[SCSI] bnx2fc: Improve error recovery by handling parity errors

During parity errors, the ramrods are not issued to FW. bnx2fc waits for the
timeout value, and proceeds with cleaning up the IOs. Since we are already
out-of-sync with FW, cleanup commands timeout too, and do not get the
completion.  This operation takes 36 secs for each session to upload causing
huge delays. To fix this, bnx2fc now gets a PARITY_ERROR from cnic driver, and
upon failure, the driver does not issue any commands to the FW and finishes the
upload process sooner.

Signed-off-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>

3 files changed, 24 insertions, 6 deletions

diff --git a/drivers/scsi/bnx2fc/bnx2fc.h b/drivers/scsi/bnx2fc/bnx2fc.h
index 2f1bea42336..baac6e96809 100644
--- a/drivers/scsi/bnx2fc/bnx2fc.h
+++ b/drivers/scsi/bnx2fc/bnx2fc.h
@@ -275,6 +275,7 @@ struct bnx2fc_rport {
 #define BNX2FC_FLAG_CTX_ALLOC_FAILURE	0x6
 #define BNX2FC_FLAG_UPLD_REQ_COMPL	0x7
 #define BNX2FC_FLAG_EXPL_LOGO		0x8
+#define BNX2FC_FLAG_DISABLE_FAILED	0x9
 
 	u8 src_addr[ETH_ALEN];
 	u32 max_sqes;
diff --git a/drivers/scsi/bnx2fc/bnx2fc_hwi.c b/drivers/scsi/bnx2fc/bnx2fc_hwi.c
index 2ca6bfe4ce5..6d6eee42ac7 100644
--- a/drivers/scsi/bnx2fc/bnx2fc_hwi.c
+++ b/drivers/scsi/bnx2fc/bnx2fc_hwi.c
@@ -1244,7 +1244,9 @@ static void bnx2fc_process_conn_disable_cmpl(struct bnx2fc_hba *hba,
 	if (disable_kcqe->completion_status) {
 		printk(KERN_ERR PFX "Disable failed with cmpl status %d\n",
 			disable_kcqe->completion_status);
-		return;
+		set_bit(BNX2FC_FLAG_DISABLE_FAILED, &tgt->flags);
+		set_bit(BNX2FC_FLAG_UPLD_REQ_COMPL, &tgt->flags);
+		wake_up_interruptible(&tgt->upld_wait);
 	} else {
 		/* disable successful */
 		BNX2FC_TGT_DBG(tgt, "disable successful\n");
diff --git a/drivers/scsi/bnx2fc/bnx2fc_tgt.c b/drivers/scsi/bnx2fc/bnx2fc_tgt.c
index 082a25c3117..85ffdeaa3b8 100644
--- a/drivers/scsi/bnx2fc/bnx2fc_tgt.c
+++ b/drivers/scsi/bnx2fc/bnx2fc_tgt.c
@@ -181,8 +181,14 @@ void bnx2fc_flush_active_ios(struct bnx2fc_rport *tgt)
 
 		set_bit(BNX2FC_FLAG_IO_COMPL, &io_req->req_flags);
 		set_bit(BNX2FC_FLAG_IO_CLEANUP, &io_req->req_flags);
-		rc = bnx2fc_initiate_cleanup(io_req);
-		BUG_ON(rc);
+
+		/* Do not issue cleanup when disable request failed */
+		if (test_bit(BNX2FC_FLAG_DISABLE_FAILED, &tgt->flags))
+			bnx2fc_process_cleanup_compl(io_req, io_req->task, 0);
+		else {
+			rc = bnx2fc_initiate_cleanup(io_req);
+			BUG_ON(rc);
+		}
 	}
 
 	list_for_each_safe(list, tmp, &tgt->active_tm_queue) {
@@ -212,8 +218,13 @@ void bnx2fc_flush_active_ios(struct bnx2fc_rport *tgt)
 			io_req->cb_arg = NULL;
 		}
 
-		rc = bnx2fc_initiate_cleanup(io_req);
-		BUG_ON(rc);
+		/* Do not issue cleanup when disable request failed */
+		if (test_bit(BNX2FC_FLAG_DISABLE_FAILED, &tgt->flags))
+			bnx2fc_process_cleanup_compl(io_req, io_req->task, 0);
+		else {
+			rc = bnx2fc_initiate_cleanup(io_req);
+			BUG_ON(rc);
+		}
 	}
 
 	list_for_each_safe(list, tmp, &tgt->io_retire_queue) {
@@ -321,9 +332,13 @@ static void bnx2fc_upload_session(struct fcoe_port *port,
 
 		del_timer_sync(&tgt->upld_timer);
 
-	} else
+	} else if (test_bit(BNX2FC_FLAG_DISABLE_FAILED, &tgt->flags)) {
+		printk(KERN_ERR PFX "ERROR!! DISABLE req failed, destroy"
+				" not sent to FW\n");
+	} else {
 		printk(KERN_ERR PFX "ERROR!! DISABLE req timed out, destroy"
 				" not sent to FW\n");
+	}
 
 	/* Free session resources */
 	bnx2fc_free_session_resc(hba, tgt);