]> err.no Git - linux-2.6/commitdiff
[SCSI] lpfc 8.1.12 : Don't process ERATT interrupts when issuing KILL_BOARD mbx command
authorJames Smart <James.Smart@Emulex.Com>
Wed, 25 Apr 2007 13:53:15 +0000 (09:53 -0400)
committerJames Bottomley <jejb@mulgrave.il.steeleye.com>
Sun, 6 May 2007 14:33:15 +0000 (09:33 -0500)
Don't process ERATT interrupts when issuing KILL_BOARD mbx command

Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
drivers/scsi/lpfc/lpfc.h
drivers/scsi/lpfc/lpfc_sli.c

index ccc4ca194c60827af988086ccaf68c80e8b975f8..913e08ee0b4a982743820080e9aef9dedacd92a1 100644 (file)
@@ -250,6 +250,7 @@ struct lpfc_hba {
 #define FC_LOOPBACK_MODE        0x40000        /* NPort is in Loopback mode */
                                        /* This flag is set while issuing */
                                        /* INIT_LINK mailbox command */
+#define FC_IGNORE_ERATT         0x80000        /* intr handler should ignore ERATT */
 
        uint32_t fc_topology;   /* link topology, from LINK INIT */
 
index 645291efce0849db90d1d91b935a8fef1ec9e171..1508b3f6639f99e7a1d62f4cb23349ae5d9d5aef 100644 (file)
@@ -1584,6 +1584,7 @@ void lpfc_reset_barrier(struct lpfc_hba * phba)
        hc_copy = readl(phba->HCregaddr);
        writel((hc_copy & ~HC_ERINT_ENA), phba->HCregaddr);
        readl(phba->HCregaddr); /* flush */
+       phba->fc_flag |= FC_IGNORE_ERATT;
 
        if (readl(phba->HAregaddr) & HA_ERATT) {
                /* Clear Chip error bit */
@@ -1626,6 +1627,7 @@ clear_errat:
        }
 
 restore_hc:
+       phba->fc_flag &= ~FC_IGNORE_ERATT;
        writel(hc_copy, phba->HCregaddr);
        readl(phba->HCregaddr); /* flush */
 }
@@ -1661,6 +1663,7 @@ lpfc_sli_brdkill(struct lpfc_hba * phba)
        status &= ~HC_ERINT_ENA;
        writel(status, phba->HCregaddr);
        readl(phba->HCregaddr); /* flush */
+       phba->fc_flag |= FC_IGNORE_ERATT;
        spin_unlock_irq(phba->host->host_lock);
 
        lpfc_kill_board(phba, pmb);
@@ -1670,6 +1673,9 @@ lpfc_sli_brdkill(struct lpfc_hba * phba)
        if (retval != MBX_SUCCESS) {
                if (retval != MBX_BUSY)
                        mempool_free(pmb, phba->mbox_mem_pool);
+               spin_lock_irq(phba->host->host_lock);
+               phba->fc_flag &= ~FC_IGNORE_ERATT;
+               spin_unlock_irq(phba->host->host_lock);
                return 1;
        }
 
@@ -1696,6 +1702,7 @@ lpfc_sli_brdkill(struct lpfc_hba * phba)
        }
        spin_lock_irq(phba->host->host_lock);
        psli->sli_flag &= ~LPFC_SLI_MBOX_ACTIVE;
+       phba->fc_flag &= ~FC_IGNORE_ERATT;
        spin_unlock_irq(phba->host->host_lock);
 
        psli->mbox_active = NULL;
@@ -3207,6 +3214,11 @@ lpfc_intr_handler(int irq, void *dev_id)
         */
        spin_lock(phba->host->host_lock);
        ha_copy = readl(phba->HAregaddr);
+       /* If somebody is waiting to handle an eratt don't process it
+        * here.  The brdkill function will do this.
+        */
+       if (phba->fc_flag & FC_IGNORE_ERATT)
+               ha_copy &= ~HA_ERATT;
        writel((ha_copy & ~(HA_LATT | HA_ERATT)), phba->HAregaddr);
        readl(phba->HAregaddr); /* flush */
        spin_unlock(phba->host->host_lock);