qedf: Misc fixes for the driver.
Subsystems (2)
Javed Hasan
maintainercover qedf: Misc fixes for the driver.
This series has misc bug fixes and code enhancements.
Kindly apply this series to scsi-queue at your earliest convenience.
Thanks,
~Javed
Javed Hasan (3):
qedf: Changed the debug parameter permission from read to read &
write.
qedf: Fix for the session’s E_D_TOV value.
qedf: FDMI Enhancement.
Saurav Kashyap (5):
qedf: Correct the comment in qedf_initiate_els.
qedf: Return SUCCESS if stale rport in encounteredon eh_abort.
qedf: Add schedule_hw_err_handler callback for fan failure.
qedf: Retry qed->probe during recovery.
qedf: Changes the %p to %px to print pointers.
drivers/scsi/qedf/qedf.h | 9 +++ drivers/scsi/qedf/qedf_els.c | 24 +++--- drivers/scsi/qedf/qedf_fip.c | 2 +- drivers/scsi/qedf/qedf_io.c | 56 ++++++------- drivers/scsi/qedf/qedf_main.c | 184 ++++++++++++++++++++++++++++++------------ 5 files changed, 184 insertions(+), 91 deletions(-)
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 1/8 qedf: Changed the debug parameter permission to read & write
Changed the debug parameter permission to read & write. Gives flexibility to change the debug verbosity dynamically.
--- drivers/scsi/qedf/qedf_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/qedf/qedf_main.c +1 -1
@@ -40,7 +40,7 @@ "remote ports (default 60)");uint qedf_debug =QEDF_LOG_INFO ; -module_param_named (debug, qedf_debug, uint,S_IRUGO ); +module_param_named (debug, qedf_debug, uint,S_IRUGO |S_IWUSR );MODULE_PARM_DESC (debug," Debug mask. Pass '1' to enable default debugging" " mask" );
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 2/8 qedf: Correct the comment in qedf_initiate_els
In reply to Javed HasanFrom: Saurav Kashyap <[email protected]>
Correction of the comment in qedf_initiate_els(). Comment was misleading, hence corrected it.
--- drivers/scsi/qedf/qedf_els.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/qedf/qedf_els.c +1 -1
@@ -124,7 +124,7 @@ static int qedf_initiate_els(struct qedf_rport *fcport, unsigned int op, task =qedf_get_task_mem (&qedf->tasks, xid);qedf_init_mp_task (els_req, task, sqe); -/* Put timer on original I/O request */ +/* Put timer on els request */ if (timer_msec)qedf_cmd_timer_set (qedf, els_req, timer_msec);
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 3/8 qedf: Fix for the session’s E_D_TOV value
In reply to Javed Hasan FW expects E_D_TOV field in connection offload parameters as “msec”. Earlier incorrect value(100ms), was leading to abort from driver in the case when data frames for read take more than 100ms from target side, resulting in FW reporting E_D_TOV expiration.
--- drivers/scsi/qedf/qedf_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/qedf/qedf_main.c +1 -1
@@ -1311,7 +1311,7 @@ static int qedf_offload_connection(struct qedf_ctx *qedf,ether_addr_copy (conn_info.dst_mac, qedf->ctlr.dest_addr); conn_info.tx_max_fc_pay_len = fcport->rdata->maxframe_size; - conn_info.e_d_tov_timer_val = qedf->lport->e_d_tov /20 ; + conn_info.e_d_tov_timer_val = qedf->lport->e_d_tov; conn_info.rec_tov_timer_val =3 ;/* I think this is what E3 was */ conn_info.rx_max_fc_pay_len = fcport->rdata->maxframe_size;
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 4/8 qedf: FDMI attributes correction
In reply to Javed HasanCorrection in the FDMI attributes required for RHBA and RPA registration.
--- drivers/scsi/qedf/qedf.h | 5 +++ drivers/scsi/qedf/qedf_main.c | 78 +++++++++++++++++++++++++++++-------------- 2 files changed, 58 insertions(+), 25 deletions(-)
drivers/scsi/qedf/qedf.h +5 -0
@@ -544,6 +544,11 @@ extern void qedf_process_seq_cleanup_compl(struct qedf_ctx *qedf,#define FCOE_WORD_TO_BYTE 4#define QEDF_MAX_TASK_NUM 0xFFFF +#define QL45xxx 0x165C +#define QL41xxx 0x8080 +#define MAX_CT_PAYLOAD 2048 +#define DISCOVERED_PORTS 4 +#define NUMBER_OF_PORTS 1struct fip_vlan {struct ethhdr eth;
drivers/scsi/qedf/qedf_main.c +53 -25
104 lines not on this page · raw message
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 5/8 qedf: Return SUCCESS if stale rport in encountered
In reply to Javed HasanFrom: Saurav Kashyap <[email protected]>
If SUCCESS is not return, it can get escalated. Hence, return
SUCCESS similar to other conditions.
--- drivers/scsi/qedf/qedf_main.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/qedf/qedf_main.c +1 -1
@@ -704,7 +704,7 @@ static int qedf_eh_abort(struct scsi_cmnd *sc_cmd) rdata = fcport->rdata;if (!rdata || !kref_get_unless_zero (&rdata->kref)) {QEDF_ERR (&qedf->dbg_ctx,"stale rport, sc_cmd=%p\n" , sc_cmd); - rc =1 ; + rc =SUCCESS ; goto out; }
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 6/8 qedf: Add schedule_hw_err_handler callback for fan failure
In reply to Javed HasanFrom: Saurav Kashyap <[email protected]>
On fan failure, disable the PCI function and initiate recovery for ramrod failure.
--- drivers/scsi/qedf/qedf.h | 4 ++++ drivers/scsi/qedf/qedf_main.c | 45 +++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 49 insertions(+)
drivers/scsi/qedf/qedf.h +4 -0
@@ -389,6 +389,7 @@ struct qedf_ctx {mempool_t *io_mempool;struct workqueue_struct *dpc_wq;struct delayed_work recovery_work; +struct delayed_work board_disable_work;struct delayed_work grcdump_work;struct delayed_work stag_work;
@@ -541,6 +542,9 @@ extern void qedf_process_seq_cleanup_compl(struct qedf_ctx *qedf,extern void qedf_wq_grcdump (struct work_struct *work);void qedf_stag_change_work (struct work_struct *work);void qedf_ctx_soft_reset (struct fc_lport *lport); +extern void qedf_board_disable_work (struct work_struct *work); +extern void qedf_schedule_hw_err_handler (void *dev, +enum qed_hw_err_type err_type);#define FCOE_WORD_TO_BYTE 4#define QEDF_MAX_TASK_NUM 0xFFFF
drivers/scsi/qedf/qedf_main.c +45 -0
63 lines not on this page · raw message
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 7/8 qedf: Retry qed->probe during recovery
In reply to Javed HasanFrom: Saurav Kashyap <[email protected]>
During recovery due to fcoe fn ramrod failure, we wait for 2 sec and then call qed->probe, if probe fails then re-try max 10 times.
--- drivers/scsi/qedf/qedf_main.c | 11 +++++++++++ 1 file changed, 11 insertions(+)
drivers/scsi/qedf/qedf_main.c +11 -0
@@ -3245,11 +3245,16 @@ static int __qedf_probe(struct pci_dev *pdev, int mode)void *task_start, *task_end;struct qed_slowpath_params slowpath_params;struct qed_probe_params qed_params; +u16 retry_cnt =10 ;/* * When doing error recovery we didn't reap the lport so don't try * to reallocate it. */ +retry_probe: +if (mode ==QEDF_MODE_RECOVERY ) +msleep (2000 ); +if (mode !=QEDF_MODE_RECOVERY ) { lport =libfc_host_alloc (&qedf_host_template,sizeof (struct qedf_ctx ));
@@ -3336,6 +3341,12 @@ static int __qedf_probe(struct pci_dev *pdev, int mode) qed_params.is_vf = is_vf; qedf->cdev = qed_ops->common->probe (pdev, &qed_params);if (!qedf->cdev) { +if ((mode ==QEDF_MODE_RECOVERY ) && retry_cnt) { +QEDF_ERR (&qedf->dbg_ctx, +"Retry %d initialize hardware\n" , retry_cnt); + retry_cnt--; + goto retry_probe; + }QEDF_ERR (&qedf->dbg_ctx,"common probe failed.\n" ); rc = -ENODEV ; goto err1;
Signature · 1 line
1.8.3.1
Javed Hasan
maintainerPATCH 8/8 qedf: Changes the %p to %px to print pointers
In reply to Javed HasanFrom: Saurav Kashyap <[email protected]>
Changed all the pointer address printing method from
%p to %px to print unmodified addresses.
--- drivers/scsi/qedf/qedf_els.c | 22 ++++++++--------- drivers/scsi/qedf/qedf_fip.c | 2 +- drivers/scsi/qedf/qedf_io.c | 56 +++++++++++++++++++++---------------------- drivers/scsi/qedf/qedf_main.c | 44 +++++++++++++++++----------------- 4 files changed, 62 insertions(+), 62 deletions(-)
drivers/scsi/qedf/qedf_els.c +11 -11
@@ -63,7 +63,7 @@ static int qedf_initiate_els(struct qedf_rport *fcport, unsigned int op, }QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"initiate_els els_req = " -"0x%p cb_arg = %p xid = %x\n" , els_req, cb_arg, +"0x%px cb_arg = %px xid = %x\n" , els_req, cb_arg, els_req->xid); els_req->sc_cmd =NULL ; els_req->cmd_type =QEDF_ELS ;
@@ -204,12 +204,12 @@ static void qedf_rrq_compl(struct qedf_els_cb_arg *cb_arg)if (!orig_io_req) {QEDF_ERR (&qedf->dbg_ctx, -"Original io_req is NULL, rrq_req = %p.\n" , rrq_req); +"Original io_req is NULL, rrq_req = %px.\n" , rrq_req); goto out_free; } refcount =kref_read (&orig_io_req->refcount); -QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"rrq_compl: orig io = %p," +QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"rrq_compl: orig io = %px," " orig xid = 0x%x, rrq_xid = 0x%x, refcount=%d\n" , orig_io_req, orig_io_req->xid, rrq_req->xid, refcount);
@@ -283,7 +283,7 @@ int qedf_send_rrq(struct qedf_ioreq *aborted_io_req) refcount =kref_read (&aborted_io_req->refcount);if (refcount !=1 ) {QEDF_INFO (&qedf->dbg_ctx,QEDF_LOG_ELS , -"refcount for xid=%x io_req=%p refcount=%d is not 1.\n" , +"refcount for xid=%x io_req=%px refcount=%d is not 1.\n" , aborted_io_req->xid, aborted_io_req, refcount);return -EINVAL ; }
@@ -293,7 +293,7 @@ int qedf_send_rrq(struct qedf_ioreq *aborted_io_req) r_a_tov = lport->r_a_tov;QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Sending RRQ orig " -"io = %p, orig_xid = 0x%x\n" , aborted_io_req, +"io = %px, orig_xid = 0x%x\n" , aborted_io_req, aborted_io_req->xid);memset (&rrq,0 ,sizeof (rrq));
@@ -381,7 +381,7 @@ void qedf_restart_rport(struct qedf_rport *fcport)if (test_bit (QEDF_RPORT_IN_RESET , &fcport->flags) || !test_bit (QEDF_RPORT_SESSION_READY , &fcport->flags) ||test_bit (QEDF_RPORT_UPLOADING_CONNECTION , &fcport->flags)) { -QEDF_ERR (&(fcport->qedf->dbg_ctx),"fcport %p already in reset or not offloaded.\n" , +QEDF_ERR (&(fcport->qedf->dbg_ctx),"fcport %px already in reset or not offloaded.\n" , fcport);spin_unlock_irqrestore (&fcport->rport_lock, flags);return ;
@@ -567,7 +567,7 @@ static void qedf_srr_compl(struct qedf_els_cb_arg *cb_arg)cancel_delayed_work_sync (&orig_io_req->timeout_work); refcount =kref_read (&orig_io_req->refcount); -QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Entered: orig_io=%p," +QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Entered: orig_io=%px," " orig_io_xid=0x%x, rec_xid=0x%x, refcount=%d\n" , orig_io_req, orig_io_req->xid, srr_req->xid, refcount);
@@ -655,7 +655,7 @@ static int qedf_send_srr(struct qedf_ioreq *orig_io_req, u32 offset, u8 r_ctl) lport = qedf->lport; r_a_tov = lport->r_a_tov; -QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Sending SRR orig_io=%p, " +QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Sending SRR orig_io=%px, " "orig_xid=0x%x\n" , orig_io_req, orig_io_req->xid);memset (&srr,0 ,sizeof (srr));
@@ -866,14 +866,14 @@ static void qedf_rec_compl(struct qedf_els_cb_arg *cb_arg)cancel_delayed_work_sync (&orig_io_req->timeout_work); refcount =kref_read (&orig_io_req->refcount); -QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Entered: orig_io=%p," +QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Entered: orig_io=%px," " orig_io_xid=0x%x, rec_xid=0x%x, refcount=%d\n" , orig_io_req, orig_io_req->xid, rec_req->xid, refcount);/* If a REC times out, free resources */ if (rec_req->event ==QEDF_IOREQ_EV_ELS_TMO ) {QEDF_ERR (&qedf->dbg_ctx, -"Got TMO event, orig_io_req %p orig_io_xid=0x%x.\n" , +"Got TMO event, orig_io_req %px orig_io_xid=0x%x.\n" , orig_io_req, orig_io_req->xid); goto out_put; }
@@ -1049,7 +1049,7 @@ int qedf_send_rec(struct qedf_ioreq *orig_io_req) rec.rec_rx_id =htons (orig_io_req->task->tstorm_st_context.read_write.rx_id); -QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Sending REC orig_io=%p, " +QEDF_INFO (&(qedf->dbg_ctx),QEDF_LOG_ELS ,"Sending REC orig_io=%px, " "orig_xid=0x%x rx_id=0x%x\n" , orig_io_req, orig_io_req->xid, rec.rec_rx_id); rc = qedf_initiate_els(fcport, ELS_REC, &rec, sizeof(rec),
drivers/scsi/qedf/qedf_fip.c +1 -1
@@ -189,7 +189,7 @@ void qedf_fip_recv(struct qedf_ctx *qedf, struct sk_buff *skb) sub = fiph->fip_subcode;QEDF_INFO (&qedf->dbg_ctx,QEDF_LOG_LL2 , -"FIP frame received: skb=%p fiph=%p source=%pM destn=%pM op=%x sub=%x vlan=%04x" , +"FIP frame received: skb=%px fiph=%px source=%pM destn=%pM op=%x sub=%x vlan=%04x" , skb, fiph, eth_hdr->h_source, eth_hdr->h_dest, op, sub, vlan);if (qedf_dump_frames)
drivers/scsi/qedf/qedf_io.c +28 -28
218 lines not on this page · raw message
drivers/scsi/qedf/qedf_main.c +22 -22
166 lines not on this page · raw message
Signature · 1 line
1.8.3.1
Martin K. Petersen
AppliedJaved,
Quoted 1 line
This series has misc bug fixes and code enhancements.
Applied to the 5.10 SCSI staging tree. Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Martin K. Petersen
AppliedQuoted 8 lines · Javed Hasan
This series has misc bug fixes and code enhancements.
Kindly apply this series to scsi-queue at your earliest convenience.
Thanks,
~Javed[...]
Applied to 5.10/scsi-queue, thanks!
[1/8] scsi: qedf: Change the debug parameter permission to read & write
https://git.kernel.org/mkp/scsi/c/066664645d9a [2/8] scsi: qedf: Correct the comment in qedf_initiate_els https://git.kernel.org/mkp/scsi/c/31fc82d7fbd8 [3/8] scsi: qedf: Fix for the session’s E_D_TOV value https://git.kernel.org/mkp/scsi/c/f78f8126264b [4/8] scsi: qedf: FDMI attributes correction https://git.kernel.org/mkp/scsi/c/41715c6292b6 [5/8] scsi: qedf: Return SUCCESS if stale rport is encountered https://git.kernel.org/mkp/scsi/c/10aff62fab26 [6/8] scsi: qedf: Add schedule_hw_err_handler callback for fan failure https://git.kernel.org/mkp/scsi/c/55e049910e08 [7/8] scsi: qedf: Retry qed->probe during recovery https://git.kernel.org/mkp/scsi/c/988100a7de0f
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering