe->raweth_qp1_flags2 = le32_to_cpu(hwcqe->raweth_qp1_flags2); > + cqe->raweth_qp1_metadata = le32_to_cpu(hwcqe- > >raweth_qp1_metadata); > > rq = &qp->rq; > if (wr_id_idx > rq->hwq.max_elements) { > -- > 1.8.3.1 -- 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, Nov 8, 2017 at 11:57 AM, Parav Pandit <parav@mellanox.com> wrote: > > >> -----Original Message----- >> From: Devesh Sharma [mailto:devesh.sharma@broadcom.com] >> Sent: Tuesday, November 07, 2017 11:45 PM >> To: dledford@redhat.com >> Cc: linux-rdma@vger.kernel.org; Parav Pandit <parav@mellanox.com>; >> leon@kernel.org; Devesh Sharma <devesh.sharma@broadcom.com> >> Subject: [PATCH] RDMA/bnxt_re: report vlan_id and sl in qp1 recv completion >> >> In a real RoCE v2 network it is possible to have two Sections of network have >> same IP hence same gid. However those may have different vlans. During >> connection resolution it is important to report the actual vlan on which the MAD >> packet was received instead of relying on other means to resolve vlan-id. >> ib_find_gid_index should not be used to resolve the vlan-id using sgid of the >> local system where the packet was received. >> >> Our device has the capability to report the actual VLAN-ID in the GSI qp >> completions. Since we have the capability our driver should move away from >> resolving the vlan-id with the help of SGID at the destination port. >> >> Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com> >> Reported-by: Parav Pandit <parav@mellanox.com> >> --- >> drivers/infiniband/hw/bnxt_re/ib_verbs.c | 33 >> ++++++++++++++++++++++++++++++++ >> drivers/infiniband/hw/bnxt_re/qplib_fp.c | 1 + >> 2 files changed, 34 insertions(+) >> >> diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c >> b/drivers/infiniband/hw/bnxt_re/ib_verbs.c >> index ebcdfb4..558d786 100644 >> --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c >> +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c >> @@ -2771,6 +2771,32 @@ static void bnxt_re_process_res_rawqp1_wc(struct >> ib_wc *wc, >> wc->wc_flags |= IB_WC_GRH; >> } >> >> +static bool bnxt_re_is_vlan_pkt(struct bnxt_qplib_cqe *orig_cqe, >> + u16 *vid, u8 *sl) >> +{ >> + u32 metadata; >> + u16 tpid; >> + bool ret = false; >> + > Large number of places reverse Christmas tree reverse order is not followed, but let's try to follow for newly added code like below. > bool ret = false; > u32 metadata; > u16 tpid; > Okay sending a rev. >> + metadata = orig_cqe->raweth_qp1_metadata; >> + if (orig_cqe->raweth_qp1_flags2 & >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_FLAGS2_META_FORMAT_VLAN) >> { >> + tpid = ((metadata & >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_MASK) >> >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_SFT); >> + if (tpid == ETH_P_8021Q) { >> + *vid = metadata & >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_VID_MASK; >> + *sl = (metadata & >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_MASK) >> >> + >> CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_SFT; >> + ret = true; >> + } >> + } >> + >> + return ret; >> +} >> + >> static void bnxt_re_process_res_rc_wc(struct ib_wc *wc, >> struct bnxt_qplib_cqe *cqe) >> { >> @@ -2796,6 +2822,8 @@ static void >> bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, >> struct bnxt_qplib_cqe *orig_cqe = NULL; >> struct bnxt_re_sqp_entries *sqp_entry = NULL; >> int nw_type; >> + u16 vlan_id; >> + u8 sl; >> >> tbl_idx = cqe->wr_id; >> >> @@ -2810,6 +2838,11 @@ static void >> bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, >> wc->ex.imm_data = orig_cqe->immdata; >> wc->src_qp = orig_cqe->src_qp; >> memcpy(wc->smac, orig_cqe->smac, ETH_ALEN); >> + if (bnxt_re_is_vlan_pkt(orig_cqe, &vlan_id, &sl)) { >> + wc->vlan_id = vlan_id; >> + wc->sl = sl; >> + wc->wc_flags |= IB_WC_WITH_VLAN; >> + } >> wc->port_num = 1; >> wc->vendor_err = orig_cqe->status; >> >> diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c >> b/drivers/infiniband/hw/bnxt_re/qplib_fp.c >> index a82044d..8962621 100644 >> --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c >> +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c >> @@ -2236,6 +2236,7 @@ static int >> bnxt_qplib_cq_process_res_raweth_qp1(struct bnxt_qplib_cq *cq, >> >> cqe->raweth_qp1_flags = le16_to_cpu(hwcqe->raweth_qp1_flags); >> cqe->raweth_qp1_flags2 = le32_to_cpu(hwcqe->raweth_qp1_flags2); >> + cqe->raweth_qp1_metadata = le32_to_cpu(hwcqe- >> >raweth_qp1_metadata); >> >> rq = &qp->rq; >> if (wr_id_idx > rq->hwq.max_elements) { >> -- >> 1.8.3.1 > -- 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, Nov 08, 2017 at 06:27:26AM +0000, Parav Pandit wrote: > > > > +static bool bnxt_re_is_vlan_pkt(struct bnxt_qplib_cqe *orig_cqe, > > + u16 *vid, u8 *sl) > > +{ > > + u32 metadata; > > + u16 tpid; > > + bool ret = false; > > + > Large number of places reverse Christmas tree reverse order is not followed, but let's try to follow for newly added code like below. > bool ret = false; > u32 metadata; > u16 tpid; Reversed Christmas tree is a netdev requirement and not other subsystems, so it a little bit too much to request from the people to do it without actual updating CodingStyle guide. I personally require it in our internal reviews to ensure that every patch can go through netdev if needed (for example shared code) and not because I like it. Thanks
diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c index ebcdfb4..558d786 100644 --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c @@ -2771,6 +2771,32 @@ static void bnxt_re_process_res_rawqp1_wc(struct ib_wc *wc, wc->wc_flags |= IB_WC_GRH; } +static bool bnxt_re_is_vlan_pkt(struct bnxt_qplib_cqe *orig_cqe, + u16 *vid, u8 *sl) +{ + u32 metadata; + u16 tpid; + bool ret = false; + + metadata = orig_cqe->raweth_qp1_metadata; + if (orig_cqe->raweth_qp1_flags2 & + CQ_RES_RAWETH_QP1_RAWETH_QP1_FLAGS2_META_FORMAT_VLAN) { + tpid = ((metadata & + CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_MASK) >> + CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_TPID_SFT); + if (tpid == ETH_P_8021Q) { + *vid = metadata & + CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_VID_MASK; + *sl = (metadata & + CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_MASK) >> + CQ_RES_RAWETH_QP1_RAWETH_QP1_METADATA_PRI_SFT; + ret = true; + } + } + + return ret; +} + static void bnxt_re_process_res_rc_wc(struct ib_wc *wc, struct bnxt_qplib_cqe *cqe) { @@ -2796,6 +2822,8 @@ static void bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, struct bnxt_qplib_cqe *orig_cqe = NULL; struct bnxt_re_sqp_entries *sqp_entry = NULL; int nw_type; + u16 vlan_id; + u8 sl; tbl_idx = cqe->wr_id; @@ -2810,6 +2838,11 @@ static void bnxt_re_process_res_shadow_qp_wc(struct bnxt_re_qp *qp, wc->ex.imm_data = orig_cqe->immdata; wc->src_qp = orig_cqe->src_qp; memcpy(wc->smac, orig_cqe->smac, ETH_ALEN); + if (bnxt_re_is_vlan_pkt(orig_cqe, &vlan_id, &sl)) { + wc->vlan_id = vlan_id; + wc->sl = sl; + wc->wc_flags |= IB_WC_WITH_VLAN; + } wc->port_num = 1; wc->vendor_err = orig_cqe->status; diff --git a/drivers/infiniband/hw/bnxt_re/qplib_fp.c b/drivers/infiniband/hw/bnxt_re/qplib_fp.c index a82044d..8962621 100644 --- a/drivers/infiniband/hw/bnxt_re/qplib_fp.c +++ b/drivers/infiniband/hw/bnxt_re/qplib_fp.c @@ -2236,6 +2236,7 @@ static int bnxt_qplib_cq_process_res_raweth_qp1(struct bnxt_qplib_cq *cq, cqe->raweth_qp1_flags = le16_to_cpu(hwcqe->raweth_qp1_flags); cqe->raweth_qp1_flags2 = le32_to_cpu(hwcqe->raweth_qp1_flags2); + cqe->raweth_qp1_metadata = le32_to_cpu(hwcqe->raweth_qp1_metadata); rq = &qp->rq; if (wr_id_idx > rq->hwq.max_elements) {
In a real RoCE v2 network it is possible to have two Sections of network have same IP hence same gid. However those may have different vlans. During connection resolution it is important to report the actual vlan on which the MAD packet was received instead of relying on other means to resolve vlan-id. ib_find_gid_index should not be used to resolve the vlan-id using sgid of the local system where the packet was received. Our device has the capability to report the actual VLAN-ID in the GSI qp completions. Since we have the capability our driver should move away from resolving the vlan-id with the help of SGID at the destination port. Signed-off-by: Devesh Sharma <devesh.sharma@broadcom.com> Reported-by: Parav Pandit <parav@mellanox.com> --- drivers/infiniband/hw/bnxt_re/ib_verbs.c | 33 ++++++++++++++++++++++++++++++++ drivers/infiniband/hw/bnxt_re/qplib_fp.c | 1 + 2 files changed, 34 insertions(+)