Merge patch series "Introduce support for multiqueue (MQ) in fnic"

Karan Tilak Kumar <kartilak@cisco.com> says:

Hi Martin, reviewers,

This cover letter describes the feature: add support for
multiqueue (MQ) to fnic driver.

Background: The Virtual Interface Card (VIC) firmware exposes several
queues that can be configured for sending IOs and receiving IO
responses. Unified Computing System Manager (UCSM) and Intersight
Manager (IMM) allows users to configure the number of queues to be
used for IOs.

The number of IO queues to be used is stored in a configuration file
by the VIC firmware. The fNIC driver reads the configuration file and
sets the number of queues to be used. Previously, the driver was
hard-coded to use only one queue. With this set of changes, the fNIC
driver will configure itself to use multiple queues. This feature
takes advantage of the block multiqueue layer to parallelize IOs being
sent out of the VIC card.

Here's a brief description of some of the salient patches:

- vnic_scsi.h needs to be in sync with VIC firmware to be able to read
the number of queues from the firmware config file. A patch has been
created for this.
- In an environment with many fnics (like we see in our customer
environments), it is hard to distinguish which fnic is printing logs.
Therefore, an fnic number has been included in the logs.
- read the number of queues from the firmware config file.
- include definitions in fnic.h to support multiqueue.
- modify the interrupt service routines (ISRs) to read from the
correct registers. The numbers that are used here come from discussions
with the VIC firmware team.
- track IO statistics for different queues.
- remove usage of host_lock, and only use fnic_lock in the fnic driver.
- use a hardware queue based spinlock to protect io_req.
- replace the hard-coded zeroth queue with a hardware queue number.
This presents a bulk of the changes.
- modify the definition of fnic_queuecommand to accept multiqueue tags.
- improve log messages, and indicate fnic number and multiqueue tags for
effective debugging.

Even though the patches have been made into a series, some patches are
heavier than others.
But, every effort has been made to keep the purpose of each patch as
a single-purpose, and to compile cleanly.

This patchset has been tested as a whole. Therefore, the tested-by fields
have been added only to two patches
in the set. All the individual patches compile cleanly. However,
I've refrained from adding tested-by to
most of the patches, so as to not mislead the reviewer/reader.

A brief note on the unit tests:

1. Increase number of queues to 64. Load driver. Run IOs via Medusa.
12+ hour run successful.
2. Configure multipathing, and run link flaps on single link.
IOs drop briefly, but pick up as expected.
3. Configure multipathing, and run link flaps on two links, with a
30 second delay in between. IOs drop briefly, but pick up as expected.

Repeat the above tests with 1 queue and 32 queues.  All tests were
successful.

Please consider this patch series for the next merge window.

Link: https://lore.kernel.org/r/20231211173617.932990-1-kartilak@cisco.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1 files changed, 45 insertions, 23 deletions

diff --git a/drivers/scsi/fnic/fnic.h b/drivers/scsi/fnic/fnic.h
index 22cef283b2b9..2074937c05bc 100644
--- a/drivers/scsi/fnic/fnic.h
+++ b/drivers/scsi/fnic/fnic.h
@@ -27,7 +27,7 @@
 
 #define DRV_NAME		"fnic"
 #define DRV_DESCRIPTION		"Cisco FCoE HBA Driver"
-#define DRV_VERSION		"1.6.0.57"
+#define DRV_VERSION		"1.7.0.0"
 #define PFX			DRV_NAME ": "
 #define DFX                     DRV_NAME "%d: "
 
@@ -36,7 +36,6 @@
 #define FNIC_MIN_IO_REQ			256 /* Min IO throttle count */
 #define FNIC_MAX_IO_REQ		1024 /* scsi_cmnd tag map entries */
 #define FNIC_DFLT_IO_REQ        256 /* Default scsi_cmnd tag map entries */
-#define	FNIC_IO_LOCKS		64 /* IO locks: power of 2 */
 #define FNIC_DFLT_QUEUE_DEPTH	256
 #define	FNIC_STATS_RATE_LIMIT	4 /* limit rate at which stats are pulled up */
 
@@ -109,7 +108,7 @@ static inline u64 fnic_flags_and_state(struct scsi_cmnd *cmd)
 #define FNIC_ABT_TERM_DELAY_TIMEOUT  500        /* mSec */
 
 #define FNIC_MAX_FCP_TARGET     256
-
+#define FNIC_PCI_OFFSET		2
 /**
  * state_flags to identify host state along along with fnic's state
  **/
@@ -144,31 +143,48 @@ do {								\
 		} while (0);					\
 } while (0)
 
-#define FNIC_MAIN_DBG(kern_level, host, fmt, args...)		\
+#define FNIC_MAIN_DBG(kern_level, host, fnic_num, fmt, args...)		\
 	FNIC_CHECK_LOGGING(FNIC_MAIN_LOGGING,			\
-			 shost_printk(kern_level, host, fmt, ##args);)
+			 shost_printk(kern_level, host,			\
+				"fnic<%d>: %s: %d: " fmt, fnic_num,\
+				__func__, __LINE__, ##args);)
 
-#define FNIC_FCS_DBG(kern_level, host, fmt, args...)		\
+#define FNIC_FCS_DBG(kern_level, host, fnic_num, fmt, args...)		\
 	FNIC_CHECK_LOGGING(FNIC_FCS_LOGGING,			\
-			 shost_printk(kern_level, host, fmt, ##args);)
+			 shost_printk(kern_level, host,			\
+				"fnic<%d>: %s: %d: " fmt, fnic_num,\
+				__func__, __LINE__, ##args);)
 
-#define FNIC_SCSI_DBG(kern_level, host, fmt, args...)		\
+#define FNIC_SCSI_DBG(kern_level, host, fnic_num, fmt, args...)		\
 	FNIC_CHECK_LOGGING(FNIC_SCSI_LOGGING,			\
-			 shost_printk(kern_level, host, fmt, ##args);)
+			 shost_printk(kern_level, host,			\
+				"fnic<%d>: %s: %d: " fmt, fnic_num,\
+				__func__, __LINE__, ##args);)
 
-#define FNIC_ISR_DBG(kern_level, host, fmt, args...)		\
+#define FNIC_ISR_DBG(kern_level, host, fnic_num, fmt, args...)		\
 	FNIC_CHECK_LOGGING(FNIC_ISR_LOGGING,			\
-			 shost_printk(kern_level, host, fmt, ##args);)
+			 shost_printk(kern_level, host,			\
+				"fnic<%d>: %s: %d: " fmt, fnic_num,\
+				__func__, __LINE__, ##args);)
 
 #define FNIC_MAIN_NOTE(kern_level, host, fmt, args...)          \
 	shost_printk(kern_level, host, fmt, ##args)
 
+#define FNIC_WQ_COPY_MAX 64
+#define FNIC_WQ_MAX 1
+#define FNIC_RQ_MAX 1
+#define FNIC_CQ_MAX (FNIC_WQ_COPY_MAX + FNIC_WQ_MAX + FNIC_RQ_MAX)
+#define FNIC_DFLT_IO_COMPLETIONS 256
+
+#define FNIC_MQ_CQ_INDEX        2
+
 extern const char *fnic_state_str[];
 
 enum fnic_intx_intr_index {
 	FNIC_INTX_WQ_RQ_COPYWQ,
-	FNIC_INTX_ERR,
+	FNIC_INTX_DUMMY,
 	FNIC_INTX_NOTIFY,
+	FNIC_INTX_ERR,
 	FNIC_INTX_INTR_MAX,
 };
 
@@ -176,7 +192,7 @@ enum fnic_msix_intr_index {
 	FNIC_MSIX_RQ,
 	FNIC_MSIX_WQ,
 	FNIC_MSIX_WQ_COPY,
-	FNIC_MSIX_ERR_NOTIFY,
+	FNIC_MSIX_ERR_NOTIFY = FNIC_MSIX_WQ_COPY + FNIC_WQ_COPY_MAX,
 	FNIC_MSIX_INTR_MAX,
 };
 
@@ -185,6 +201,7 @@ struct fnic_msix_entry {
 	char devname[IFNAMSIZ + 11];
 	irqreturn_t (*isr)(int, void *);
 	void *devid;
+	int irq_num;
 };
 
 enum fnic_state {
@@ -194,12 +211,6 @@ enum fnic_state {
 	FNIC_IN_ETH_TRANS_FC_MODE,
 };
 
-#define FNIC_WQ_COPY_MAX 1
-#define FNIC_WQ_MAX 1
-#define FNIC_RQ_MAX 1
-#define FNIC_CQ_MAX (FNIC_WQ_COPY_MAX + FNIC_WQ_MAX + FNIC_RQ_MAX)
-#define FNIC_DFLT_IO_COMPLETIONS 256
-
 struct mempool;
 
 enum fnic_evt {
@@ -214,8 +225,16 @@ struct fnic_event {
 	enum fnic_evt event;
 };
 
+struct fnic_cpy_wq {
+	unsigned long hw_lock_flags;
+	u16 active_ioreq_count;
+	u16 ioreq_table_size;
+	____cacheline_aligned struct fnic_io_req **io_req_table;
+};
+
 /* Per-instance private data structure */
 struct fnic {
+	int fnic_num;
 	struct fc_lport *lport;
 	struct fcoe_ctlr ctlr;		/* FIP FCoE controller structure */
 	struct vnic_dev_bar bar0;
@@ -282,8 +301,8 @@ struct fnic {
 	struct fnic_host_tag *tags;
 	mempool_t *io_req_pool;
 	mempool_t *io_sgl_pool[FNIC_SGL_NUM_CACHES];
-	spinlock_t io_req_lock[FNIC_IO_LOCKS];	/* locks for scsi cmnds */
 
+	unsigned int copy_wq_base;
 	struct work_struct link_work;
 	struct work_struct frame_work;
 	struct sk_buff_head frame_queue;
@@ -302,7 +321,9 @@ struct fnic {
 	/*** FIP related data members  -- end ***/
 
 	/* copy work queue cache line section */
-	____cacheline_aligned struct vnic_wq_copy wq_copy[FNIC_WQ_COPY_MAX];
+	____cacheline_aligned struct vnic_wq_copy hw_copy_wq[FNIC_WQ_COPY_MAX];
+	____cacheline_aligned struct fnic_cpy_wq sw_copy_wq[FNIC_WQ_COPY_MAX];
+
 	/* completion queue cache line section */
 	____cacheline_aligned struct vnic_cq cq[FNIC_CQ_MAX];
 
@@ -330,6 +351,7 @@ extern const struct attribute_group *fnic_host_groups[];
 
 void fnic_clear_intr_mode(struct fnic *fnic);
 int fnic_set_intr_mode(struct fnic *fnic);
+int fnic_set_intr_mode_msix(struct fnic *fnic);
 void fnic_free_intr(struct fnic *fnic);
 int fnic_request_intr(struct fnic *fnic);
 
@@ -356,7 +378,7 @@ void fnic_scsi_cleanup(struct fc_lport *);
 void fnic_scsi_abort_io(struct fc_lport *);
 void fnic_empty_scsi_cleanup(struct fc_lport *);
 void fnic_exch_mgr_reset(struct fc_lport *, u32, u32);
-int fnic_wq_copy_cmpl_handler(struct fnic *fnic, int);
+int fnic_wq_copy_cmpl_handler(struct fnic *fnic, int copy_work_to_do, unsigned int cq_index);
 int fnic_wq_cmpl_handler(struct fnic *fnic, int);
 int fnic_flogi_reg_handler(struct fnic *fnic, u32);
 void fnic_wq_copy_cleanup_handler(struct vnic_wq_copy *wq,
@@ -364,7 +386,7 @@ void fnic_wq_copy_cleanup_handler(struct vnic_wq_copy *wq,
 int fnic_fw_reset_handler(struct fnic *fnic);
 void fnic_terminate_rport_io(struct fc_rport *);
 const char *fnic_state_to_str(unsigned int state);
-
+void fnic_mq_map_queues_cpus(struct Scsi_Host *host);
 void fnic_log_q_error(struct fnic *fnic);
 void fnic_handle_link_event(struct fnic *fnic);