Additional EDiF bug fixes
Subsystems (2)
Nilesh Javali
maintainercover Additional EDiF bug fixes
Martin,
Please apply the qla2xxx driver additional EDiF bug fixes to the scsi tree at your earliest convenience.
Thanks,
Nilesh
Nilesh Javali (1):
qla2xxx: Update version to 10.02.07.600-k
Quinn Tran (9):
qla2xxx: edif: Fix IO timeout due to over subscription
qla2xxx: edif: send logo for unexpected IKE message
qla2xxx: edif: reduce disruption due to multiple app start
qla2xxx: edif: fix no login after app start
qla2xxx: edif: tear down session if keys has been removed
qla2xxx: edif: fix session thrash
qla2xxx: edif: Fix no logout on delete for n2n
qla2xxx: edif: Reduce N2N thrashing at app_start time
qla2xxx: edif: Fix slow session tear down
drivers/scsi/qla2xxx/qla_def.h | 5 ++ drivers/scsi/qla2xxx/qla_edif.c | 80 +++++++++++++++++++++++------- drivers/scsi/qla2xxx/qla_edif.h | 4 ++ drivers/scsi/qla2xxx/qla_fw.h | 2 +- drivers/scsi/qla2xxx/qla_init.c | 10 +++- drivers/scsi/qla2xxx/qla_iocb.c | 3 ++ drivers/scsi/qla2xxx/qla_isr.c | 35 +++++++------ drivers/scsi/qla2xxx/qla_version.h | 4 +- 8 files changed, 107 insertions(+), 36 deletions(-)
base-commit: 3fd3a52ca672fea71ff6ebaded2e2ddbbfb3a397
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 1/10 qla2xxx: edif: Fix IO timeout due to over subscription
From: Quinn Tran <[email protected]>
Current edif code does not keep track of FW IOCB resources. This led to IOCB queue full on error recovery (IO timeout). This patch make use of the existing code that track IOCB resources to prevent over subscription of IOCB resources.
--- drivers/scsi/qla2xxx/qla_edif.c | 7 +++++++ 1 file changed, 7 insertions(+)
drivers/scsi/qla2xxx/qla_edif.c +7 -0
@@ -2951,6 +2951,12 @@ qla28xx_start_scsi_edif(srb_t *sp) tot_dsds = nseg; req_cnt =qla24xx_calc_iocbs (vha, tot_dsds); + + sp->iores.res_type =RESOURCE_INI ; + sp->iores.iocb_cnt = req_cnt; +if (qla_get_iocbs (sp->qpair, &sp->iores)) + goto queuing_error; +if (req->cnt < (req_cnt +2 )) { cnt =IS_SHADOW_REG_CAPABLE (ha) ? *req->out_ptr :rd_reg_dword (req->req_q_out);
@@ -3142,6 +3148,7 @@ qla28xx_start_scsi_edif(srb_t *sp)mempool_free (sp->u.scmd.ct6_ctx, ha->ctx_mempool); sp->u.scmd.ct6_ctx =NULL ; } +qla_put_iocbs (sp->qpair, &sp->iores);spin_unlock_irqrestore (lock, flags);return QLA_FUNCTION_FAILED ;
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 2/10 qla2xxx: edif: send logo for unexpected IKE message
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
If session is down and local port continue to receive
AUTH ELS messages for the down session, driver need to
send back LOGO so that remote device knows to tear down
its session. This patch terminate/cleanup the AUTH ELS exchange followed with a pass through LOGO.
--- drivers/scsi/qla2xxx/qla_edif.c | 19 +++++++++++++++++-- drivers/scsi/qla2xxx/qla_fw.h | 2 +- 2 files changed, 18 insertions(+), 3 deletions(-)
drivers/scsi/qla2xxx/qla_edif.c +17 -2
@@ -2565,8 +2565,7 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp) fcport =qla2x00_find_fcport_by_pid (host, &purex->pur_info.pur_sid); -if (DBELL_INACTIVE (vha) || - (fcport &&EDIF_SESSION_DOWN (fcport))) { +if (DBELL_INACTIVE (vha)) {ql_dbg (ql_dbg_edif, host,0x0910c ,"%s e_dbell.db_flags =%x %06x\n" , __func__, host->e_dbell.db_flags, fcport ? fcport->d_id.b24 :0 );
@@ -2576,6 +2575,22 @@ void qla24xx_auth_els(scsi_qla_host_t *vha, void **pkt, struct rsp_que **rsp)return ; } +if (fcport &&EDIF_SESSION_DOWN (fcport)) { +ql_dbg (ql_dbg_edif, host,0x13b6 , +"%s terminate exchange. Send logo to 0x%x\n" , + __func__, a.did.b24); + + a.tx_byte_count = a.tx_len =0 ; + a.tx_addr =0 ; + a.control_flags =EPD_RX_XCHG ;/* EPD_RX_XCHG = terminate cmd */ +qla_els_reject_iocb (host, (*rsp)->qpair, &a); +qla_enode_free (host, ptr); +/* send logo to let remote port knows to tear down session */ + fcport->send_els_logo =1 ; +qlt_schedule_sess_for_deletion (fcport); +return ; + } +/* add the local enode to the list */ qla_enode_add (host, ptr);
drivers/scsi/qla2xxx/qla_fw.h +1 -1
@@ -807,7 +807,7 @@ struct els_entry_24xx {#define EPD_ELS_COMMAND (0 << 13)#define EPD_ELS_ACC (1 << 13)#define EPD_ELS_RJT (2 << 13) -#define EPD_RX_XCHG (3 << 13) +#define EPD_RX_XCHG (3 << 13)/* terminate exchange */ #define ECF_CLR_PASSTHRU_PEND BIT_12#define ECF_INCL_FRAME_HDR BIT_11#define ECF_SEC_LOGIN BIT_3
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 3/10 qla2xxx: edif: reduce disruption due to multiple app start
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
Multiple app start can trigger session bounce. Driver will skip over session tear down if app start is seen more than once.
--- drivers/scsi/qla2xxx/qla_edif.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/scsi/qla2xxx/qla_edif.c +2 -2
@@ -510,8 +510,7 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)/* mark doorbell as active since an app is now present */ vha->e_dbell.db_flags |=EDB_ACTIVE ; }else { -ql_dbg (ql_dbg_edif, vha,0x911e ,"%s doorbell already active\n" , - __func__); + goto out; }if (N2N_TOPO (vha->hw)) {
@@ -578,6 +577,7 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job) __func__); } +out: appreply.host_support_edif = vha->hw->flags.edif_enabled; appreply.edif_enode_active = vha->pur_cinfo.enode_flags; appreply.edif_edb_active = vha->e_dbell.db_flags;
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 4/10 qla2xxx: edif: fix no login after app start
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
The scenario is user loaded driver but has not start authentication app. All sessions to secure device will exhaust all login attempt, fail and in stay in delete state.
Some time later, app started. Driver will:
replenish the login retry count,
trigger delete to prepare for secure login,
after deletion, relogin is triggered.
For the session that is already deleted, the delete trigger is a no-op. If none of the session trigger the relogin needed, then there's no movement.
This patch add a re-login trigger to relogin.
--- drivers/scsi/qla2xxx/qla_edif.c | 1 + 1 file changed, 1 insertion(+)
drivers/scsi/qla2xxx/qla_edif.c +1 -0
@@ -567,6 +567,7 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)qlt_schedule_sess_for_deletion (fcport);qla_edif_sa_ctl_init (vha, fcport); } +set_bit (RELOGIN_NEEDED , &vha->dpc_flags); }if (vha->pur_cinfo.enode_flags !=ENODE_ACTIVE ) {
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 5/10 qla2xxx: edif: tear down session if keys has been removed
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
If all keys for a session has been deleted then trigger a session tear down.
--- drivers/scsi/qla2xxx/qla_def.h | 5 +++++ drivers/scsi/qla2xxx/qla_isr.c | 1 + 2 files changed, 6 insertions(+)
drivers/scsi/qla2xxx/qla_def.h +5 -0
@@ -2158,6 +2158,11 @@ typedef struct {#define CS_IOCB_ERROR 0x31/* Generic error for IOCB request failure */ #define CS_REJECT_RECEIVED 0x4E/* Reject received */ +#define CS_EDIF_AUTH_ERROR 0x63/* decrypt error */ +#define CS_EDIF_PAD_LEN_ERROR 0x65/* pad > frame size, not 4byte align */ +#define CS_EDIF_INV_REQ 0x66/* invalid request */ +#define CS_EDIF_SPI_ERROR 0x67/* rx frame unable to locate sa */ +#define CS_EDIF_HDR_ERROR 0x69/* data frame != expected len */ #define CS_BAD_PAYLOAD 0x80/* Driver defined */ #define CS_UNKNOWN 0x81/* Driver defined */ #define CS_RETRY 0x82/* Driver defined */
drivers/scsi/qla2xxx/qla_isr.c +1 -0
@@ -3426,6 +3426,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)case CS_PORT_UNAVAILABLE :case CS_TIMEOUT :case CS_RESET : +case CS_EDIF_INV_REQ :/* * We are going to have the fc class block the rport
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 6/10 qla2xxx: edif: fix session thrash
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
Current code prematurely sends out prli before authentication application has given the ok to do it. This cause prli failure and session tear down. This patch prevents prli from going out before authentication app gives the ok.
--- drivers/scsi/qla2xxx/qla_edif.c | 2 +- drivers/scsi/qla2xxx/qla_edif.h | 4 ++++ drivers/scsi/qla2xxx/qla_init.c | 10 +++++++++- 3 files changed, 14 insertions(+), 2 deletions(-)
drivers/scsi/qla2xxx/qla_edif.c +1 -1
@@ -3517,7 +3517,7 @@ int qla_edif_process_els(scsi_qla_host_t *vha, struct bsg_job *bsg_job)if (qla_bsg_check (vha, bsg_job, fcport))return 0 ; -if (fcport->loop_id ==FC_NO_LOOP_ID ) { +if (EDIF_SESS_DELETE (fcport)) {ql_dbg (ql_dbg_edif, vha,0x910d ,"%s ELS code %x, no loop id.\n" , __func__, bsg_request->rqst_data.r_els.els_code);
drivers/scsi/qla2xxx/qla_edif.h +4 -0
@@ -141,4 +141,8 @@ struct enode { (DBELL_ACTIVE (_fcport->vha) && \ (_fcport->disc_state == DSC_LOGIN_AUTH_PEND)) +#define EDIF_SESS_DELETE (_s) \ + (qla_ini_mode_enabled(_s->vha) && (_s->disc_state == DSC_DELETE_PEND || \ + _s->disc_state == DSC_DELETED)) +#endif /* __QLA_EDIF_H */
drivers/scsi/qla2xxx/qla_init.c +9 -1
@@ -1762,8 +1762,16 @@ int qla24xx_fcport_handle_login(struct scsi_qla_host *vha, fc_port_t *fcport)break ;case DSC_LOGIN_PEND : -if (fcport->fw_login_state ==DSC_LS_PLOGI_COMP ) +if (vha->hw->flags.edif_enabled) +break ; + +if (fcport->fw_login_state ==DSC_LS_PLOGI_COMP ) { +ql_dbg (ql_dbg_disc, vha,0x2118 , +"%s %d %8phC post %s PRLI\n" , + __func__, __LINE__, fcport->port_name, +NVME_TARGET (vha->hw, fcport) ?"NVME" :"FC" );qla24xx_post_prli_work (vha, fcport); + }break ;case DSC_UPD_FCPORT :
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 7/10 qla2xxx: edif: Fix no logout on delete for n2n
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
Driver failed to send implicit logout on session delete. For edif, this failed to flush any lingering SA index in FW. This patch set a flag to turn on implicit logout early in the session recovery to make sure it would go out in case of error.
--- drivers/scsi/qla2xxx/qla_iocb.c | 3 +++ 1 file changed, 3 insertions(+)
drivers/scsi/qla2xxx/qla_iocb.c +3 -0
@@ -2882,6 +2882,9 @@ static void qla2x00_els_dcmd2_sp_done(srb_t *sp, int res) sp->name, res, sp->handle, fcport->d_id.b24, fcport->port_name); fcport->flags &= ~(FCF_ASYNC_SENT |FCF_ASYNC_ACTIVE ); +/* For edif, set logout on delete to ensure any residual key from FW is flushed.*/ + fcport->logout_on_delete =1 ; + fcport->chip_reset = vha->hw->base_qpair->chip_reset;if (sp->flags &SRB_WAKEUP_ON_COMP )complete (&lio->u.els_plogi.comp);
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 8/10 qla2xxx: edif: Reduce N2N thrashing at app_start time
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
For n2n + remote wwpn is bigger than local adapter, remote adapter will login to local adapter while authentication application is not running. When authentication application do starts, the current session in FW needs to to be invalidate. This patch make sure the old session is torn down before triggering a relogin.
--- drivers/scsi/qla2xxx/qla_edif.c | 47 ++++++++++++++++++++++++--------- 1 file changed, 34 insertions(+), 13 deletions(-)
drivers/scsi/qla2xxx/qla_edif.c +34 -13
@@ -517,11 +517,28 @@ qla_edif_app_start(scsi_qla_host_t *vha, struct bsg_job *bsg_job)list_for_each_entry_safe (fcport, tf, &vha->vp_fcports, list) fcport->n2n_link_reset_cnt =0 ; -if (vha->hw->flags.n2n_fw_acc_sec) -set_bit (N2N_LINK_RESET , &vha->dpc_flags); -else +if (vha->hw->flags.n2n_fw_acc_sec) { +list_for_each_entry_safe (fcport, tf, &vha->vp_fcports, list) +qla_edif_sa_ctl_init (vha, fcport); + +/* +* While authentication app was not running, remote device +* could still try to login with this local port. Let's +* clear the state and try again. +*/ +qla2x00_wait_for_sess_deletion (vha); + +/* bounce the link to get the other guy to relogin */ +if (!vha->hw->flags.n2n_bigger) { +set_bit (N2N_LINK_RESET , &vha->dpc_flags); +qla2xxx_wake_dpc (vha); + } + }else { +qla2x00_wait_for_hba_online (vha);set_bit (ISP_ABORT_NEEDED , &vha->dpc_flags); -qla2xxx_wake_dpc (vha); +qla2xxx_wake_dpc (vha); +qla2x00_wait_for_hba_online (vha); + } }else {list_for_each_entry_safe (fcport, tf, &vha->vp_fcports, list) { ql_dbg(ql_dbg_edif, vha, 0x2058,
@@ -920,17 +937,21 @@ qla_edif_app_getfcinfo(scsi_qla_host_t *vha, struct bsg_job *bsg_job)if (tdid.b24 !=0 && tdid.b24 != fcport->d_id.b24)continue ; -if (fcport->scan_state !=QLA_FCPORT_FOUND ) -continue ; +if (!N2N_TOPO (vha->hw)) { +if (fcport->scan_state !=QLA_FCPORT_FOUND ) +continue ; -if (fcport->port_type ==FCT_UNKNOWN && !fcport->fc4_features) - rval =qla24xx_async_gffid (vha, fcport, true); +if (fcport->port_type ==FCT_UNKNOWN && + !fcport->fc4_features) + rval =qla24xx_async_gffid (vha, fcport, + true); -if (!rval && - !(fcport->fc4_features &FC4_FF_TARGET || - fcport->port_type & - (FCT_TARGET |FCT_NVME_TARGET ))) -continue ; +if (!rval && + !(fcport->fc4_features &FC4_FF_TARGET || + fcport->port_type & + (FCT_TARGET |FCT_NVME_TARGET ))) +continue ; + } rval =0 ;
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 9/10 qla2xxx: edif: Fix slow session tear down
In reply to Nilesh JavaliFrom: Quinn Tran <[email protected]>
User experience slow recovery when target device went through a stop/start of the authentication application (app_stop/app_start).
Between the period of app_stop and app_start on the target device, target device choose to send ELS Reject for any receive AUTH ELS command. At this time, authentication application does not do els reject if it encounters error.
Therefore, AUTH ELS reject signify authentication application is not running. If driver pass up the AUTH ELS Reject to the authentication application, then it would result in authentication application retrying/resending the same AUTH ELS command again + delay.
As a work around, driver would trigger a session tear down where it tell the local authentication application to also tear down. At the next relogin, both side would be synchronize.
--- drivers/scsi/qla2xxx/qla_isr.c | 34 ++++++++++++++++++++-------------- 1 file changed, 20 insertions(+), 14 deletions(-)
drivers/scsi/qla2xxx/qla_isr.c +20 -14
@@ -2245,9 +2245,9 @@ qla24xx_els_ct_entry(scsi_qla_host_t *v, struct req_que *req, res =DID_ERROR <<16 ; } -if (logit) { -if (sp->remap.remapped && - ((u8 *)sp->remap.rsp.buf)[0 ] ==ELS_LS_RJT ) { +if (sp->remap.remapped && + ((u8 *)sp->remap.rsp.buf)[0 ] ==ELS_LS_RJT ) { +if (logit) {ql_dbg (ql_dbg_user, vha,0x503f ,"%s IOCB Done LS_RJT hdl=%x comp_status=0x%x\n" , type, sp->handle, comp_status);
@@ -2259,18 +2259,24 @@ qla24xx_els_ct_entry(scsi_qla_host_t *v, struct req_que *req, pkt)->total_byte_count), e->s_id[0 ], e->s_id[2 ], e->s_id[1 ], e->d_id[2], e->d_id[1], e->d_id[0]); - }else { -ql_log (ql_log_info, vha,0x503f , -"%s IOCB Done hdl=%x comp_status=0x%x\n" , - type, sp->handle, comp_status); -ql_log (ql_log_info, vha,0x503f , -"subcode 1=0x%x subcode 2=0x%x bytes=0x%x %02x%02x%02x -> %02x%02x%02x\n" , - fw_status[1 ], fw_status[2 ], -le32_to_cpu (((struct els_sts_entry_24xx *) - pkt)->total_byte_count), - e->s_id[0 ], e->s_id[2 ], e->s_id[1 ], - e->d_id[2 ], e->d_id[1 ], e->d_id[0 ]); } +if (sp->fcport && sp->fcport->flags &FCF_FCSP_DEVICE && + sp->type ==SRB_ELS_CMD_HST_NOLOGIN ) { +ql_dbg (ql_dbg_edif, vha,0x911e , +"%s rcv reject. Sched delete\n" , __func__); +qlt_schedule_sess_for_deletion (sp->fcport); + } + }else if (logit) { +ql_log (ql_log_info, vha,0x503f , +"%s IOCB Done hdl=%x comp_status=0x%x\n" , + type, sp->handle, comp_status); +ql_log (ql_log_info, vha,0x503f , +"subcode 1=0x%x subcode 2=0x%x bytes=0x%x %02x%02x%02x -> %02x%02x%02x\n" , + fw_status[1 ], fw_status[2 ], +le32_to_cpu (((struct els_sts_entry_24xx *) + pkt)->total_byte_count), + e->s_id[0 ], e->s_id[2 ], e->s_id[1 ], + e->d_id[2 ], e->d_id[1 ], e->d_id[0 ]); } } goto els_ct_done;
Signature · 1 line
2.19.0.rc0
Nilesh Javali
maintainerPATCH 10/10 qla2xxx: Update version to 10.02.07.600-k
In reply to Nilesh Javali--- drivers/scsi/qla2xxx/qla_version.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/scsi/qla2xxx/qla_version.h +2 -2
@@ -6,9 +6,9 @@/* * Driver version */ -#define QLA2XXX_VERSION "10.02.07.500-k" +#define QLA2XXX_VERSION "10.02.07.600-k"#define QLA_DRIVER_MAJOR_VER 10#define QLA_DRIVER_MINOR_VER 2#define QLA_DRIVER_PATCH_VER 7 -#define QLA_DRIVER_BETA_VER 500 +#define QLA_DRIVER_BETA_VER 600
Signature · 1 line
2.19.0.rc0
Quoted 37 lines · Nilesh Javali
Martin,
Please apply the qla2xxx driver additional EDiF bug fixes to the scsi tree at your earliest convenience.
Thanks,
NileshNilesh Javali (1):
qla2xxx: Update version to 10.02.07.600-kQuinn Tran (9):
qla2xxx: edif: Fix IO timeout due to over subscription
qla2xxx: edif: send logo for unexpected IKE message
qla2xxx: edif: reduce disruption due to multiple app start
qla2xxx: edif: fix no login after app start
qla2xxx: edif: tear down session if keys has been removed
qla2xxx: edif: fix session thrash
qla2xxx: edif: Fix no logout on delete for n2n
qla2xxx: edif: Reduce N2N thrashing atapp_starttime
qla2xxx: edif: Fix slow session tear downdrivers/scsi/qla2xxx/qla_def.h | 5 ++ drivers/scsi/qla2xxx/qla_edif.c | 80 +++++++++++++++++++++++------- drivers/scsi/qla2xxx/qla_edif.h | 4 ++ drivers/scsi/qla2xxx/qla_fw.h | 2 +- drivers/scsi/qla2xxx/qla_init.c | 10 +++- drivers/scsi/qla2xxx/qla_iocb.c | 3 ++ drivers/scsi/qla2xxx/qla_isr.c | 35 +++++++------ drivers/scsi/qla2xxx/qla_version.h | 4 +- 8 files changed, 107 insertions(+), 36 deletions(-)
Quoted 4 lines
base-commit:
3fd3a52ca672fea71ff6ebaded2e2ddbbfb3a397Signature · 1 line
2.19.0.rc0
The Series looks Good.
--
Himanshu Madhani Oracle Linux Engineering
Martin K. Petersen
AppliedNilesh,
Quoted 2 lines
Please apply the qla2xxx driver additional EDiF bug fixes to the scsi tree at your earliest convenience.
Applied to 5.20/scsi-staging, thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Martin K. Petersen
AppliedQuoted 9 lines · Nilesh Javali
Martin,
Please apply the qla2xxx driver additional EDiF bug fixes to the scsi tree at your earliest convenience.
Thanks,
Nilesh[...]
Applied to 5.20/scsi-queue, thanks!
[01/10] qla2xxx: edif: Fix IO timeout due to over subscription
https://git.kernel.org/mkp/scsi/c/63ab6cb582fa [02/10] qla2xxx: edif: send logo for unexpected IKE message https://git.kernel.org/mkp/scsi/c/2b659ed67a12 [03/10] qla2xxx: edif: reduce disruption due to multiple app start https://git.kernel.org/mkp/scsi/c/0dbfce5255fe [04/10] qla2xxx: edif: fix no login after app start https://git.kernel.org/mkp/scsi/c/24c796098f53 [05/10] qla2xxx: edif: tear down session if keys has been removed https://git.kernel.org/mkp/scsi/c/d7e2e4a68fc0 [06/10] qla2xxx: edif: fix session thrash https://git.kernel.org/mkp/scsi/c/a8fdfb0b39c2 [07/10] qla2xxx: edif: Fix no logout on delete for n2n https://git.kernel.org/mkp/scsi/c/ec538eb838f3 [08/10] qla2xxx: edif: Reduce N2N thrashing at app_start time https://git.kernel.org/mkp/scsi/c/37be3f9d6993 [09/10] qla2xxx: edif: Fix slow session tear down https://git.kernel.org/mkp/scsi/c/bcf536072f74 [10/10] qla2xxx: Update version to 10.02.07.600-k https://git.kernel.org/mkp/scsi/c/0f4d7d556125
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering