2005-04-18 01:05:31 +04:00
/*******************************************************************
* This file is part of the Emulex Linux Device Driver for *
2005-06-25 18:34:39 +04:00
* Fibre Channel Host Bus Adapters . *
2006-03-01 03:25:27 +03:00
* Copyright ( C ) 2004 - 2006 Emulex . All rights reserved . *
2005-06-25 18:34:39 +04:00
* EMULEX and SLI are trademarks of Emulex . *
2005-04-18 01:05:31 +04:00
* www . emulex . com *
2005-06-25 18:34:39 +04:00
* Portions Copyright ( C ) 2004 - 2005 Christoph Hellwig *
2005-04-18 01:05:31 +04:00
* *
* This program is free software ; you can redistribute it and / or *
2005-06-25 18:34:39 +04:00
* modify it under the terms of version 2 of the GNU General *
* Public License as published by the Free Software Foundation . *
* This program is distributed in the hope that it will be useful . *
* ALL EXPRESS OR IMPLIED CONDITIONS , REPRESENTATIONS AND *
* WARRANTIES , INCLUDING ANY IMPLIED WARRANTY OF MERCHANTABILITY , *
* FITNESS FOR A PARTICULAR PURPOSE , OR NON - INFRINGEMENT , ARE *
* DISCLAIMED , EXCEPT TO THE EXTENT THAT SUCH DISCLAIMERS ARE HELD *
* TO BE LEGALLY INVALID . See the GNU General Public License for *
* more details , a copy of which can be found in the file COPYING *
* included with this package . *
2005-04-18 01:05:31 +04:00
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * */
struct lpfc_sli2_slim ;
2006-07-06 23:49:25 +04:00
# define LPFC_MAX_TARGET 256 /* max number of targets supported */
# define LPFC_MAX_DISC_THREADS 64 / * max outstanding discovery els
requests */
# define LPFC_MAX_NS_RETRY 3 / * Number of retry attempts to contact
the NameServer before giving up . */
2005-04-18 01:05:31 +04:00
# define LPFC_DFT_HBA_Q_DEPTH 2048 /* max cmds per hba */
# define LPFC_LC_HBA_Q_DEPTH 1024 /* max cmds per low cost hba */
# define LPFC_LP101_HBA_Q_DEPTH 128 /* max cmds per low cost hba */
2005-11-28 19:42:38 +03:00
# define LPFC_CMD_PER_LUN 3 /* max outstanding cmds per lun */
2005-04-18 01:05:31 +04:00
# define LPFC_SG_SEG_CNT 64 /* sg element count per scsi cmnd */
# define LPFC_IOCB_LIST_CNT 2250 /* list of IOCBs for fast-path usage. */
2005-11-28 19:42:38 +03:00
# define LPFC_Q_RAMP_UP_INTERVAL 120 /* lun q_depth ramp up interval */
2005-04-18 01:05:31 +04:00
/* Define macros for 64 bit support */
# define putPaddrLow(addr) ((uint32_t) (0xffffffff & (u64)(addr)))
# define putPaddrHigh(addr) ((uint32_t) (0xffffffff & (((u64)(addr))>>32)))
# define getPaddr(high, low) ((dma_addr_t)( \
( ( ( u64 ) ( high ) < < 16 ) < < 16 ) | ( ( u64 ) ( low ) ) ) )
/* Provide maximum configuration definitions. */
# define LPFC_DRVR_TIMEOUT 16 /* driver iocb timeout value in sec */
# define FC_MAX_ADPTMSG 64
# define MAX_HBAEVT 32
[SCSI] lpfc 8.1.1 : Add polled-mode support
- Add functionality to run in polled mode only. Includes run time
attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters
Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.
This patch adds a new "poll" attribute, and the following behavior:
if value is 0 (default):
The driver uses the normal method for i/o completion. It uses the
firmware feature of interrupt coalesing. The firmware allows a
minimum number of i/o completions before an interrupt, or a maximum
time delay between interrupts. By default, the driver sets these
to no delay (disabled) or 1 i/o - meaning coalescing is disabled.
Attributes were provided to change the coalescing values, but it was
a module-load time only and global across all adapters.
This patch allows them to be writable on a per-adapter basis.
if value is 1 :
Interrupts are left enabled, expecting that the user has tuned the
interrupt coalescing values. When this setting is enabled, the driver
will attempt to service completed i/o whenever new i/o is submitted
to the adapter. If the coalescing values are large, and the i/o
generation rate steady, an interrupt will be avoided by servicing
completed i/o prior to the coalescing thresholds kicking in. However,
if the i/o completion load is high enough or i/o generation slow, the
coalescion values will ensure that completed i/o is serviced in a timely
fashion.
if value is 3 :
Turns off FCP i/o interrupts altogether. The coalescing values now have
no effect. A new attribute "poll_tmo" (default 10ms) exists to set
the polling interval for i/o completion. When this setting is enabled,
the driver will attempt to service completed i/o and restart the
interval timer whenever new i/o is submitted. This behavior allows for
servicing of completed i/o sooner than the interval timer, but ensures
that if no i/o is being issued, then the interval timer will kick in
to service the outstanding i/o.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-11-30 00:32:13 +03:00
enum lpfc_polling_flags {
ENABLE_FCP_RING_POLLING = 0x1 ,
DISABLE_FCP_RING_INT = 0x2
} ;
2005-04-18 01:05:31 +04:00
/* Provide DMA memory definitions the driver uses per port instance. */
struct lpfc_dmabuf {
struct list_head list ;
void * virt ; /* virtual address ptr */
dma_addr_t phys ; /* mapped address */
} ;
struct lpfc_dma_pool {
struct lpfc_dmabuf * elements ;
uint32_t max_count ;
uint32_t current_count ;
} ;
/* Priority bit. Set value to exceed low water mark in lpfc_mem. */
# define MEM_PRI 0x100
/****************************************************************************/
/* Device VPD save area */
/****************************************************************************/
typedef struct lpfc_vpd {
uint32_t status ; /* vpd status value */
uint32_t length ; /* number of bytes actually returned */
struct {
uint32_t rsvd1 ; /* Revision numbers */
uint32_t biuRev ;
uint32_t smRev ;
uint32_t smFwRev ;
uint32_t endecRev ;
uint16_t rBit ;
uint8_t fcphHigh ;
uint8_t fcphLow ;
uint8_t feaLevelHigh ;
uint8_t feaLevelLow ;
uint32_t postKernRev ;
uint32_t opFwRev ;
uint8_t opFwName [ 16 ] ;
uint32_t sli1FwRev ;
uint8_t sli1FwName [ 16 ] ;
uint32_t sli2FwRev ;
uint8_t sli2FwName [ 16 ] ;
} rev ;
} lpfc_vpd_t ;
struct lpfc_scsi_buf ;
/*
* lpfc stat counters
*/
struct lpfc_stats {
/* Statistics for ELS commands */
uint32_t elsLogiCol ;
uint32_t elsRetryExceeded ;
uint32_t elsXmitRetry ;
uint32_t elsDelayRetry ;
uint32_t elsRcvDrop ;
uint32_t elsRcvFrame ;
uint32_t elsRcvRSCN ;
uint32_t elsRcvRNID ;
uint32_t elsRcvFARP ;
uint32_t elsRcvFARPR ;
uint32_t elsRcvFLOGI ;
uint32_t elsRcvPLOGI ;
uint32_t elsRcvADISC ;
uint32_t elsRcvPDISC ;
uint32_t elsRcvFAN ;
uint32_t elsRcvLOGO ;
uint32_t elsRcvPRLO ;
uint32_t elsRcvPRLI ;
2006-03-01 03:25:15 +03:00
uint32_t elsRcvLIRR ;
uint32_t elsRcvRPS ;
uint32_t elsRcvRPL ;
2005-04-18 01:05:31 +04:00
uint32_t elsXmitFLOGI ;
uint32_t elsXmitPLOGI ;
uint32_t elsXmitPRLI ;
uint32_t elsXmitADISC ;
uint32_t elsXmitLOGO ;
uint32_t elsXmitSCR ;
uint32_t elsXmitRNID ;
uint32_t elsXmitFARP ;
uint32_t elsXmitFARPR ;
uint32_t elsXmitACC ;
uint32_t elsXmitLSRJT ;
uint32_t frameRcvBcast ;
uint32_t frameRcvMulti ;
uint32_t strayXmitCmpl ;
uint32_t frameXmitDelay ;
uint32_t xriCmdCmpl ;
uint32_t xriStatErr ;
uint32_t LinkUp ;
uint32_t LinkDown ;
uint32_t LinkMultiEvent ;
uint32_t NoRcvBuf ;
uint32_t fcpCmd ;
uint32_t fcpCmpl ;
uint32_t fcpRspErr ;
uint32_t fcpRemoteStop ;
uint32_t fcpPortRjt ;
uint32_t fcpPortBusy ;
uint32_t fcpError ;
uint32_t fcpLocalErr ;
} ;
enum sysfs_mbox_state {
SMBOX_IDLE ,
SMBOX_WRITING ,
SMBOX_READING
} ;
struct lpfc_sysfs_mbox {
enum sysfs_mbox_state state ;
size_t offset ;
struct lpfcMboxq * mbox ;
} ;
struct lpfc_hba {
struct lpfc_sli sli ;
struct lpfc_sli2_slim * slim2p ;
dma_addr_t slim2p_mapping ;
uint16_t pci_cfg_value ;
2006-03-01 03:25:27 +03:00
int32_t hba_state ;
# define LPFC_STATE_UNKNOWN 0 /* HBA state is unknown */
# define LPFC_WARM_START 1 /* HBA state after selective reset */
# define LPFC_INIT_START 2 /* Initial state after board reset */
# define LPFC_INIT_MBX_CMDS 3 /* Initialize HBA with mbox commands */
# define LPFC_LINK_DOWN 4 /* HBA initialized, link is down */
# define LPFC_LINK_UP 5 /* Link is up - issue READ_LA */
# define LPFC_LOCAL_CFG_LINK 6 /* local NPORT Id configured */
# define LPFC_FLOGI 7 /* FLOGI sent to Fabric */
# define LPFC_FABRIC_CFG_LINK 8 / * Fabric assigned NPORT Id
2005-04-18 01:05:31 +04:00
configured */
2006-03-01 03:25:27 +03:00
# define LPFC_NS_REG 9 /* Register with NameServer */
# define LPFC_NS_QRY 10 /* Query NameServer for NPort ID list */
# define LPFC_BUILD_DISC_LIST 11 / * Build ADISC and PLOGI lists for
2005-04-18 01:05:31 +04:00
* device authentication / discovery */
2006-03-01 03:25:27 +03:00
# define LPFC_DISC_AUTH 12 /* Processing ADISC list */
# define LPFC_CLEAR_LA 13 / * authentication cmplt - issue
2005-04-18 01:05:31 +04:00
CLEAR_LA */
# define LPFC_HBA_READY 32
2006-03-01 03:25:27 +03:00
# define LPFC_HBA_ERROR -1
2005-04-18 01:05:31 +04:00
2006-03-07 23:04:13 +03:00
int32_t stopped ; /* HBA has not been restarted since last ERATT */
2005-04-18 01:05:31 +04:00
uint8_t fc_linkspeed ; /* Link speed after last READ_LA */
uint32_t fc_eventTag ; /* event tag for link attention */
uint32_t fc_prli_sent ; /* cntr for outstanding PRLIs */
uint32_t num_disc_nodes ; /*in addition to hba_state */
struct timer_list fc_estabtmo ; /* link establishment timer */
struct timer_list fc_disctmo ; /* Discovery rescue timer */
struct timer_list fc_fdmitmo ; /* fdmi timer */
/* These fields used to be binfo */
struct lpfc_name fc_nodename ; /* fc nodename */
struct lpfc_name fc_portname ; /* fc portname */
uint32_t fc_pref_DID ; /* preferred D_ID */
uint8_t fc_pref_ALPA ; /* preferred AL_PA */
uint32_t fc_edtov ; /* E_D_TOV timer value */
uint32_t fc_arbtov ; /* ARB_TOV timer value */
uint32_t fc_ratov ; /* R_A_TOV timer value */
uint32_t fc_rttov ; /* R_T_TOV timer value */
uint32_t fc_altov ; /* AL_TOV timer value */
uint32_t fc_crtov ; /* C_R_TOV timer value */
uint32_t fc_citov ; /* C_I_TOV timer value */
uint32_t fc_myDID ; /* fibre channel S_ID */
uint32_t fc_prevDID ; /* previous fibre channel S_ID */
struct serv_parm fc_sparam ; /* buffer for our service parameters */
struct serv_parm fc_fabparam ; /* fabric service parameters buffer */
uint8_t alpa_map [ 128 ] ; /* AL_PA map from READ_LA */
uint8_t fc_ns_retry ; /* retries for fabric nameserver */
uint32_t fc_nlp_cnt ; /* outstanding NODELIST requests */
uint32_t fc_rscn_id_cnt ; /* count of RSCNs payloads in list */
struct lpfc_dmabuf * fc_rscn_id_list [ FC_MAX_HOLD_RSCN ] ;
uint32_t lmt ;
uint32_t fc_flag ; /* FC flags */
# define FC_PT2PT 0x1 /* pt2pt with no fabric */
# define FC_PT2PT_PLOGI 0x2 /* pt2pt initiate PLOGI */
# define FC_DISC_TMO 0x4 /* Discovery timer running */
# define FC_PUBLIC_LOOP 0x8 /* Public loop */
# define FC_LBIT 0x10 /* LOGIN bit in loopinit set */
# define FC_RSCN_MODE 0x20 /* RSCN cmd rcv'ed */
# define FC_NLP_MORE 0x40 /* More node to process in node tbl */
# define FC_OFFLINE_MODE 0x80 /* Interface is offline for diag */
# define FC_FABRIC 0x100 /* We are fabric attached */
# define FC_ESTABLISH_LINK 0x200 /* Reestablish Link */
# define FC_RSCN_DISCOVERY 0x400 /* Authenticate all devices after RSCN*/
# define FC_LOADING 0x1000 /* HBA in process of loading drvr */
# define FC_UNLOADING 0x2000 /* HBA in process of unloading drvr */
# define FC_SCSI_SCAN_TMO 0x4000 /* scsi scan timer running */
# define FC_ABORT_DISCOVERY 0x8000 /* we want to abort discovery */
# define FC_NDISC_ACTIVE 0x10000 /* NPort discovery active */
2006-03-01 03:25:23 +03:00
# define FC_BYPASSED_MODE 0x20000 /* NPort is in bypassed mode */
2005-04-18 01:05:31 +04:00
uint32_t fc_topology ; /* link topology, from LINK INIT */
struct lpfc_stats fc_stat ;
/* These are the head/tail pointers for the bind, plogi, adisc, unmap,
* and map lists . Their counters are immediately following .
*/
struct list_head fc_plogi_list ;
struct list_head fc_adisc_list ;
struct list_head fc_reglogin_list ;
struct list_head fc_prli_list ;
struct list_head fc_nlpunmap_list ;
struct list_head fc_nlpmap_list ;
struct list_head fc_npr_list ;
struct list_head fc_unused_list ;
/* Keep counters for the number of entries in each list. */
uint16_t fc_plogi_cnt ;
uint16_t fc_adisc_cnt ;
uint16_t fc_reglogin_cnt ;
uint16_t fc_prli_cnt ;
uint16_t fc_unmap_cnt ;
uint16_t fc_map_cnt ;
uint16_t fc_npr_cnt ;
uint16_t fc_unused_cnt ;
struct lpfc_nodelist fc_fcpnodev ; /* nodelist entry for no device */
uint32_t nport_event_cnt ; /* timestamp for nlplist entry */
uint32_t wwnn [ 2 ] ;
uint32_t RandomData [ 7 ] ;
uint32_t cfg_log_verbose ;
uint32_t cfg_lun_queue_depth ;
uint32_t cfg_nodev_tmo ;
2006-08-19 01:47:08 +04:00
uint32_t cfg_devloss_tmo ;
2005-04-18 01:05:31 +04:00
uint32_t cfg_hba_queue_depth ;
uint32_t cfg_fcp_class ;
uint32_t cfg_use_adisc ;
uint32_t cfg_ack0 ;
uint32_t cfg_topology ;
uint32_t cfg_scan_down ;
uint32_t cfg_link_speed ;
uint32_t cfg_cr_delay ;
uint32_t cfg_cr_count ;
2006-03-01 06:33:08 +03:00
uint32_t cfg_multi_ring_support ;
2005-04-18 01:05:31 +04:00
uint32_t cfg_fdmi_on ;
uint32_t cfg_discovery_threads ;
uint32_t cfg_max_luns ;
[SCSI] lpfc 8.1.1 : Add polled-mode support
- Add functionality to run in polled mode only. Includes run time
attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters
Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.
This patch adds a new "poll" attribute, and the following behavior:
if value is 0 (default):
The driver uses the normal method for i/o completion. It uses the
firmware feature of interrupt coalesing. The firmware allows a
minimum number of i/o completions before an interrupt, or a maximum
time delay between interrupts. By default, the driver sets these
to no delay (disabled) or 1 i/o - meaning coalescing is disabled.
Attributes were provided to change the coalescing values, but it was
a module-load time only and global across all adapters.
This patch allows them to be writable on a per-adapter basis.
if value is 1 :
Interrupts are left enabled, expecting that the user has tuned the
interrupt coalescing values. When this setting is enabled, the driver
will attempt to service completed i/o whenever new i/o is submitted
to the adapter. If the coalescing values are large, and the i/o
generation rate steady, an interrupt will be avoided by servicing
completed i/o prior to the coalescing thresholds kicking in. However,
if the i/o completion load is high enough or i/o generation slow, the
coalescion values will ensure that completed i/o is serviced in a timely
fashion.
if value is 3 :
Turns off FCP i/o interrupts altogether. The coalescing values now have
no effect. A new attribute "poll_tmo" (default 10ms) exists to set
the polling interval for i/o completion. When this setting is enabled,
the driver will attempt to service completed i/o and restart the
interval timer whenever new i/o is submitted. This behavior allows for
servicing of completed i/o sooner than the interval timer, but ensures
that if no i/o is being issued, then the interval timer will kick in
to service the outstanding i/o.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-11-30 00:32:13 +03:00
uint32_t cfg_poll ;
uint32_t cfg_poll_tmo ;
2005-04-18 01:05:31 +04:00
uint32_t cfg_sg_seg_cnt ;
uint32_t cfg_sg_dma_buf_size ;
2006-08-19 01:47:18 +04:00
uint64_t cfg_soft_wwpn ;
2005-04-18 01:05:31 +04:00
2006-08-19 01:47:08 +04:00
uint32_t dev_loss_tmo_changed ;
2005-04-18 01:05:31 +04:00
lpfc_vpd_t vpd ; /* vital product data */
struct Scsi_Host * host ;
struct pci_dev * pcidev ;
struct list_head work_list ;
uint32_t work_ha ; /* Host Attention Bits for WT */
uint32_t work_ha_mask ; /* HA Bits owned by WT */
uint32_t work_hs ; /* HS stored in case of ERRAT */
uint32_t work_status [ 2 ] ; /* Extra status from SLIM */
uint32_t work_hba_events ; /* Timeout to be handled */
# define WORKER_DISC_TMO 0x1 /* Discovery timeout */
# define WORKER_ELS_TMO 0x2 /* ELS timeout */
# define WORKER_MBOX_TMO 0x4 /* MBOX timeout */
# define WORKER_FDMI_TMO 0x8 /* FDMI timeout */
wait_queue_head_t * work_wait ;
struct task_struct * worker_thread ;
unsigned long pci_bar0_map ; /* Physical address for PCI BAR0 */
unsigned long pci_bar2_map ; /* Physical address for PCI BAR2 */
void __iomem * slim_memmap_p ; /* Kernel memory mapped address for
PCI BAR0 */
void __iomem * ctrl_regs_memmap_p ; /* Kernel memory mapped address for
PCI BAR2 */
void __iomem * MBslimaddr ; /* virtual address for mbox cmds */
void __iomem * HAregaddr ; /* virtual address for host attn reg */
void __iomem * CAregaddr ; /* virtual address for chip attn reg */
void __iomem * HSregaddr ; /* virtual address for host status
reg */
void __iomem * HCregaddr ; /* virtual address for host ctl reg */
int brd_no ; /* FC board number */
char SerialNumber [ 32 ] ; /* adapter Serial Number */
char OptionROMVersion [ 32 ] ; /* adapter BIOS / Fcode version */
char ModelDesc [ 256 ] ; /* Model Description */
char ModelName [ 80 ] ; /* Model Name */
char ProgramType [ 256 ] ; /* Program Type */
char Port [ 20 ] ; /* Port No */
uint8_t vpd_flag ; /* VPD data flag */
# define VPD_MODEL_DESC 0x1 /* valid vpd model description */
# define VPD_MODEL_NAME 0x2 /* valid vpd model name */
# define VPD_PROGRAM_TYPE 0x4 /* valid vpd program type */
# define VPD_PORT 0x8 /* valid vpd port data */
# define VPD_MASK 0xf /* mask for any vpd data */
2006-08-19 01:47:18 +04:00
uint8_t soft_wwpn_enable ;
[SCSI] lpfc 8.1.1 : Add polled-mode support
- Add functionality to run in polled mode only. Includes run time
attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters
Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.
This patch adds a new "poll" attribute, and the following behavior:
if value is 0 (default):
The driver uses the normal method for i/o completion. It uses the
firmware feature of interrupt coalesing. The firmware allows a
minimum number of i/o completions before an interrupt, or a maximum
time delay between interrupts. By default, the driver sets these
to no delay (disabled) or 1 i/o - meaning coalescing is disabled.
Attributes were provided to change the coalescing values, but it was
a module-load time only and global across all adapters.
This patch allows them to be writable on a per-adapter basis.
if value is 1 :
Interrupts are left enabled, expecting that the user has tuned the
interrupt coalescing values. When this setting is enabled, the driver
will attempt to service completed i/o whenever new i/o is submitted
to the adapter. If the coalescing values are large, and the i/o
generation rate steady, an interrupt will be avoided by servicing
completed i/o prior to the coalescing thresholds kicking in. However,
if the i/o completion load is high enough or i/o generation slow, the
coalescion values will ensure that completed i/o is serviced in a timely
fashion.
if value is 3 :
Turns off FCP i/o interrupts altogether. The coalescing values now have
no effect. A new attribute "poll_tmo" (default 10ms) exists to set
the polling interval for i/o completion. When this setting is enabled,
the driver will attempt to service completed i/o and restart the
interval timer whenever new i/o is submitted. This behavior allows for
servicing of completed i/o sooner than the interval timer, but ensures
that if no i/o is being issued, then the interval timer will kick in
to service the outstanding i/o.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-11-30 00:32:13 +03:00
struct timer_list fcp_poll_timer ;
2005-04-18 01:05:31 +04:00
struct timer_list els_tmofunc ;
[SCSI] lpfc 8.1.1 : Add polled-mode support
- Add functionality to run in polled mode only. Includes run time
attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters
Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.
This patch adds a new "poll" attribute, and the following behavior:
if value is 0 (default):
The driver uses the normal method for i/o completion. It uses the
firmware feature of interrupt coalesing. The firmware allows a
minimum number of i/o completions before an interrupt, or a maximum
time delay between interrupts. By default, the driver sets these
to no delay (disabled) or 1 i/o - meaning coalescing is disabled.
Attributes were provided to change the coalescing values, but it was
a module-load time only and global across all adapters.
This patch allows them to be writable on a per-adapter basis.
if value is 1 :
Interrupts are left enabled, expecting that the user has tuned the
interrupt coalescing values. When this setting is enabled, the driver
will attempt to service completed i/o whenever new i/o is submitted
to the adapter. If the coalescing values are large, and the i/o
generation rate steady, an interrupt will be avoided by servicing
completed i/o prior to the coalescing thresholds kicking in. However,
if the i/o completion load is high enough or i/o generation slow, the
coalescion values will ensure that completed i/o is serviced in a timely
fashion.
if value is 3 :
Turns off FCP i/o interrupts altogether. The coalescing values now have
no effect. A new attribute "poll_tmo" (default 10ms) exists to set
the polling interval for i/o completion. When this setting is enabled,
the driver will attempt to service completed i/o and restart the
interval timer whenever new i/o is submitted. This behavior allows for
servicing of completed i/o sooner than the interval timer, but ensures
that if no i/o is being issued, then the interval timer will kick in
to service the outstanding i/o.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-11-30 00:32:13 +03:00
2005-04-18 01:05:31 +04:00
/*
* stat counters
*/
uint64_t fc4InputRequests ;
uint64_t fc4OutputRequests ;
uint64_t fc4ControlRequests ;
struct lpfc_sysfs_mbox sysfs_mbox ;
/* fastpath list. */
[SCSI] lpfc 8.1.1 : Add polled-mode support
- Add functionality to run in polled mode only. Includes run time
attribute to enable mode.
- Enable runtime writable hba settings for coallescing and delay parameters
Customers have requested a mode in the driver to run strictly polled.
This is generally to support an environment where the server is extremely
loaded and is looking to reclaim some cpu cycles from adapter interrupt
handling.
This patch adds a new "poll" attribute, and the following behavior:
if value is 0 (default):
The driver uses the normal method for i/o completion. It uses the
firmware feature of interrupt coalesing. The firmware allows a
minimum number of i/o completions before an interrupt, or a maximum
time delay between interrupts. By default, the driver sets these
to no delay (disabled) or 1 i/o - meaning coalescing is disabled.
Attributes were provided to change the coalescing values, but it was
a module-load time only and global across all adapters.
This patch allows them to be writable on a per-adapter basis.
if value is 1 :
Interrupts are left enabled, expecting that the user has tuned the
interrupt coalescing values. When this setting is enabled, the driver
will attempt to service completed i/o whenever new i/o is submitted
to the adapter. If the coalescing values are large, and the i/o
generation rate steady, an interrupt will be avoided by servicing
completed i/o prior to the coalescing thresholds kicking in. However,
if the i/o completion load is high enough or i/o generation slow, the
coalescion values will ensure that completed i/o is serviced in a timely
fashion.
if value is 3 :
Turns off FCP i/o interrupts altogether. The coalescing values now have
no effect. A new attribute "poll_tmo" (default 10ms) exists to set
the polling interval for i/o completion. When this setting is enabled,
the driver will attempt to service completed i/o and restart the
interval timer whenever new i/o is submitted. This behavior allows for
servicing of completed i/o sooner than the interval timer, but ensures
that if no i/o is being issued, then the interval timer will kick in
to service the outstanding i/o.
Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-11-30 00:32:13 +03:00
spinlock_t scsi_buf_list_lock ;
2005-04-18 01:05:31 +04:00
struct list_head lpfc_scsi_buf_list ;
uint32_t total_scsi_bufs ;
struct list_head lpfc_iocb_list ;
uint32_t total_iocbq_bufs ;
/* pci_mem_pools */
struct pci_pool * lpfc_scsi_dma_buf_pool ;
struct pci_pool * lpfc_mbuf_pool ;
struct lpfc_dma_pool lpfc_mbuf_safety_pool ;
mempool_t * mbox_mem_pool ;
mempool_t * nlp_mem_pool ;
struct list_head freebufList ;
struct list_head ctrspbuflist ;
struct list_head rnidrspbuflist ;
2005-08-10 23:03:01 +04:00
struct fc_host_statistics link_stats ;
2005-04-18 01:05:31 +04:00
} ;
struct rnidrsp {
void * buf ;
uint32_t uniqueid ;
struct list_head list ;
uint32_t data ;
} ;
2006-08-19 01:46:43 +04:00
# define FC_REG_DUMP_EVENT 0x10 /* Register for Dump events */