]> git.kernelconcepts.de Git - karo-tx-linux.git/commitdiff
[SCSI] qla2xxx: Cleanup FCP-command-status processing debug statements.
authorAndrew Vasquez <andrew.vasquez@qlogic.com>
Tue, 4 May 2010 22:01:27 +0000 (15:01 -0700)
committerJames Bottomley <James.Bottomley@suse.de>
Sun, 16 May 2010 22:21:54 +0000 (18:21 -0400)
Migrate to a consistent set of debug entries during status-IOCB
handling:

* group CS_TIMEOUT handling with CS_PORT_UNAVAILABLE and the like
  (more regrouping of common behaviour).

* drop CS_DATA_OVERRUN handling as it now falls into the
  'default' case (still returns DID_ERROR).

* consolidate CS_RESET and CS_ABORTED handling, as we the only
  functional difference was a printk() (still returns DID_RESET).

* dropped all the earlier inconsistent [qla_]printk()s sprinkled
  throught the needlessly large case-statement.  Failure case
  I/Os are now logged with a 'standard' format:

<command failure details>
<command generic details>

  so, for example a dropped-frame is logged as:

qla2xxx 0000:13:00.0: scsi(16:0:0) Dropped frame(s) detected (4000 of 4000 bytes).
qla2xxx 0000:13:00.0: scsi(16:0:0) FCP command status: 0x15-0x18 (70018) \
    oxid=78 ser=76 cdb=280000 len=4000 rsp_info=0 resid=0 fw_resid=4000

  which should now convey all relevant information.

Signed-off-by: Giridhar Malavali <giridhar.malavali@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>
drivers/scsi/qla2xxx/qla_isr.c

index a1968eb5e9ce24454006534ad37ba82a8943cd90..b92f9a6ee4fb3cfba0de1a7b440cbb3f923d2004 100644 (file)
@@ -1283,6 +1283,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
        struct sts_entry_24xx *sts24;
        uint16_t        comp_status;
        uint16_t        scsi_status;
+       uint16_t        ox_id;
        uint8_t         lscsi_status;
        int32_t         resid;
        uint32_t        sense_len, rsp_info_len, resid_len, fw_resid_len;
@@ -1291,6 +1292,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
        uint32_t handle;
        uint16_t que;
        struct req_que *req;
+       int logit = 1;
 
        sts = (sts_entry_t *) pkt;
        sts24 = (struct sts_entry_24xx *) pkt;
@@ -1320,9 +1322,9 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                sp = NULL;
 
        if (sp == NULL) {
-               DEBUG2(printk("scsi(%ld): Status Entry invalid handle.\n",
-                   vha->host_no));
-               qla_printk(KERN_WARNING, ha, "Status Entry invalid handle.\n");
+               qla_printk(KERN_WARNING, ha,
+                   "scsi(%ld): Invalid status handle (0x%x).\n", vha->host_no,
+                   sts->handle);
 
                set_bit(ISP_ABORT_NEEDED, &vha->dpc_flags);
                qla2xxx_wake_dpc(vha);
@@ -1330,10 +1332,9 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
        }
        cp = sp->cmd;
        if (cp == NULL) {
-               DEBUG2(printk("scsi(%ld): Command already returned back to OS "
-                   "pkt->handle=%d sp=%p.\n", vha->host_no, handle, sp));
                qla_printk(KERN_WARNING, ha,
-                   "Command is NULL: already returned to OS (sp=%p)\n", sp);
+                   "scsi(%ld): Command already returned (0x%x/%p).\n",
+                   vha->host_no, sts->handle, sp);
 
                return;
        }
@@ -1342,6 +1343,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
 
        fcport = sp->fcport;
 
+       ox_id = 0;
        sense_len = rsp_info_len = resid_len = fw_resid_len = 0;
        if (IS_FWI2_CAPABLE(ha)) {
                if (scsi_status & SS_SENSE_LEN_VALID)
@@ -1355,6 +1357,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                rsp_info = sts24->data;
                sense_data = sts24->data;
                host_to_fcp_swap(sts24->data, sizeof(sts24->data));
+               ox_id = le16_to_cpu(sts24->ox_id);
        } else {
                if (scsi_status & SS_SENSE_LEN_VALID)
                        sense_len = le16_to_cpu(sts->req_sense_length);
@@ -1371,17 +1374,13 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                if (IS_FWI2_CAPABLE(ha))
                        sense_data += rsp_info_len;
                if (rsp_info_len > 3 && rsp_info[3]) {
-                       DEBUG2(printk("scsi(%ld:%d:%d:%d) FCP I/O protocol "
-                           "failure (%x/%02x%02x%02x%02x%02x%02x%02x%02x)..."
-                           "retrying command\n", vha->host_no,
-                           cp->device->channel, cp->device->id,
-                           cp->device->lun, rsp_info_len, rsp_info[0],
-                           rsp_info[1], rsp_info[2], rsp_info[3], rsp_info[4],
-                           rsp_info[5], rsp_info[6], rsp_info[7]));
+                       DEBUG2(qla_printk(KERN_INFO, ha,
+                           "scsi(%ld:%d:%d): FCP I/O protocol failure "
+                           "(0x%x/0x%x).\n", vha->host_no, cp->device->id,
+                           cp->device->lun, rsp_info_len, rsp_info[3]));
 
                        cp->result = DID_BUS_BUSY << 16;
-                       qla2x00_sp_compl(ha, sp);
-                       return;
+                       goto out;
                }
        }
 
@@ -1408,12 +1407,10 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                            ((unsigned)(scsi_bufflen(cp) - resid) <
                             cp->underflow)) {
                                qla_printk(KERN_INFO, ha,
-                                          "scsi(%ld:%d:%d:%d): Mid-layer underflow "
-                                          "detected (%x of %x bytes)...returning "
-                                          "error status.\n", vha->host_no,
-                                          cp->device->channel, cp->device->id,
-                                          cp->device->lun, resid,
-                                          scsi_bufflen(cp));
+                                   "scsi(%ld:%d:%d): Mid-layer underflow "
+                                   "detected (0x%x of 0x%x bytes).\n",
+                                   vha->host_no, cp->device->id,
+                                   cp->device->lun, resid, scsi_bufflen(cp));
 
                                cp->result = DID_ERROR << 16;
                                break;
@@ -1422,12 +1419,12 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                cp->result = DID_OK << 16 | lscsi_status;
 
                if (lscsi_status == SAM_STAT_TASK_SET_FULL) {
-                       DEBUG2(printk(KERN_INFO
-                           "scsi(%ld): QUEUE FULL status detected "
-                           "0x%x-0x%x.\n", vha->host_no, comp_status,
-                           scsi_status));
+                       DEBUG2(qla_printk(KERN_INFO, ha,
+                           "scsi(%ld:%d:%d) QUEUE FULL detected.\n",
+                           vha->host_no, cp->device->id, cp->device->lun));
                        break;
                }
+               logit = 0;
                if (lscsi_status != SS_CHECK_CONDITION)
                        break;
 
@@ -1439,23 +1436,14 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                break;
 
        case CS_DATA_UNDERRUN:
-               DEBUG2(printk(KERN_INFO
-                   "scsi(%ld:%d:%d) UNDERRUN status detected 0x%x-0x%x. "
-                   "resid=0x%x fw_resid=0x%x cdb=0x%x os_underflow=0x%x\n",
-                   vha->host_no, cp->device->id, cp->device->lun, comp_status,
-                   scsi_status, resid_len, fw_resid_len, cp->cmnd[0],
-                   cp->underflow));
-
                /* Use F/W calculated residual length. */
                resid = IS_FWI2_CAPABLE(ha) ? fw_resid_len : resid_len;
                scsi_set_resid(cp, resid);
                if (scsi_status & SS_RESIDUAL_UNDER) {
                        if (IS_FWI2_CAPABLE(ha) && fw_resid_len != resid_len) {
-                               DEBUG2(printk(
-                                   "scsi(%ld:%d:%d:%d) Dropped frame(s) "
-                                   "detected (%x of %x bytes)...residual "
-                                   "length mismatch...retrying command.\n",
-                                   vha->host_no, cp->device->channel,
+                               DEBUG2(qla_printk(KERN_INFO, ha,
+                                   "scsi(%ld:%d:%d) Dropped frame(s) detected "
+                                   "(0x%x of 0x%x bytes).\n", vha->host_no,
                                    cp->device->id, cp->device->lun, resid,
                                    scsi_bufflen(cp)));
 
@@ -1467,21 +1455,18 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                            ((unsigned)(scsi_bufflen(cp) - resid) <
                            cp->underflow)) {
                                qla_printk(KERN_INFO, ha,
-                                   "scsi(%ld:%d:%d:%d): Mid-layer underflow "
-                                   "detected (%x of %x bytes)...returning "
-                                   "error status.\n", vha->host_no,
-                                   cp->device->channel, cp->device->id,
+                                   "scsi(%ld:%d:%d): Mid-layer underflow "
+                                   "detected (0x%x of 0x%x bytes).\n",
+                                   vha->host_no, cp->device->id,
                                    cp->device->lun, resid, scsi_bufflen(cp));
 
                                cp->result = DID_ERROR << 16;
                                break;
                        }
                } else if (!lscsi_status) {
-                       DEBUG2(printk(
-                           "scsi(%ld:%d:%d:%d) Dropped frame(s) detected "
-                           "(%x of %x bytes)...firmware reported underrun..."
-                           "retrying command.\n", vha->host_no,
-                           cp->device->channel, cp->device->id,
+                       DEBUG2(qla_printk(KERN_INFO, ha,
+                           "scsi(%ld:%d:%d) Dropped frame(s) detected (0x%x "
+                           "of 0x%x bytes).\n", vha->host_no, cp->device->id,
                            cp->device->lun, resid, scsi_bufflen(cp)));
 
                        cp->result = DID_ERROR << 16;
@@ -1489,6 +1474,7 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                }
 
                cp->result = DID_OK << 16 | lscsi_status;
+               logit = 0;
 
                /*
                 * Check to see if SCSI Status is non zero. If so report SCSI
@@ -1496,10 +1482,11 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                 */
                if (lscsi_status != 0) {
                        if (lscsi_status == SAM_STAT_TASK_SET_FULL) {
-                               DEBUG2(printk(KERN_INFO
-                                   "scsi(%ld): QUEUE FULL status detected "
-                                   "0x%x-0x%x.\n", vha->host_no, comp_status,
-                                   scsi_status));
+                               DEBUG2(qla_printk(KERN_INFO, ha,
+                                   "scsi(%ld:%d:%d) QUEUE FULL detected.\n",
+                                   vha->host_no, cp->device->id,
+                                   cp->device->lun));
+                               logit = 1;
                                break;
                        }
                        if (lscsi_status != SS_CHECK_CONDITION)
@@ -1513,109 +1500,56 @@ qla2x00_status_entry(scsi_qla_host_t *vha, struct rsp_que *rsp, void *pkt)
                }
                break;
 
-       case CS_DATA_OVERRUN:
-               DEBUG2(printk(KERN_INFO
-                   "scsi(%ld:%d:%d): OVERRUN status detected 0x%x-0x%x\n",
-                   vha->host_no, cp->device->id, cp->device->lun, comp_status,
-                   scsi_status));
-               DEBUG2(printk(KERN_INFO
-                   "CDB: 0x%x 0x%x 0x%x 0x%x 0x%x 0x%x\n",
-                   cp->cmnd[0], cp->cmnd[1], cp->cmnd[2], cp->cmnd[3],
-                   cp->cmnd[4], cp->cmnd[5]));
-               DEBUG2(printk(KERN_INFO
-                   "PID=0x%lx req=0x%x xtra=0x%x -- returning DID_ERROR "
-                   "status!\n",
-                   cp->serial_number, scsi_bufflen(cp), resid_len));
-
-               cp->result = DID_ERROR << 16;
-               break;
-
        case CS_PORT_LOGGED_OUT:
        case CS_PORT_CONFIG_CHG:
        case CS_PORT_BUSY:
        case CS_INCOMPLETE:
        case CS_PORT_UNAVAILABLE:
-               /*
-                * If the port is in Target Down state, return all IOs for this
-                * Target with DID_NO_CONNECT ELSE Queue the IOs in the
-                * retry_queue.
-                */
-               DEBUG2(printk("scsi(%ld:%d:%d): status_entry: Port Down "
-                   "pid=%ld, compl status=0x%x, port state=0x%x\n",
-                   vha->host_no, cp->device->id, cp->device->lun,
-                   cp->serial_number, comp_status,
-                   atomic_read(&fcport->state)));
-
+       case CS_TIMEOUT:
                /*
                 * We are going to have the fc class block the rport
                 * while we try to recover so instruct the mid layer
                 * to requeue until the class decides how to handle this.
                 */
                cp->result = DID_TRANSPORT_DISRUPTED << 16;
+
+               if (comp_status == CS_TIMEOUT) {
+                       if (IS_FWI2_CAPABLE(ha))
+                               break;
+                       else if ((le16_to_cpu(sts->status_flags) &
+                           SF_LOGOUT_SENT) == 0)
+                               break;
+               }
+
+               DEBUG2(qla_printk(KERN_INFO, ha,
+                       "scsi(%ld:%d:%d) Port down status: port-state=0x%x\n",
+                       vha->host_no, cp->device->id, cp->device->lun,
+                       atomic_read(&fcport->state)));
+
                if (atomic_read(&fcport->state) == FCS_ONLINE)
                        qla2x00_mark_device_lost(fcport->vha, fcport, 1, 1);
                break;
 
        case CS_RESET:
-               DEBUG2(printk(KERN_INFO
-                   "scsi(%ld): RESET status detected 0x%x-0x%x.\n",
-                   vha->host_no, comp_status, scsi_status));
-
-               cp->result = DID_RESET << 16;
-               break;
-
        case CS_ABORTED:
-               /*
-                * hv2.19.12 - DID_ABORT does not retry the request if we
-                * aborted this request then abort otherwise it must be a
-                * reset.
-                */
-               DEBUG2(printk(KERN_INFO
-                   "scsi(%ld): ABORT status detected 0x%x-0x%x.\n",
-                   vha->host_no, comp_status, scsi_status));
-
                cp->result = DID_RESET << 16;
                break;
-
-       case CS_TIMEOUT:
-               /*
-                * We are going to have the fc class block the rport
-                * while we try to recover so instruct the mid layer
-                * to requeue until the class decides how to handle this.
-                */
-               cp->result = DID_TRANSPORT_DISRUPTED << 16;
-
-               if (IS_FWI2_CAPABLE(ha)) {
-                       DEBUG2(printk(KERN_INFO
-                           "scsi(%ld:%d:%d:%d): TIMEOUT status detected "
-                           "0x%x-0x%x\n", vha->host_no, cp->device->channel,
-                           cp->device->id, cp->device->lun, comp_status,
-                           scsi_status));
-                       break;
-               }
-               DEBUG2(printk(KERN_INFO
-                   "scsi(%ld:%d:%d:%d): TIMEOUT status detected 0x%x-0x%x "
-                   "sflags=%x.\n", vha->host_no, cp->device->channel,
-                   cp->device->id, cp->device->lun, comp_status, scsi_status,
-                   le16_to_cpu(sts->status_flags)));
-
-               /* Check to see if logout occurred. */
-               if ((le16_to_cpu(sts->status_flags) & SF_LOGOUT_SENT))
-                       qla2x00_mark_device_lost(fcport->vha, fcport, 1, 1);
-               break;
-
        default:
-               DEBUG3(printk("scsi(%ld): Error detected (unknown status) "
-                   "0x%x-0x%x.\n", vha->host_no, comp_status, scsi_status));
-               qla_printk(KERN_INFO, ha,
-                   "Unknown status detected 0x%x-0x%x.\n",
-                   comp_status, scsi_status);
-
                cp->result = DID_ERROR << 16;
                break;
        }
 
-       /* Place command on done queue. */
+out:
+       if (logit)
+               DEBUG2(qla_printk(KERN_INFO, ha,
+                   "scsi(%ld:%d:%d) FCP command status: 0x%x-0x%x (0x%x) "
+                   "oxid=0x%x ser=0x%lx cdb=%02x%02x%02x len=0x%x "
+                   "rsp_info=0x%x resid=0x%x fw_resid=0x%x\n", vha->host_no,
+                   cp->device->id, cp->device->lun, comp_status, scsi_status,
+                   cp->result, ox_id, cp->serial_number, cp->cmnd[0],
+                   cp->cmnd[1], cp->cmnd[2], scsi_bufflen(cp), rsp_info_len,
+                   resid_len, fw_resid_len));
+
        if (rsp->status_srb == NULL)
                qla2x00_sp_compl(ha, sp);
 }