WSL2-Linux-Kernel/include/scsi/scsi_transport_fc.h

883 строки
28 KiB
C
Исходник Обычный вид История

/* SPDX-License-Identifier: GPL-2.0-or-later */
/*
* FiberChannel transport specific attributes exported to sysfs.
*
* Copyright (c) 2003 Silicon Graphics, Inc. All rights reserved.
* Copyright (C) 2004-2007 James Smart, Emulex Corporation
* Rewrite for host, target, device, and remote port attributes,
* statistics, and service functions...
*/
#ifndef SCSI_TRANSPORT_FC_H
#define SCSI_TRANSPORT_FC_H
#include <linux/sched.h>
#include <linux/bsg-lib.h>
#include <asm/unaligned.h>
[SCSI] update fc_transport for removal of block/unblock functions We recently went back to implement a board reset. When we perform the reset, we wanted to tear down the internal data structures and rebuild them. Unfortunately, when it came to the rport structure, things were odd. If we deleted them, the scsi targets and sdevs would be torn down. Not a good thing for a temporary reset. We could block the rports, but we either maintain the internal structures to keep the rport reference (perhaps even replicating what's in the transport), or we have to fatten the fc transport with new search routines to find the rport (and deal with a case of a dangling rport that the driver forgets). It dawned on me that we had actually reached this state incorrectly. When the fc transport first started, we did the block/unblock first, then added the rport interface. The purpose of block/unblock is to hide the temporary disappearance of the rport (e.g. being deleted, then readded). Why are we making the driver do the block/unblock ? We should be making the transport have only an rport add/delete, and the let the transport handle the block/unblock. So... This patch removes the existing fc_remote_port_block/unblock functions. It moves the block/unblock functionality into the fc_remote_port_add/delete functions. Updates for the lpfc driver are included. Qlogic driver updates are also enclosed, thanks to the contributions of Andrew Vasquez. [Note: the qla2xxx changes are relative to the scsi-misc-2.6 tree as of this morning - which does not include the recent patches sent by Andrew]. The zfcp driver does not use the block/unblock functions. One last comment: The resulting behavior feels very clean. The LLDD is concerned only with add/delete, which corresponds to the physical disappearance. However, the fact that the scsi target and sdevs are not immediately torn down after the LLDD calls delete causes an interesting scenario... the midlayer can call the xxx_slave_alloc and xxx_queuecommand functions with a sdev that is at the location the rport used to be. The driver must validate the device exists when it first enters these functions. In thinking about it, this has always been the case for the LLDD and these routines. The existing drivers already check for existence. However, this highlights that simple validation via data structure dereferencing needs to be watched. To deal with this, a new transport function, fc_remote_port_chkready() was created that LLDDs should call when they first enter these two routines. It validates the rport state, and returns a scsi result which could be returned. In addition to solving the above, it also creates consistent behavior from the LLDD's when the block and deletes are occuring. Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-10-18 20:03:35 +04:00
#include <scsi/scsi.h>
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
#include <scsi/scsi_netlink.h>
#include <scsi/scsi_host.h>
struct scsi_transport_template;
/*
* FC Port definitions - Following FC HBAAPI guidelines
*
* Note: Not all binary values for the different fields match HBAAPI.
* Instead, we use densely packed ordinal values or enums.
* We get away with this as we never present the actual binary values
* externally. For sysfs, we always present the string that describes
* the value. Thus, an admin doesn't need a magic HBAAPI decoder ring
* to understand the values. The HBAAPI user-space library is free to
* convert the strings into the HBAAPI-specified binary values.
*
* Note: Not all HBAAPI-defined values are contained in the definitions
* below. Those not appropriate to an fc_host (e.g. FCP initiator) have
* been removed.
*/
/*
* fc_port_type: If you alter this, you also need to alter scsi_transport_fc.c
* (for the ascii descriptions).
*/
enum fc_port_type {
FC_PORTTYPE_UNKNOWN,
FC_PORTTYPE_OTHER,
FC_PORTTYPE_NOTPRESENT,
FC_PORTTYPE_NPORT, /* Attached to FPort */
FC_PORTTYPE_NLPORT, /* (Public) Loop w/ FLPort */
FC_PORTTYPE_LPORT, /* (Private) Loop w/o FLPort */
FC_PORTTYPE_PTP, /* Point to Point w/ another NPort */
FC_PORTTYPE_NPIV, /* VPORT based on NPIV */
};
/*
* fc_port_state: If you alter this, you also need to alter scsi_transport_fc.c
* (for the ascii descriptions).
*/
enum fc_port_state {
FC_PORTSTATE_UNKNOWN,
FC_PORTSTATE_NOTPRESENT,
FC_PORTSTATE_ONLINE,
FC_PORTSTATE_OFFLINE, /* User has taken Port Offline */
FC_PORTSTATE_BLOCKED,
FC_PORTSTATE_BYPASSED,
FC_PORTSTATE_DIAGNOSTICS,
FC_PORTSTATE_LINKDOWN,
FC_PORTSTATE_ERROR,
FC_PORTSTATE_LOOPBACK,
FC_PORTSTATE_DELETED,
FC_PORTSTATE_MARGINAL,
};
/*
* fc_vport_state: If you alter this, you also need to alter
* scsi_transport_fc.c (for the ascii descriptions).
*/
enum fc_vport_state {
FC_VPORT_UNKNOWN,
FC_VPORT_ACTIVE,
FC_VPORT_DISABLED,
FC_VPORT_LINKDOWN,
FC_VPORT_INITIALIZING,
FC_VPORT_NO_FABRIC_SUPP,
FC_VPORT_NO_FABRIC_RSCS,
FC_VPORT_FABRIC_LOGOUT,
FC_VPORT_FABRIC_REJ_WWN,
FC_VPORT_FAILED,
};
/*
* FC Classes of Service
* Note: values are not enumerated, as they can be "or'd" together
* for reporting (e.g. report supported_classes). If you alter this list,
* you also need to alter scsi_transport_fc.c (for the ascii descriptions).
*/
#define FC_COS_UNSPECIFIED 0
#define FC_COS_CLASS1 2
#define FC_COS_CLASS2 4
#define FC_COS_CLASS3 8
#define FC_COS_CLASS4 0x10
#define FC_COS_CLASS6 0x40
/*
* FC Port Speeds
* Note: values are not enumerated, as they can be "or'd" together
* for reporting (e.g. report supported_speeds). If you alter this list,
* you also need to alter scsi_transport_fc.c (for the ascii descriptions).
*/
#define FC_PORTSPEED_UNKNOWN 0 /* Unknown - transceiver
incapable of reporting */
#define FC_PORTSPEED_1GBIT 1
#define FC_PORTSPEED_2GBIT 2
#define FC_PORTSPEED_10GBIT 4
#define FC_PORTSPEED_4GBIT 8
#define FC_PORTSPEED_8GBIT 0x10
#define FC_PORTSPEED_16GBIT 0x20
#define FC_PORTSPEED_32GBIT 0x40
#define FC_PORTSPEED_20GBIT 0x80
#define FC_PORTSPEED_40GBIT 0x100
#define FC_PORTSPEED_50GBIT 0x200
#define FC_PORTSPEED_100GBIT 0x400
#define FC_PORTSPEED_25GBIT 0x800
#define FC_PORTSPEED_64GBIT 0x1000
#define FC_PORTSPEED_128GBIT 0x2000
#define FC_PORTSPEED_256GBIT 0x4000
#define FC_PORTSPEED_NOT_NEGOTIATED (1 << 15) /* Speed not established */
/*
* fc_tgtid_binding_type: If you alter this, you also need to alter
* scsi_transport_fc.c (for the ascii descriptions).
*/
enum fc_tgtid_binding_type {
FC_TGTID_BIND_NONE,
FC_TGTID_BIND_BY_WWPN,
FC_TGTID_BIND_BY_WWNN,
FC_TGTID_BIND_BY_ID,
};
/*
* FC Port Roles
* Note: values are not enumerated, as they can be "or'd" together
* for reporting (e.g. report roles). If you alter this list,
* you also need to alter scsi_transport_fc.c (for the ascii descriptions).
*/
#define FC_PORT_ROLE_UNKNOWN 0x00
#define FC_PORT_ROLE_FCP_TARGET 0x01
#define FC_PORT_ROLE_FCP_INITIATOR 0x02
#define FC_PORT_ROLE_IP_PORT 0x04
#define FC_PORT_ROLE_FCP_DUMMY_INITIATOR 0x08
#define FC_PORT_ROLE_NVME_INITIATOR 0x10
#define FC_PORT_ROLE_NVME_TARGET 0x20
#define FC_PORT_ROLE_NVME_DISCOVERY 0x40
/* The following are for compatibility */
#define FC_RPORT_ROLE_UNKNOWN FC_PORT_ROLE_UNKNOWN
#define FC_RPORT_ROLE_FCP_TARGET FC_PORT_ROLE_FCP_TARGET
#define FC_RPORT_ROLE_FCP_INITIATOR FC_PORT_ROLE_FCP_INITIATOR
#define FC_RPORT_ROLE_IP_PORT FC_PORT_ROLE_IP_PORT
/* Macro for use in defining Virtual Port attributes */
#define FC_VPORT_ATTR(_name,_mode,_show,_store) \
struct device_attribute dev_attr_vport_##_name = \
__ATTR(_name,_mode,_show,_store)
/*
* fc_vport_identifiers: This set of data contains all elements
* to uniquely identify and instantiate a FC virtual port.
*
* Notes:
* symbolic_name: The driver is to append the symbolic_name string data
* to the symbolic_node_name data that it generates by default.
* the resulting combination should then be registered with the switch.
* It is expected that things like Xen may stuff a VM title into
* this field.
*/
#define FC_VPORT_SYMBOLIC_NAMELEN 64
struct fc_vport_identifiers {
u64 node_name;
u64 port_name;
u32 roles;
bool disable;
enum fc_port_type vport_type; /* only FC_PORTTYPE_NPIV allowed */
char symbolic_name[FC_VPORT_SYMBOLIC_NAMELEN];
};
/*
* FC Virtual Port Attributes
*
* This structure exists for each FC port is a virtual FC port. Virtual
* ports share the physical link with the Physical port. Each virtual
* ports has a unique presence on the SAN, and may be instantiated via
* NPIV, Virtual Fabrics, or via additional ALPAs. As the vport is a
* unique presence, each vport has it's own view of the fabric,
* authentication privilege, and priorities.
*
* A virtual port may support 1 or more FC4 roles. Typically it is a
* FCP Initiator. It could be a FCP Target, or exist sole for an IP over FC
* roles. FC port attributes for the vport will be reported on any
* fc_host class object allocated for an FCP Initiator.
*
* --
*
* Fixed attributes are not expected to change. The driver is
* expected to set these values after receiving the fc_vport structure
* via the vport_create() call from the transport.
* The transport fully manages all get functions w/o driver interaction.
*
* Dynamic attributes are expected to change. The driver participates
* in all get/set operations via functions provided by the driver.
*
* Private attributes are transport-managed values. They are fully
* managed by the transport w/o driver interaction.
*/
struct fc_vport {
/* Fixed Attributes */
/* Dynamic Attributes */
/* Private (Transport-managed) Attributes */
enum fc_vport_state vport_state;
enum fc_vport_state vport_last_state;
u64 node_name;
u64 port_name;
u32 roles;
u32 vport_id; /* Admin Identifier for the vport */
enum fc_port_type vport_type;
char symbolic_name[FC_VPORT_SYMBOLIC_NAMELEN];
/* exported data */
void *dd_data; /* Used for driver-specific storage */
/* internal data */
struct Scsi_Host *shost; /* Physical Port Parent */
unsigned int channel;
u32 number;
u8 flags;
struct list_head peers;
struct device dev;
struct work_struct vport_delete_work;
} __attribute__((aligned(sizeof(unsigned long))));
/* bit field values for struct fc_vport "flags" field: */
#define FC_VPORT_CREATING 0x01
#define FC_VPORT_DELETING 0x02
#define FC_VPORT_DELETED 0x04
#define FC_VPORT_DEL 0x06 /* Any DELETE state */
#define dev_to_vport(d) \
container_of(d, struct fc_vport, dev)
#define transport_class_to_vport(dev) \
dev_to_vport(dev->parent)
#define vport_to_shost(v) \
(v->shost)
#define vport_to_shost_channel(v) \
(v->channel)
#define vport_to_parent(v) \
(v->dev.parent)
/* Error return codes for vport_create() callback */
#define VPCERR_UNSUPPORTED -ENOSYS /* no driver/adapter
support */
#define VPCERR_BAD_WWN -ENOTUNIQ /* driver validation
of WWNs failed */
#define VPCERR_NO_FABRIC_SUPP -EOPNOTSUPP /* Fabric connection
is loop or the
Fabric Port does
not support NPIV */
/*
* fc_rport_identifiers: This set of data contains all elements
* to uniquely identify a remote FC port. The driver uses this data
* to report the existence of a remote FC port in the topology. Internally,
* the transport uses this data for attributes and to manage consistent
* target id bindings.
*/
struct fc_rport_identifiers {
u64 node_name;
u64 port_name;
u32 port_id;
u32 roles;
};
/*
* Fabric Performance Impact Notification Statistics
*/
struct fc_fpin_stats {
/* Delivery */
u64 dn;
u64 dn_unknown;
u64 dn_timeout;
u64 dn_unable_to_route;
u64 dn_device_specific;
/* Link Integrity */
u64 li;
u64 li_failure_unknown;
u64 li_link_failure_count;
u64 li_loss_of_sync_count;
u64 li_loss_of_signals_count;
u64 li_prim_seq_err_count;
u64 li_invalid_tx_word_count;
u64 li_invalid_crc_count;
u64 li_device_specific;
/* Congestion/Peer Congestion */
u64 cn;
u64 cn_clear;
u64 cn_lost_credit;
u64 cn_credit_stall;
u64 cn_oversubscription;
u64 cn_device_specific;
};
/* Macro for use in defining Remote Port attributes */
#define FC_RPORT_ATTR(_name,_mode,_show,_store) \
struct device_attribute dev_attr_rport_##_name = \
__ATTR(_name,_mode,_show,_store)
/*
* FC Remote Port Attributes
*
* This structure exists for each remote FC port that a LLDD notifies
* the subsystem of. A remote FC port may or may not be a SCSI Target,
* also be a SCSI initiator, IP endpoint, etc. As such, the remote
* port is considered a separate entity, independent of "role" (such
* as scsi target).
*
* --
*
* Attributes are based on HBAAPI V2.0 definitions. Only those
* attributes that are determinable by the local port (aka Host)
* are contained.
*
* Fixed attributes are not expected to change. The driver is
* expected to set these values after successfully calling
* fc_remote_port_add(). The transport fully manages all get functions
* w/o driver interaction.
*
* Dynamic attributes are expected to change. The driver participates
* in all get/set operations via functions provided by the driver.
*
* Private attributes are transport-managed values. They are fully
* managed by the transport w/o driver interaction.
*/
struct fc_rport { /* aka fc_starget_attrs */
/* Fixed Attributes */
u32 maxframe_size;
u32 supported_classes;
/* Dynamic Attributes */
u32 dev_loss_tmo; /* Remote Port loss timeout in seconds. */
struct fc_fpin_stats fpin_stats;
/* Private (Transport-managed) Attributes */
u64 node_name;
u64 port_name;
u32 port_id;
u32 roles;
enum fc_port_state port_state; /* Will only be ONLINE or UNKNOWN */
u32 scsi_target_id;
u32 fast_io_fail_tmo;
/* exported data */
void *dd_data; /* Used for driver-specific storage */
/* internal data */
unsigned int channel;
u32 number;
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
u8 flags;
struct list_head peers;
struct device dev;
struct delayed_work dev_loss_work;
struct work_struct scan_work;
struct delayed_work fail_io_work;
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
struct work_struct stgt_delete_work;
struct work_struct rport_delete_work;
[SCSI] FC Pass Thru support Attached is the ELS/CT pass-thru patch for the FC Transport. The patch creates a generic framework that lays on top of bsg and the SGIO v4 ioctl in order to pass transaction requests to LLDD's. The interface supports the following operations: On an fc_host basis: Request login to the specified N_Port_ID, creating an fc_rport. Request logout of the specified N_Port_ID, deleting an fc_rport Send ELS request to specified N_Port_ID w/o requiring a login, and wait for ELS response. Send CT request to specified N_Port_ID and wait for CT response. Login is required, but LLDD is allowed to manage login and decide whether it stays in place after the request is satisfied. Vendor-Unique request. Allows a LLDD-specific request to be passed to the LLDD, and the passing of a response back to the application. On an fc_rport basis: Send ELS request to nport and wait for ELS response. Send CT request to nport and wait for CT response. The patch also exports several headers from include/scsi such that they can be available to user-space applications: include/scsi/scsi.h include/scsi/scsi_netlink.h include/scsi/scsi_netlink_fc.h include/scsi/scsi_bsg_fc.h For further information, refer to the last RFC: http://marc.info/?l=linux-scsi&m=123436574018579&w=2 Note: Documentation is still spotty and will be added later. [bharrosh@panasas.com: update for new block API] Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-03-26 20:33:19 +03:00
struct request_queue *rqst_q; /* bsg support */
} __attribute__((aligned(sizeof(unsigned long))));
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
/* bit field values for struct fc_rport "flags" field: */
#define FC_RPORT_DEVLOSS_PENDING 0x01
#define FC_RPORT_SCAN_PENDING 0x02
#define FC_RPORT_FAST_FAIL_TIMEDOUT 0x04
#define FC_RPORT_DEVLOSS_CALLBK_DONE 0x08
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
#define dev_to_rport(d) \
container_of(d, struct fc_rport, dev)
#define transport_class_to_rport(dev) \
dev_to_rport(dev->parent)
#define rport_to_shost(r) \
dev_to_shost(r->dev.parent)
/*
* FC SCSI Target Attributes
*
* The SCSI Target is considered an extension of a remote port (as
* a remote port can be more than a SCSI Target). Within the scsi
* subsystem, we leave the Target as a separate entity. Doing so
* provides backward compatibility with prior FC transport api's,
* and lets remote ports be handled entirely within the FC transport
* and independently from the scsi subsystem. The drawback is that
* some data will be duplicated.
*/
struct fc_starget_attrs { /* aka fc_target_attrs */
/* Dynamic Attributes */
u64 node_name;
u64 port_name;
u32 port_id;
};
#define fc_starget_node_name(x) \
(((struct fc_starget_attrs *)&(x)->starget_data)->node_name)
#define fc_starget_port_name(x) \
(((struct fc_starget_attrs *)&(x)->starget_data)->port_name)
#define fc_starget_port_id(x) \
(((struct fc_starget_attrs *)&(x)->starget_data)->port_id)
#define starget_to_rport(s) \
scsi_is_fc_rport(s->dev.parent) ? dev_to_rport(s->dev.parent) : NULL
/*
* FC Local Port (Host) Statistics
*/
/* FC Statistics - Following FC HBAAPI v2.0 guidelines */
struct fc_host_statistics {
/* port statistics */
u64 seconds_since_last_reset;
u64 tx_frames;
u64 tx_words;
u64 rx_frames;
u64 rx_words;
u64 lip_count;
u64 nos_count;
u64 error_frames;
u64 dumped_frames;
u64 link_failure_count;
u64 loss_of_sync_count;
u64 loss_of_signal_count;
u64 prim_seq_protocol_err_count;
u64 invalid_tx_word_count;
u64 invalid_crc_count;
/* fc4 statistics (only FCP supported currently) */
u64 fcp_input_requests;
u64 fcp_output_requests;
u64 fcp_control_requests;
u64 fcp_input_megabytes;
u64 fcp_output_megabytes;
u64 fcp_packet_alloc_failures; /* fcp packet allocation failures */
u64 fcp_packet_aborts; /* fcp packet aborted */
u64 fcp_frame_alloc_failures; /* fcp frame allocation failures */
/* fc exches statistics */
u64 fc_no_free_exch; /* no free exch memory */
u64 fc_no_free_exch_xid; /* no free exch id */
u64 fc_xid_not_found; /* exch not found for a response */
u64 fc_xid_busy; /* exch exist for new a request */
u64 fc_seq_not_found; /* seq is not found for exchange */
u64 fc_non_bls_resp; /* a non BLS response frame with
a sequence responder in new exch */
/* Host Congestion Signals */
u64 cn_sig_warn;
u64 cn_sig_alarm;
};
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
/*
* FC Event Codes - Polled and Async, following FC HBAAPI v2.0 guidelines
*/
/*
* fc_host_event_code: If you alter this, you also need to alter
* scsi_transport_fc.c (for the ascii descriptions).
*/
enum fc_host_event_code {
FCH_EVT_LIP = 0x1,
FCH_EVT_LINKUP = 0x2,
FCH_EVT_LINKDOWN = 0x3,
FCH_EVT_LIPRESET = 0x4,
FCH_EVT_RSCN = 0x5,
FCH_EVT_ADAPTER_CHANGE = 0x103,
FCH_EVT_PORT_UNKNOWN = 0x200,
FCH_EVT_PORT_OFFLINE = 0x201,
FCH_EVT_PORT_ONLINE = 0x202,
FCH_EVT_PORT_FABRIC = 0x204,
FCH_EVT_LINK_UNKNOWN = 0x500,
FCH_EVT_LINK_FPIN = 0x501,
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
FCH_EVT_VENDOR_UNIQUE = 0xffff,
};
/*
* FC Local Port (Host) Attributes
*
* Attributes are based on HBAAPI V2.0 definitions.
* Note: OSDeviceName is determined by user-space library
*
* Fixed attributes are not expected to change. The driver is
* expected to set these values after successfully calling scsi_add_host().
* The transport fully manages all get functions w/o driver interaction.
*
* Dynamic attributes are expected to change. The driver participates
* in all get/set operations via functions provided by the driver.
*
* Private attributes are transport-managed values. They are fully
* managed by the transport w/o driver interaction.
*/
#define FC_VENDOR_IDENTIFIER 8
#define FC_FC4_LIST_SIZE 32
#define FC_SYMBOLIC_NAME_SIZE 256
#define FC_VERSION_STRING_SIZE 64
#define FC_SERIAL_NUMBER_SIZE 64
struct fc_host_attrs {
/* Fixed Attributes */
u64 node_name;
u64 port_name;
u64 permanent_port_name;
u32 supported_classes;
u8 supported_fc4s[FC_FC4_LIST_SIZE];
u32 supported_speeds;
u32 maxframe_size;
u16 max_npiv_vports;
u32 max_ct_payload;
u32 num_ports;
u32 num_discovered_ports;
u32 bootbios_state;
char serial_number[FC_SERIAL_NUMBER_SIZE];
char manufacturer[FC_SERIAL_NUMBER_SIZE];
char model[FC_SYMBOLIC_NAME_SIZE];
char model_description[FC_SYMBOLIC_NAME_SIZE];
char hardware_version[FC_VERSION_STRING_SIZE];
char driver_version[FC_VERSION_STRING_SIZE];
char firmware_version[FC_VERSION_STRING_SIZE];
char optionrom_version[FC_VERSION_STRING_SIZE];
char vendor_identifier[FC_VENDOR_IDENTIFIER];
char bootbios_version[FC_SYMBOLIC_NAME_SIZE];
/* Dynamic Attributes */
u32 port_id;
enum fc_port_type port_type;
enum fc_port_state port_state;
u8 active_fc4s[FC_FC4_LIST_SIZE];
u32 speed;
u64 fabric_name;
char symbolic_name[FC_SYMBOLIC_NAME_SIZE];
char system_hostname[FC_SYMBOLIC_NAME_SIZE];
u32 dev_loss_tmo;
struct fc_fpin_stats fpin_stats;
/* Private (Transport-managed) Attributes */
enum fc_tgtid_binding_type tgtid_bind_type;
/* internal data */
struct list_head rports;
struct list_head rport_bindings;
struct list_head vports;
u32 next_rport_number;
u32 next_target_id;
u32 next_vport_number;
u16 npiv_vports_inuse;
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
/* work queues for rport state manipulation */
char work_q_name[20];
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
struct workqueue_struct *work_q;
char devloss_work_q_name[20];
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
struct workqueue_struct *devloss_work_q;
[SCSI] FC Pass Thru support Attached is the ELS/CT pass-thru patch for the FC Transport. The patch creates a generic framework that lays on top of bsg and the SGIO v4 ioctl in order to pass transaction requests to LLDD's. The interface supports the following operations: On an fc_host basis: Request login to the specified N_Port_ID, creating an fc_rport. Request logout of the specified N_Port_ID, deleting an fc_rport Send ELS request to specified N_Port_ID w/o requiring a login, and wait for ELS response. Send CT request to specified N_Port_ID and wait for CT response. Login is required, but LLDD is allowed to manage login and decide whether it stays in place after the request is satisfied. Vendor-Unique request. Allows a LLDD-specific request to be passed to the LLDD, and the passing of a response back to the application. On an fc_rport basis: Send ELS request to nport and wait for ELS response. Send CT request to nport and wait for CT response. The patch also exports several headers from include/scsi such that they can be available to user-space applications: include/scsi/scsi.h include/scsi/scsi_netlink.h include/scsi/scsi_netlink_fc.h include/scsi/scsi_bsg_fc.h For further information, refer to the last RFC: http://marc.info/?l=linux-scsi&m=123436574018579&w=2 Note: Documentation is still spotty and will be added later. [bharrosh@panasas.com: update for new block API] Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-03-26 20:33:19 +03:00
/* bsg support */
struct request_queue *rqst_q;
/* FDMI support version*/
u8 fdmi_version;
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
};
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
#define shost_to_fc_host(x) \
((struct fc_host_attrs *)(x)->shost_data)
#define fc_host_node_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->node_name)
#define fc_host_port_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->port_name)
#define fc_host_permanent_port_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->permanent_port_name)
#define fc_host_supported_classes(x) \
(((struct fc_host_attrs *)(x)->shost_data)->supported_classes)
#define fc_host_supported_fc4s(x) \
(((struct fc_host_attrs *)(x)->shost_data)->supported_fc4s)
#define fc_host_supported_speeds(x) \
(((struct fc_host_attrs *)(x)->shost_data)->supported_speeds)
#define fc_host_maxframe_size(x) \
(((struct fc_host_attrs *)(x)->shost_data)->maxframe_size)
#define fc_host_max_npiv_vports(x) \
(((struct fc_host_attrs *)(x)->shost_data)->max_npiv_vports)
#define fc_host_serial_number(x) \
(((struct fc_host_attrs *)(x)->shost_data)->serial_number)
#define fc_host_manufacturer(x) \
(((struct fc_host_attrs *)(x)->shost_data)->manufacturer)
#define fc_host_model(x) \
(((struct fc_host_attrs *)(x)->shost_data)->model)
#define fc_host_model_description(x) \
(((struct fc_host_attrs *)(x)->shost_data)->model_description)
#define fc_host_hardware_version(x) \
(((struct fc_host_attrs *)(x)->shost_data)->hardware_version)
#define fc_host_driver_version(x) \
(((struct fc_host_attrs *)(x)->shost_data)->driver_version)
#define fc_host_firmware_version(x) \
(((struct fc_host_attrs *)(x)->shost_data)->firmware_version)
#define fc_host_optionrom_version(x) \
(((struct fc_host_attrs *)(x)->shost_data)->optionrom_version)
#define fc_host_port_id(x) \
(((struct fc_host_attrs *)(x)->shost_data)->port_id)
#define fc_host_port_type(x) \
(((struct fc_host_attrs *)(x)->shost_data)->port_type)
#define fc_host_port_state(x) \
(((struct fc_host_attrs *)(x)->shost_data)->port_state)
#define fc_host_active_fc4s(x) \
(((struct fc_host_attrs *)(x)->shost_data)->active_fc4s)
#define fc_host_speed(x) \
(((struct fc_host_attrs *)(x)->shost_data)->speed)
#define fc_host_fabric_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->fabric_name)
#define fc_host_symbolic_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->symbolic_name)
#define fc_host_system_hostname(x) \
(((struct fc_host_attrs *)(x)->shost_data)->system_hostname)
#define fc_host_tgtid_bind_type(x) \
(((struct fc_host_attrs *)(x)->shost_data)->tgtid_bind_type)
#define fc_host_rports(x) \
(((struct fc_host_attrs *)(x)->shost_data)->rports)
#define fc_host_rport_bindings(x) \
(((struct fc_host_attrs *)(x)->shost_data)->rport_bindings)
#define fc_host_vports(x) \
(((struct fc_host_attrs *)(x)->shost_data)->vports)
#define fc_host_next_rport_number(x) \
(((struct fc_host_attrs *)(x)->shost_data)->next_rport_number)
#define fc_host_next_target_id(x) \
(((struct fc_host_attrs *)(x)->shost_data)->next_target_id)
#define fc_host_next_vport_number(x) \
(((struct fc_host_attrs *)(x)->shost_data)->next_vport_number)
#define fc_host_npiv_vports_inuse(x) \
(((struct fc_host_attrs *)(x)->shost_data)->npiv_vports_inuse)
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
#define fc_host_work_q_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->work_q_name)
#define fc_host_work_q(x) \
(((struct fc_host_attrs *)(x)->shost_data)->work_q)
#define fc_host_devloss_work_q_name(x) \
(((struct fc_host_attrs *)(x)->shost_data)->devloss_work_q_name)
#define fc_host_devloss_work_q(x) \
(((struct fc_host_attrs *)(x)->shost_data)->devloss_work_q)
#define fc_host_dev_loss_tmo(x) \
(((struct fc_host_attrs *)(x)->shost_data)->dev_loss_tmo)
#define fc_host_max_ct_payload(x) \
(((struct fc_host_attrs *)(x)->shost_data)->max_ct_payload)
#define fc_host_vendor_identifier(x) \
(((struct fc_host_attrs *)(x)->shost_data)->vendor_identifier)
#define fc_host_num_discovered_ports(x) \
(((struct fc_host_attrs *)(x)->shost_data)->num_discovered_ports)
#define fc_host_num_ports(x) \
(((struct fc_host_attrs *)(x)->shost_data)->num_ports)
#define fc_host_bootbios_version(x) \
(((struct fc_host_attrs *)(x)->shost_data)->bootbios_version)
#define fc_host_bootbios_state(x) \
(((struct fc_host_attrs *)(x)->shost_data)->bootbios_state)
/* The functions by which the transport class and the driver communicate */
struct fc_function_template {
void (*get_rport_dev_loss_tmo)(struct fc_rport *);
void (*set_rport_dev_loss_tmo)(struct fc_rport *, u32);
void (*get_starget_node_name)(struct scsi_target *);
void (*get_starget_port_name)(struct scsi_target *);
void (*get_starget_port_id)(struct scsi_target *);
void (*get_host_port_id)(struct Scsi_Host *);
void (*get_host_port_type)(struct Scsi_Host *);
void (*get_host_port_state)(struct Scsi_Host *);
void (*get_host_active_fc4s)(struct Scsi_Host *);
void (*get_host_speed)(struct Scsi_Host *);
void (*get_host_fabric_name)(struct Scsi_Host *);
void (*get_host_symbolic_name)(struct Scsi_Host *);
void (*set_host_system_hostname)(struct Scsi_Host *);
struct fc_host_statistics * (*get_fc_host_stats)(struct Scsi_Host *);
void (*reset_fc_host_stats)(struct Scsi_Host *);
int (*issue_fc_host_lip)(struct Scsi_Host *);
void (*dev_loss_tmo_callbk)(struct fc_rport *);
void (*terminate_rport_io)(struct fc_rport *);
void (*set_vport_symbolic_name)(struct fc_vport *);
int (*vport_create)(struct fc_vport *, bool);
int (*vport_disable)(struct fc_vport *, bool);
int (*vport_delete)(struct fc_vport *);
[SCSI] FC Pass Thru support Attached is the ELS/CT pass-thru patch for the FC Transport. The patch creates a generic framework that lays on top of bsg and the SGIO v4 ioctl in order to pass transaction requests to LLDD's. The interface supports the following operations: On an fc_host basis: Request login to the specified N_Port_ID, creating an fc_rport. Request logout of the specified N_Port_ID, deleting an fc_rport Send ELS request to specified N_Port_ID w/o requiring a login, and wait for ELS response. Send CT request to specified N_Port_ID and wait for CT response. Login is required, but LLDD is allowed to manage login and decide whether it stays in place after the request is satisfied. Vendor-Unique request. Allows a LLDD-specific request to be passed to the LLDD, and the passing of a response back to the application. On an fc_rport basis: Send ELS request to nport and wait for ELS response. Send CT request to nport and wait for CT response. The patch also exports several headers from include/scsi such that they can be available to user-space applications: include/scsi/scsi.h include/scsi/scsi_netlink.h include/scsi/scsi_netlink_fc.h include/scsi/scsi_bsg_fc.h For further information, refer to the last RFC: http://marc.info/?l=linux-scsi&m=123436574018579&w=2 Note: Documentation is still spotty and will be added later. [bharrosh@panasas.com: update for new block API] Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-03-26 20:33:19 +03:00
/* bsg support */
int (*bsg_request)(struct bsg_job *);
int (*bsg_timeout)(struct bsg_job *);
[SCSI] FC Pass Thru support Attached is the ELS/CT pass-thru patch for the FC Transport. The patch creates a generic framework that lays on top of bsg and the SGIO v4 ioctl in order to pass transaction requests to LLDD's. The interface supports the following operations: On an fc_host basis: Request login to the specified N_Port_ID, creating an fc_rport. Request logout of the specified N_Port_ID, deleting an fc_rport Send ELS request to specified N_Port_ID w/o requiring a login, and wait for ELS response. Send CT request to specified N_Port_ID and wait for CT response. Login is required, but LLDD is allowed to manage login and decide whether it stays in place after the request is satisfied. Vendor-Unique request. Allows a LLDD-specific request to be passed to the LLDD, and the passing of a response back to the application. On an fc_rport basis: Send ELS request to nport and wait for ELS response. Send CT request to nport and wait for CT response. The patch also exports several headers from include/scsi such that they can be available to user-space applications: include/scsi/scsi.h include/scsi/scsi_netlink.h include/scsi/scsi_netlink_fc.h include/scsi/scsi_bsg_fc.h For further information, refer to the last RFC: http://marc.info/?l=linux-scsi&m=123436574018579&w=2 Note: Documentation is still spotty and will be added later. [bharrosh@panasas.com: update for new block API] Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-03-26 20:33:19 +03:00
/* allocation lengths for host-specific data */
u32 dd_fcrport_size;
u32 dd_fcvport_size;
[SCSI] FC Pass Thru support Attached is the ELS/CT pass-thru patch for the FC Transport. The patch creates a generic framework that lays on top of bsg and the SGIO v4 ioctl in order to pass transaction requests to LLDD's. The interface supports the following operations: On an fc_host basis: Request login to the specified N_Port_ID, creating an fc_rport. Request logout of the specified N_Port_ID, deleting an fc_rport Send ELS request to specified N_Port_ID w/o requiring a login, and wait for ELS response. Send CT request to specified N_Port_ID and wait for CT response. Login is required, but LLDD is allowed to manage login and decide whether it stays in place after the request is satisfied. Vendor-Unique request. Allows a LLDD-specific request to be passed to the LLDD, and the passing of a response back to the application. On an fc_rport basis: Send ELS request to nport and wait for ELS response. Send CT request to nport and wait for CT response. The patch also exports several headers from include/scsi such that they can be available to user-space applications: include/scsi/scsi.h include/scsi/scsi_netlink.h include/scsi/scsi_netlink_fc.h include/scsi/scsi_bsg_fc.h For further information, refer to the last RFC: http://marc.info/?l=linux-scsi&m=123436574018579&w=2 Note: Documentation is still spotty and will be added later. [bharrosh@panasas.com: update for new block API] Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-03-26 20:33:19 +03:00
u32 dd_bsg_size;
/*
* The driver sets these to tell the transport class it
* wants the attributes displayed in sysfs. If the show_ flag
* is not set, the attribute will be private to the transport
* class
*/
/* remote port fixed attributes */
unsigned long show_rport_maxframe_size:1;
unsigned long show_rport_supported_classes:1;
unsigned long show_rport_dev_loss_tmo:1;
/*
* target dynamic attributes
* These should all be "1" if the driver uses the remote port
* add/delete functions (so attributes reflect rport values).
*/
unsigned long show_starget_node_name:1;
unsigned long show_starget_port_name:1;
unsigned long show_starget_port_id:1;
/* host fixed attributes */
unsigned long show_host_node_name:1;
unsigned long show_host_port_name:1;
unsigned long show_host_permanent_port_name:1;
unsigned long show_host_supported_classes:1;
unsigned long show_host_supported_fc4s:1;
unsigned long show_host_supported_speeds:1;
unsigned long show_host_maxframe_size:1;
unsigned long show_host_serial_number:1;
unsigned long show_host_manufacturer:1;
unsigned long show_host_model:1;
unsigned long show_host_model_description:1;
unsigned long show_host_hardware_version:1;
unsigned long show_host_driver_version:1;
unsigned long show_host_firmware_version:1;
unsigned long show_host_optionrom_version:1;
/* host dynamic attributes */
unsigned long show_host_port_id:1;
unsigned long show_host_port_type:1;
unsigned long show_host_port_state:1;
unsigned long show_host_active_fc4s:1;
unsigned long show_host_speed:1;
unsigned long show_host_fabric_name:1;
unsigned long show_host_symbolic_name:1;
unsigned long show_host_system_hostname:1;
unsigned long disable_target_scan:1;
};
[SCSI] update fc_transport for removal of block/unblock functions We recently went back to implement a board reset. When we perform the reset, we wanted to tear down the internal data structures and rebuild them. Unfortunately, when it came to the rport structure, things were odd. If we deleted them, the scsi targets and sdevs would be torn down. Not a good thing for a temporary reset. We could block the rports, but we either maintain the internal structures to keep the rport reference (perhaps even replicating what's in the transport), or we have to fatten the fc transport with new search routines to find the rport (and deal with a case of a dangling rport that the driver forgets). It dawned on me that we had actually reached this state incorrectly. When the fc transport first started, we did the block/unblock first, then added the rport interface. The purpose of block/unblock is to hide the temporary disappearance of the rport (e.g. being deleted, then readded). Why are we making the driver do the block/unblock ? We should be making the transport have only an rport add/delete, and the let the transport handle the block/unblock. So... This patch removes the existing fc_remote_port_block/unblock functions. It moves the block/unblock functionality into the fc_remote_port_add/delete functions. Updates for the lpfc driver are included. Qlogic driver updates are also enclosed, thanks to the contributions of Andrew Vasquez. [Note: the qla2xxx changes are relative to the scsi-misc-2.6 tree as of this morning - which does not include the recent patches sent by Andrew]. The zfcp driver does not use the block/unblock functions. One last comment: The resulting behavior feels very clean. The LLDD is concerned only with add/delete, which corresponds to the physical disappearance. However, the fact that the scsi target and sdevs are not immediately torn down after the LLDD calls delete causes an interesting scenario... the midlayer can call the xxx_slave_alloc and xxx_queuecommand functions with a sdev that is at the location the rport used to be. The driver must validate the device exists when it first enters these functions. In thinking about it, this has always been the case for the LLDD and these routines. The existing drivers already check for existence. However, this highlights that simple validation via data structure dereferencing needs to be watched. To deal with this, a new transport function, fc_remote_port_chkready() was created that LLDDs should call when they first enter these two routines. It validates the rport state, and returns a scsi result which could be returned. In addition to solving the above, it also creates consistent behavior from the LLDD's when the block and deletes are occuring. Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-10-18 20:03:35 +04:00
/**
* fc_remote_port_chkready - called to validate the remote port state
* prior to initiating io to the port.
*
* Returns a scsi result code that can be returned by the LLDD.
*
* @rport: remote port to be checked
**/
static inline int
fc_remote_port_chkready(struct fc_rport *rport)
{
int result;
switch (rport->port_state) {
case FC_PORTSTATE_ONLINE:
case FC_PORTSTATE_MARGINAL:
if (rport->roles & FC_PORT_ROLE_FCP_TARGET)
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
result = 0;
else if (rport->flags & FC_RPORT_DEVLOSS_PENDING)
result = DID_IMM_RETRY << 16;
[SCSI] FC transport: fixes for workq deadlocks As previously reported via Michael Reed, the FC transport took a hit in 2.6.15 (perhaps a little earlier) when we solved a recursion error. There are 2 deadlocks occurring: - With scan and the delete items sharing the same workq, flushing the workq for the delete code was getting it stalled behind a very long running scan code path. - There's a deadlock where scsi_remove_target() has to sit behind scsi_scan_target() due to contention over the scan_lock(). This patch resolves the 1st deadlock and significantly reduces the odds of the second. So far, we have only replicated the 2nd deadlock on a highly-parallel SMP system. More on the 2nd deadlock in a following email. This patch reworks the transport to: - Only use the scsi host workq for scanning - Use 2 other workq's internally. One for deletions, the other for scheduled deletions. Originally, we tried this with a single workq, but the occassional flushes of the scheduled queues was hitting the second deadlock with a slightly higher frequency. In the future, we'll look at the LLDD's and the transport to see if we can get rid of this extra overhead. - When moving to the other workq's we tightened up some object states and some lock handling. - Properly syncs adds/deletes - minor code cleanups - directly reference fc_host_attrs, rather than through attribute macros - flush the right workq on delayed work cancel failures. Large kudos to Michael Reed who has been working this issue for the last month. Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-04-10 18:14:05 +04:00
else
result = DID_NO_CONNECT << 16;
[SCSI] update fc_transport for removal of block/unblock functions We recently went back to implement a board reset. When we perform the reset, we wanted to tear down the internal data structures and rebuild them. Unfortunately, when it came to the rport structure, things were odd. If we deleted them, the scsi targets and sdevs would be torn down. Not a good thing for a temporary reset. We could block the rports, but we either maintain the internal structures to keep the rport reference (perhaps even replicating what's in the transport), or we have to fatten the fc transport with new search routines to find the rport (and deal with a case of a dangling rport that the driver forgets). It dawned on me that we had actually reached this state incorrectly. When the fc transport first started, we did the block/unblock first, then added the rport interface. The purpose of block/unblock is to hide the temporary disappearance of the rport (e.g. being deleted, then readded). Why are we making the driver do the block/unblock ? We should be making the transport have only an rport add/delete, and the let the transport handle the block/unblock. So... This patch removes the existing fc_remote_port_block/unblock functions. It moves the block/unblock functionality into the fc_remote_port_add/delete functions. Updates for the lpfc driver are included. Qlogic driver updates are also enclosed, thanks to the contributions of Andrew Vasquez. [Note: the qla2xxx changes are relative to the scsi-misc-2.6 tree as of this morning - which does not include the recent patches sent by Andrew]. The zfcp driver does not use the block/unblock functions. One last comment: The resulting behavior feels very clean. The LLDD is concerned only with add/delete, which corresponds to the physical disappearance. However, the fact that the scsi target and sdevs are not immediately torn down after the LLDD calls delete causes an interesting scenario... the midlayer can call the xxx_slave_alloc and xxx_queuecommand functions with a sdev that is at the location the rport used to be. The driver must validate the device exists when it first enters these functions. In thinking about it, this has always been the case for the LLDD and these routines. The existing drivers already check for existence. However, this highlights that simple validation via data structure dereferencing needs to be watched. To deal with this, a new transport function, fc_remote_port_chkready() was created that LLDDs should call when they first enter these two routines. It validates the rport state, and returns a scsi result which could be returned. In addition to solving the above, it also creates consistent behavior from the LLDD's when the block and deletes are occuring. Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-10-18 20:03:35 +04:00
break;
case FC_PORTSTATE_BLOCKED:
[SCSI] fc class: unblock target after calling terminate callback (take 2) When we block a rport and the driver implements the terminate callback we will fail IO that was running quickly. However IO that was in the scsi_device/block queue sits there until the dev_loss_tmo fires, and this can make it look like IO is lost because new IO will get executed but that IO stuck in the blocked queue sits there for some time longer. With this patch when the fast io fail tmo fires, we will fail the blocked IO and any new IO. This patch also allows all drivers to partially support the fast io fail tmo. If the terminate io callback is not implemented, we will still fail blocked IO and any new IO, so multipath can handle that. This patch also allows the fc and iscsi classes to implement the same behavior. The timers are just unfornately named differently. This patch also fixes the problem where drivers were unblocking the target in their terminate callback, which was needed for rport removal, but for fast io fail timeout it would cause IO to bounce arround the scsi/block layer and the LLD queuecommand. And it for drivers that could have IO stuck but did not have a terminate callback the unblock calls in the class will fix them. v2. - fix up bit setting style to meet JamesS's pref. - Broke out new host byte error changes to make it easier to read. - added JamesS's ack from list. v1 - initial patch Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Acked-by: James Smart <James.Smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2008-08-20 03:45:23 +04:00
if (rport->flags & FC_RPORT_FAST_FAIL_TIMEDOUT)
result = DID_TRANSPORT_FAILFAST << 16;
[SCSI] fc class: unblock target after calling terminate callback (take 2) When we block a rport and the driver implements the terminate callback we will fail IO that was running quickly. However IO that was in the scsi_device/block queue sits there until the dev_loss_tmo fires, and this can make it look like IO is lost because new IO will get executed but that IO stuck in the blocked queue sits there for some time longer. With this patch when the fast io fail tmo fires, we will fail the blocked IO and any new IO. This patch also allows all drivers to partially support the fast io fail tmo. If the terminate io callback is not implemented, we will still fail blocked IO and any new IO, so multipath can handle that. This patch also allows the fc and iscsi classes to implement the same behavior. The timers are just unfornately named differently. This patch also fixes the problem where drivers were unblocking the target in their terminate callback, which was needed for rport removal, but for fast io fail timeout it would cause IO to bounce arround the scsi/block layer and the LLD queuecommand. And it for drivers that could have IO stuck but did not have a terminate callback the unblock calls in the class will fix them. v2. - fix up bit setting style to meet JamesS's pref. - Broke out new host byte error changes to make it easier to read. - added JamesS's ack from list. v1 - initial patch Signed-off-by: Mike Christie <michaelc@cs.wisc.edu> Acked-by: James Smart <James.Smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2008-08-20 03:45:23 +04:00
else
result = DID_IMM_RETRY << 16;
[SCSI] update fc_transport for removal of block/unblock functions We recently went back to implement a board reset. When we perform the reset, we wanted to tear down the internal data structures and rebuild them. Unfortunately, when it came to the rport structure, things were odd. If we deleted them, the scsi targets and sdevs would be torn down. Not a good thing for a temporary reset. We could block the rports, but we either maintain the internal structures to keep the rport reference (perhaps even replicating what's in the transport), or we have to fatten the fc transport with new search routines to find the rport (and deal with a case of a dangling rport that the driver forgets). It dawned on me that we had actually reached this state incorrectly. When the fc transport first started, we did the block/unblock first, then added the rport interface. The purpose of block/unblock is to hide the temporary disappearance of the rport (e.g. being deleted, then readded). Why are we making the driver do the block/unblock ? We should be making the transport have only an rport add/delete, and the let the transport handle the block/unblock. So... This patch removes the existing fc_remote_port_block/unblock functions. It moves the block/unblock functionality into the fc_remote_port_add/delete functions. Updates for the lpfc driver are included. Qlogic driver updates are also enclosed, thanks to the contributions of Andrew Vasquez. [Note: the qla2xxx changes are relative to the scsi-misc-2.6 tree as of this morning - which does not include the recent patches sent by Andrew]. The zfcp driver does not use the block/unblock functions. One last comment: The resulting behavior feels very clean. The LLDD is concerned only with add/delete, which corresponds to the physical disappearance. However, the fact that the scsi target and sdevs are not immediately torn down after the LLDD calls delete causes an interesting scenario... the midlayer can call the xxx_slave_alloc and xxx_queuecommand functions with a sdev that is at the location the rport used to be. The driver must validate the device exists when it first enters these functions. In thinking about it, this has always been the case for the LLDD and these routines. The existing drivers already check for existence. However, this highlights that simple validation via data structure dereferencing needs to be watched. To deal with this, a new transport function, fc_remote_port_chkready() was created that LLDDs should call when they first enter these two routines. It validates the rport state, and returns a scsi result which could be returned. In addition to solving the above, it also creates consistent behavior from the LLDD's when the block and deletes are occuring. Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-10-18 20:03:35 +04:00
break;
default:
result = DID_NO_CONNECT << 16;
break;
}
return result;
}
static inline u64 wwn_to_u64(const u8 *wwn)
{
return get_unaligned_be64(wwn);
}
static inline void u64_to_wwn(u64 inm, u8 *wwn)
{
put_unaligned_be64(inm, wwn);
}
[SCSI] update fc_transport for removal of block/unblock functions We recently went back to implement a board reset. When we perform the reset, we wanted to tear down the internal data structures and rebuild them. Unfortunately, when it came to the rport structure, things were odd. If we deleted them, the scsi targets and sdevs would be torn down. Not a good thing for a temporary reset. We could block the rports, but we either maintain the internal structures to keep the rport reference (perhaps even replicating what's in the transport), or we have to fatten the fc transport with new search routines to find the rport (and deal with a case of a dangling rport that the driver forgets). It dawned on me that we had actually reached this state incorrectly. When the fc transport first started, we did the block/unblock first, then added the rport interface. The purpose of block/unblock is to hide the temporary disappearance of the rport (e.g. being deleted, then readded). Why are we making the driver do the block/unblock ? We should be making the transport have only an rport add/delete, and the let the transport handle the block/unblock. So... This patch removes the existing fc_remote_port_block/unblock functions. It moves the block/unblock functionality into the fc_remote_port_add/delete functions. Updates for the lpfc driver are included. Qlogic driver updates are also enclosed, thanks to the contributions of Andrew Vasquez. [Note: the qla2xxx changes are relative to the scsi-misc-2.6 tree as of this morning - which does not include the recent patches sent by Andrew]. The zfcp driver does not use the block/unblock functions. One last comment: The resulting behavior feels very clean. The LLDD is concerned only with add/delete, which corresponds to the physical disappearance. However, the fact that the scsi target and sdevs are not immediately torn down after the LLDD calls delete causes an interesting scenario... the midlayer can call the xxx_slave_alloc and xxx_queuecommand functions with a sdev that is at the location the rport used to be. The driver must validate the device exists when it first enters these functions. In thinking about it, this has always been the case for the LLDD and these routines. The existing drivers already check for existence. However, this highlights that simple validation via data structure dereferencing needs to be watched. To deal with this, a new transport function, fc_remote_port_chkready() was created that LLDDs should call when they first enter these two routines. It validates the rport state, and returns a scsi result which could be returned. In addition to solving the above, it also creates consistent behavior from the LLDD's when the block and deletes are occuring. Rejections fixed up and Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2005-10-18 20:03:35 +04:00
/**
* fc_vport_set_state() - called to set a vport's state. Saves the old state,
* excepting the transitory states of initializing and sending the ELS
* traffic to instantiate the vport on the link.
*
* Assumes the driver has surrounded this with the proper locking to ensure
* a coherent state change.
*
* @vport: virtual port whose state is changing
* @new_state: new state
**/
static inline void
fc_vport_set_state(struct fc_vport *vport, enum fc_vport_state new_state)
{
if ((new_state != FC_VPORT_UNKNOWN) &&
(new_state != FC_VPORT_INITIALIZING))
vport->vport_last_state = vport->vport_state;
vport->vport_state = new_state;
}
struct scsi_transport_template *fc_attach_transport(
struct fc_function_template *);
void fc_release_transport(struct scsi_transport_template *);
void fc_remove_host(struct Scsi_Host *);
struct fc_rport *fc_remote_port_add(struct Scsi_Host *shost,
int channel, struct fc_rport_identifiers *ids);
void fc_remote_port_delete(struct fc_rport *rport);
void fc_remote_port_rolechg(struct fc_rport *rport, u32 roles);
int scsi_is_fc_rport(const struct device *);
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
u32 fc_get_event_number(void);
void fc_host_post_event(struct Scsi_Host *shost, u32 event_number,
enum fc_host_event_code event_code, u32 event_data);
void fc_host_post_vendor_event(struct Scsi_Host *shost, u32 event_number,
u32 data_len, char *data_buf, u64 vendor_id);
struct fc_rport *fc_find_rport_by_wwpn(struct Scsi_Host *shost, u64 wwpn);
void fc_host_post_fc_event(struct Scsi_Host *shost, u32 event_number,
enum fc_host_event_code event_code,
u32 data_len, char *data_buf, u64 vendor_id);
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
/* Note: when specifying vendor_id to fc_host_post_vendor_event()
* or fc_host_post_fc_event(), be sure to read the Vendor Type
* and ID formatting requirements specified in scsi_netlink.h
* Note: when calling fc_host_post_fc_event(), vendor_id may be
* specified as 0.
[SCSI] SCSI and FC Transport: add netlink support for posting of transport events This patch formally adds support for the posting of FC events via netlink. It is a followup to the original RFC at: http://marc.theaimsgroup.com/?l=linux-scsi&m=114530667923464&w=2 and the initial posting at: http://marc.theaimsgroup.com/?l=linux-scsi&m=115507374832500&w=2 The patch has been updated to optimize the send path, per the discussions in the initial posting. Per discussions at the Storage Summit and at OLS, we are to use netlink for async events from transports. Also per discussions, to avoid a netlink protocol per transport, I've create a single NETLINK_SCSITRANSPORT protocol, which can then be used by all transports. This patch: - Creates new files scsi_netlink.c and scsi_netlink.h, which contains the single and shared definitions for the SCSI Transport. It is tied into the base SCSI subsystem intialization. Contains a single interface routine, scsi_send_transport_event(), for a transport to send an event (via multicast to a protocol specific group). - Creates a new scsi_netlink_fc.h file, which contains the FC netlink event messages - Adds 3 new routines to the fc transport: fc_get_event_number() - to get a FC event # fc_host_post_event() - to send a simple FC event (32 bits of data) fc_host_post_vendor_event() - to send a Vendor unique event, with arbitrary amounts of data. Note: the separation of event number allows for a LLD to send a standard event, followed by vendor-specific data for the event. Note: This patch assumes 2 prior fc transport patches have been installed: http://marc.theaimsgroup.com/?l=linux-scsi&m=115555807316329&w=2 http://marc.theaimsgroup.com/?l=linux-scsi&m=115581614930261&w=2 Sorry - next time I'll do something like making these individual patches of the same posting when I know they'll be posted closely together. Signed-off-by: James Smart <James.Smart@emulex.com> Tidy up configuration not to make SCSI always select NET Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
2006-08-19 01:30:09 +04:00
*/
void fc_host_fpin_rcv(struct Scsi_Host *shost, u32 fpin_len, char *fpin_buf);
struct fc_vport *fc_vport_create(struct Scsi_Host *shost, int channel,
struct fc_vport_identifiers *);
int fc_vport_terminate(struct fc_vport *vport);
int fc_block_rport(struct fc_rport *rport);
int fc_block_scsi_eh(struct scsi_cmnd *cmnd);
enum blk_eh_timer_return fc_eh_timed_out(struct scsi_cmnd *scmd);
bool fc_eh_should_retry_cmd(struct scsi_cmnd *scmd);
static inline struct Scsi_Host *fc_bsg_to_shost(struct bsg_job *job)
{
if (scsi_is_host_device(job->dev))
return dev_to_shost(job->dev);
return rport_to_shost(dev_to_rport(job->dev));
}
static inline struct fc_rport *fc_bsg_to_rport(struct bsg_job *job)
{
if (scsi_is_fc_rport(job->dev))
return dev_to_rport(job->dev);
return NULL;
}
#endif /* SCSI_TRANSPORT_FC_H */