RDMA/erdma: cleanups and updates 9-9-2022
Subsystems (2)
Cheng Xu
maintainercover RDMA/erdma: cleanups and updates 9-9-2022
Hi,
This series has new updates for erdma driver:
- #1~#3 aim at making code more clean, including eliminating unnecessary castings, removing redundant include headers, and hiding hardware internal opcodes.
- #4 introcuces dynamic mtu support.
Thanks,
Cheng Xu
Cheng Xu (4):
RDMA/erdma: Eliminate unnecessary casting for erdma_post_cmd_wait RDMA/erdma: Remove redundant includes
RDMA/erdma: Make hardware internal opcodes invisible to driver RDMA/erdma: Support dynamic mtu
drivers/infiniband/hw/erdma/erdma.h | 4 ++- drivers/infiniband/hw/erdma/erdma_cm.c | 8 ------ drivers/infiniband/hw/erdma/erdma_cmdq.c | 8 +----- drivers/infiniband/hw/erdma/erdma_cq.c | 4 --- drivers/infiniband/hw/erdma/erdma_eq.c | 13 ++------- drivers/infiniband/hw/erdma/erdma_hw.h | 14 ++++++--- drivers/infiniband/hw/erdma/erdma_main.c | 17 +++++------ drivers/infiniband/hw/erdma/erdma_qp.c | 15 ++-------- drivers/infiniband/hw/erdma/erdma_verbs.c | 35 +++++++++++------------ drivers/infiniband/hw/erdma/erdma_verbs.h | 9 +----- 10 files changed, 43 insertions(+), 84 deletions(-)
Signature · 1 line
2.27.0
Cheng Xu
maintainerPATCH 1/4 RDMA/erdma: Eliminate unnecessary casting for erdma_post_cmd_wait
erdma_post_cmd_wait does not use the 'u64 *req' input parameter directly. So it is better to define it to 'void *req', and by this we can eliminate the casting when calling erdma_post_cmd_wait.
--- drivers/infiniband/hw/erdma/erdma.h | 2 +- drivers/infiniband/hw/erdma/erdma_cmdq.c | 2 +- drivers/infiniband/hw/erdma/erdma_eq.c | 7 ++----- drivers/infiniband/hw/erdma/erdma_qp.c | 6 ++---- drivers/infiniband/hw/erdma/erdma_verbs.c | 17 ++++++----------- 5 files changed, 12 insertions(+), 22 deletions(-)
drivers/infiniband/hw/erdma/erdma.h +1 -1
@@ -269,7 +269,7 @@ void erdma_finish_cmdq_init(struct erdma_dev *dev);void erdma_cmdq_destroy (struct erdma_dev *dev);void erdma_cmdq_build_reqhdr (u64 *hdr,u32 mod,u32 op); -int erdma_post_cmd_wait (struct erdma_cmdq *cmdq,u64 *req,u32 req_size, +int erdma_post_cmd_wait (struct erdma_cmdq *cmdq,void *req,u32 req_size,u64 *resp0,u64 *resp1);void erdma_cmdq_completion_handler (struct erdma_cmdq *cmdq);
drivers/infiniband/hw/erdma/erdma_cmdq.c +1 -1
@@ -441,7 +441,7 @@ void erdma_cmdq_build_reqhdr(u64 *hdr, u32 mod, u32 op)FIELD_PREP (ERDMA_CMD_HDR_OPCODE_MASK , op); } -int erdma_post_cmd_wait (struct erdma_cmdq *cmdq,u64 *req,u32 req_size, +int erdma_post_cmd_wait (struct erdma_cmdq *cmdq,void *req,u32 req_size,u64 *resp0,u64 *resp1) {struct erdma_comp_wait *comp_wait;
drivers/infiniband/hw/erdma/erdma_eq.c +2 -5
@@ -229,9 +229,7 @@ static int create_eq_cmd(struct erdma_dev *dev, u32 eqn, struct erdma_eq *eq) req.db_dma_addr_l =lower_32_bits (db_info_dma_addr); req.db_dma_addr_h =upper_32_bits (db_info_dma_addr); -return erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req, -sizeof (struct erdma_cmdq_create_eq_req ), -NULL ,NULL ); +return erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); }static int erdma_ceq_init_one (struct erdma_dev *dev,u16 ceqn)
@@ -281,8 +279,7 @@ static void erdma_ceq_uninit_one(struct erdma_dev *dev, u16 ceqn) req.qtype =ERDMA_EQ_TYPE_CEQ ; req.vector_idx = ceqn +1 ; - err =erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); + err =erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL );if (err)return ;
drivers/infiniband/hw/erdma/erdma_qp.c +2 -4
@@ -105,8 +105,7 @@ static int erdma_modify_qp_state_to_rts(struct erdma_qp *qp, req.send_nxt +=MPA_DEFAULT_HDR_LEN + qp->attrs.pd_len; req.recv_nxt = tp->rcv_nxt; -return erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); +return erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); } static int erdma_modify_qp_state_to_stop(struct erdma_qp *qp,
@@ -124,8 +123,7 @@ static int erdma_modify_qp_state_to_stop(struct erdma_qp *qp, req.cfg =FIELD_PREP (ERDMA_CMD_MODIFY_QP_STATE_MASK , attrs->state) |FIELD_PREP (ERDMA_CMD_MODIFY_QP_QPN_MASK ,QP_ID (qp)); -return erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); +return erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); } int erdma_modify_qp_internal(struct erdma_qp *qp, struct erdma_qp_attrs *attrs,
drivers/infiniband/hw/erdma/erdma_verbs.c +6 -11
@@ -102,7 +102,7 @@ static int create_qp_cmd(struct erdma_dev *dev, struct erdma_qp *qp) req.rq_db_info_dma_addr = user_qp->rq_db_info_dma_addr; } - err =erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req), &resp0, + err =erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req), &resp0, &resp1);if (!err) qp->attrs.cookie =
@@ -151,8 +151,7 @@ static int regmr_cmd(struct erdma_dev *dev, struct erdma_mr *mr) } post_cmd: -return erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); +return erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); }static int create_cq_cmd (struct erdma_dev *dev,struct erdma_cq *cq)
@@ -202,8 +201,7 @@ static int create_cq_cmd(struct erdma_dev *dev, struct erdma_cq *cq) req.cq_db_info_addr = cq->user_cq.db_info_dma_addr; } -return erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); +return erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); }static int erdma_alloc_idx (struct erdma_resource_cb *res_cb)
@@ -976,8 +974,7 @@ int erdma_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) req.cfg =FIELD_PREP (ERDMA_CMD_MR_MPT_IDX_MASK , ibmr->lkey >>8 ) |FIELD_PREP (ERDMA_CMD_MR_KEY_MASK , ibmr->lkey &0xFF ); - ret =erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); + ret =erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL );if (ret)return ret;
@@ -1002,8 +999,7 @@ int erdma_destroy_cq(struct ib_cq *ibcq, struct ib_udata *udata) CMDQ_OPCODE_DESTROY_CQ); req.cqn = cq->cqn; - err =erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); + err =erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL );if (err)return err;
@@ -1040,8 +1036,7 @@ int erdma_destroy_qp(struct ib_qp *ibqp, struct ib_udata *udata) CMDQ_OPCODE_DESTROY_QP); req.qpn =QP_ID (qp); - err =erdma_post_cmd_wait (&dev->cmdq, (u64 *)&req,sizeof (req),NULL , -NULL ); + err =erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL );if (err)return err;
Signature · 1 line
2.27.0
Cheng Xu
maintainerPATCH 2/4 RDMA/erdma: Remove redundant includes
In reply to Cheng XuMany of erdma's includes are redundant, because they are already included indirectly by kernel headers or custom headers. So we remove all the unnecessary direct-includes. Besides, add linux/pci.h to erdma.h because it's also used in the file.
--- drivers/infiniband/hw/erdma/erdma.h | 1 + drivers/infiniband/hw/erdma/erdma_cm.c | 8 -------- drivers/infiniband/hw/erdma/erdma_cmdq.c | 6 ------ drivers/infiniband/hw/erdma/erdma_cq.c | 3 --- drivers/infiniband/hw/erdma/erdma_eq.c | 6 ------ drivers/infiniband/hw/erdma/erdma_main.c | 9 --------- drivers/infiniband/hw/erdma/erdma_qp.c | 9 --------- drivers/infiniband/hw/erdma/erdma_verbs.c | 7 ------- drivers/infiniband/hw/erdma/erdma_verbs.h | 8 -------- 9 files changed, 1 insertion(+), 56 deletions(-)
drivers/infiniband/hw/erdma/erdma.h +1 -0
@@ -9,6 +9,7 @@#include <linux/bitfield.h> #include <linux/netdevice.h> +#include <linux/pci.h> #include <linux/xarray.h> #include <rdma/ib_verbs.h>
drivers/infiniband/hw/erdma/erdma_cm.c +0 -8
@@ -10,15 +10,7 @@/* Copyright (c) 2008-2019, IBM Corporation */ /* Copyright (c) 2017, Open Grid Computing, Inc. */ -#include <linux/errno.h> -#include <linux/inetdevice.h> -#include <linux/net.h> -#include <linux/types.h> #include <linux/workqueue.h> -#include <net/addrconf.h> - -#include <rdma/ib_user_verbs.h> -#include <rdma/ib_verbs.h> #include "erdma.h" #include "erdma_cm.h"
drivers/infiniband/hw/erdma/erdma_cmdq.c +0 -6
@@ -4,13 +4,7 @@/* Kai Shen <[email protected]> */ /* Copyright (c) 2020-2022, Alibaba Group. */ -#include <linux/kernel.h> -#include <linux/pci.h> -#include <linux/types.h> -#include "erdma.h" -#include "erdma_hw.h" -#include "erdma_verbs.h" static void arm_cmdq_cq (struct erdma_cmdq *cmdq) {
drivers/infiniband/hw/erdma/erdma_cq.c +0 -3
@@ -4,9 +4,6 @@/* Kai Shen <[email protected]> */ /* Copyright (c) 2020-2022, Alibaba Group. */ -#include <rdma/ib_verbs.h> - -#include "erdma_hw.h" #include "erdma_verbs.h" static void *get_next_valid_cqe (struct erdma_cq *cq)
drivers/infiniband/hw/erdma/erdma_eq.c +0 -6
@@ -4,12 +4,6 @@/* Kai Shen <[email protected]> */ /* Copyright (c) 2020-2022, Alibaba Group. */ -#include <linux/errno.h> -#include <linux/pci.h> -#include <linux/types.h> - -#include "erdma.h" -#include "erdma_hw.h" #include "erdma_verbs.h" #define MAX_POLL_CHUNK_SIZE 16
drivers/infiniband/hw/erdma/erdma_main.c +0 -9
@@ -4,21 +4,12 @@/* Kai Shen <[email protected]> */ /* Copyright (c) 2020-2022, Alibaba Group. */ -#include <linux/errno.h> -#include <linux/init.h> -#include <linux/kernel.h> -#include <linux/list.h> #include <linux/module.h> -#include <linux/netdevice.h> -#include <linux/pci.h> #include <net/addrconf.h> #include <rdma/erdma-abi.h> -#include <rdma/ib_verbs.h> -#include <rdma/ib_user_verbs.h> #include "erdma.h" #include "erdma_cm.h" -#include "erdma_hw.h" #include "erdma_verbs.h" MODULE_AUTHOR ("Cheng Xu <[email protected]>" );
drivers/infiniband/hw/erdma/erdma_qp.c +0 -9
@@ -6,15 +6,6 @@/* Authors: Bernard Metzler <[email protected]> */ /* Copyright (c) 2008-2019, IBM Corporation */ -#include <linux/errno.h> -#include <linux/pci.h> -#include <linux/scatterlist.h> -#include <linux/types.h> - -#include <rdma/ib_user_verbs.h> -#include <rdma/ib_verbs.h> - -#include "erdma.h" #include "erdma_cm.h" #include "erdma_verbs.h"
drivers/infiniband/hw/erdma/erdma_verbs.c +0 -7
@@ -9,21 +9,14 @@/* Copyright (c) 2013-2015, Mellanox Technologies. All rights reserved. */ -#include <linux/errno.h> -#include <linux/pci.h> -#include <linux/types.h> -#include <linux/uaccess.h> #include <linux/vmalloc.h> #include <net/addrconf.h> #include <rdma/erdma-abi.h> #include <rdma/ib_umem.h> -#include <rdma/ib_user_verbs.h> -#include <rdma/ib_verbs.h> #include <rdma/uverbs_ioctl.h> #include "erdma.h" #include "erdma_cm.h" -#include "erdma_hw.h" #include "erdma_verbs.h" static int create_qp_cmd (struct erdma_dev *dev,struct erdma_qp *qp)
drivers/infiniband/hw/erdma/erdma_verbs.h +0 -8
@@ -7,15 +7,7 @@#ifndef __ERDMA_VERBS_H__#define __ERDMA_VERBS_H__ -#include <linux/errno.h> - -#include <rdma/ib_verbs.h> -#include <rdma/ib_user_verbs.h> -#include <rdma/iw_cm.h> -#include "erdma.h" -#include "erdma_cm.h" -#include "erdma_hw.h" /* RDMA Capability. */ #define ERDMA_MAX_PD (128 * 1024)
Signature · 1 line
2.27.0
Cheng Xu
maintainerPATCH 3/4 RDMA/erdma: Make hardware internal opcodes invisible to driver
In reply to Cheng XuSome opcodes are used in hardware internally, and driver does not care about them. So, we change them to reserved opcodes in driver.
--- drivers/infiniband/hw/erdma/erdma_cq.c | 1 - drivers/infiniband/hw/erdma/erdma_hw.h | 8 ++++---- 2 files changed, 4 insertions(+), 5 deletions(-)
drivers/infiniband/hw/erdma/erdma_cq.c +0 -1
@@ -59,7 +59,6 @@ static const enum ib_wc_opcode wc_mapping_table[ERDMA_NUM_OPCODES] = { [ERDMA_OP_RECV_IMM] = IB_WC_RECV_RDMA_WITH_IMM, [ERDMA_OP_RECV_INV] = IB_WC_RECV, [ERDMA_OP_WRITE_WITH_IMM] = IB_WC_RDMA_WRITE, - [ERDMA_OP_INVALIDATE] = IB_WC_LOCAL_INV, [ERDMA_OP_RSP_SEND_IMM] = IB_WC_RECV, [ERDMA_OP_SEND_WITH_INV] = IB_WC_SEND, [ERDMA_OP_REG_MR] = IB_WC_REG_MR,
drivers/infiniband/hw/erdma/erdma_hw.h +4 -4
@@ -450,13 +450,13 @@ enum erdma_opcode {ERDMA_OP_RECV_IMM =5 ,ERDMA_OP_RECV_INV =6 , -ERDMA_OP_REQ_ERR =7 , -ERDMA_OP_READ_RESPONSE =8 , +ERDMA_OP_RSVD0 =7 , +ERDMA_OP_RSVD1 =8 ,ERDMA_OP_WRITE_WITH_IMM =9 , -ERDMA_OP_RECV_ERR =10 , +ERDMA_OP_RSVD2 =10 , +ERDMA_OP_RSVD3 =11 , -ERDMA_OP_INVALIDATE =11 ,ERDMA_OP_RSP_SEND_IMM =12 , ERDMA_OP_SEND_WITH_INV = 13,
Signature · 1 line
2.27.0
Cheng Xu
maintainerPATCH 4/4 RDMA/erdma: Support dynamic mtu
In reply to Cheng XuHardware now support jumbo frame for RDMA. So we introduce a new CMDQ message to support mtu change notification.
--- drivers/infiniband/hw/erdma/erdma.h | 1 + drivers/infiniband/hw/erdma/erdma_hw.h | 6 ++++++ drivers/infiniband/hw/erdma/erdma_main.c | 8 +++++++- drivers/infiniband/hw/erdma/erdma_verbs.c | 11 +++++++++++ drivers/infiniband/hw/erdma/erdma_verbs.h | 1 + 5 files changed, 26 insertions(+), 1 deletion(-)
drivers/infiniband/hw/erdma/erdma.h +1 -0
@@ -197,6 +197,7 @@ struct erdma_dev {struct erdma_devattr attrs;/* physical port state (only one port per device) */ enum ib_port_state state; +u32 mtu;/* cmdq and aeq use the same msix vector */ struct erdma_irq comm_irq;
drivers/infiniband/hw/erdma/erdma_hw.h +6 -0
@@ -153,6 +153,7 @@ enum CMDQ_COMMON_OPCODE {CMDQ_OPCODE_CREATE_EQ =0 ,CMDQ_OPCODE_DESTROY_EQ =1 ,CMDQ_OPCODE_QUERY_FW_INFO =2 , + CMDQ_OPCODE_CONF_MTU = 3, };/* cmdq-SQE HDR */
@@ -190,6 +191,11 @@ struct erdma_cmdq_destroy_eq_req {u8 qtype; }; +struct erdma_cmdq_config_mtu_req { +u64 hdr; +u32 mtu; +}; +/* create_cq cfg0 */ #define ERDMA_CMD_CREATE_CQ_DEPTH_MASK GENMASK(31, 24)#define ERDMA_CMD_CREATE_CQ_PAGESIZE_MASK GENMASK(23, 20)
drivers/infiniband/hw/erdma/erdma_main.c +7 -1
@@ -34,10 +34,15 @@ static int erdma_netdev_event(struct notifier_block *nb, unsigned long event, dev->state =IB_PORT_DOWN ;erdma_port_event (dev,IB_EVENT_PORT_ERR );break ; +case NETDEV_CHANGEMTU : +if (dev->mtu != netdev->mtu) { +erdma_set_mtu (dev, netdev->mtu); + dev->mtu = netdev->mtu; + } +break ;case NETDEV_REGISTER :case NETDEV_UNREGISTER :case NETDEV_CHANGEADDR : -case NETDEV_CHANGEMTU :case NETDEV_GOING_DOWN :case NETDEV_CHANGE :default :
@@ -95,6 +100,7 @@ static int erdma_device_register(struct erdma_dev *dev)if (ret)return ret; + dev->mtu = dev->netdev->mtu;addrconf_addr_eui48 ((u8 *)&ibdev->node_guid, dev->netdev->dev_addr); ret =ib_register_device (ibdev,"erdma_%d" , &dev->pdev->dev);
drivers/infiniband/hw/erdma/erdma_verbs.c +11 -0
@@ -1436,6 +1436,17 @@ int erdma_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr,return ret; } +void erdma_set_mtu (struct erdma_dev *dev,u32 mtu) +{ +struct erdma_cmdq_config_mtu_req req; + +erdma_cmdq_build_reqhdr (&req.hdr,CMDQ_SUBMOD_COMMON , +CMDQ_OPCODE_CONF_MTU ); + req.mtu = mtu; + +erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); +} +void erdma_port_event (struct erdma_dev *dev,enum ib_event_type reason) {struct ib_event event;
drivers/infiniband/hw/erdma/erdma_verbs.h +1 -0
@@ -330,5 +330,6 @@ struct ib_mr *erdma_ib_alloc_mr(struct ib_pd *ibpd, enum ib_mr_type mr_type,int erdma_map_mr_sg (struct ib_mr *ibmr,struct scatterlist *sg,int sg_nents,unsigned int *sg_offset);void erdma_port_event (struct erdma_dev *dev,enum ib_event_type reason); +void erdma_set_mtu (struct erdma_dev *dev,u32 mtu);#endif
Signature · 1 line
2.27.0
Re: [PATCH for-next 4/4] RDMA/erdma: Support dynamic mtu
Quoted 11 lines · Cheng Xu
Hardware now support jumbo frame for RDMA. So we introduce a new CMDQ message to support mtu change notification.
Signed-off-by Cheng Xu <[email protected]>--- drivers/infiniband/hw/erdma/erdma.h | 1 + drivers/infiniband/hw/erdma/erdma_hw.h | 6 ++++++ drivers/infiniband/hw/erdma/erdma_main.c | 8 +++++++- drivers/infiniband/hw/erdma/erdma_verbs.c | 11 +++++++++++ drivers/infiniband/hw/erdma/erdma_verbs.h | 1 + 5 files changed, 26 insertions(+), 1 deletion(-)
<...>
Quoted 35 lines
drivers/infiniband/hw/erdma/erdma_main.c +7
-1@@ -34,10 +34,15 @@ static int erdma_netdev_event(struct notifier_block *nb, unsigned long event, dev->state =IB_PORT_DOWN ;erdma_port_event (dev,IB_EVENT_PORT_ERR );break ; +case NETDEV_CHANGEMTU : +if (dev->mtu != netdev->mtu) { +erdma_set_mtu (dev, netdev->mtu); + dev->mtu = netdev->mtu; + } +break ;case NETDEV_REGISTER :case NETDEV_UNREGISTER :case NETDEV_CHANGEADDR : -case NETDEV_CHANGEMTU :case NETDEV_GOING_DOWN :case NETDEV_CHANGE :default :@@ -95,6 +100,7 @@ static int erdma_device_register(struct erdma_dev *dev)if (ret)return ret; + dev->mtu = dev->netdev->mtu;addrconf_addr_eui48 ((u8 *)&ibdev->node_guid, dev->netdev->dev_addr); ret =ib_register_device (ibdev,"erdma_%d" , &dev->pdev->dev);drivers/infiniband/hw/erdma/erdma_verbs.c +0
-0@@ -1436,6 +1436,17 @@ int erdma_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr,return ret; }
drivers/infiniband/hw/erdma/erdma_verbs.c int erdma_create_cq(struct ib_cq *ibcq, const struct ib_cq_init_attr *attr, · line 1448
Quoted 10 lines
drivers/infiniband/hw/erdma/erdma_verbs.c +10
-0+void erdma_set_mtu (struct erdma_dev *dev,u32 mtu) +{ +struct erdma_cmdq_config_mtu_req req; + +erdma_cmdq_build_reqhdr (&req.hdr,CMDQ_SUBMOD_COMMON , +CMDQ_OPCODE_CONF_MTU ); + req.mtu = mtu; + +erdma_post_cmd_wait (&dev->cmdq, &req,sizeof (req),NULL ,NULL ); +}
I don't see any backward compatibility here. How can you make sure that new code that supports MTU change works correctly on old FW/device?
Thanks
Cheng Xu
maintainerRe: [PATCH for-next 4/4] RDMA/erdma: Support dynamic mtu
Quoted 4 lines · Leon Romanovsky
Quoted 3 lines · Cheng Xu
Hardware now support jumbo frame for RDMA. So we introduce a new CMDQ message to support mtu change notification.
<...>
Quoted 4 lines
I don't see any backward compatibility here. How can you make sure that new code that supports MTU change works correctly on old FW/device?
In this case, driver needn't to consider backward compatibility.
ERDMA hardware is programmable part of our iaas infrastructure, and can be hot-update without BMs/VMs awareness.Before I submitted this patch, all the FWs has been updated, and support this feature, no old FWs exist.
Thanks,
Cheng Xu
Re: [PATCH for-next 4/4] RDMA/erdma: Support dynamic mtu
Quoted 13 lines · Cheng Xu
Quoted 4 lines · Leon Romanovsky
Quoted 3 lines · Cheng Xu
Hardware now support jumbo frame for RDMA. So we introduce a new CMDQ message to support mtu change notification.
<...>
Quoted 4 lines
I don't see any backward compatibility here. How can you make sure that new code that supports MTU change works correctly on old FW/device?
Quoted 5 lines
In this case, driver needn't to consider backward compatibility.
ERDMA hardware is programmable part of our iaas infrastructure, and can be hot-update without BMs/VMs awareness.Before I submitted this patch, all the FWs has been updated, and support this feature, no old FWs exist.
I'm not big fan of such answers, but ok, applied.
Quoted 5 lines
Thanks,
Cheng Xu
Leon Romanovsky
AppliedRe: (subset) [PATCH for-next 4/4] RDMA/erdma: Support dynamic mtu
In reply to Cheng XuQuoted 4 lines · Cheng Xu
Hardware now support jumbo frame for RDMA. So we introduce a new CMDQ message to support mtu change notification.
Applied, thanks!
[4/4] RDMA/erdma: Support dynamic mtu
https://git.kernel.org/rdma/rdma/c/9bdb9350f3808b
Best regards,
Signature · 1 line
Leon Romanovsky <[email protected]>
Leon Romanovsky
AppliedQuoted 8 lines · Cheng Xu
Hi,
This series has new updates for erdma driver:
- #1~#3 aim at making code more clean, including eliminating unnecessary castings, removing redundant include headers, and hiding hardware internal opcodes.
- #4 introcuces dynamic mtu support.
Quoted 7 lines
Thanks,
Cheng XuCheng Xu (4):
RDMA/erdma: Eliminate unnecessary casting forerdma_post_cmd_waitRDMA/erdma: Remove redundant includes
RDMA/erdma: Make hardware internal opcodes invisible to driver
I applied these three to -next.
Quoted 1 line
RDMA/erdma: Support dynamic mtu
This patch needs clarification.
Thanks