linux_dsm_epyc7002/drivers/scsi/lpfc/lpfc_attr.c

6319 lines
188 KiB
C
Raw Normal View History

/*******************************************************************
* This file is part of the Emulex Linux Device Driver for *
* Fibre Channel Host Bus Adapters. *
* Copyright (C) 2017 Broadcom. All Rights Reserved. The term *
* Broadcom refers to Broadcom Limited and/or its subsidiaries. *
* Copyright (C) 2004-2016 Emulex. All rights reserved. *
* EMULEX and SLI are trademarks of Emulex. *
* www.broadcom.com *
* Portions Copyright (C) 2004-2005 Christoph Hellwig *
* *
* This program is free software; you can redistribute it and/or *
* 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. *
*******************************************************************/
#include <linux/ctype.h>
#include <linux/delay.h>
#include <linux/pci.h>
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/aer.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 15:04:11 +07:00
#include <linux/gfp.h>
#include <linux/kernel.h>
#include <scsi/scsi.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_tcq.h>
#include <scsi/scsi_transport_fc.h>
#include <scsi/fc/fc_fs.h>
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
#include <linux/nvme-fc-driver.h>
#include "lpfc_hw4.h"
#include "lpfc_hw.h"
#include "lpfc_sli.h"
#include "lpfc_sli4.h"
#include "lpfc_nl.h"
#include "lpfc_disc.h"
#include "lpfc.h"
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
#include "lpfc_scsi.h"
#include "lpfc_nvme.h"
#include "lpfc_nvmet.h"
#include "lpfc_logmsg.h"
#include "lpfc_version.h"
#include "lpfc_compat.h"
#include "lpfc_crtn.h"
#include "lpfc_vport.h"
#include "lpfc_attr.h"
#define LPFC_DEF_DEVLOSS_TMO 30
#define LPFC_MIN_DEVLOSS_TMO 1
#define LPFC_MAX_DEVLOSS_TMO 255
#define LPFC_DEF_MRQ_POST 256
#define LPFC_MIN_MRQ_POST 32
#define LPFC_MAX_MRQ_POST 512
/*
* Write key size should be multiple of 4. If write key is changed
* make sure that library write key is also changed.
*/
#define LPFC_REG_WRITE_KEY_SIZE 4
#define LPFC_REG_WRITE_KEY "EMLX"
/**
* lpfc_jedec_to_ascii - Hex to ascii convertor according to JEDEC rules
* @incr: integer to convert.
* @hdw: ascii string holding converted integer plus a string terminator.
*
* Description:
* JEDEC Joint Electron Device Engineering Council.
* Convert a 32 bit integer composed of 8 nibbles into an 8 byte ascii
* character string. The string is then terminated with a NULL in byte 9.
* Hex 0-9 becomes ascii '0' to '9'.
* Hex a-f becomes ascii '=' to 'B' capital B.
*
* Notes:
* Coded for 32 bit integers only.
**/
static void
lpfc_jedec_to_ascii(int incr, char hdw[])
{
int i, j;
for (i = 0; i < 8; i++) {
j = (incr & 0xf);
if (j <= 9)
hdw[7 - i] = 0x30 + j;
else
hdw[7 - i] = 0x61 + j - 10;
incr = (incr >> 4);
}
hdw[8] = 0;
return;
}
/**
* lpfc_drvr_version_show - Return the Emulex driver string with version number
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_drvr_version_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
return snprintf(buf, PAGE_SIZE, LPFC_MODULE_DESC "\n");
}
/**
* lpfc_enable_fip_show - Return the fip mode of the HBA
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_enable_fip_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (phba->hba_flag & HBA_FIP_SUPPORT)
return snprintf(buf, PAGE_SIZE, "1\n");
else
return snprintf(buf, PAGE_SIZE, "0\n");
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
static ssize_t
lpfc_nvme_info_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = shost_priv(shost);
struct lpfc_hba *phba = vport->phba;
struct lpfc_nvmet_tgtport *tgtp;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
struct nvme_fc_local_port *localport;
struct lpfc_nvme_lport *lport;
struct lpfc_nvme_rport *rport;
struct nvme_fc_remote_port *nrport;
char *statep;
int len = 0;
if (!(phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME)) {
len += snprintf(buf, PAGE_SIZE, "NVME Disabled\n");
return len;
}
if (phba->nvmet_support) {
if (!phba->targetport) {
len = snprintf(buf, PAGE_SIZE,
"NVME Target: x%llx is not allocated\n",
wwn_to_u64(vport->fc_portname.u.wwn));
return len;
}
/* Port state is only one of two values for now. */
if (phba->targetport->port_id)
statep = "REGISTERED";
else
statep = "INIT";
len += snprintf(buf + len, PAGE_SIZE - len,
"NVME Target: Enabled State %s\n",
statep);
len += snprintf(buf + len, PAGE_SIZE - len,
"%s%d WWPN x%llx WWNN x%llx DID x%06x\n",
"NVME Target: lpfc",
phba->brd_no,
wwn_to_u64(vport->fc_portname.u.wwn),
wwn_to_u64(vport->fc_nodename.u.wwn),
phba->targetport->port_id);
len += snprintf(buf + len, PAGE_SIZE,
"\nNVME Target: Statistics\n");
tgtp = (struct lpfc_nvmet_tgtport *)phba->targetport->private;
len += snprintf(buf+len, PAGE_SIZE-len,
"LS: Rcv %08x Drop %08x Abort %08x\n",
atomic_read(&tgtp->rcv_ls_req_in),
atomic_read(&tgtp->rcv_ls_req_drop),
atomic_read(&tgtp->xmt_ls_abort));
if (atomic_read(&tgtp->rcv_ls_req_in) !=
atomic_read(&tgtp->rcv_ls_req_out)) {
len += snprintf(buf+len, PAGE_SIZE-len,
"Rcv LS: in %08x != out %08x\n",
atomic_read(&tgtp->rcv_ls_req_in),
atomic_read(&tgtp->rcv_ls_req_out));
}
len += snprintf(buf+len, PAGE_SIZE-len,
"LS: Xmt %08x Drop %08x Cmpl %08x Err %08x\n",
atomic_read(&tgtp->xmt_ls_rsp),
atomic_read(&tgtp->xmt_ls_drop),
atomic_read(&tgtp->xmt_ls_rsp_cmpl),
atomic_read(&tgtp->xmt_ls_rsp_error));
len += snprintf(buf+len, PAGE_SIZE-len,
"FCP: Rcv %08x Drop %08x\n",
atomic_read(&tgtp->rcv_fcp_cmd_in),
atomic_read(&tgtp->rcv_fcp_cmd_drop));
if (atomic_read(&tgtp->rcv_fcp_cmd_in) !=
atomic_read(&tgtp->rcv_fcp_cmd_out)) {
len += snprintf(buf+len, PAGE_SIZE-len,
"Rcv FCP: in %08x != out %08x\n",
atomic_read(&tgtp->rcv_fcp_cmd_in),
atomic_read(&tgtp->rcv_fcp_cmd_out));
}
len += snprintf(buf+len, PAGE_SIZE-len,
"FCP Rsp: RD %08x rsp %08x WR %08x rsp %08x\n",
atomic_read(&tgtp->xmt_fcp_read),
atomic_read(&tgtp->xmt_fcp_read_rsp),
atomic_read(&tgtp->xmt_fcp_write),
atomic_read(&tgtp->xmt_fcp_rsp));
len += snprintf(buf+len, PAGE_SIZE-len,
"FCP Rsp: abort %08x drop %08x\n",
atomic_read(&tgtp->xmt_fcp_abort),
atomic_read(&tgtp->xmt_fcp_drop));
len += snprintf(buf+len, PAGE_SIZE-len,
"FCP Rsp Cmpl: %08x err %08x drop %08x\n",
atomic_read(&tgtp->xmt_fcp_rsp_cmpl),
atomic_read(&tgtp->xmt_fcp_rsp_error),
atomic_read(&tgtp->xmt_fcp_rsp_drop));
len += snprintf(buf+len, PAGE_SIZE-len,
"ABORT: Xmt %08x Err %08x Cmpl %08x",
atomic_read(&tgtp->xmt_abort_rsp),
atomic_read(&tgtp->xmt_abort_rsp_error),
atomic_read(&tgtp->xmt_abort_cmpl));
len += snprintf(buf+len, PAGE_SIZE-len, "\n");
return len;
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
localport = vport->localport;
if (!localport) {
len = snprintf(buf, PAGE_SIZE,
"NVME Initiator x%llx is not allocated\n",
wwn_to_u64(vport->fc_portname.u.wwn));
return len;
}
len = snprintf(buf, PAGE_SIZE, "NVME Initiator Enabled\n");
spin_lock_irq(shost->host_lock);
lport = (struct lpfc_nvme_lport *)localport->private;
/* Port state is only one of two values for now. */
if (localport->port_id)
statep = "ONLINE";
else
statep = "UNKNOWN ";
len += snprintf(buf + len, PAGE_SIZE - len,
"%s%d WWPN x%llx WWNN x%llx DID x%06x %s\n",
"NVME LPORT lpfc",
phba->brd_no,
wwn_to_u64(vport->fc_portname.u.wwn),
wwn_to_u64(vport->fc_nodename.u.wwn),
localport->port_id, statep);
list_for_each_entry(rport, &lport->rport_list, list) {
/* local short-hand pointer. */
nrport = rport->remoteport;
/* Port state is only one of two values for now. */
switch (nrport->port_state) {
case FC_OBJSTATE_ONLINE:
statep = "ONLINE";
break;
case FC_OBJSTATE_UNKNOWN:
statep = "UNKNOWN ";
break;
default:
statep = "UNSUPPORTED";
break;
}
/* Tab in to show lport ownership. */
len += snprintf(buf + len, PAGE_SIZE - len,
"NVME RPORT ");
if (phba->brd_no >= 10)
len += snprintf(buf + len, PAGE_SIZE - len, " ");
len += snprintf(buf + len, PAGE_SIZE - len, "WWPN x%llx ",
nrport->port_name);
len += snprintf(buf + len, PAGE_SIZE - len, "WWNN x%llx ",
nrport->node_name);
len += snprintf(buf + len, PAGE_SIZE - len, "DID x%06x ",
nrport->port_id);
switch (nrport->port_role) {
case FC_PORT_ROLE_NVME_INITIATOR:
len += snprintf(buf + len, PAGE_SIZE - len,
"INITIATOR ");
break;
case FC_PORT_ROLE_NVME_TARGET:
len += snprintf(buf + len, PAGE_SIZE - len,
"TARGET ");
break;
case FC_PORT_ROLE_NVME_DISCOVERY:
len += snprintf(buf + len, PAGE_SIZE - len,
"DISCOVERY ");
break;
default:
len += snprintf(buf + len, PAGE_SIZE - len,
"UNKNOWN_ROLE x%x",
nrport->port_role);
break;
}
len += snprintf(buf + len, PAGE_SIZE - len, "%s ", statep);
/* Terminate the string. */
len += snprintf(buf + len, PAGE_SIZE - len, "\n");
}
spin_unlock_irq(shost->host_lock);
len += snprintf(buf + len, PAGE_SIZE, "\nNVME Statistics\n");
len += snprintf(buf+len, PAGE_SIZE-len,
"LS: Xmt %016llx Cmpl %016llx\n",
phba->fc4NvmeLsRequests,
phba->fc4NvmeLsCmpls);
len += snprintf(buf+len, PAGE_SIZE-len,
"FCP: Rd %016llx Wr %016llx IO %016llx\n",
phba->fc4NvmeInputRequests,
phba->fc4NvmeOutputRequests,
phba->fc4NvmeControlRequests);
len += snprintf(buf+len, PAGE_SIZE-len,
" Cmpl %016llx\n", phba->fc4NvmeIoCmpls);
return len;
}
static ssize_t
lpfc_bg_info_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (phba->cfg_enable_bg)
if (phba->sli3_options & LPFC_SLI3_BG_ENABLED)
return snprintf(buf, PAGE_SIZE, "BlockGuard Enabled\n");
else
return snprintf(buf, PAGE_SIZE,
"BlockGuard Not Supported\n");
else
return snprintf(buf, PAGE_SIZE,
"BlockGuard Disabled\n");
}
static ssize_t
lpfc_bg_guard_err_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%llu\n",
(unsigned long long)phba->bg_guard_err_cnt);
}
static ssize_t
lpfc_bg_apptag_err_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%llu\n",
(unsigned long long)phba->bg_apptag_err_cnt);
}
static ssize_t
lpfc_bg_reftag_err_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%llu\n",
(unsigned long long)phba->bg_reftag_err_cnt);
}
/**
* lpfc_info_show - Return some pci info about the host in ascii
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the formatted text from lpfc_info().
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_info_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *host = class_to_shost(dev);
return snprintf(buf, PAGE_SIZE, "%s\n",lpfc_info(host));
}
/**
* lpfc_serialnum_show - Return the hba serial number in ascii
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the formatted text serial number.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_serialnum_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s\n",phba->SerialNumber);
}
/**
* lpfc_temp_sensor_show - Return the temperature sensor level
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the formatted support level.
*
* Description:
* Returns a number indicating the temperature sensor level currently
* supported, zero or one in ascii.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_temp_sensor_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n",phba->temp_sensor_support);
}
/**
* lpfc_modeldesc_show - Return the model description of the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the scsi vpd model description.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_modeldesc_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s\n",phba->ModelDesc);
}
/**
* lpfc_modelname_show - Return the model name of the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the scsi vpd model name.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_modelname_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s\n",phba->ModelName);
}
/**
* lpfc_programtype_show - Return the program type of the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the scsi vpd program type.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_programtype_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s\n",phba->ProgramType);
}
/**
* lpfc_mlomgmt_show - Return the Menlo Maintenance sli flag
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the Menlo Maintenance sli flag.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_mlomgmt_show(struct device *dev, struct device_attribute *attr, char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n",
(phba->sli.sli_flag & LPFC_MENLO_MAINT));
}
/**
* lpfc_vportnum_show - Return the port number in ascii of the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains scsi vpd program type.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_vportnum_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s\n",phba->Port);
}
/**
* lpfc_fwrev_show - Return the firmware rev running in the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the scsi vpd program type.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_fwrev_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t if_type;
uint8_t sli_family;
char fwrev[FW_REV_STR_SIZE];
int len;
lpfc_decode_firmware_rev(phba, fwrev, 1);
if_type = phba->sli4_hba.pc_sli4_params.if_type;
sli_family = phba->sli4_hba.pc_sli4_params.sli_family;
if (phba->sli_rev < LPFC_SLI_REV4)
len = snprintf(buf, PAGE_SIZE, "%s, sli-%d\n",
fwrev, phba->sli_rev);
else
len = snprintf(buf, PAGE_SIZE, "%s, sli-%d:%d:%x\n",
fwrev, phba->sli_rev, if_type, sli_family);
return len;
}
/**
* lpfc_hdw_show - Return the jedec information about the hba
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the scsi vpd program type.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_hdw_show(struct device *dev, struct device_attribute *attr, char *buf)
{
char hdw[9];
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
lpfc_vpd_t *vp = &phba->vpd;
lpfc_jedec_to_ascii(vp->rev.biuRev, hdw);
return snprintf(buf, PAGE_SIZE, "%s\n", hdw);
}
/**
* lpfc_option_rom_version_show - Return the adapter ROM FCode version
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the ROM and FCode ascii strings.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_option_rom_version_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
char fwrev[FW_REV_STR_SIZE];
if (phba->sli_rev < LPFC_SLI_REV4)
return snprintf(buf, PAGE_SIZE, "%s\n", phba->OptionROMVersion);
lpfc_decode_firmware_rev(phba, fwrev, 1);
return snprintf(buf, PAGE_SIZE, "%s\n", fwrev);
}
/**
* lpfc_state_show - Return the link state of the port
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains text describing the state of the link.
*
* Notes:
* The switch statement has no default so zero will be returned.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_link_state_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int len = 0;
switch (phba->link_state) {
case LPFC_LINK_UNKNOWN:
case LPFC_WARM_START:
case LPFC_INIT_START:
case LPFC_INIT_MBX_CMDS:
case LPFC_LINK_DOWN:
case LPFC_HBA_ERROR:
if (phba->hba_flag & LINK_DISABLED)
len += snprintf(buf + len, PAGE_SIZE-len,
"Link Down - User disabled\n");
else
len += snprintf(buf + len, PAGE_SIZE-len,
"Link Down\n");
break;
case LPFC_LINK_UP:
case LPFC_CLEAR_LA:
case LPFC_HBA_READY:
len += snprintf(buf + len, PAGE_SIZE-len, "Link Up - ");
switch (vport->port_state) {
case LPFC_LOCAL_CFG_LINK:
len += snprintf(buf + len, PAGE_SIZE-len,
"Configuring Link\n");
break;
case LPFC_FDISC:
case LPFC_FLOGI:
case LPFC_FABRIC_CFG_LINK:
case LPFC_NS_REG:
case LPFC_NS_QRY:
case LPFC_BUILD_DISC_LIST:
case LPFC_DISC_AUTH:
len += snprintf(buf + len, PAGE_SIZE - len,
"Discovery\n");
break;
case LPFC_VPORT_READY:
len += snprintf(buf + len, PAGE_SIZE - len, "Ready\n");
break;
case LPFC_VPORT_FAILED:
len += snprintf(buf + len, PAGE_SIZE - len, "Failed\n");
break;
case LPFC_VPORT_UNKNOWN:
len += snprintf(buf + len, PAGE_SIZE - len,
"Unknown\n");
break;
}
if (phba->sli.sli_flag & LPFC_MENLO_MAINT)
len += snprintf(buf + len, PAGE_SIZE-len,
" Menlo Maint Mode\n");
2010-11-21 11:11:48 +07:00
else if (phba->fc_topology == LPFC_TOPOLOGY_LOOP) {
if (vport->fc_flag & FC_PUBLIC_LOOP)
len += snprintf(buf + len, PAGE_SIZE-len,
" Public Loop\n");
else
len += snprintf(buf + len, PAGE_SIZE-len,
" Private Loop\n");
} else {
if (vport->fc_flag & FC_FABRIC)
len += snprintf(buf + len, PAGE_SIZE-len,
" Fabric\n");
else
len += snprintf(buf + len, PAGE_SIZE-len,
" Point-2-Point\n");
}
}
return len;
}
/**
* lpfc_sli4_protocol_show - Return the fip mode of the HBA
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_sli4_protocol_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (phba->sli_rev < LPFC_SLI_REV4)
return snprintf(buf, PAGE_SIZE, "fc\n");
if (phba->sli4_hba.lnk_info.lnk_dv == LPFC_LNK_DAT_VAL) {
if (phba->sli4_hba.lnk_info.lnk_tp == LPFC_LNK_TYPE_GE)
return snprintf(buf, PAGE_SIZE, "fcoe\n");
if (phba->sli4_hba.lnk_info.lnk_tp == LPFC_LNK_TYPE_FC)
return snprintf(buf, PAGE_SIZE, "fc\n");
}
return snprintf(buf, PAGE_SIZE, "unknown\n");
}
/**
* lpfc_oas_supported_show - Return whether or not Optimized Access Storage
* (OAS) is supported.
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_oas_supported_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n",
phba->sli4_hba.pc_sli4_params.oas_supported);
}
/**
* lpfc_link_state_store - Transition the link_state on an HBA port
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: one or more lpfc_polling_flags values.
* @count: not used.
*
* Returns:
* -EINVAL if the buffer is not "up" or "down"
* return from link state change function if non-zero
* length of the buf on success
**/
static ssize_t
lpfc_link_state_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int status = -EINVAL;
if ((strncmp(buf, "up", sizeof("up") - 1) == 0) &&
(phba->link_state == LPFC_LINK_DOWN))
status = phba->lpfc_hba_init_link(phba, MBX_NOWAIT);
else if ((strncmp(buf, "down", sizeof("down") - 1) == 0) &&
(phba->link_state >= LPFC_LINK_UP))
status = phba->lpfc_hba_down_link(phba, MBX_NOWAIT);
if (status == 0)
return strlen(buf);
else
return status;
}
/**
* lpfc_num_discovered_ports_show - Return sum of mapped and unmapped vports
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the sum of fc mapped and unmapped.
*
* Description:
* Returns the ascii text number of the sum of the fc mapped and unmapped
* vport counts.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_num_discovered_ports_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
return snprintf(buf, PAGE_SIZE, "%d\n",
vport->fc_map_cnt + vport->fc_unmap_cnt);
}
/**
* lpfc_issue_lip - Misnomer, name carried over from long ago
* @shost: Scsi_Host pointer.
*
* Description:
* Bring the link down gracefully then re-init the link. The firmware will
* re-init the fiber channel interface as required. Does not issue a LIP.
*
* Returns:
* -EPERM port offline or management commands are being blocked
* -ENOMEM cannot allocate memory for the mailbox command
* -EIO error sending the mailbox command
* zero for success
**/
static int
lpfc_issue_lip(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
LPFC_MBOXQ_t *pmboxq;
int mbxstatus = MBXERR_ERROR;
if ((vport->fc_flag & FC_OFFLINE_MODE) ||
(phba->sli.sli_flag & LPFC_BLOCK_MGMT_IO))
return -EPERM;
pmboxq = mempool_alloc(phba->mbox_mem_pool,GFP_KERNEL);
if (!pmboxq)
return -ENOMEM;
memset((void *)pmboxq, 0, sizeof (LPFC_MBOXQ_t));
pmboxq->u.mb.mbxCommand = MBX_DOWN_LINK;
pmboxq->u.mb.mbxOwner = OWN_HOST;
mbxstatus = lpfc_sli_issue_mbox_wait(phba, pmboxq, LPFC_MBOX_TMO * 2);
if ((mbxstatus == MBX_SUCCESS) &&
(pmboxq->u.mb.mbxStatus == 0 ||
pmboxq->u.mb.mbxStatus == MBXERR_LINK_DOWN)) {
memset((void *)pmboxq, 0, sizeof (LPFC_MBOXQ_t));
lpfc_init_link(phba, pmboxq, phba->cfg_topology,
phba->cfg_link_speed);
mbxstatus = lpfc_sli_issue_mbox_wait(phba, pmboxq,
phba->fc_ratov * 2);
if ((mbxstatus == MBX_SUCCESS) &&
(pmboxq->u.mb.mbxStatus == MBXERR_SEC_NO_PERMISSION))
lpfc_printf_log(phba, KERN_ERR, LOG_MBOX | LOG_SLI,
"2859 SLI authentication is required "
"for INIT_LINK but has not done yet\n");
}
lpfc_set_loopback_flag(phba);
if (mbxstatus != MBX_TIMEOUT)
mempool_free(pmboxq, phba->mbox_mem_pool);
if (mbxstatus == MBXERR_ERROR)
return -EIO;
return 0;
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
int
lpfc_emptyq_wait(struct lpfc_hba *phba, struct list_head *q, spinlock_t *lock)
{
int cnt = 0;
spin_lock_irq(lock);
while (!list_empty(q)) {
spin_unlock_irq(lock);
msleep(20);
if (cnt++ > 250) { /* 5 secs */
lpfc_printf_log(phba, KERN_WARNING, LOG_INIT,
"0466 %s %s\n",
"Outstanding IO when ",
"bringing Adapter offline\n");
return 0;
}
spin_lock_irq(lock);
}
spin_unlock_irq(lock);
return 1;
}
/**
* lpfc_do_offline - Issues a mailbox command to bring the link down
* @phba: lpfc_hba pointer.
* @type: LPFC_EVT_OFFLINE, LPFC_EVT_WARM_START, LPFC_EVT_KILL.
*
* Notes:
* Assumes any error from lpfc_do_offline() will be negative.
* Can wait up to 5 seconds for the port ring buffers count
* to reach zero, prints a warning if it is not zero and continues.
* lpfc_workq_post_event() returns a non-zero return code if call fails.
*
* Returns:
* -EIO error posting the event
* zero for success
**/
static int
lpfc_do_offline(struct lpfc_hba *phba, uint32_t type)
{
struct completion online_compl;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
struct lpfc_queue *qp = NULL;
struct lpfc_sli_ring *pring;
struct lpfc_sli *psli;
int status = 0;
int i;
int rc;
init_completion(&online_compl);
rc = lpfc_workq_post_event(phba, &status, &online_compl,
LPFC_EVT_OFFLINE_PREP);
if (rc == 0)
return -ENOMEM;
wait_for_completion(&online_compl);
if (status != 0)
return -EIO;
psli = &phba->sli;
[SCSI] lpfc 8.2.4 : Miscellaneous Fixes Miscellaneous Fixes: - Fix a couple of sparse complaints - Reset the FCP recovery flag when the node is not a FCP2 device. - Speed up offline prep delays - Fixed a memory leak in lpfc_mem_alloc failure path - Fixed external loopback test. - Fixed error code returned from the driver when HBA is over heated. - Correct Max NPIV vport to limits read from adapter - Add missing locks around fc_flag and FC_NEEDS_REG_VPI - Add missing hba ids for device identification - Added support for SET_VARIABLE and MBX_WRITE_WWN mailbox commands - Changed all temperature event messages from warning to error - Fix reporting of link speed when link is down - Added support for MBX_WRITE_WWN mailbox command - Change del_timer_sync() in ISR to del_timer() in interrupt handler - Correct instances of beXX_to_cpu() that should be cpu_to_beXX() - Perform target flush before releasing node references on module unload - Avoid bogus devloss_tmo messages when driver unloads - Fix panic when HBA generates ERATT interupt - Fix mbox race condition and a workaround on back-to-back mailbox commands - Force NPIV off for pt2pt mode between 2 NPorts - Stop worker thread before removing fc_host. - Fix up discovery timeout error case due to missing clear_la - Tighten mailbox polling code to speed up detection of fast completions - Only allow DUMP_MEMORY if adapter offline due to overtemp errors - Added extended error information to the log messages in chip init. Signed-off-by: James Smart <James.Smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2008-01-11 13:52:54 +07:00
/* Wait a little for things to settle down, but not
* long enough for dev loss timeout to expire.
*/
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (phba->sli_rev != LPFC_SLI_REV4) {
for (i = 0; i < psli->num_rings; i++) {
pring = &psli->sli3_ring[i];
if (!lpfc_emptyq_wait(phba, &pring->txcmplq,
&phba->hbalock))
goto out;
}
} else {
list_for_each_entry(qp, &phba->sli4_hba.lpfc_wq_list, wq_list) {
pring = qp->pring;
if (!pring)
continue;
if (!lpfc_emptyq_wait(phba, &pring->txcmplq,
&pring->ring_lock))
goto out;
}
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
out:
init_completion(&online_compl);
rc = lpfc_workq_post_event(phba, &status, &online_compl, type);
if (rc == 0)
return -ENOMEM;
wait_for_completion(&online_compl);
if (status != 0)
return -EIO;
return 0;
}
/**
* lpfc_selective_reset - Offline then onlines the port
* @phba: lpfc_hba pointer.
*
* Description:
* If the port is configured to allow a reset then the hba is brought
* offline then online.
*
* Notes:
* Assumes any error from lpfc_do_offline() will be negative.
* Do not make this function static.
*
* Returns:
* lpfc_do_offline() return code if not zero
* -EIO reset not configured or error posting the event
* zero for success
**/
int
lpfc_selective_reset(struct lpfc_hba *phba)
{
struct completion online_compl;
int status = 0;
int rc;
if (!phba->cfg_enable_hba_reset)
return -EACCES;
if (!(phba->pport->fc_flag & FC_OFFLINE_MODE)) {
status = lpfc_do_offline(phba, LPFC_EVT_OFFLINE);
if (status != 0)
return status;
}
init_completion(&online_compl);
rc = lpfc_workq_post_event(phba, &status, &online_compl,
LPFC_EVT_ONLINE);
if (rc == 0)
return -ENOMEM;
wait_for_completion(&online_compl);
if (status != 0)
return -EIO;
return 0;
}
/**
* lpfc_issue_reset - Selectively resets an adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing the string "selective".
* @count: unused variable.
*
* Description:
* If the buf contains the string "selective" then lpfc_selective_reset()
* is called to perform the reset.
*
* Notes:
* Assumes any error from lpfc_selective_reset() will be negative.
* If lpfc_selective_reset() returns zero then the length of the buffer
* is returned which indicates success
*
* Returns:
* -EINVAL if the buffer does not contain the string "selective"
* length of buf if lpfc-selective_reset() if the call succeeds
* return value of lpfc_selective_reset() if the call fails
**/
static ssize_t
lpfc_issue_reset(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int status = -EINVAL;
2011-10-11 08:32:10 +07:00
if (!phba->cfg_enable_hba_reset)
return -EACCES;
if (strncmp(buf, "selective", sizeof("selective") - 1) == 0)
status = phba->lpfc_selective_reset(phba);
if (status == 0)
return strlen(buf);
else
return status;
}
/**
* lpfc_sli4_pdev_status_reg_wait - Wait for pdev status register for readyness
* @phba: lpfc_hba pointer.
*
* Description:
* SLI4 interface type-2 device to wait on the sliport status register for
* the readyness after performing a firmware reset.
*
* Returns:
* zero for success, -EPERM when port does not have privilege to perform the
* reset, -EIO when port timeout from recovering from the reset.
*
* Note:
* As the caller will interpret the return code by value, be careful in making
* change or addition to return codes.
**/
2011-10-11 08:32:10 +07:00
int
lpfc_sli4_pdev_status_reg_wait(struct lpfc_hba *phba)
{
struct lpfc_register portstat_reg = {0};
int i;
msleep(100);
lpfc_readl(phba->sli4_hba.u.if_type2.STATUSregaddr,
&portstat_reg.word0);
/* verify if privileged for the request operation */
if (!bf_get(lpfc_sliport_status_rn, &portstat_reg) &&
!bf_get(lpfc_sliport_status_err, &portstat_reg))
return -EPERM;
/* wait for the SLI port firmware ready after firmware reset */
for (i = 0; i < LPFC_FW_RESET_MAXIMUM_WAIT_10MS_CNT; i++) {
msleep(10);
lpfc_readl(phba->sli4_hba.u.if_type2.STATUSregaddr,
&portstat_reg.word0);
if (!bf_get(lpfc_sliport_status_err, &portstat_reg))
continue;
if (!bf_get(lpfc_sliport_status_rn, &portstat_reg))
continue;
if (!bf_get(lpfc_sliport_status_rdy, &portstat_reg))
continue;
break;
}
if (i < LPFC_FW_RESET_MAXIMUM_WAIT_10MS_CNT)
return 0;
else
return -EIO;
}
/**
* lpfc_sli4_pdev_reg_request - Request physical dev to perform a register acc
* @phba: lpfc_hba pointer.
*
* Description:
* Request SLI4 interface type-2 device to perform a physical register set
* access.
*
* Returns:
* zero for success
**/
static ssize_t
lpfc_sli4_pdev_reg_request(struct lpfc_hba *phba, uint32_t opcode)
{
struct completion online_compl;
struct pci_dev *pdev = phba->pcidev;
uint32_t before_fc_flag;
uint32_t sriov_nr_virtfn;
uint32_t reg_val;
int status = 0, rc = 0;
int job_posted = 1, sriov_err;
if (!phba->cfg_enable_hba_reset)
return -EACCES;
if ((phba->sli_rev < LPFC_SLI_REV4) ||
(bf_get(lpfc_sli_intf_if_type, &phba->sli4_hba.sli_intf) !=
LPFC_SLI_INTF_IF_TYPE_2))
return -EPERM;
/* Keep state if we need to restore back */
before_fc_flag = phba->pport->fc_flag;
sriov_nr_virtfn = phba->cfg_sriov_nr_virtfn;
/* Disable SR-IOV virtual functions if enabled */
if (phba->cfg_sriov_nr_virtfn) {
pci_disable_sriov(pdev);
phba->cfg_sriov_nr_virtfn = 0;
}
if (opcode == LPFC_FW_DUMP)
phba->hba_flag |= HBA_FW_DUMP_OP;
status = lpfc_do_offline(phba, LPFC_EVT_OFFLINE);
if (status != 0) {
phba->hba_flag &= ~HBA_FW_DUMP_OP;
return status;
}
/* wait for the device to be quiesced before firmware reset */
msleep(100);
reg_val = readl(phba->sli4_hba.conf_regs_memmap_p +
LPFC_CTL_PDEV_CTL_OFFSET);
if (opcode == LPFC_FW_DUMP)
reg_val |= LPFC_FW_DUMP_REQUEST;
else if (opcode == LPFC_FW_RESET)
reg_val |= LPFC_CTL_PDEV_CTL_FRST;
else if (opcode == LPFC_DV_RESET)
reg_val |= LPFC_CTL_PDEV_CTL_DRST;
writel(reg_val, phba->sli4_hba.conf_regs_memmap_p +
LPFC_CTL_PDEV_CTL_OFFSET);
/* flush */
readl(phba->sli4_hba.conf_regs_memmap_p + LPFC_CTL_PDEV_CTL_OFFSET);
/* delay driver action following IF_TYPE_2 reset */
rc = lpfc_sli4_pdev_status_reg_wait(phba);
if (rc == -EPERM) {
/* no privilege for reset */
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
"3150 No privilege to perform the requested "
"access: x%x\n", reg_val);
} else if (rc == -EIO) {
/* reset failed, there is nothing more we can do */
lpfc_printf_log(phba, KERN_ERR, LOG_SLI,
"3153 Fail to perform the requested "
"access: x%x\n", reg_val);
return rc;
}
/* keep the original port state */
if (before_fc_flag & FC_OFFLINE_MODE)
goto out;
init_completion(&online_compl);
job_posted = lpfc_workq_post_event(phba, &status, &online_compl,
LPFC_EVT_ONLINE);
if (!job_posted)
goto out;
wait_for_completion(&online_compl);
out:
/* in any case, restore the virtual functions enabled as before */
if (sriov_nr_virtfn) {
sriov_err =
lpfc_sli_probe_sriov_nr_virtfn(phba, sriov_nr_virtfn);
if (!sriov_err)
phba->cfg_sriov_nr_virtfn = sriov_nr_virtfn;
}
/* return proper error code */
if (!rc) {
if (!job_posted)
rc = -ENOMEM;
else if (status)
rc = -EIO;
}
return rc;
}
/**
* lpfc_nport_evt_cnt_show - Return the number of nport events
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the ascii number of nport events.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_nport_evt_cnt_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->nport_event_cnt);
}
/**
* lpfc_board_mode_show - Return the state of the board
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the state of the adapter.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_board_mode_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
char * state;
if (phba->link_state == LPFC_HBA_ERROR)
state = "error";
else if (phba->link_state == LPFC_WARM_START)
state = "warm start";
else if (phba->link_state == LPFC_INIT_START)
state = "offline";
else
state = "online";
return snprintf(buf, PAGE_SIZE, "%s\n", state);
}
/**
* lpfc_board_mode_store - Puts the hba in online, offline, warm or error state
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing one of the strings "online", "offline", "warm" or "error".
* @count: unused variable.
*
* Returns:
* -EACCES if enable hba reset not enabled
* -EINVAL if the buffer does not contain a valid string (see above)
* -EIO if lpfc_workq_post_event() or lpfc_do_offline() fails
* buf length greater than zero indicates success
**/
static ssize_t
lpfc_board_mode_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct completion online_compl;
char *board_mode_str = NULL;
int status = 0;
int rc;
if (!phba->cfg_enable_hba_reset) {
status = -EACCES;
goto board_mode_out;
}
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3050 lpfc_board_mode set to %s\n", buf);
init_completion(&online_compl);
if(strncmp(buf, "online", sizeof("online") - 1) == 0) {
rc = lpfc_workq_post_event(phba, &status, &online_compl,
LPFC_EVT_ONLINE);
if (rc == 0) {
status = -ENOMEM;
goto board_mode_out;
}
wait_for_completion(&online_compl);
} else if (strncmp(buf, "offline", sizeof("offline") - 1) == 0)
status = lpfc_do_offline(phba, LPFC_EVT_OFFLINE);
else if (strncmp(buf, "warm", sizeof("warm") - 1) == 0)
if (phba->sli_rev == LPFC_SLI_REV4)
status = -EINVAL;
else
status = lpfc_do_offline(phba, LPFC_EVT_WARM_START);
else if (strncmp(buf, "error", sizeof("error") - 1) == 0)
if (phba->sli_rev == LPFC_SLI_REV4)
status = -EINVAL;
else
status = lpfc_do_offline(phba, LPFC_EVT_KILL);
else if (strncmp(buf, "dump", sizeof("dump") - 1) == 0)
status = lpfc_sli4_pdev_reg_request(phba, LPFC_FW_DUMP);
else if (strncmp(buf, "fw_reset", sizeof("fw_reset") - 1) == 0)
status = lpfc_sli4_pdev_reg_request(phba, LPFC_FW_RESET);
else if (strncmp(buf, "dv_reset", sizeof("dv_reset") - 1) == 0)
status = lpfc_sli4_pdev_reg_request(phba, LPFC_DV_RESET);
else
status = -EINVAL;
board_mode_out:
if (!status)
return strlen(buf);
else {
board_mode_str = strchr(buf, '\n');
if (board_mode_str)
*board_mode_str = '\0';
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3097 Failed \"%s\", status(%d), "
"fc_flag(x%x)\n",
buf, status, phba->pport->fc_flag);
return status;
}
}
/**
* lpfc_get_hba_info - Return various bits of informaton about the adapter
* @phba: pointer to the adapter structure.
* @mxri: max xri count.
* @axri: available xri count.
* @mrpi: max rpi count.
* @arpi: available rpi count.
* @mvpi: max vpi count.
* @avpi: available vpi count.
*
* Description:
* If an integer pointer for an count is not null then the value for the
* count is returned.
*
* Returns:
* zero on error
* one for success
**/
static int
lpfc_get_hba_info(struct lpfc_hba *phba,
uint32_t *mxri, uint32_t *axri,
uint32_t *mrpi, uint32_t *arpi,
uint32_t *mvpi, uint32_t *avpi)
{
struct lpfc_mbx_read_config *rd_config;
LPFC_MBOXQ_t *pmboxq;
MAILBOX_t *pmb;
int rc = 0;
uint32_t max_vpi;
/*
* prevent udev from issuing mailbox commands until the port is
* configured.
*/
if (phba->link_state < LPFC_LINK_DOWN ||
!phba->mbox_mem_pool ||
(phba->sli.sli_flag & LPFC_SLI_ACTIVE) == 0)
return 0;
if (phba->sli.sli_flag & LPFC_BLOCK_MGMT_IO)
return 0;
pmboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
if (!pmboxq)
return 0;
memset(pmboxq, 0, sizeof (LPFC_MBOXQ_t));
pmb = &pmboxq->u.mb;
pmb->mbxCommand = MBX_READ_CONFIG;
pmb->mbxOwner = OWN_HOST;
pmboxq->context1 = NULL;
if (phba->pport->fc_flag & FC_OFFLINE_MODE)
rc = MBX_NOT_FINISHED;
else
rc = lpfc_sli_issue_mbox_wait(phba, pmboxq, phba->fc_ratov * 2);
if (rc != MBX_SUCCESS) {
if (rc != MBX_TIMEOUT)
mempool_free(pmboxq, phba->mbox_mem_pool);
return 0;
}
if (phba->sli_rev == LPFC_SLI_REV4) {
rd_config = &pmboxq->u.mqe.un.rd_config;
if (mrpi)
*mrpi = bf_get(lpfc_mbx_rd_conf_rpi_count, rd_config);
if (arpi)
*arpi = bf_get(lpfc_mbx_rd_conf_rpi_count, rd_config) -
phba->sli4_hba.max_cfg_param.rpi_used;
if (mxri)
*mxri = bf_get(lpfc_mbx_rd_conf_xri_count, rd_config);
if (axri)
*axri = bf_get(lpfc_mbx_rd_conf_xri_count, rd_config) -
phba->sli4_hba.max_cfg_param.xri_used;
/* Account for differences with SLI-3. Get vpi count from
* mailbox data and subtract one for max vpi value.
*/
max_vpi = (bf_get(lpfc_mbx_rd_conf_vpi_count, rd_config) > 0) ?
(bf_get(lpfc_mbx_rd_conf_vpi_count, rd_config) - 1) : 0;
if (mvpi)
*mvpi = max_vpi;
if (avpi)
*avpi = max_vpi - phba->sli4_hba.max_cfg_param.vpi_used;
} else {
if (mrpi)
*mrpi = pmb->un.varRdConfig.max_rpi;
if (arpi)
*arpi = pmb->un.varRdConfig.avail_rpi;
if (mxri)
*mxri = pmb->un.varRdConfig.max_xri;
if (axri)
*axri = pmb->un.varRdConfig.avail_xri;
if (mvpi)
*mvpi = pmb->un.varRdConfig.max_vpi;
if (avpi)
*avpi = pmb->un.varRdConfig.avail_vpi;
}
mempool_free(pmboxq, phba->mbox_mem_pool);
return 1;
}
/**
* lpfc_max_rpi_show - Return maximum rpi
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the maximum rpi count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mrpi count.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_max_rpi_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt;
if (lpfc_get_hba_info(phba, NULL, NULL, &cnt, NULL, NULL, NULL))
return snprintf(buf, PAGE_SIZE, "%d\n", cnt);
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_used_rpi_show - Return maximum rpi minus available rpi
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing the used rpi count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mrpi and arpi counts.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_used_rpi_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt, acnt;
if (lpfc_get_hba_info(phba, NULL, NULL, &cnt, &acnt, NULL, NULL))
return snprintf(buf, PAGE_SIZE, "%d\n", (cnt - acnt));
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_max_xri_show - Return maximum xri
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the maximum xri count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mrpi count.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_max_xri_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt;
if (lpfc_get_hba_info(phba, &cnt, NULL, NULL, NULL, NULL, NULL))
return snprintf(buf, PAGE_SIZE, "%d\n", cnt);
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_used_xri_show - Return maximum xpi minus the available xpi
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the used xri count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mxri and axri counts.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_used_xri_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt, acnt;
if (lpfc_get_hba_info(phba, &cnt, &acnt, NULL, NULL, NULL, NULL))
return snprintf(buf, PAGE_SIZE, "%d\n", (cnt - acnt));
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_max_vpi_show - Return maximum vpi
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the maximum vpi count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mvpi count.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_max_vpi_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt;
if (lpfc_get_hba_info(phba, NULL, NULL, NULL, NULL, &cnt, NULL))
return snprintf(buf, PAGE_SIZE, "%d\n", cnt);
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_used_vpi_show - Return maximum vpi minus the available vpi
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the used vpi count in decimal or "Unknown".
*
* Description:
* Calls lpfc_get_hba_info() asking for just the mvpi and avpi counts.
* If lpfc_get_hba_info() returns zero (failure) the buffer text is set
* to "Unknown" and the buffer length is returned, therefore the caller
* must check for "Unknown" in the buffer to detect a failure.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_used_vpi_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint32_t cnt, acnt;
if (lpfc_get_hba_info(phba, NULL, NULL, NULL, NULL, &cnt, &acnt))
return snprintf(buf, PAGE_SIZE, "%d\n", (cnt - acnt));
return snprintf(buf, PAGE_SIZE, "Unknown\n");
}
/**
* lpfc_npiv_info_show - Return text about NPIV support for the adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: text that must be interpreted to determine if npiv is supported.
*
* Description:
* Buffer will contain text indicating npiv is not suppoerted on the port,
* the port is an NPIV physical port, or it is an npiv virtual port with
* the id of the vport.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_npiv_info_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (!(phba->max_vpi))
return snprintf(buf, PAGE_SIZE, "NPIV Not Supported\n");
if (vport->port_type == LPFC_PHYSICAL_PORT)
return snprintf(buf, PAGE_SIZE, "NPIV Physical\n");
return snprintf(buf, PAGE_SIZE, "NPIV Virtual (VPI %d)\n", vport->vpi);
}
/**
* lpfc_poll_show - Return text about poll support for the adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the cfg_poll in hex.
*
* Notes:
* cfg_poll should be a lpfc_polling_flags type.
*
* Returns: size of formatted string.
**/
[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 04:32:13 +07:00
static ssize_t
lpfc_poll_show(struct device *dev, struct device_attribute *attr,
char *buf)
[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 04:32:13 +07:00
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
[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 04:32:13 +07:00
return snprintf(buf, PAGE_SIZE, "%#x\n", phba->cfg_poll);
}
/**
* lpfc_poll_store - Set the value of cfg_poll for the adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: one or more lpfc_polling_flags values.
* @count: not used.
*
* Notes:
* buf contents converted to integer and checked for a valid value.
*
* Returns:
* -EINVAL if the buffer connot be converted or is out of range
* length of the buf on success
**/
[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 04:32:13 +07:00
static ssize_t
lpfc_poll_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
[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 04:32:13 +07:00
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
[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 04:32:13 +07:00
uint32_t creg_val;
uint32_t old_val;
int val=0;
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
if ((val & 0x3) != val)
return -EINVAL;
if (phba->sli_rev == LPFC_SLI_REV4)
val = 0;
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3051 lpfc_poll changed from %d to %d\n",
phba->cfg_poll, val);
spin_lock_irq(&phba->hbalock);
[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 04:32:13 +07:00
old_val = phba->cfg_poll;
if (val & ENABLE_FCP_RING_POLLING) {
if ((val & DISABLE_FCP_RING_INT) &&
!(old_val & DISABLE_FCP_RING_INT)) {
if (lpfc_readl(phba->HCregaddr, &creg_val)) {
spin_unlock_irq(&phba->hbalock);
return -EINVAL;
}
[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 04:32:13 +07:00
creg_val &= ~(HC_R0INT_ENA << LPFC_FCP_RING);
writel(creg_val, phba->HCregaddr);
readl(phba->HCregaddr); /* flush */
lpfc_poll_start_timer(phba);
}
} else if (val != 0x0) {
spin_unlock_irq(&phba->hbalock);
[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 04:32:13 +07:00
return -EINVAL;
}
if (!(val & DISABLE_FCP_RING_INT) &&
(old_val & DISABLE_FCP_RING_INT))
{
spin_unlock_irq(&phba->hbalock);
[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 04:32:13 +07:00
del_timer(&phba->fcp_poll_timer);
spin_lock_irq(&phba->hbalock);
if (lpfc_readl(phba->HCregaddr, &creg_val)) {
spin_unlock_irq(&phba->hbalock);
return -EINVAL;
}
[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 04:32:13 +07:00
creg_val |= (HC_R0INT_ENA << LPFC_FCP_RING);
writel(creg_val, phba->HCregaddr);
readl(phba->HCregaddr); /* flush */
}
phba->cfg_poll = val;
spin_unlock_irq(&phba->hbalock);
[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 04:32:13 +07:00
return strlen(buf);
}
/**
* lpfc_fips_level_show - Return the current FIPS level for the HBA
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_fips_level_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->fips_level);
}
/**
* lpfc_fips_rev_show - Return the FIPS Spec revision for the HBA
* @dev: class unused variable.
* @attr: device attribute, not used.
* @buf: on return contains the module description text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_fips_rev_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->fips_spec_rev);
}
/**
* lpfc_dss_show - Return the current state of dss and the configured state
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the formatted text.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_dss_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "%s - %sOperational\n",
(phba->cfg_enable_dss) ? "Enabled" : "Disabled",
(phba->sli3_options & LPFC_SLI3_DSS_ENABLED) ?
"" : "Not ");
}
/**
* lpfc_sriov_hw_max_virtfn_show - Return maximum number of virtual functions
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the formatted support level.
*
* Description:
* Returns the maximum number of virtual functions a physical function can
* support, 0 will be returned if called on virtual function.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_sriov_hw_max_virtfn_show(struct device *dev,
struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
uint16_t max_nr_virtfn;
max_nr_virtfn = lpfc_sli_sriov_nr_virtfn_get(phba);
return snprintf(buf, PAGE_SIZE, "%d\n", max_nr_virtfn);
}
static inline bool lpfc_rangecheck(uint val, uint min, uint max)
{
return val >= min && val <= max;
}
/**
* lpfc_param_show - Return a cfg attribute value in decimal
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_show.
*
* lpfc_##attr##_show: Return the decimal value of an adapters cfg_xxx field.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the attribute value in decimal.
*
* Returns: size of formatted string.
**/
#define lpfc_param_show(attr) \
static ssize_t \
lpfc_##attr##_show(struct device *dev, struct device_attribute *attr, \
char *buf) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
struct lpfc_hba *phba = vport->phba;\
return snprintf(buf, PAGE_SIZE, "%d\n",\
phba->cfg_##attr);\
}
/**
* lpfc_param_hex_show - Return a cfg attribute value in hex
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_show
*
* lpfc_##attr##_show: Return the hex value of an adapters cfg_xxx field.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the attribute value in hexadecimal.
*
* Returns: size of formatted string.
**/
#define lpfc_param_hex_show(attr) \
static ssize_t \
lpfc_##attr##_show(struct device *dev, struct device_attribute *attr, \
char *buf) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
struct lpfc_hba *phba = vport->phba;\
uint val = 0;\
val = phba->cfg_##attr;\
return snprintf(buf, PAGE_SIZE, "%#x\n",\
phba->cfg_##attr);\
}
/**
* lpfc_param_init - Initializes a cfg attribute
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_init. The macro also
* takes a default argument, a minimum and maximum argument.
*
* lpfc_##attr##_init: Initializes an attribute.
* @phba: pointer the the adapter structure.
* @val: integer attribute value.
*
* Validates the min and max values then sets the adapter config field
* accordingly, or uses the default if out of range and prints an error message.
*
* Returns:
* zero on success
* -EINVAL if default used
**/
#define lpfc_param_init(attr, default, minval, maxval) \
static int \
lpfc_##attr##_init(struct lpfc_hba *phba, uint val) \
{ \
if (lpfc_rangecheck(val, minval, maxval)) {\
phba->cfg_##attr = val;\
return 0;\
}\
lpfc_printf_log(phba, KERN_ERR, LOG_INIT, \
"0449 lpfc_"#attr" attribute cannot be set to %d, "\
"allowed range is ["#minval", "#maxval"]\n", val); \
phba->cfg_##attr = default;\
return -EINVAL;\
}
/**
* lpfc_param_set - Set a cfg attribute value
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_set
*
* lpfc_##attr##_set: Sets an attribute value.
* @phba: pointer the the adapter structure.
* @val: integer attribute value.
*
* Description:
* Validates the min and max values then sets the
* adapter config field if in the valid range. prints error message
* and does not set the parameter if invalid.
*
* Returns:
* zero on success
* -EINVAL if val is invalid
**/
#define lpfc_param_set(attr, default, minval, maxval) \
static int \
lpfc_##attr##_set(struct lpfc_hba *phba, uint val) \
{ \
if (lpfc_rangecheck(val, minval, maxval)) {\
lpfc_printf_log(phba, KERN_ERR, LOG_INIT, \
"3052 lpfc_" #attr " changed from %d to %d\n", \
phba->cfg_##attr, val); \
phba->cfg_##attr = val;\
return 0;\
}\
lpfc_printf_log(phba, KERN_ERR, LOG_INIT, \
"0450 lpfc_"#attr" attribute cannot be set to %d, "\
"allowed range is ["#minval", "#maxval"]\n", val); \
return -EINVAL;\
}
/**
* lpfc_param_store - Set a vport attribute value
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_store.
*
* lpfc_##attr##_store: Set an sttribute value.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: contains the attribute value in ascii.
* @count: not used.
*
* Description:
* Convert the ascii text number to an integer, then
* use the lpfc_##attr##_set function to set the value.
*
* Returns:
* -EINVAL if val is invalid or lpfc_##attr##_set() fails
* length of buffer upon success.
**/
#define lpfc_param_store(attr) \
static ssize_t \
lpfc_##attr##_store(struct device *dev, struct device_attribute *attr, \
const char *buf, size_t count) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
struct lpfc_hba *phba = vport->phba;\
uint val = 0;\
if (!isdigit(buf[0]))\
return -EINVAL;\
if (sscanf(buf, "%i", &val) != 1)\
return -EINVAL;\
if (lpfc_##attr##_set(phba, val) == 0) \
return strlen(buf);\
else \
return -EINVAL;\
}
/**
* lpfc_vport_param_show - Return decimal formatted cfg attribute value
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_show
*
* lpfc_##attr##_show: prints the attribute value in decimal.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the attribute value in decimal.
*
* Returns: length of formatted string.
**/
#define lpfc_vport_param_show(attr) \
static ssize_t \
lpfc_##attr##_show(struct device *dev, struct device_attribute *attr, \
char *buf) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
return snprintf(buf, PAGE_SIZE, "%d\n", vport->cfg_##attr);\
}
/**
* lpfc_vport_param_hex_show - Return hex formatted attribute value
*
* Description:
* Macro that given an attr e.g.
* hba_queue_depth expands into a function with the name
* lpfc_hba_queue_depth_show
*
* lpfc_##attr##_show: prints the attribute value in hexadecimal.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the attribute value in hexadecimal.
*
* Returns: length of formatted string.
**/
#define lpfc_vport_param_hex_show(attr) \
static ssize_t \
lpfc_##attr##_show(struct device *dev, struct device_attribute *attr, \
char *buf) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
return snprintf(buf, PAGE_SIZE, "%#x\n", vport->cfg_##attr);\
}
/**
* lpfc_vport_param_init - Initialize a vport cfg attribute
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_init. The macro also
* takes a default argument, a minimum and maximum argument.
*
* lpfc_##attr##_init: validates the min and max values then sets the
* adapter config field accordingly, or uses the default if out of range
* and prints an error message.
* @phba: pointer the the adapter structure.
* @val: integer attribute value.
*
* Returns:
* zero on success
* -EINVAL if default used
**/
#define lpfc_vport_param_init(attr, default, minval, maxval) \
static int \
lpfc_##attr##_init(struct lpfc_vport *vport, uint val) \
{ \
if (lpfc_rangecheck(val, minval, maxval)) {\
vport->cfg_##attr = val;\
return 0;\
}\
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT, \
"0423 lpfc_"#attr" attribute cannot be set to %d, "\
"allowed range is ["#minval", "#maxval"]\n", val); \
vport->cfg_##attr = default;\
return -EINVAL;\
}
/**
* lpfc_vport_param_set - Set a vport cfg attribute
*
* Description:
* Macro that given an attr e.g. hba_queue_depth expands
* into a function with the name lpfc_hba_queue_depth_set
*
* lpfc_##attr##_set: validates the min and max values then sets the
* adapter config field if in the valid range. prints error message
* and does not set the parameter if invalid.
* @phba: pointer the the adapter structure.
* @val: integer attribute value.
*
* Returns:
* zero on success
* -EINVAL if val is invalid
**/
#define lpfc_vport_param_set(attr, default, minval, maxval) \
static int \
lpfc_##attr##_set(struct lpfc_vport *vport, uint val) \
{ \
if (lpfc_rangecheck(val, minval, maxval)) {\
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT, \
"3053 lpfc_" #attr \
" changed from %d (x%x) to %d (x%x)\n", \
vport->cfg_##attr, vport->cfg_##attr, \
val, val); \
vport->cfg_##attr = val;\
return 0;\
}\
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT, \
"0424 lpfc_"#attr" attribute cannot be set to %d, "\
"allowed range is ["#minval", "#maxval"]\n", val); \
return -EINVAL;\
}
/**
* lpfc_vport_param_store - Set a vport attribute
*
* Description:
* Macro that given an attr e.g. hba_queue_depth
* expands into a function with the name lpfc_hba_queue_depth_store
*
* lpfc_##attr##_store: convert the ascii text number to an integer, then
* use the lpfc_##attr##_set function to set the value.
* @cdev: class device that is converted into a Scsi_host.
* @buf: contains the attribute value in decimal.
* @count: not used.
*
* Returns:
* -EINVAL if val is invalid or lpfc_##attr##_set() fails
* length of buffer upon success.
**/
#define lpfc_vport_param_store(attr) \
static ssize_t \
lpfc_##attr##_store(struct device *dev, struct device_attribute *attr, \
const char *buf, size_t count) \
{ \
struct Scsi_Host *shost = class_to_shost(dev);\
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;\
uint val = 0;\
if (!isdigit(buf[0]))\
return -EINVAL;\
if (sscanf(buf, "%i", &val) != 1)\
return -EINVAL;\
if (lpfc_##attr##_set(vport, val) == 0) \
return strlen(buf);\
else \
return -EINVAL;\
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
static DEVICE_ATTR(nvme_info, 0444, lpfc_nvme_info_show, NULL);
static DEVICE_ATTR(bg_info, S_IRUGO, lpfc_bg_info_show, NULL);
static DEVICE_ATTR(bg_guard_err, S_IRUGO, lpfc_bg_guard_err_show, NULL);
static DEVICE_ATTR(bg_apptag_err, S_IRUGO, lpfc_bg_apptag_err_show, NULL);
static DEVICE_ATTR(bg_reftag_err, S_IRUGO, lpfc_bg_reftag_err_show, NULL);
static DEVICE_ATTR(info, S_IRUGO, lpfc_info_show, NULL);
static DEVICE_ATTR(serialnum, S_IRUGO, lpfc_serialnum_show, NULL);
static DEVICE_ATTR(modeldesc, S_IRUGO, lpfc_modeldesc_show, NULL);
static DEVICE_ATTR(modelname, S_IRUGO, lpfc_modelname_show, NULL);
static DEVICE_ATTR(programtype, S_IRUGO, lpfc_programtype_show, NULL);
static DEVICE_ATTR(portnum, S_IRUGO, lpfc_vportnum_show, NULL);
static DEVICE_ATTR(fwrev, S_IRUGO, lpfc_fwrev_show, NULL);
static DEVICE_ATTR(hdw, S_IRUGO, lpfc_hdw_show, NULL);
static DEVICE_ATTR(link_state, S_IRUGO | S_IWUSR, lpfc_link_state_show,
lpfc_link_state_store);
static DEVICE_ATTR(option_rom_version, S_IRUGO,
lpfc_option_rom_version_show, NULL);
static DEVICE_ATTR(num_discovered_ports, S_IRUGO,
lpfc_num_discovered_ports_show, NULL);
static DEVICE_ATTR(menlo_mgmt_mode, S_IRUGO, lpfc_mlomgmt_show, NULL);
static DEVICE_ATTR(nport_evt_cnt, S_IRUGO, lpfc_nport_evt_cnt_show, NULL);
static DEVICE_ATTR(lpfc_drvr_version, S_IRUGO, lpfc_drvr_version_show, NULL);
static DEVICE_ATTR(lpfc_enable_fip, S_IRUGO, lpfc_enable_fip_show, NULL);
static DEVICE_ATTR(board_mode, S_IRUGO | S_IWUSR,
lpfc_board_mode_show, lpfc_board_mode_store);
static DEVICE_ATTR(issue_reset, S_IWUSR, NULL, lpfc_issue_reset);
static DEVICE_ATTR(max_vpi, S_IRUGO, lpfc_max_vpi_show, NULL);
static DEVICE_ATTR(used_vpi, S_IRUGO, lpfc_used_vpi_show, NULL);
static DEVICE_ATTR(max_rpi, S_IRUGO, lpfc_max_rpi_show, NULL);
static DEVICE_ATTR(used_rpi, S_IRUGO, lpfc_used_rpi_show, NULL);
static DEVICE_ATTR(max_xri, S_IRUGO, lpfc_max_xri_show, NULL);
static DEVICE_ATTR(used_xri, S_IRUGO, lpfc_used_xri_show, NULL);
static DEVICE_ATTR(npiv_info, S_IRUGO, lpfc_npiv_info_show, NULL);
static DEVICE_ATTR(lpfc_temp_sensor, S_IRUGO, lpfc_temp_sensor_show, NULL);
static DEVICE_ATTR(lpfc_fips_level, S_IRUGO, lpfc_fips_level_show, NULL);
static DEVICE_ATTR(lpfc_fips_rev, S_IRUGO, lpfc_fips_rev_show, NULL);
static DEVICE_ATTR(lpfc_dss, S_IRUGO, lpfc_dss_show, NULL);
static DEVICE_ATTR(lpfc_sriov_hw_max_virtfn, S_IRUGO,
lpfc_sriov_hw_max_virtfn_show, NULL);
static DEVICE_ATTR(protocol, S_IRUGO, lpfc_sli4_protocol_show, NULL);
static DEVICE_ATTR(lpfc_xlane_supported, S_IRUGO, lpfc_oas_supported_show,
NULL);
static char *lpfc_soft_wwn_key = "C99G71SL8032A";
#define WWN_SZ 8
/**
* lpfc_wwn_set - Convert string to the 8 byte WWN value.
* @buf: WWN string.
* @cnt: Length of string.
* @wwn: Array to receive converted wwn value.
*
* Returns:
* -EINVAL if the buffer does not contain a valid wwn
* 0 success
**/
static size_t
lpfc_wwn_set(const char *buf, size_t cnt, char wwn[])
{
unsigned int i, j;
/* Count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
if ((cnt < 16) || (cnt > 18) || ((cnt == 17) && (*buf++ != 'x')) ||
((cnt == 18) && ((*buf++ != '0') || (*buf++ != 'x'))))
return -EINVAL;
memset(wwn, 0, WWN_SZ);
/* Validate and store the new name */
for (i = 0, j = 0; i < 16; i++) {
if ((*buf >= 'a') && (*buf <= 'f'))
j = ((j << 4) | ((*buf++ - 'a') + 10));
else if ((*buf >= 'A') && (*buf <= 'F'))
j = ((j << 4) | ((*buf++ - 'A') + 10));
else if ((*buf >= '0') && (*buf <= '9'))
j = ((j << 4) | (*buf++ - '0'));
else
return -EINVAL;
if (i % 2) {
wwn[i/2] = j & 0xff;
j = 0;
}
}
return 0;
}
/**
* lpfc_soft_wwn_enable_store - Allows setting of the wwn if the key is valid
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing the string lpfc_soft_wwn_key.
* @count: must be size of lpfc_soft_wwn_key.
*
* Returns:
* -EINVAL if the buffer does not contain lpfc_soft_wwn_key
* length of buf indicates success
**/
static ssize_t
lpfc_soft_wwn_enable_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
unsigned int cnt = count;
/*
* We're doing a simple sanity check for soft_wwpn setting.
* We require that the user write a specific key to enable
* the soft_wwpn attribute to be settable. Once the attribute
* is written, the enable key resets. If further updates are
* desired, the key must be written again to re-enable the
* attribute.
*
* The "key" is not secret - it is a hardcoded string shown
* here. The intent is to protect against the random user or
* application that is just writing attributes.
*/
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
if ((cnt != strlen(lpfc_soft_wwn_key)) ||
(strncmp(buf, lpfc_soft_wwn_key, strlen(lpfc_soft_wwn_key)) != 0))
return -EINVAL;
phba->soft_wwn_enable = 1;
dev_printk(KERN_WARNING, &phba->pcidev->dev,
"lpfc%d: soft_wwpn assignment has been enabled.\n",
phba->brd_no);
dev_printk(KERN_WARNING, &phba->pcidev->dev,
" The soft_wwpn feature is not supported by Broadcom.");
return count;
}
static DEVICE_ATTR(lpfc_soft_wwn_enable, S_IWUSR, NULL,
lpfc_soft_wwn_enable_store);
/**
* lpfc_soft_wwpn_show - Return the cfg soft ww port name of the adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the wwpn in hexadecimal.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_soft_wwpn_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
return snprintf(buf, PAGE_SIZE, "0x%llx\n",
(unsigned long long)phba->cfg_soft_wwpn);
}
/**
* lpfc_soft_wwpn_store - Set the ww port name of the adapter
* @dev class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: contains the wwpn in hexadecimal.
* @count: number of wwpn bytes in buf
*
* Returns:
* -EACCES hba reset not enabled, adapter over temp
* -EINVAL soft wwn not enabled, count is invalid, invalid wwpn byte invalid
* -EIO error taking adapter offline or online
* value of count on success
**/
static ssize_t
lpfc_soft_wwpn_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct completion online_compl;
int stat1 = 0, stat2 = 0;
unsigned int cnt = count;
u8 wwpn[WWN_SZ];
int rc;
if (!phba->cfg_enable_hba_reset)
return -EACCES;
spin_lock_irq(&phba->hbalock);
if (phba->over_temp_state == HBA_OVER_TEMP) {
spin_unlock_irq(&phba->hbalock);
return -EACCES;
}
spin_unlock_irq(&phba->hbalock);
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
if (!phba->soft_wwn_enable)
return -EINVAL;
/* lock setting wwpn, wwnn down */
phba->soft_wwn_enable = 0;
rc = lpfc_wwn_set(buf, cnt, wwpn);
if (rc) {
/* not able to set wwpn, unlock it */
phba->soft_wwn_enable = 1;
return rc;
}
phba->cfg_soft_wwpn = wwn_to_u64(wwpn);
fc_host_port_name(shost) = phba->cfg_soft_wwpn;
if (phba->cfg_soft_wwnn)
fc_host_node_name(shost) = phba->cfg_soft_wwnn;
dev_printk(KERN_NOTICE, &phba->pcidev->dev,
"lpfc%d: Reinitializing to use soft_wwpn\n", phba->brd_no);
stat1 = lpfc_do_offline(phba, LPFC_EVT_OFFLINE);
if (stat1)
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"0463 lpfc_soft_wwpn attribute set failed to "
"reinit adapter - %d\n", stat1);
init_completion(&online_compl);
rc = lpfc_workq_post_event(phba, &stat2, &online_compl,
LPFC_EVT_ONLINE);
if (rc == 0)
return -ENOMEM;
wait_for_completion(&online_compl);
if (stat2)
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"0464 lpfc_soft_wwpn attribute set failed to "
"reinit adapter - %d\n", stat2);
return (stat1 || stat2) ? -EIO : count;
}
static DEVICE_ATTR(lpfc_soft_wwpn, S_IRUGO | S_IWUSR,
lpfc_soft_wwpn_show, lpfc_soft_wwpn_store);
/**
* lpfc_soft_wwnn_show - Return the cfg soft ww node name for the adapter
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: on return contains the wwnn in hexadecimal.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_soft_wwnn_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "0x%llx\n",
(unsigned long long)phba->cfg_soft_wwnn);
}
/**
* lpfc_soft_wwnn_store - sets the ww node name of the adapter
* @cdev: class device that is converted into a Scsi_host.
* @buf: contains the ww node name in hexadecimal.
* @count: number of wwnn bytes in buf.
*
* Returns:
* -EINVAL soft wwn not enabled, count is invalid, invalid wwnn byte invalid
* value of count on success
**/
static ssize_t
lpfc_soft_wwnn_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
unsigned int cnt = count;
u8 wwnn[WWN_SZ];
int rc;
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
if (!phba->soft_wwn_enable)
return -EINVAL;
rc = lpfc_wwn_set(buf, cnt, wwnn);
if (rc) {
/* Allow wwnn to be set many times, as long as the enable
* is set. However, once the wwpn is set, everything locks.
*/
return rc;
}
phba->cfg_soft_wwnn = wwn_to_u64(wwnn);
dev_printk(KERN_NOTICE, &phba->pcidev->dev,
"lpfc%d: soft_wwnn set. Value will take effect upon "
"setting of the soft_wwpn\n", phba->brd_no);
return count;
}
static DEVICE_ATTR(lpfc_soft_wwnn, S_IRUGO | S_IWUSR,
lpfc_soft_wwnn_show, lpfc_soft_wwnn_store);
/**
* lpfc_oas_tgt_show - Return wwpn of target whose luns maybe enabled for
* Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* Returns:
* value of count
**/
static ssize_t
lpfc_oas_tgt_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "0x%llx\n",
wwn_to_u64(phba->cfg_oas_tgt_wwpn));
}
/**
* lpfc_oas_tgt_store - Store wwpn of target whose luns maybe enabled for
* Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
* @count: Size of the data buffer.
*
* Returns:
* -EINVAL count is invalid, invalid wwpn byte invalid
* -EPERM oas is not supported by hba
* value of count on success
**/
static ssize_t
lpfc_oas_tgt_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
unsigned int cnt = count;
uint8_t wwpn[WWN_SZ];
int rc;
if (!phba->cfg_fof)
return -EPERM;
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
rc = lpfc_wwn_set(buf, cnt, wwpn);
if (rc)
return rc;
memcpy(phba->cfg_oas_tgt_wwpn, wwpn, (8 * sizeof(uint8_t)));
memcpy(phba->sli4_hba.oas_next_tgt_wwpn, wwpn, (8 * sizeof(uint8_t)));
if (wwn_to_u64(wwpn) == 0)
phba->cfg_oas_flags |= OAS_FIND_ANY_TARGET;
else
phba->cfg_oas_flags &= ~OAS_FIND_ANY_TARGET;
phba->cfg_oas_flags &= ~OAS_LUN_VALID;
phba->sli4_hba.oas_next_lun = FIND_FIRST_OAS_LUN;
return count;
}
static DEVICE_ATTR(lpfc_xlane_tgt, S_IRUGO | S_IWUSR,
lpfc_oas_tgt_show, lpfc_oas_tgt_store);
/**
* lpfc_oas_priority_show - Return wwpn of target whose luns maybe enabled for
* Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* Returns:
* value of count
**/
static ssize_t
lpfc_oas_priority_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->cfg_oas_priority);
}
/**
* lpfc_oas_priority_store - Store wwpn of target whose luns maybe enabled for
* Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
* @count: Size of the data buffer.
*
* Returns:
* -EINVAL count is invalid, invalid wwpn byte invalid
* -EPERM oas is not supported by hba
* value of count on success
**/
static ssize_t
lpfc_oas_priority_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
unsigned int cnt = count;
unsigned long val;
int ret;
if (!phba->cfg_fof)
return -EPERM;
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
ret = kstrtoul(buf, 0, &val);
if (ret || (val > 0x7f))
return -EINVAL;
if (val)
phba->cfg_oas_priority = (uint8_t)val;
else
phba->cfg_oas_priority = phba->cfg_XLanePriority;
return count;
}
static DEVICE_ATTR(lpfc_xlane_priority, S_IRUGO | S_IWUSR,
lpfc_oas_priority_show, lpfc_oas_priority_store);
/**
* lpfc_oas_vpt_show - Return wwpn of vport whose targets maybe enabled
* for Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* Returns:
* value of count on success
**/
static ssize_t
lpfc_oas_vpt_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "0x%llx\n",
wwn_to_u64(phba->cfg_oas_vpt_wwpn));
}
/**
* lpfc_oas_vpt_store - Store wwpn of vport whose targets maybe enabled
* for Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
* @count: Size of the data buffer.
*
* Returns:
* -EINVAL count is invalid, invalid wwpn byte invalid
* -EPERM oas is not supported by hba
* value of count on success
**/
static ssize_t
lpfc_oas_vpt_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
unsigned int cnt = count;
uint8_t wwpn[WWN_SZ];
int rc;
if (!phba->cfg_fof)
return -EPERM;
/* count may include a LF at end of string */
if (buf[cnt-1] == '\n')
cnt--;
rc = lpfc_wwn_set(buf, cnt, wwpn);
if (rc)
return rc;
memcpy(phba->cfg_oas_vpt_wwpn, wwpn, (8 * sizeof(uint8_t)));
memcpy(phba->sli4_hba.oas_next_vpt_wwpn, wwpn, (8 * sizeof(uint8_t)));
if (wwn_to_u64(wwpn) == 0)
phba->cfg_oas_flags |= OAS_FIND_ANY_VPORT;
else
phba->cfg_oas_flags &= ~OAS_FIND_ANY_VPORT;
phba->cfg_oas_flags &= ~OAS_LUN_VALID;
if (phba->cfg_oas_priority == 0)
phba->cfg_oas_priority = phba->cfg_XLanePriority;
phba->sli4_hba.oas_next_lun = FIND_FIRST_OAS_LUN;
return count;
}
static DEVICE_ATTR(lpfc_xlane_vpt, S_IRUGO | S_IWUSR,
lpfc_oas_vpt_show, lpfc_oas_vpt_store);
/**
* lpfc_oas_lun_state_show - Return the current state (enabled or disabled)
* of whether luns will be enabled or disabled
* for Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* Returns:
* size of formatted string.
**/
static ssize_t
lpfc_oas_lun_state_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->cfg_oas_lun_state);
}
/**
* lpfc_oas_lun_state_store - Store the state (enabled or disabled)
* of whether luns will be enabled or disabled
* for Optimized Access Storage (OAS) operations.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
* @count: Size of the data buffer.
*
* Returns:
* -EINVAL count is invalid, invalid wwpn byte invalid
* -EPERM oas is not supported by hba
* value of count on success
**/
static ssize_t
lpfc_oas_lun_state_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
int val = 0;
if (!phba->cfg_fof)
return -EPERM;
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
if ((val != 0) && (val != 1))
return -EINVAL;
phba->cfg_oas_lun_state = val;
return strlen(buf);
}
static DEVICE_ATTR(lpfc_xlane_lun_state, S_IRUGO | S_IWUSR,
lpfc_oas_lun_state_show, lpfc_oas_lun_state_store);
/**
* lpfc_oas_lun_status_show - Return the status of the Optimized Access
* Storage (OAS) lun returned by the
* lpfc_oas_lun_show function.
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* Returns:
* size of formatted string.
**/
static ssize_t
lpfc_oas_lun_status_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
if (!(phba->cfg_oas_flags & OAS_LUN_VALID))
return -EFAULT;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->cfg_oas_lun_status);
}
static DEVICE_ATTR(lpfc_xlane_lun_status, S_IRUGO,
lpfc_oas_lun_status_show, NULL);
/**
* lpfc_oas_lun_state_set - enable or disable a lun for Optimized Access Storage
* (OAS) operations.
* @phba: lpfc_hba pointer.
* @ndlp: pointer to fcp target node.
* @lun: the fc lun for setting oas state.
* @oas_state: the oas state to be set to the lun.
*
* Returns:
* SUCCESS : 0
* -EPERM OAS is not enabled or not supported by this port.
*
*/
static size_t
lpfc_oas_lun_state_set(struct lpfc_hba *phba, uint8_t vpt_wwpn[],
uint8_t tgt_wwpn[], uint64_t lun,
uint32_t oas_state, uint8_t pri)
{
int rc = 0;
if (!phba->cfg_fof)
return -EPERM;
if (oas_state) {
if (!lpfc_enable_oas_lun(phba, (struct lpfc_name *)vpt_wwpn,
(struct lpfc_name *)tgt_wwpn,
lun, pri))
rc = -ENOMEM;
} else {
lpfc_disable_oas_lun(phba, (struct lpfc_name *)vpt_wwpn,
(struct lpfc_name *)tgt_wwpn, lun, pri);
}
return rc;
}
/**
* lpfc_oas_lun_get_next - get the next lun that has been enabled for Optimized
* Access Storage (OAS) operations.
* @phba: lpfc_hba pointer.
* @vpt_wwpn: wwpn of the vport associated with the returned lun
* @tgt_wwpn: wwpn of the target associated with the returned lun
* @lun_status: status of the lun returned lun
*
* Returns the first or next lun enabled for OAS operations for the vport/target
* specified. If a lun is found, its vport wwpn, target wwpn and status is
* returned. If the lun is not found, NOT_OAS_ENABLED_LUN is returned.
*
* Return:
* lun that is OAS enabled for the vport/target
* NOT_OAS_ENABLED_LUN when no oas enabled lun found.
*/
static uint64_t
lpfc_oas_lun_get_next(struct lpfc_hba *phba, uint8_t vpt_wwpn[],
uint8_t tgt_wwpn[], uint32_t *lun_status,
uint32_t *lun_pri)
{
uint64_t found_lun;
if (unlikely(!phba) || !vpt_wwpn || !tgt_wwpn)
return NOT_OAS_ENABLED_LUN;
if (lpfc_find_next_oas_lun(phba, (struct lpfc_name *)
phba->sli4_hba.oas_next_vpt_wwpn,
(struct lpfc_name *)
phba->sli4_hba.oas_next_tgt_wwpn,
&phba->sli4_hba.oas_next_lun,
(struct lpfc_name *)vpt_wwpn,
(struct lpfc_name *)tgt_wwpn,
&found_lun, lun_status, lun_pri))
return found_lun;
else
return NOT_OAS_ENABLED_LUN;
}
/**
* lpfc_oas_lun_state_change - enable/disable a lun for OAS operations
* @phba: lpfc_hba pointer.
* @vpt_wwpn: vport wwpn by reference.
* @tgt_wwpn: target wwpn by reference.
* @lun: the fc lun for setting oas state.
* @oas_state: the oas state to be set to the oas_lun.
*
* This routine enables (OAS_LUN_ENABLE) or disables (OAS_LUN_DISABLE)
* a lun for OAS operations.
*
* Return:
* SUCCESS: 0
* -ENOMEM: failed to enable an lun for OAS operations
* -EPERM: OAS is not enabled
*/
static ssize_t
lpfc_oas_lun_state_change(struct lpfc_hba *phba, uint8_t vpt_wwpn[],
uint8_t tgt_wwpn[], uint64_t lun,
uint32_t oas_state, uint8_t pri)
{
int rc;
rc = lpfc_oas_lun_state_set(phba, vpt_wwpn, tgt_wwpn, lun,
oas_state, pri);
return rc;
}
/**
* lpfc_oas_lun_show - Return oas enabled luns from a chosen target
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* This routine returns a lun enabled for OAS each time the function
* is called.
*
* Returns:
* SUCCESS: size of formatted string.
* -EFAULT: target or vport wwpn was not set properly.
* -EPERM: oas is not enabled.
**/
static ssize_t
lpfc_oas_lun_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
uint64_t oas_lun;
int len = 0;
if (!phba->cfg_fof)
return -EPERM;
if (wwn_to_u64(phba->cfg_oas_vpt_wwpn) == 0)
if (!(phba->cfg_oas_flags & OAS_FIND_ANY_VPORT))
return -EFAULT;
if (wwn_to_u64(phba->cfg_oas_tgt_wwpn) == 0)
if (!(phba->cfg_oas_flags & OAS_FIND_ANY_TARGET))
return -EFAULT;
oas_lun = lpfc_oas_lun_get_next(phba, phba->cfg_oas_vpt_wwpn,
phba->cfg_oas_tgt_wwpn,
&phba->cfg_oas_lun_status,
&phba->cfg_oas_priority);
if (oas_lun != NOT_OAS_ENABLED_LUN)
phba->cfg_oas_flags |= OAS_LUN_VALID;
len += snprintf(buf + len, PAGE_SIZE-len, "0x%llx", oas_lun);
return len;
}
/**
* lpfc_oas_lun_store - Sets the OAS state for lun
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: buffer for passing information.
*
* This function sets the OAS state for lun. Before this function is called,
* the vport wwpn, target wwpn, and oas state need to be set.
*
* Returns:
* SUCCESS: size of formatted string.
* -EFAULT: target or vport wwpn was not set properly.
* -EPERM: oas is not enabled.
* size of formatted string.
**/
static ssize_t
lpfc_oas_lun_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
uint64_t scsi_lun;
uint32_t pri;
ssize_t rc;
if (!phba->cfg_fof)
return -EPERM;
if (wwn_to_u64(phba->cfg_oas_vpt_wwpn) == 0)
return -EFAULT;
if (wwn_to_u64(phba->cfg_oas_tgt_wwpn) == 0)
return -EFAULT;
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "0x%llx", &scsi_lun) != 1)
return -EINVAL;
pri = phba->cfg_oas_priority;
if (pri == 0)
pri = phba->cfg_XLanePriority;
lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
"3372 Try to set vport 0x%llx target 0x%llx lun:0x%llx "
"priority 0x%x with oas state %d\n",
wwn_to_u64(phba->cfg_oas_vpt_wwpn),
wwn_to_u64(phba->cfg_oas_tgt_wwpn), scsi_lun,
pri, phba->cfg_oas_lun_state);
rc = lpfc_oas_lun_state_change(phba, phba->cfg_oas_vpt_wwpn,
phba->cfg_oas_tgt_wwpn, scsi_lun,
phba->cfg_oas_lun_state, pri);
if (rc)
return rc;
return count;
}
static DEVICE_ATTR(lpfc_xlane_lun, S_IRUGO | S_IWUSR,
lpfc_oas_lun_show, lpfc_oas_lun_store);
int lpfc_enable_nvmet_cnt;
unsigned long long lpfc_enable_nvmet[LPFC_NVMET_MAX_PORTS] = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
module_param_array(lpfc_enable_nvmet, ullong, &lpfc_enable_nvmet_cnt, 0444);
MODULE_PARM_DESC(lpfc_enable_nvmet, "Enable HBA port(s) WWPN as a NVME Target");
[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 04:32:13 +07:00
static int lpfc_poll = 0;
module_param(lpfc_poll, int, S_IRUGO);
[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 04:32:13 +07:00
MODULE_PARM_DESC(lpfc_poll, "FCP ring polling mode control:"
" 0 - none,"
" 1 - poll with interrupts enabled"
" 3 - poll and disable FCP ring interrupts");
static DEVICE_ATTR(lpfc_poll, S_IRUGO | S_IWUSR,
lpfc_poll_show, lpfc_poll_store);
scsi: lpfc: Fix eh_deadline setting for sli3 adapters. A previous change unilaterally removed the hba reset entry point from the sli3 host template. This was done to allow tape devices being used for back up from being removed. Why was this done ? When there was non-responding device on the fabric, the error escalation policy would escalate to the reset handler. When the reset handler was called, it would reset the adapter, dropping link, thus logging out and terminating all i/o's - on any target. If there was a tape device on the same adapter that wasn't in error, it would kill the tape i/o's, effectively killing the tape device state. With the reset point removed, the adapter reset avoided the fabric logout, allowing the other devices to continue to operate unaffected. A hack - yes. Hint: we really need a transport I_T nexus reset callback added to the eh process (in between the SCSI target reset and hba reset points), so a fc logout could occur to the one bad target only and stop the error escalation process. This patch commonizes the approach so it can be used for sli3 and sli4 adapters, but mandates the admin, via module parameter, specifically identify which adapters the resets are to be removed for. Additionally, bus_reset, which sends Target Reset TMFs to all targets, is also removed from the template as it too has the same effect as the adapter reset. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> Reviewed-by: Laurence Oberman <loberman@redhat.com> Tested-by: Laurence Oberman <loberman@redhat.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-03-05 00:30:31 +07:00
int lpfc_no_hba_reset_cnt;
unsigned long lpfc_no_hba_reset[MAX_HBAS_NO_RESET] = {
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0};
module_param_array(lpfc_no_hba_reset, ulong, &lpfc_no_hba_reset_cnt, 0444);
MODULE_PARM_DESC(lpfc_no_hba_reset, "WWPN of HBAs that should not be reset");
LPFC_ATTR(sli_mode, 0, 0, 3,
"SLI mode selector:"
" 0 - auto (SLI-3 if supported),"
" 2 - select SLI-2 even on SLI-3 capable HBAs,"
" 3 - select SLI-3");
LPFC_ATTR_R(enable_npiv, 1, 0, 1,
"Enable NPIV functionality");
LPFC_ATTR_R(fcf_failover_policy, 1, 1, 2,
"FCF Fast failover=1 Priority failover=2");
/*
# lpfc_enable_rrq: Track XRI/OXID reuse after IO failures
# 0x0 = disabled, XRI/OXID use not tracked.
# 0x1 = XRI/OXID reuse is timed with ratov, RRQ sent.
# 0x2 = XRI/OXID reuse is timed with ratov, No RRQ sent.
*/
LPFC_ATTR_R(enable_rrq, 2, 0, 2,
"Enable RRQ functionality");
/*
# lpfc_suppress_link_up: Bring link up at initialization
# 0x0 = bring link up (issue MBX_INIT_LINK)
# 0x1 = do NOT bring link up at initialization(MBX_INIT_LINK)
# 0x2 = never bring up link
# Default value is 0.
*/
LPFC_ATTR_R(suppress_link_up, LPFC_INITIALIZE_LINK, LPFC_INITIALIZE_LINK,
LPFC_DELAY_INIT_LINK_INDEFINITELY,
"Suppress Link Up at initialization");
/*
# lpfc_cnt: Number of IOCBs allocated for ELS, CT, and ABTS
# 1 - (1024)
# 2 - (2048)
# 3 - (3072)
# 4 - (4096)
# 5 - (5120)
*/
static ssize_t
lpfc_iocb_hw_show(struct device *dev, struct device_attribute *attr, char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *) shost->hostdata)->phba;
return snprintf(buf, PAGE_SIZE, "%d\n", phba->iocb_max);
}
static DEVICE_ATTR(iocb_hw, S_IRUGO,
lpfc_iocb_hw_show, NULL);
static ssize_t
lpfc_txq_hw_show(struct device *dev, struct device_attribute *attr, char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *) shost->hostdata)->phba;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
struct lpfc_sli_ring *pring = lpfc_phba_elsring(phba);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
return snprintf(buf, PAGE_SIZE, "%d\n", pring->txq_max);
}
static DEVICE_ATTR(txq_hw, S_IRUGO,
lpfc_txq_hw_show, NULL);
static ssize_t
lpfc_txcmplq_hw_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_hba *phba = ((struct lpfc_vport *) shost->hostdata)->phba;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
struct lpfc_sli_ring *pring = lpfc_phba_elsring(phba);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
return snprintf(buf, PAGE_SIZE, "%d\n", pring->txcmplq_max);
}
static DEVICE_ATTR(txcmplq_hw, S_IRUGO,
lpfc_txcmplq_hw_show, NULL);
LPFC_ATTR_R(iocb_cnt, 2, 1, 5,
"Number of IOCBs alloc for ELS, CT, and ABTS: 1k to 5k IOCBs");
/*
# lpfc_nodev_tmo: If set, it will hold all I/O errors on devices that disappear
# until the timer expires. Value range is [0,255]. Default value is 30.
*/
static int lpfc_nodev_tmo = LPFC_DEF_DEVLOSS_TMO;
static int lpfc_devloss_tmo = LPFC_DEF_DEVLOSS_TMO;
module_param(lpfc_nodev_tmo, int, 0);
MODULE_PARM_DESC(lpfc_nodev_tmo,
"Seconds driver will hold I/O waiting "
"for a device to come back");
/**
* lpfc_nodev_tmo_show - Return the hba dev loss timeout value
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains the dev loss timeout in decimal.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_nodev_tmo_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
return snprintf(buf, PAGE_SIZE, "%d\n", vport->cfg_devloss_tmo);
}
/**
* lpfc_nodev_tmo_init - Set the hba nodev timeout value
* @vport: lpfc vport structure pointer.
* @val: contains the nodev timeout value.
*
* Description:
* If the devloss tmo is already set then nodev tmo is set to devloss tmo,
* a kernel error message is printed and zero is returned.
* Else if val is in range then nodev tmo and devloss tmo are set to val.
* Otherwise nodev tmo is set to the default value.
*
* Returns:
* zero if already set or if val is in range
* -EINVAL val out of range
**/
static int
lpfc_nodev_tmo_init(struct lpfc_vport *vport, int val)
{
if (vport->cfg_devloss_tmo != LPFC_DEF_DEVLOSS_TMO) {
vport->cfg_nodev_tmo = vport->cfg_devloss_tmo;
if (val != LPFC_DEF_DEVLOSS_TMO)
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0407 Ignoring lpfc_nodev_tmo module "
"parameter because lpfc_devloss_tmo "
"is set.\n");
return 0;
}
if (val >= LPFC_MIN_DEVLOSS_TMO && val <= LPFC_MAX_DEVLOSS_TMO) {
vport->cfg_nodev_tmo = val;
vport->cfg_devloss_tmo = val;
return 0;
}
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0400 lpfc_nodev_tmo attribute cannot be set to"
" %d, allowed range is [%d, %d]\n",
val, LPFC_MIN_DEVLOSS_TMO, LPFC_MAX_DEVLOSS_TMO);
vport->cfg_nodev_tmo = LPFC_DEF_DEVLOSS_TMO;
return -EINVAL;
}
/**
* lpfc_update_rport_devloss_tmo - Update dev loss tmo value
* @vport: lpfc vport structure pointer.
*
* Description:
* Update all the ndlp's dev loss tmo with the vport devloss tmo value.
**/
static void
lpfc_update_rport_devloss_tmo(struct lpfc_vport *vport)
{
struct Scsi_Host *shost;
struct lpfc_nodelist *ndlp;
shost = lpfc_shost_from_vport(vport);
spin_lock_irq(shost->host_lock);
list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp)
if (NLP_CHK_NODE_ACT(ndlp) && ndlp->rport)
ndlp->rport->dev_loss_tmo = vport->cfg_devloss_tmo;
spin_unlock_irq(shost->host_lock);
}
/**
* lpfc_nodev_tmo_set - Set the vport nodev tmo and devloss tmo values
* @vport: lpfc vport structure pointer.
* @val: contains the tmo value.
*
* Description:
* If the devloss tmo is already set or the vport dev loss tmo has changed
* then a kernel error message is printed and zero is returned.
* Else if val is in range then nodev tmo and devloss tmo are set to val.
* Otherwise nodev tmo is set to the default value.
*
* Returns:
* zero if already set or if val is in range
* -EINVAL val out of range
**/
static int
lpfc_nodev_tmo_set(struct lpfc_vport *vport, int val)
{
if (vport->dev_loss_tmo_changed ||
(lpfc_devloss_tmo != LPFC_DEF_DEVLOSS_TMO)) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0401 Ignoring change to lpfc_nodev_tmo "
"because lpfc_devloss_tmo is set.\n");
return 0;
}
if (val >= LPFC_MIN_DEVLOSS_TMO && val <= LPFC_MAX_DEVLOSS_TMO) {
vport->cfg_nodev_tmo = val;
vport->cfg_devloss_tmo = val;
/*
* For compat: set the fc_host dev loss so new rports
* will get the value.
*/
fc_host_dev_loss_tmo(lpfc_shost_from_vport(vport)) = val;
lpfc_update_rport_devloss_tmo(vport);
return 0;
}
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0403 lpfc_nodev_tmo attribute cannot be set to "
"%d, allowed range is [%d, %d]\n",
val, LPFC_MIN_DEVLOSS_TMO, LPFC_MAX_DEVLOSS_TMO);
return -EINVAL;
}
lpfc_vport_param_store(nodev_tmo)
static DEVICE_ATTR(lpfc_nodev_tmo, S_IRUGO | S_IWUSR,
lpfc_nodev_tmo_show, lpfc_nodev_tmo_store);
/*
# lpfc_devloss_tmo: If set, it will hold all I/O errors on devices that
# disappear until the timer expires. Value range is [0,255]. Default
# value is 30.
*/
module_param(lpfc_devloss_tmo, int, S_IRUGO);
MODULE_PARM_DESC(lpfc_devloss_tmo,
"Seconds driver will hold I/O waiting "
"for a device to come back");
lpfc_vport_param_init(devloss_tmo, LPFC_DEF_DEVLOSS_TMO,
LPFC_MIN_DEVLOSS_TMO, LPFC_MAX_DEVLOSS_TMO)
lpfc_vport_param_show(devloss_tmo)
/**
* lpfc_devloss_tmo_set - Sets vport nodev tmo, devloss tmo values, changed bit
* @vport: lpfc vport structure pointer.
* @val: contains the tmo value.
*
* Description:
* If val is in a valid range then set the vport nodev tmo,
* devloss tmo, also set the vport dev loss tmo changed flag.
* Else a kernel error message is printed.
*
* Returns:
* zero if val is in range
* -EINVAL val out of range
**/
static int
lpfc_devloss_tmo_set(struct lpfc_vport *vport, int val)
{
if (val >= LPFC_MIN_DEVLOSS_TMO && val <= LPFC_MAX_DEVLOSS_TMO) {
vport->cfg_nodev_tmo = val;
vport->cfg_devloss_tmo = val;
vport->dev_loss_tmo_changed = 1;
fc_host_dev_loss_tmo(lpfc_shost_from_vport(vport)) = val;
lpfc_update_rport_devloss_tmo(vport);
return 0;
}
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0404 lpfc_devloss_tmo attribute cannot be set to "
"%d, allowed range is [%d, %d]\n",
val, LPFC_MIN_DEVLOSS_TMO, LPFC_MAX_DEVLOSS_TMO);
return -EINVAL;
}
lpfc_vport_param_store(devloss_tmo)
static DEVICE_ATTR(lpfc_devloss_tmo, S_IRUGO | S_IWUSR,
lpfc_devloss_tmo_show, lpfc_devloss_tmo_store);
/*
* lpfc_suppress_rsp: Enable suppress rsp feature is firmware supports it
* lpfc_suppress_rsp = 0 Disable
* lpfc_suppress_rsp = 1 Enable (default)
*
*/
LPFC_ATTR_R(suppress_rsp, 1, 0, 1,
"Enable suppress rsp feature is firmware supports it");
/*
* lpfc_nvmet_mrq: Specify number of RQ pairs for processing NVMET cmds
* lpfc_nvmet_mrq = 1 use a single RQ pair
* lpfc_nvmet_mrq >= 2 use specified RQ pairs for MRQ
*
*/
LPFC_ATTR_R(nvmet_mrq,
1, 1, 16,
"Specify number of RQ pairs for processing NVMET cmds");
/*
* lpfc_nvmet_mrq_post: Specify number buffers to post on every MRQ
*
*/
LPFC_ATTR_R(nvmet_mrq_post, LPFC_DEF_MRQ_POST,
LPFC_MIN_MRQ_POST, LPFC_MAX_MRQ_POST,
"Specify number of buffers to post on every MRQ");
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
/*
* lpfc_enable_fc4_type: Defines what FC4 types are supported.
* Supported Values: 1 - register just FCP
* 3 - register both FCP and NVME
* Supported values are [1,3]. Default value is 3
*/
LPFC_ATTR_R(enable_fc4_type, LPFC_ENABLE_BOTH,
LPFC_ENABLE_FCP, LPFC_ENABLE_BOTH,
"Define fc4 type to register with fabric.");
/*
* lpfc_xri_split: Defines the division of XRI resources between SCSI and NVME
* This parameter is only used if:
* lpfc_enable_fc4_type is 3 - register both FCP and NVME and
* port is not configured for NVMET.
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
*
* ELS/CT always get 10% of XRIs, up to a maximum of 250
* The remaining XRIs get split up based on lpfc_xri_split per port:
*
* Supported Values are in percentages
* the xri_split value is the percentage the SCSI port will get. The remaining
* percentage will go to NVME.
*/
LPFC_ATTR_R(xri_split, 50, 10, 90,
"Division of XRI resources between SCSI and NVME");
/*
# lpfc_log_verbose: Only turn this flag on if you are willing to risk being
# deluged with LOTS of information.
# You can set a bit mask to record specific types of verbose messages:
# See lpfc_logmsh.h for definitions.
*/
LPFC_VPORT_ATTR_HEX_RW(log_verbose, 0x0, 0x0, 0xffffffff,
"Verbose logging bit-mask");
/*
# lpfc_enable_da_id: This turns on the DA_ID CT command that deregisters
# objects that have been registered with the nameserver after login.
*/
LPFC_VPORT_ATTR_R(enable_da_id, 1, 0, 1,
"Deregister nameserver objects before LOGO");
/*
# lun_queue_depth: This parameter is used to limit the number of outstanding
# commands per FCP LUN. Value range is [1,512]. Default value is 30.
# If this parameter value is greater than 1/8th the maximum number of exchanges
# supported by the HBA port, then the lun queue depth will be reduced to
# 1/8th the maximum number of exchanges.
*/
LPFC_VPORT_ATTR_R(lun_queue_depth, 30, 1, 512,
"Max number of FCP commands we can queue to a specific LUN");
/*
# tgt_queue_depth: This parameter is used to limit the number of outstanding
# commands per target port. Value range is [10,65535]. Default value is 65535.
*/
LPFC_VPORT_ATTR_R(tgt_queue_depth, 65535, 10, 65535,
"Max number of FCP commands we can queue to a specific target port");
/*
# hba_queue_depth: This parameter is used to limit the number of outstanding
# commands per lpfc HBA. Value range is [32,8192]. If this parameter
# value is greater than the maximum number of exchanges supported by the HBA,
# then maximum number of exchanges supported by the HBA is used to determine
# the hba_queue_depth.
*/
LPFC_ATTR_R(hba_queue_depth, 8192, 32, 8192,
"Max number of FCP commands we can queue to a lpfc HBA");
/*
# peer_port_login: This parameter allows/prevents logins
# between peer ports hosted on the same physical port.
# When this parameter is set 0 peer ports of same physical port
# are not allowed to login to each other.
# When this parameter is set 1 peer ports of same physical port
# are allowed to login to each other.
# Default value of this parameter is 0.
*/
LPFC_VPORT_ATTR_R(peer_port_login, 0, 0, 1,
"Allow peer ports on the same physical port to login to each "
"other.");
/*
# restrict_login: This parameter allows/prevents logins
# between Virtual Ports and remote initiators.
# When this parameter is not set (0) Virtual Ports will accept PLOGIs from
# other initiators and will attempt to PLOGI all remote ports.
# When this parameter is set (1) Virtual Ports will reject PLOGIs from
# remote ports and will not attempt to PLOGI to other initiators.
# This parameter does not restrict to the physical port.
# This parameter does not restrict logins to Fabric resident remote ports.
# Default value of this parameter is 1.
*/
static int lpfc_restrict_login = 1;
module_param(lpfc_restrict_login, int, S_IRUGO);
MODULE_PARM_DESC(lpfc_restrict_login,
"Restrict virtual ports login to remote initiators.");
lpfc_vport_param_show(restrict_login);
/**
* lpfc_restrict_login_init - Set the vport restrict login flag
* @vport: lpfc vport structure pointer.
* @val: contains the restrict login value.
*
* Description:
* If val is not in a valid range then log a kernel error message and set
* the vport restrict login to one.
* If the port type is physical clear the restrict login flag and return.
* Else set the restrict login flag to val.
*
* Returns:
* zero if val is in range
* -EINVAL val out of range
**/
static int
lpfc_restrict_login_init(struct lpfc_vport *vport, int val)
{
if (val < 0 || val > 1) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0422 lpfc_restrict_login attribute cannot "
"be set to %d, allowed range is [0, 1]\n",
val);
vport->cfg_restrict_login = 1;
return -EINVAL;
}
if (vport->port_type == LPFC_PHYSICAL_PORT) {
vport->cfg_restrict_login = 0;
return 0;
}
vport->cfg_restrict_login = val;
return 0;
}
/**
* lpfc_restrict_login_set - Set the vport restrict login flag
* @vport: lpfc vport structure pointer.
* @val: contains the restrict login value.
*
* Description:
* If val is not in a valid range then log a kernel error message and set
* the vport restrict login to one.
* If the port type is physical and the val is not zero log a kernel
* error message, clear the restrict login flag and return zero.
* Else set the restrict login flag to val.
*
* Returns:
* zero if val is in range
* -EINVAL val out of range
**/
static int
lpfc_restrict_login_set(struct lpfc_vport *vport, int val)
{
if (val < 0 || val > 1) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0425 lpfc_restrict_login attribute cannot "
"be set to %d, allowed range is [0, 1]\n",
val);
vport->cfg_restrict_login = 1;
return -EINVAL;
}
if (vport->port_type == LPFC_PHYSICAL_PORT && val != 0) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"0468 lpfc_restrict_login must be 0 for "
"Physical ports.\n");
vport->cfg_restrict_login = 0;
return 0;
}
vport->cfg_restrict_login = val;
return 0;
}
lpfc_vport_param_store(restrict_login);
static DEVICE_ATTR(lpfc_restrict_login, S_IRUGO | S_IWUSR,
lpfc_restrict_login_show, lpfc_restrict_login_store);
/*
# Some disk devices have a "select ID" or "select Target" capability.
# From a protocol standpoint "select ID" usually means select the
# Fibre channel "ALPA". In the FC-AL Profile there is an "informative
# annex" which contains a table that maps a "select ID" (a number
# between 0 and 7F) to an ALPA. By default, for compatibility with
# older drivers, the lpfc driver scans this table from low ALPA to high
# ALPA.
#
# Turning on the scan-down variable (on = 1, off = 0) will
# cause the lpfc driver to use an inverted table, effectively
# scanning ALPAs from high to low. Value range is [0,1]. Default value is 1.
#
# (Note: This "select ID" functionality is a LOOP ONLY characteristic
# and will not work across a fabric. Also this parameter will take
# effect only in the case when ALPA map is not available.)
*/
LPFC_VPORT_ATTR_R(scan_down, 1, 0, 1,
"Start scanning for devices from highest ALPA to lowest");
/*
# lpfc_topology: link topology for init link
# 0x0 = attempt loop mode then point-to-point
# 0x01 = internal loopback mode
# 0x02 = attempt point-to-point mode only
# 0x04 = attempt loop mode only
# 0x06 = attempt point-to-point mode then loop
# Set point-to-point mode if you want to run as an N_Port.
# Set loop mode if you want to run as an NL_Port. Value range is [0,0x6].
# Default value is 0.
*/
LPFC_ATTR(topology, 0, 0, 6,
"Select Fibre Channel topology");
/**
* lpfc_topology_set - Set the adapters topology field
* @phba: lpfc_hba pointer.
* @val: topology value.
*
* Description:
* If val is in a valid range then set the adapter's topology field and
* issue a lip; if the lip fails reset the topology to the old value.
*
* If the value is not in range log a kernel error message and return an error.
*
* Returns:
* zero if val is in range and lip okay
* non-zero return value from lpfc_issue_lip()
* -EINVAL val out of range
**/
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
static ssize_t
lpfc_topology_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int val = 0;
int nolip = 0;
const char *val_buf = buf;
int err;
uint32_t prev_val;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (!strncmp(buf, "nolip ", strlen("nolip "))) {
nolip = 1;
val_buf = &buf[strlen("nolip ")];
}
if (!isdigit(val_buf[0]))
return -EINVAL;
if (sscanf(val_buf, "%i", &val) != 1)
return -EINVAL;
if (val >= 0 && val <= 6) {
prev_val = phba->cfg_topology;
if (phba->cfg_link_speed == LPFC_USER_LINK_SPEED_16G &&
val == 4) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3113 Loop mode not supported at speed %d\n",
val);
return -EINVAL;
}
if (phba->pcidev->device == PCI_DEVICE_ID_LANCER_G6_FC &&
val == 4) {
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3114 Loop mode not supported\n");
return -EINVAL;
}
phba->cfg_topology = val;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (nolip)
return strlen(buf);
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3054 lpfc_topology changed from %d to %d\n",
prev_val, val);
if (prev_val != val && phba->sli_rev == LPFC_SLI_REV4)
phba->fc_topology_changed = 1;
err = lpfc_issue_lip(lpfc_shost_from_vport(phba->pport));
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (err) {
phba->cfg_topology = prev_val;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
return -EINVAL;
} else
return strlen(buf);
}
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"%d:0467 lpfc_topology attribute cannot be set to %d, "
"allowed range is [0, 6]\n",
phba->brd_no, val);
return -EINVAL;
}
lpfc_param_show(topology)
static DEVICE_ATTR(lpfc_topology, S_IRUGO | S_IWUSR,
lpfc_topology_show, lpfc_topology_store);
/**
* lpfc_static_vport_show: Read callback function for
* lpfc_static_vport sysfs file.
* @dev: Pointer to class device object.
* @attr: device attribute structure.
* @buf: Data buffer.
*
* This function is the read call back function for
* lpfc_static_vport sysfs file. The lpfc_static_vport
* sysfs file report the mageability of the vport.
**/
static ssize_t
lpfc_static_vport_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
if (vport->vport_flag & STATIC_VPORT)
sprintf(buf, "1\n");
else
sprintf(buf, "0\n");
return strlen(buf);
}
/*
* Sysfs attribute to control the statistical data collection.
*/
static DEVICE_ATTR(lpfc_static_vport, S_IRUGO,
lpfc_static_vport_show, NULL);
/**
* lpfc_stat_data_ctrl_store - write call back for lpfc_stat_data_ctrl sysfs file
* @dev: Pointer to class device.
* @buf: Data buffer.
* @count: Size of the data buffer.
*
* This function get called when an user write to the lpfc_stat_data_ctrl
* sysfs file. This function parse the command written to the sysfs file
* and take appropriate action. These commands are used for controlling
* driver statistical data collection.
* Following are the command this function handles.
*
* setbucket <bucket_type> <base> <step>
* = Set the latency buckets.
* destroybucket = destroy all the buckets.
* start = start data collection
* stop = stop data collection
* reset = reset the collected data
**/
static ssize_t
lpfc_stat_data_ctrl_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
#define LPFC_MAX_DATA_CTRL_LEN 1024
static char bucket_data[LPFC_MAX_DATA_CTRL_LEN];
unsigned long i;
char *str_ptr, *token;
struct lpfc_vport **vports;
struct Scsi_Host *v_shost;
char *bucket_type_str, *base_str, *step_str;
unsigned long base, step, bucket_type;
if (!strncmp(buf, "setbucket", strlen("setbucket"))) {
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (strlen(buf) > (LPFC_MAX_DATA_CTRL_LEN - 1))
return -EINVAL;
strncpy(bucket_data, buf, LPFC_MAX_DATA_CTRL_LEN);
str_ptr = &bucket_data[0];
/* Ignore this token - this is command token */
token = strsep(&str_ptr, "\t ");
if (!token)
return -EINVAL;
bucket_type_str = strsep(&str_ptr, "\t ");
if (!bucket_type_str)
return -EINVAL;
if (!strncmp(bucket_type_str, "linear", strlen("linear")))
bucket_type = LPFC_LINEAR_BUCKET;
else if (!strncmp(bucket_type_str, "power2", strlen("power2")))
bucket_type = LPFC_POWER2_BUCKET;
else
return -EINVAL;
base_str = strsep(&str_ptr, "\t ");
if (!base_str)
return -EINVAL;
base = simple_strtoul(base_str, NULL, 0);
step_str = strsep(&str_ptr, "\t ");
if (!step_str)
return -EINVAL;
step = simple_strtoul(step_str, NULL, 0);
if (!step)
return -EINVAL;
/* Block the data collection for every vport */
vports = lpfc_create_vport_work_array(phba);
if (vports == NULL)
return -ENOMEM;
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
v_shost = lpfc_shost_from_vport(vports[i]);
spin_lock_irq(v_shost->host_lock);
/* Block and reset data collection */
vports[i]->stat_data_blocked = 1;
if (vports[i]->stat_data_enabled)
lpfc_vport_reset_stat_data(vports[i]);
spin_unlock_irq(v_shost->host_lock);
}
/* Set the bucket attributes */
phba->bucket_type = bucket_type;
phba->bucket_base = base;
phba->bucket_step = step;
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
v_shost = lpfc_shost_from_vport(vports[i]);
/* Unblock data collection */
spin_lock_irq(v_shost->host_lock);
vports[i]->stat_data_blocked = 0;
spin_unlock_irq(v_shost->host_lock);
}
lpfc_destroy_vport_work_array(phba, vports);
return strlen(buf);
}
if (!strncmp(buf, "destroybucket", strlen("destroybucket"))) {
vports = lpfc_create_vport_work_array(phba);
if (vports == NULL)
return -ENOMEM;
for (i = 0; i <= phba->max_vports && vports[i] != NULL; i++) {
v_shost = lpfc_shost_from_vport(vports[i]);
spin_lock_irq(shost->host_lock);
vports[i]->stat_data_blocked = 1;
lpfc_free_bucket(vport);
vport->stat_data_enabled = 0;
vports[i]->stat_data_blocked = 0;
spin_unlock_irq(shost->host_lock);
}
lpfc_destroy_vport_work_array(phba, vports);
phba->bucket_type = LPFC_NO_BUCKET;
phba->bucket_base = 0;
phba->bucket_step = 0;
return strlen(buf);
}
if (!strncmp(buf, "start", strlen("start"))) {
/* If no buckets configured return error */
if (phba->bucket_type == LPFC_NO_BUCKET)
return -EINVAL;
spin_lock_irq(shost->host_lock);
if (vport->stat_data_enabled) {
spin_unlock_irq(shost->host_lock);
return strlen(buf);
}
lpfc_alloc_bucket(vport);
vport->stat_data_enabled = 1;
spin_unlock_irq(shost->host_lock);
return strlen(buf);
}
if (!strncmp(buf, "stop", strlen("stop"))) {
spin_lock_irq(shost->host_lock);
if (vport->stat_data_enabled == 0) {
spin_unlock_irq(shost->host_lock);
return strlen(buf);
}
lpfc_free_bucket(vport);
vport->stat_data_enabled = 0;
spin_unlock_irq(shost->host_lock);
return strlen(buf);
}
if (!strncmp(buf, "reset", strlen("reset"))) {
if ((phba->bucket_type == LPFC_NO_BUCKET)
|| !vport->stat_data_enabled)
return strlen(buf);
spin_lock_irq(shost->host_lock);
vport->stat_data_blocked = 1;
lpfc_vport_reset_stat_data(vport);
vport->stat_data_blocked = 0;
spin_unlock_irq(shost->host_lock);
return strlen(buf);
}
return -EINVAL;
}
/**
* lpfc_stat_data_ctrl_show - Read function for lpfc_stat_data_ctrl sysfs file
* @dev: Pointer to class device object.
* @buf: Data buffer.
*
* This function is the read call back function for
* lpfc_stat_data_ctrl sysfs file. This function report the
* current statistical data collection state.
**/
static ssize_t
lpfc_stat_data_ctrl_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int index = 0;
int i;
char *bucket_type;
unsigned long bucket_value;
switch (phba->bucket_type) {
case LPFC_LINEAR_BUCKET:
bucket_type = "linear";
break;
case LPFC_POWER2_BUCKET:
bucket_type = "power2";
break;
default:
bucket_type = "No Bucket";
break;
}
sprintf(&buf[index], "Statistical Data enabled :%d, "
"blocked :%d, Bucket type :%s, Bucket base :%d,"
" Bucket step :%d\nLatency Ranges :",
vport->stat_data_enabled, vport->stat_data_blocked,
bucket_type, phba->bucket_base, phba->bucket_step);
index = strlen(buf);
if (phba->bucket_type != LPFC_NO_BUCKET) {
for (i = 0; i < LPFC_MAX_BUCKET_COUNT; i++) {
if (phba->bucket_type == LPFC_LINEAR_BUCKET)
bucket_value = phba->bucket_base +
phba->bucket_step * i;
else
bucket_value = phba->bucket_base +
(1 << i) * phba->bucket_step;
if (index + 10 > PAGE_SIZE)
break;
sprintf(&buf[index], "%08ld ", bucket_value);
index = strlen(buf);
}
}
sprintf(&buf[index], "\n");
return strlen(buf);
}
/*
* Sysfs attribute to control the statistical data collection.
*/
static DEVICE_ATTR(lpfc_stat_data_ctrl, S_IRUGO | S_IWUSR,
lpfc_stat_data_ctrl_show, lpfc_stat_data_ctrl_store);
/*
* lpfc_drvr_stat_data: sysfs attr to get driver statistical data.
*/
/*
* Each Bucket takes 11 characters and 1 new line + 17 bytes WWN
* for each target.
*/
#define STAT_DATA_SIZE_PER_TARGET(NUM_BUCKETS) ((NUM_BUCKETS) * 11 + 18)
#define MAX_STAT_DATA_SIZE_PER_TARGET \
STAT_DATA_SIZE_PER_TARGET(LPFC_MAX_BUCKET_COUNT)
/**
* sysfs_drvr_stat_data_read - Read function for lpfc_drvr_stat_data attribute
* @filp: sysfs file
* @kobj: Pointer to the kernel object
* @bin_attr: Attribute object
* @buff: Buffer pointer
* @off: File offset
* @count: Buffer size
*
* This function is the read call back function for lpfc_drvr_stat_data
* sysfs file. This function export the statistical data to user
* applications.
**/
static ssize_t
sysfs_drvr_stat_data_read(struct file *filp, struct kobject *kobj,
struct bin_attribute *bin_attr,
char *buf, loff_t off, size_t count)
{
struct device *dev = container_of(kobj, struct device,
kobj);
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int i = 0, index = 0;
unsigned long nport_index;
struct lpfc_nodelist *ndlp = NULL;
nport_index = (unsigned long)off /
MAX_STAT_DATA_SIZE_PER_TARGET;
if (!vport->stat_data_enabled || vport->stat_data_blocked
|| (phba->bucket_type == LPFC_NO_BUCKET))
return 0;
spin_lock_irq(shost->host_lock);
list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) {
if (!NLP_CHK_NODE_ACT(ndlp) || !ndlp->lat_data)
continue;
if (nport_index > 0) {
nport_index--;
continue;
}
if ((index + MAX_STAT_DATA_SIZE_PER_TARGET)
> count)
break;
if (!ndlp->lat_data)
continue;
/* Print the WWN */
sprintf(&buf[index], "%02x%02x%02x%02x%02x%02x%02x%02x:",
ndlp->nlp_portname.u.wwn[0],
ndlp->nlp_portname.u.wwn[1],
ndlp->nlp_portname.u.wwn[2],
ndlp->nlp_portname.u.wwn[3],
ndlp->nlp_portname.u.wwn[4],
ndlp->nlp_portname.u.wwn[5],
ndlp->nlp_portname.u.wwn[6],
ndlp->nlp_portname.u.wwn[7]);
index = strlen(buf);
for (i = 0; i < LPFC_MAX_BUCKET_COUNT; i++) {
sprintf(&buf[index], "%010u,",
ndlp->lat_data[i].cmd_count);
index = strlen(buf);
}
sprintf(&buf[index], "\n");
index = strlen(buf);
}
spin_unlock_irq(shost->host_lock);
return index;
}
static struct bin_attribute sysfs_drvr_stat_data_attr = {
.attr = {
.name = "lpfc_drvr_stat_data",
.mode = S_IRUSR,
},
.size = LPFC_MAX_TARGET * MAX_STAT_DATA_SIZE_PER_TARGET,
.read = sysfs_drvr_stat_data_read,
.write = NULL,
};
/*
# lpfc_link_speed: Link speed selection for initializing the Fibre Channel
# connection.
2010-11-21 11:11:48 +07:00
# Value range is [0,16]. Default value is 0.
*/
/**
* lpfc_link_speed_set - Set the adapters link speed
* @phba: lpfc_hba pointer.
* @val: link speed value.
*
* Description:
* If val is in a valid range then set the adapter's link speed field and
* issue a lip; if the lip fails reset the link speed to the old value.
*
* Notes:
* If the value is not in range log a kernel error message and return an error.
*
* Returns:
* zero if val is in range and lip okay.
* non-zero return value from lpfc_issue_lip()
* -EINVAL val out of range
**/
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
static ssize_t
lpfc_link_speed_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
2010-11-21 11:11:48 +07:00
int val = LPFC_USER_LINK_SPEED_AUTO;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
int nolip = 0;
const char *val_buf = buf;
int err;
uint32_t prev_val, if_type;
if_type = bf_get(lpfc_sli_intf_if_type, &phba->sli4_hba.sli_intf);
if (if_type == LPFC_SLI_INTF_IF_TYPE_2 &&
phba->hba_flag & HBA_FORCED_LINK_SPEED)
return -EPERM;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (!strncmp(buf, "nolip ", strlen("nolip "))) {
nolip = 1;
val_buf = &buf[strlen("nolip ")];
}
if (!isdigit(val_buf[0]))
return -EINVAL;
if (sscanf(val_buf, "%i", &val) != 1)
return -EINVAL;
lpfc_printf_vlog(vport, KERN_ERR, LOG_INIT,
"3055 lpfc_link_speed changed from %d to %d %s\n",
phba->cfg_link_speed, val, nolip ? "(nolip)" : "(lip)");
2010-11-21 11:11:48 +07:00
if (((val == LPFC_USER_LINK_SPEED_1G) && !(phba->lmt & LMT_1Gb)) ||
((val == LPFC_USER_LINK_SPEED_2G) && !(phba->lmt & LMT_2Gb)) ||
((val == LPFC_USER_LINK_SPEED_4G) && !(phba->lmt & LMT_4Gb)) ||
((val == LPFC_USER_LINK_SPEED_8G) && !(phba->lmt & LMT_8Gb)) ||
((val == LPFC_USER_LINK_SPEED_10G) && !(phba->lmt & LMT_10Gb)) ||
((val == LPFC_USER_LINK_SPEED_16G) && !(phba->lmt & LMT_16Gb)) ||
((val == LPFC_USER_LINK_SPEED_32G) && !(phba->lmt & LMT_32Gb))) {
2010-11-21 11:11:48 +07:00
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"2879 lpfc_link_speed attribute cannot be set "
"to %d. Speed is not supported by this port.\n",
val);
return -EINVAL;
2010-11-21 11:11:48 +07:00
}
if (val == LPFC_USER_LINK_SPEED_16G &&
phba->fc_topology == LPFC_TOPOLOGY_LOOP) {
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3112 lpfc_link_speed attribute cannot be set "
"to %d. Speed is not supported in loop mode.\n",
val);
return -EINVAL;
}
2010-11-21 11:11:48 +07:00
if ((val >= 0) && (val <= LPFC_USER_LINK_SPEED_MAX) &&
(LPFC_USER_LINK_SPEED_BITMAP & (1 << val))) {
prev_val = phba->cfg_link_speed;
phba->cfg_link_speed = val;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (nolip)
return strlen(buf);
err = lpfc_issue_lip(lpfc_shost_from_vport(phba->pport));
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
if (err) {
phba->cfg_link_speed = prev_val;
[SCSI] lpfc 8.3.1: misc fixes/changes 8.3.1 Fixes/Changes : - Fix incorrect byte-swapping on word 4 of IOCB (data length) which caused LUNs to not be discovered on big-endian (e.g. PPC) - Remove a bad cast of MBslimaddr which loses the __iomem (sparse) - Make lpfc_debugfs_mask_disc_trc static (sparse) - Correct misspelled word BlockGuard in lpfc_logmsg.h comment - Replaced repeated code segment for canceling IOCBs from a list with a function call, lpfc_sli_cancel_iocbs(). - Increased HBQ buffers to support 40KB SSC sequences. - Added sysfs interface to update speed and topology parameter without link bounce. - Fixed bug with sysfs fc_host WWNs not being updated after changing the WWNs. - Check if the active mailbox is NULL in the beginning of the mailbox timeout handler - fixes panic in the mailbox timeout handler while running IO stress test - Fixed system panic in lpfc_pci_remove_one() due to ndlp indirect reference to phba through vport - Removed de-reference of scsi device after call to scsi_done() to fix panic in scsi completion path while accessing scsi device after scsi_done is called. - Fixed "Nodelist not empty" message when unloading the driver after target reboot test - Added LP2105 HBA model description - Added code to print all 16 words of unrecognized ASYNC events - Fixed memory leak in vport create + delete loop - Added support for handling dual error bit from HBA - Fixed a driver NULL pointer dereference in lpfc_sli_process_sol_iocb - Fixed a discovery bug with FC switch reboot in lpfc_setup_disc_node - Take NULL termintator into account when calculating available buffer space Signed-off-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2009-04-07 05:48:10 +07:00
return -EINVAL;
} else
return strlen(buf);
}
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
2010-11-21 11:11:48 +07:00
"0469 lpfc_link_speed attribute cannot be set to %d, "
"allowed values are ["LPFC_LINK_SPEED_STRING"]\n", val);
return -EINVAL;
}
static int lpfc_link_speed = 0;
module_param(lpfc_link_speed, int, S_IRUGO);
MODULE_PARM_DESC(lpfc_link_speed, "Select link speed");
lpfc_param_show(link_speed)
/**
* lpfc_link_speed_init - Set the adapters link speed
* @phba: lpfc_hba pointer.
* @val: link speed value.
*
* Description:
* If val is in a valid range then set the adapter's link speed field.
*
* Notes:
* If the value is not in range log a kernel error message, clear the link
* speed and return an error.
*
* Returns:
* zero if val saved.
* -EINVAL val out of range
**/
static int
lpfc_link_speed_init(struct lpfc_hba *phba, int val)
{
if (val == LPFC_USER_LINK_SPEED_16G && phba->cfg_topology == 4) {
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3111 lpfc_link_speed of %d cannot "
"support loop mode, setting topology to default.\n",
val);
phba->cfg_topology = 0;
}
2010-11-21 11:11:48 +07:00
if ((val >= 0) && (val <= LPFC_USER_LINK_SPEED_MAX) &&
(LPFC_USER_LINK_SPEED_BITMAP & (1 << val))) {
phba->cfg_link_speed = val;
return 0;
}
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"0405 lpfc_link_speed attribute cannot "
"be set to %d, allowed values are "
"["LPFC_LINK_SPEED_STRING"]\n", val);
2010-11-21 11:11:48 +07:00
phba->cfg_link_speed = LPFC_USER_LINK_SPEED_AUTO;
return -EINVAL;
}
static DEVICE_ATTR(lpfc_link_speed, S_IRUGO | S_IWUSR,
2010-11-21 11:11:48 +07:00
lpfc_link_speed_show, lpfc_link_speed_store);
/*
# lpfc_aer_support: Support PCIe device Advanced Error Reporting (AER)
# 0 = aer disabled or not supported
# 1 = aer supported and enabled (default)
# Value range is [0,1]. Default value is 1.
*/
LPFC_ATTR(aer_support, 1, 0, 1,
"Enable PCIe device AER support");
lpfc_param_show(aer_support)
/**
* lpfc_aer_support_store - Set the adapter for aer support
*
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing enable or disable aer flag.
* @count: unused variable.
*
* Description:
* If the val is 1 and currently the device's AER capability was not
* enabled, invoke the kernel's enable AER helper routine, trying to
* enable the device's AER capability. If the helper routine enabling
* AER returns success, update the device's cfg_aer_support flag to
* indicate AER is supported by the device; otherwise, if the device
* AER capability is already enabled to support AER, then do nothing.
*
* If the val is 0 and currently the device's AER support was enabled,
* invoke the kernel's disable AER helper routine. After that, update
* the device's cfg_aer_support flag to indicate AER is not supported
* by the device; otherwise, if the device AER capability is already
* disabled from supporting AER, then do nothing.
*
* Returns:
* length of the buf on success if val is in range the intended mode
* is supported.
* -EINVAL if val out of range or intended mode is not supported.
**/
static ssize_t
lpfc_aer_support_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int val = 0, rc = -EINVAL;
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
switch (val) {
case 0:
if (phba->hba_flag & HBA_AER_ENABLED) {
rc = pci_disable_pcie_error_reporting(phba->pcidev);
if (!rc) {
spin_lock_irq(&phba->hbalock);
phba->hba_flag &= ~HBA_AER_ENABLED;
spin_unlock_irq(&phba->hbalock);
phba->cfg_aer_support = 0;
rc = strlen(buf);
} else
rc = -EPERM;
} else {
phba->cfg_aer_support = 0;
rc = strlen(buf);
}
break;
case 1:
if (!(phba->hba_flag & HBA_AER_ENABLED)) {
rc = pci_enable_pcie_error_reporting(phba->pcidev);
if (!rc) {
spin_lock_irq(&phba->hbalock);
phba->hba_flag |= HBA_AER_ENABLED;
spin_unlock_irq(&phba->hbalock);
phba->cfg_aer_support = 1;
rc = strlen(buf);
} else
rc = -EPERM;
} else {
phba->cfg_aer_support = 1;
rc = strlen(buf);
}
break;
default:
rc = -EINVAL;
break;
}
return rc;
}
static DEVICE_ATTR(lpfc_aer_support, S_IRUGO | S_IWUSR,
lpfc_aer_support_show, lpfc_aer_support_store);
/**
* lpfc_aer_cleanup_state - Clean up aer state to the aer enabled device
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing flag 1 for aer cleanup state.
* @count: unused variable.
*
* Description:
* If the @buf contains 1 and the device currently has the AER support
* enabled, then invokes the kernel AER helper routine
* pci_cleanup_aer_uncorrect_error_status to clean up the uncorrectable
* error status register.
*
* Notes:
*
* Returns:
* -EINVAL if the buf does not contain the 1 or the device is not currently
* enabled with the AER support.
**/
static ssize_t
lpfc_aer_cleanup_state(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int val, rc = -1;
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
if (val != 1)
return -EINVAL;
if (phba->hba_flag & HBA_AER_ENABLED)
rc = pci_cleanup_aer_uncorrect_error_status(phba->pcidev);
if (rc == 0)
return strlen(buf);
else
return -EPERM;
}
static DEVICE_ATTR(lpfc_aer_state_cleanup, S_IWUSR, NULL,
lpfc_aer_cleanup_state);
/**
* lpfc_sriov_nr_virtfn_store - Enable the adapter for sr-iov virtual functions
*
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing the string the number of vfs to be enabled.
* @count: unused variable.
*
* Description:
* When this api is called either through user sysfs, the driver shall
* try to enable or disable SR-IOV virtual functions according to the
* following:
*
* If zero virtual function has been enabled to the physical function,
* the driver shall invoke the pci enable virtual function api trying
* to enable the virtual functions. If the nr_vfn provided is greater
* than the maximum supported, the maximum virtual function number will
* be used for invoking the api; otherwise, the nr_vfn provided shall
* be used for invoking the api. If the api call returned success, the
* actual number of virtual functions enabled will be set to the driver
* cfg_sriov_nr_virtfn; otherwise, -EINVAL shall be returned and driver
* cfg_sriov_nr_virtfn remains zero.
*
* If none-zero virtual functions have already been enabled to the
* physical function, as reflected by the driver's cfg_sriov_nr_virtfn,
* -EINVAL will be returned and the driver does nothing;
*
* If the nr_vfn provided is zero and none-zero virtual functions have
* been enabled, as indicated by the driver's cfg_sriov_nr_virtfn, the
* disabling virtual function api shall be invoded to disable all the
* virtual functions and driver's cfg_sriov_nr_virtfn shall be set to
* zero. Otherwise, if zero virtual function has been enabled, do
* nothing.
*
* Returns:
* length of the buf on success if val is in range the intended mode
* is supported.
* -EINVAL if val out of range or intended mode is not supported.
**/
static ssize_t
lpfc_sriov_nr_virtfn_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct pci_dev *pdev = phba->pcidev;
int val = 0, rc = -EINVAL;
/* Sanity check on user data */
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
if (val < 0)
return -EINVAL;
/* Request disabling virtual functions */
if (val == 0) {
if (phba->cfg_sriov_nr_virtfn > 0) {
pci_disable_sriov(pdev);
phba->cfg_sriov_nr_virtfn = 0;
}
return strlen(buf);
}
/* Request enabling virtual functions */
if (phba->cfg_sriov_nr_virtfn > 0) {
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3018 There are %d virtual functions "
"enabled on physical function.\n",
phba->cfg_sriov_nr_virtfn);
return -EEXIST;
}
if (val <= LPFC_MAX_VFN_PER_PFN)
phba->cfg_sriov_nr_virtfn = val;
else {
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3019 Enabling %d virtual functions is not "
"allowed.\n", val);
return -EINVAL;
}
rc = lpfc_sli_probe_sriov_nr_virtfn(phba, phba->cfg_sriov_nr_virtfn);
if (rc) {
phba->cfg_sriov_nr_virtfn = 0;
rc = -EPERM;
} else
rc = strlen(buf);
return rc;
}
LPFC_ATTR(sriov_nr_virtfn, LPFC_DEF_VFN_PER_PFN, 0, LPFC_MAX_VFN_PER_PFN,
"Enable PCIe device SR-IOV virtual fn");
lpfc_param_show(sriov_nr_virtfn)
static DEVICE_ATTR(lpfc_sriov_nr_virtfn, S_IRUGO | S_IWUSR,
lpfc_sriov_nr_virtfn_show, lpfc_sriov_nr_virtfn_store);
/**
* lpfc_request_firmware_store - Request for Linux generic firmware upgrade
*
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: containing the string the number of vfs to be enabled.
* @count: unused variable.
*
* Description:
*
* Returns:
* length of the buf on success if val is in range the intended mode
* is supported.
* -EINVAL if val out of range or intended mode is not supported.
**/
static ssize_t
lpfc_request_firmware_upgrade_store(struct device *dev,
struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int val = 0, rc = -EINVAL;
/* Sanity check on user data */
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
if (val != 1)
return -EINVAL;
rc = lpfc_sli4_request_firmware_update(phba, RUN_FW_UPGRADE);
if (rc)
rc = -EPERM;
else
rc = strlen(buf);
return rc;
}
static int lpfc_req_fw_upgrade;
module_param(lpfc_req_fw_upgrade, int, S_IRUGO|S_IWUSR);
MODULE_PARM_DESC(lpfc_req_fw_upgrade, "Enable Linux generic firmware upgrade");
lpfc_param_show(request_firmware_upgrade)
/**
* lpfc_request_firmware_upgrade_init - Enable initial linux generic fw upgrade
* @phba: lpfc_hba pointer.
* @val: 0 or 1.
*
* Description:
* Set the initial Linux generic firmware upgrade enable or disable flag.
*
* Returns:
* zero if val saved.
* -EINVAL val out of range
**/
static int
lpfc_request_firmware_upgrade_init(struct lpfc_hba *phba, int val)
{
if (val >= 0 && val <= 1) {
phba->cfg_request_firmware_upgrade = val;
return 0;
}
return -EINVAL;
}
static DEVICE_ATTR(lpfc_req_fw_upgrade, S_IRUGO | S_IWUSR,
lpfc_request_firmware_upgrade_show,
lpfc_request_firmware_upgrade_store);
/**
* lpfc_fcp_imax_store
*
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: string with the number of fast-path FCP interrupts per second.
* @count: unused variable.
*
* Description:
* If val is in a valid range [636,651042], then set the adapter's
* maximum number of fast-path FCP interrupts per second.
*
* Returns:
* length of the buf on success if val is in range the intended mode
* is supported.
* -EINVAL if val out of range or intended mode is not supported.
**/
static ssize_t
lpfc_fcp_imax_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
int val = 0, i;
/* fcp_imax is only valid for SLI4 */
if (phba->sli_rev != LPFC_SLI_REV4)
return -EINVAL;
/* Sanity check on user data */
if (!isdigit(buf[0]))
return -EINVAL;
if (sscanf(buf, "%i", &val) != 1)
return -EINVAL;
/*
* Value range for the HBA is [5000,5000000]
* The value for each EQ depends on how many EQs are configured.
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
* Allow value == 0
*/
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (val && (val < LPFC_MIN_IMAX || val > LPFC_MAX_IMAX))
return -EINVAL;
phba->cfg_fcp_imax = (uint32_t)val;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
for (i = 0; i < phba->io_channel_irqs; i++)
lpfc_modify_hba_eq_delay(phba, i);
return strlen(buf);
}
/*
# lpfc_fcp_imax: The maximum number of fast-path FCP interrupts per second
# for the HBA.
#
# Value range is [5,000 to 5,000,000]. Default value is 50,000.
*/
static int lpfc_fcp_imax = LPFC_DEF_IMAX;
module_param(lpfc_fcp_imax, int, S_IRUGO|S_IWUSR);
MODULE_PARM_DESC(lpfc_fcp_imax,
"Set the maximum number of FCP interrupts per second per HBA");
lpfc_param_show(fcp_imax)
/**
* lpfc_fcp_imax_init - Set the initial sr-iov virtual function enable
* @phba: lpfc_hba pointer.
* @val: link speed value.
*
* Description:
* If val is in a valid range [636,651042], then initialize the adapter's
* maximum number of fast-path FCP interrupts per second.
*
* Returns:
* zero if val saved.
* -EINVAL val out of range
**/
static int
lpfc_fcp_imax_init(struct lpfc_hba *phba, int val)
{
if (phba->sli_rev != LPFC_SLI_REV4) {
phba->cfg_fcp_imax = 0;
return 0;
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if ((val >= LPFC_MIN_IMAX && val <= LPFC_MAX_IMAX) ||
(val == 0)) {
phba->cfg_fcp_imax = val;
return 0;
}
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3016 lpfc_fcp_imax: %d out of range, using default\n",
val);
phba->cfg_fcp_imax = LPFC_DEF_IMAX;
return 0;
}
static DEVICE_ATTR(lpfc_fcp_imax, S_IRUGO | S_IWUSR,
lpfc_fcp_imax_show, lpfc_fcp_imax_store);
/**
* lpfc_state_show - Display current driver CPU affinity
* @dev: class converted to a Scsi_host structure.
* @attr: device attribute, not used.
* @buf: on return contains text describing the state of the link.
*
* Returns: size of formatted string.
**/
static ssize_t
lpfc_fcp_cpu_map_show(struct device *dev, struct device_attribute *attr,
char *buf)
{
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct lpfc_vector_map_info *cpup;
int len = 0;
if ((phba->sli_rev != LPFC_SLI_REV4) ||
(phba->intr_type != MSIX))
return len;
switch (phba->cfg_fcp_cpu_map) {
case 0:
len += snprintf(buf + len, PAGE_SIZE-len,
"fcp_cpu_map: No mapping (%d)\n",
phba->cfg_fcp_cpu_map);
return len;
case 1:
len += snprintf(buf + len, PAGE_SIZE-len,
"fcp_cpu_map: HBA centric mapping (%d): "
"%d online CPUs\n",
phba->cfg_fcp_cpu_map,
phba->sli4_hba.num_online_cpu);
break;
case 2:
len += snprintf(buf + len, PAGE_SIZE-len,
"fcp_cpu_map: Driver centric mapping (%d): "
"%d online CPUs\n",
phba->cfg_fcp_cpu_map,
phba->sli4_hba.num_online_cpu);
break;
}
while (phba->sli4_hba.curr_disp_cpu < phba->sli4_hba.num_present_cpu) {
cpup = &phba->sli4_hba.cpu_map[phba->sli4_hba.curr_disp_cpu];
/* margin should fit in this and the truncated message */
if (cpup->irq == LPFC_VECTOR_MAP_EMPTY)
len += snprintf(buf + len, PAGE_SIZE-len,
"CPU %02d io_chan %02d "
"physid %d coreid %d\n",
phba->sli4_hba.curr_disp_cpu,
cpup->channel_id, cpup->phys_id,
cpup->core_id);
else
len += snprintf(buf + len, PAGE_SIZE-len,
"CPU %02d io_chan %02d "
"physid %d coreid %d IRQ %d\n",
phba->sli4_hba.curr_disp_cpu,
cpup->channel_id, cpup->phys_id,
cpup->core_id, cpup->irq);
phba->sli4_hba.curr_disp_cpu++;
/* display max number of CPUs keeping some margin */
if (phba->sli4_hba.curr_disp_cpu <
phba->sli4_hba.num_present_cpu &&
(len >= (PAGE_SIZE - 64))) {
len += snprintf(buf + len, PAGE_SIZE-len, "more...\n");
break;
}
}
if (phba->sli4_hba.curr_disp_cpu == phba->sli4_hba.num_present_cpu)
phba->sli4_hba.curr_disp_cpu = 0;
return len;
}
/**
* lpfc_fcp_cpu_map_store - Change CPU affinity of driver vectors
* @dev: class device that is converted into a Scsi_host.
* @attr: device attribute, not used.
* @buf: one or more lpfc_polling_flags values.
* @count: not used.
*
* Returns:
* -EINVAL - Not implemented yet.
**/
static ssize_t
lpfc_fcp_cpu_map_store(struct device *dev, struct device_attribute *attr,
const char *buf, size_t count)
{
int status = -EINVAL;
return status;
}
/*
# lpfc_fcp_cpu_map: Defines how to map CPUs to IRQ vectors
# for the HBA.
#
# Value range is [0 to 2]. Default value is LPFC_DRIVER_CPU_MAP (2).
# 0 - Do not affinitze IRQ vectors
# 1 - Affintize HBA vectors with respect to each HBA
# (start with CPU0 for each HBA)
# 2 - Affintize HBA vectors with respect to the entire driver
# (round robin thru all CPUs across all HBAs)
*/
static int lpfc_fcp_cpu_map = LPFC_DRIVER_CPU_MAP;
module_param(lpfc_fcp_cpu_map, int, S_IRUGO|S_IWUSR);
MODULE_PARM_DESC(lpfc_fcp_cpu_map,
"Defines how to map CPUs to IRQ vectors per HBA");
/**
* lpfc_fcp_cpu_map_init - Set the initial sr-iov virtual function enable
* @phba: lpfc_hba pointer.
* @val: link speed value.
*
* Description:
* If val is in a valid range [0-2], then affinitze the adapter's
* MSIX vectors.
*
* Returns:
* zero if val saved.
* -EINVAL val out of range
**/
static int
lpfc_fcp_cpu_map_init(struct lpfc_hba *phba, int val)
{
if (phba->sli_rev != LPFC_SLI_REV4) {
phba->cfg_fcp_cpu_map = 0;
return 0;
}
if (val >= LPFC_MIN_CPU_MAP && val <= LPFC_MAX_CPU_MAP) {
phba->cfg_fcp_cpu_map = val;
return 0;
}
lpfc_printf_log(phba, KERN_ERR, LOG_INIT,
"3326 lpfc_fcp_cpu_map: %d out of range, using "
"default\n", val);
phba->cfg_fcp_cpu_map = LPFC_DRIVER_CPU_MAP;
return 0;
}
static DEVICE_ATTR(lpfc_fcp_cpu_map, S_IRUGO | S_IWUSR,
lpfc_fcp_cpu_map_show, lpfc_fcp_cpu_map_store);
/*
# lpfc_fcp_class: Determines FC class to use for the FCP protocol.
# Value range is [2,3]. Default value is 3.
*/
LPFC_VPORT_ATTR_R(fcp_class, 3, 2, 3,
"Select Fibre Channel class of service for FCP sequences");
/*
# lpfc_use_adisc: Use ADISC for FCP rediscovery instead of PLOGI. Value range
# is [0,1]. Default value is 0.
*/
LPFC_VPORT_ATTR_RW(use_adisc, 0, 0, 1,
"Use ADISC on rediscovery to authenticate FCP devices");
/*
# lpfc_first_burst_size: First burst size to use on the NPorts
# that support first burst.
# Value range is [0,65536]. Default value is 0.
*/
LPFC_VPORT_ATTR_RW(first_burst_size, 0, 0, 65536,
"First burst size for Targets that support first burst");
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
/*
* lpfc_nvmet_fb_size: NVME Target mode supported first burst size.
* When the driver is configured as an NVME target, this value is
* communicated to the NVME initiator in the PRLI response. It is
* used only when the lpfc_nvme_enable_fb and lpfc_nvmet_support
* parameters are set and the target is sending the PRLI RSP.
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
* Parameter supported on physical port only - no NPIV support.
* Value range is [0,65536]. Default value is 0.
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
*/
LPFC_ATTR_RW(nvmet_fb_size, 0, 0, 65536,
"NVME Target mode first burst size in 512B increments.");
/*
* lpfc_nvme_enable_fb: Enable NVME first burst on I and T functions.
* For the Initiator (I), enabling this parameter means that an NVMET
* PRLI response with FBA enabled and an FB_SIZE set to a nonzero value will be
* processed by the initiator for subsequent NVME FCP IO. For the target
* function (T), enabling this parameter qualifies the lpfc_nvmet_fb_size
* driver parameter as the target function's first burst size returned to the
* initiator in the target's NVME PRLI response. Parameter supported on physical
* port only - no NPIV support.
* Value range is [0,1]. Default value is 0 (disabled).
*/
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
LPFC_ATTR_RW(nvme_enable_fb, 0, 0, 1,
"Enable First Burst feature on I and T functions.");
/*
# lpfc_max_scsicmpl_time: Use scsi command completion time to control I/O queue
# depth. Default value is 0. When the value of this parameter is zero the
# SCSI command completion time is not used for controlling I/O queue depth. When
# the parameter is set to a non-zero value, the I/O queue depth is controlled
# to limit the I/O completion time to the parameter value.
# The value is set in milliseconds.
*/
LPFC_VPORT_ATTR(max_scsicmpl_time, 0, 0, 60000,
"Use command completion time to control queue depth");
lpfc_vport_param_show(max_scsicmpl_time);
static int
lpfc_max_scsicmpl_time_set(struct lpfc_vport *vport, int val)
{
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
struct lpfc_nodelist *ndlp, *next_ndlp;
if (val == vport->cfg_max_scsicmpl_time)
return 0;
if ((val < 0) || (val > 60000))
return -EINVAL;
vport->cfg_max_scsicmpl_time = val;
spin_lock_irq(shost->host_lock);
list_for_each_entry_safe(ndlp, next_ndlp, &vport->fc_nodes, nlp_listp) {
if (!NLP_CHK_NODE_ACT(ndlp))
continue;
if (ndlp->nlp_state == NLP_STE_UNUSED_NODE)
continue;
ndlp->cmd_qdepth = vport->cfg_tgt_queue_depth;
}
spin_unlock_irq(shost->host_lock);
return 0;
}
lpfc_vport_param_store(max_scsicmpl_time);
static DEVICE_ATTR(lpfc_max_scsicmpl_time, S_IRUGO | S_IWUSR,
lpfc_max_scsicmpl_time_show,
lpfc_max_scsicmpl_time_store);
/*
# lpfc_ack0: Use ACK0, instead of ACK1 for class 2 acknowledgement. Value
# range is [0,1]. Default value is 0.
*/
LPFC_ATTR_R(ack0, 0, 0, 1, "Enable ACK0 support");
/*
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
* lpfc_io_sched: Determine scheduling algrithmn for issuing FCP cmds
* range is [0,1]. Default value is 0.
* For [0], FCP commands are issued to Work Queues ina round robin fashion.
* For [1], FCP commands are issued to a Work Queue associated with the
* current CPU.
*
* LPFC_FCP_SCHED_ROUND_ROBIN == 0
* LPFC_FCP_SCHED_BY_CPU == 1
*
* The driver dynamically sets this to 1 (BY_CPU) if it's able to set up cpu
* affinity for FCP/NVME I/Os through Work Queues associated with the current
* CPU. Otherwise, the default 0 (Round Robin) scheduling of FCP/NVME I/Os
* through WQs will be used.
*/
LPFC_ATTR_RW(fcp_io_sched, LPFC_FCP_SCHED_ROUND_ROBIN,
LPFC_FCP_SCHED_ROUND_ROBIN,
LPFC_FCP_SCHED_BY_CPU,
"Determine scheduling algorithm for "
"issuing commands [0] - Round Robin, [1] - Current CPU");
/*
# lpfc_fcp2_no_tgt_reset: Determine bus reset behavior
# range is [0,1]. Default value is 0.
# For [0], bus reset issues target reset to ALL devices
# For [1], bus reset issues target reset to non-FCP2 devices
*/
LPFC_ATTR_RW(fcp2_no_tgt_reset, 0, 0, 1, "Determine bus reset behavior for "
"FCP2 devices [0] - issue tgt reset, [1] - no tgt reset");
/*
# lpfc_cr_delay & lpfc_cr_count: Default values for I/O colaesing
# cr_delay (msec) or cr_count outstanding commands. cr_delay can take
# value [0,63]. cr_count can take value [1,255]. Default value of cr_delay
# is 0. Default value of cr_count is 1. The cr_count feature is disabled if
# cr_delay is set to 0.
*/
LPFC_ATTR_RW(cr_delay, 0, 0, 63, "A count of milliseconds after which an "
"interrupt response is generated");
LPFC_ATTR_RW(cr_count, 1, 1, 255, "A count of I/O completions after which an "
"interrupt response is generated");
/*
# lpfc_multi_ring_support: Determines how many rings to spread available
# cmd/rsp IOCB entries across.
# Value range is [1,2]. Default value is 1.
*/
LPFC_ATTR_R(multi_ring_support, 1, 1, 2, "Determines number of primary "
"SLI rings to spread IOCB entries across");
/*
# lpfc_multi_ring_rctl: If lpfc_multi_ring_support is enabled, this
# identifies what rctl value to configure the additional ring for.
# Value range is [1,0xff]. Default value is 4 (Unsolicated Data).
*/
LPFC_ATTR_R(multi_ring_rctl, FC_RCTL_DD_UNSOL_DATA, 1,
255, "Identifies RCTL for additional ring configuration");
/*
# lpfc_multi_ring_type: If lpfc_multi_ring_support is enabled, this
# identifies what type value to configure the additional ring for.
# Value range is [1,0xff]. Default value is 5 (LLC/SNAP).
*/
LPFC_ATTR_R(multi_ring_type, FC_TYPE_IP, 1,
255, "Identifies TYPE for additional ring configuration");
/*
# lpfc_enable_SmartSAN: Sets up FDMI support for SmartSAN
# 0 = SmartSAN functionality disabled (default)
# 1 = SmartSAN functionality enabled
# This parameter will override the value of lpfc_fdmi_on module parameter.
# Value range is [0,1]. Default value is 0.
*/
LPFC_ATTR_R(enable_SmartSAN, 0, 0, 1, "Enable SmartSAN functionality");
/*
# lpfc_fdmi_on: Controls FDMI support.
# 0 No FDMI support (default)
# 1 Traditional FDMI support
# Traditional FDMI support means the driver will assume FDMI-2 support;
# however, if that fails, it will fallback to FDMI-1.
# If lpfc_enable_SmartSAN is set to 1, the driver ignores lpfc_fdmi_on.
# If lpfc_enable_SmartSAN is set 0, the driver uses the current value of
# lpfc_fdmi_on.
# Value range [0,1]. Default value is 0.
*/
LPFC_ATTR_R(fdmi_on, 0, 0, 1, "Enable FDMI support");
/*
# Specifies the maximum number of ELS cmds we can have outstanding (for
# discovery). Value range is [1,64]. Default value = 32.
*/
LPFC_VPORT_ATTR(discovery_threads, 32, 1, 64, "Maximum number of ELS commands "
"during discovery");
/*
# lpfc_max_luns: maximum allowed LUN ID. This is the highest LUN ID that
# will be scanned by the SCSI midlayer when sequential scanning is
# used; and is also the highest LUN ID allowed when the SCSI midlayer
# parses REPORT_LUN responses. The lpfc driver has no LUN count or
# LUN ID limit, but the SCSI midlayer requires this field for the uses
# above. The lpfc driver limits the default value to 255 for two reasons.
# As it bounds the sequential scan loop, scanning for thousands of luns
# on a target can take minutes of wall clock time. Additionally,
# there are FC targets, such as JBODs, that only recognize 8-bits of
# LUN ID. When they receive a value greater than 8 bits, they chop off
# the high order bits. In other words, they see LUN IDs 0, 256, 512,
# and so on all as LUN ID 0. This causes the linux kernel, which sees
# valid responses at each of the LUN IDs, to believe there are multiple
# devices present, when in fact, there is only 1.
# A customer that is aware of their target behaviors, and the results as
# indicated above, is welcome to increase the lpfc_max_luns value.
# As mentioned, this value is not used by the lpfc driver, only the
# SCSI midlayer.
# Value range is [0,65535]. Default value is 255.
# NOTE: The SCSI layer might probe all allowed LUN on some old targets.
*/
LPFC_VPORT_ULL_ATTR_R(max_luns, 255, 0, 65535, "Maximum allowed LUN ID");
[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 04:32:13 +07:00
/*
# lpfc_poll_tmo: .Milliseconds driver will wait between polling FCP ring.
# Value range is [1,255], default value is 10.
*/
LPFC_ATTR_RW(poll_tmo, 10, 1, 255,
"Milliseconds driver will wait between polling FCP ring");
/*
# lpfc_task_mgmt_tmo: Maximum time to wait for task management commands
# to complete in seconds. Value range is [5,180], default value is 60.
*/
LPFC_ATTR_RW(task_mgmt_tmo, 60, 5, 180,
"Maximum time to wait for task management commands to complete");
/*
# lpfc_use_msi: Use MSI (Message Signaled Interrupts) in systems that
# support this feature
[SCSI] lpfc: restore MSI-X/MSI support A Gentoo bug report [1] showed that as of 2.6.31 lpfc only uses INTx interrupts. This patch restores lpfc's ability to support MSI-X/MSI interrupts that the "Addition of SLI4 Interface - Base Support" patch [2] broke. It reestablishes MSI-X as the default interrupt method and in case MSI-X is not supported lpfc_sli{4,}_enable_intr fallbacks to MSI and then to INTx. [1]: http://bugs.gentoo.org/show_bug.cgi?id=296319 [2]: commit da0436e915a5c17ee79e72c1bf978a4ebb1cbf4d [James Smart: Background: Nothing Broke. This was intended. We had originally enabled MSI-X by default, but in qualification within the last 12 months, we encountered a major catch-22: There were at least 4 platforms, from 2 major OEMs, that : - Say they support MSI-X - platform routines work and act as if they do. - We enable it, generate a test interrupt to check they really do deliver it, and it works. - But shortly after attachment, the system hangs or loses interrupts, resulting in a bad system behavior. Given the distro's picking up the 2.6.32 kernel, we had to stick with a default of MSI-X off, with user-enabled MSI-X as these platforms couldn't get fixed. However, we're also now encountering platforms that require MSI-X and never INTx, so we must change. It's desired also for also for performance reasons. So - now (2.6.33) is the right time to re-enable MSI-X by default. ] [jejb: fix up comment on default values] Signed-off-by: George Kadianakis <desnacked@gmail.com> Acked-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-01-18 02:19:31 +07:00
# 0 = MSI disabled
# 1 = MSI enabled
[SCSI] lpfc: restore MSI-X/MSI support A Gentoo bug report [1] showed that as of 2.6.31 lpfc only uses INTx interrupts. This patch restores lpfc's ability to support MSI-X/MSI interrupts that the "Addition of SLI4 Interface - Base Support" patch [2] broke. It reestablishes MSI-X as the default interrupt method and in case MSI-X is not supported lpfc_sli{4,}_enable_intr fallbacks to MSI and then to INTx. [1]: http://bugs.gentoo.org/show_bug.cgi?id=296319 [2]: commit da0436e915a5c17ee79e72c1bf978a4ebb1cbf4d [James Smart: Background: Nothing Broke. This was intended. We had originally enabled MSI-X by default, but in qualification within the last 12 months, we encountered a major catch-22: There were at least 4 platforms, from 2 major OEMs, that : - Say they support MSI-X - platform routines work and act as if they do. - We enable it, generate a test interrupt to check they really do deliver it, and it works. - But shortly after attachment, the system hangs or loses interrupts, resulting in a bad system behavior. Given the distro's picking up the 2.6.32 kernel, we had to stick with a default of MSI-X off, with user-enabled MSI-X as these platforms couldn't get fixed. However, we're also now encountering platforms that require MSI-X and never INTx, so we must change. It's desired also for also for performance reasons. So - now (2.6.33) is the right time to re-enable MSI-X by default. ] [jejb: fix up comment on default values] Signed-off-by: George Kadianakis <desnacked@gmail.com> Acked-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-01-18 02:19:31 +07:00
# 2 = MSI-X enabled (default)
# Value range is [0,2]. Default value is 2.
*/
[SCSI] lpfc: restore MSI-X/MSI support A Gentoo bug report [1] showed that as of 2.6.31 lpfc only uses INTx interrupts. This patch restores lpfc's ability to support MSI-X/MSI interrupts that the "Addition of SLI4 Interface - Base Support" patch [2] broke. It reestablishes MSI-X as the default interrupt method and in case MSI-X is not supported lpfc_sli{4,}_enable_intr fallbacks to MSI and then to INTx. [1]: http://bugs.gentoo.org/show_bug.cgi?id=296319 [2]: commit da0436e915a5c17ee79e72c1bf978a4ebb1cbf4d [James Smart: Background: Nothing Broke. This was intended. We had originally enabled MSI-X by default, but in qualification within the last 12 months, we encountered a major catch-22: There were at least 4 platforms, from 2 major OEMs, that : - Say they support MSI-X - platform routines work and act as if they do. - We enable it, generate a test interrupt to check they really do deliver it, and it works. - But shortly after attachment, the system hangs or loses interrupts, resulting in a bad system behavior. Given the distro's picking up the 2.6.32 kernel, we had to stick with a default of MSI-X off, with user-enabled MSI-X as these platforms couldn't get fixed. However, we're also now encountering platforms that require MSI-X and never INTx, so we must change. It's desired also for also for performance reasons. So - now (2.6.33) is the right time to re-enable MSI-X by default. ] [jejb: fix up comment on default values] Signed-off-by: George Kadianakis <desnacked@gmail.com> Acked-by: James Smart <james.smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@suse.de>
2010-01-18 02:19:31 +07:00
LPFC_ATTR_R(use_msi, 2, 0, 2, "Use Message Signaled Interrupts (1) or "
"MSI-X (2), if possible");
/*
* lpfc_nvme_oas: Use the oas bit when sending NVME/NVMET IOs
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
*
* 0 = NVME OAS disabled
* 1 = NVME OAS enabled
*
* Value range is [0,1]. Default value is 0.
*/
LPFC_ATTR_RW(nvme_oas, 0, 0, 1,
"Use OAS bit on NVME IOs");
/*
* lpfc_fcp_io_channel: Set the number of FCP IO channels the driver
* will advertise it supports to the SCSI layer. This also will map to
* the number of WQs the driver will create.
*
* 0 = Configure the number of io channels to the number of active CPUs.
* 1,32 = Manually specify how many io channels to use.
*
* Value range is [0,32]. Default value is 4.
*/
LPFC_ATTR_R(fcp_io_channel,
LPFC_FCP_IO_CHAN_DEF,
LPFC_HBA_IO_CHAN_MIN, LPFC_HBA_IO_CHAN_MAX,
"Set the number of FCP I/O channels");
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
/*
* lpfc_nvme_io_channel: Set the number of IO hardware queues the driver
* will advertise it supports to the NVME layer. This also will map to
* the number of WQs the driver will create.
*
* This module parameter is valid when lpfc_enable_fc4_type is set
* to support NVME.
*
* The NVME Layer will try to create this many, plus 1 administrative
* hardware queue. The administrative queue will always map to WQ 0
* A hardware IO queue maps (qidx) to a specific driver WQ.
*
* 0 = Configure the number of io channels to the number of active CPUs.
* 1,32 = Manually specify how many io channels to use.
*
* Value range is [0,32]. Default value is 0.
*/
LPFC_ATTR_R(nvme_io_channel,
LPFC_NVME_IO_CHAN_DEF,
LPFC_HBA_IO_CHAN_MIN, LPFC_HBA_IO_CHAN_MAX,
"Set the number of NVME I/O channels");
/*
# lpfc_enable_hba_reset: Allow or prevent HBA resets to the hardware.
# 0 = HBA resets disabled
# 1 = HBA resets enabled (default)
# Value range is [0,1]. Default value is 1.
*/
LPFC_ATTR_R(enable_hba_reset, 1, 0, 1, "Enable HBA resets from the driver.");
/*
# lpfc_enable_hba_heartbeat: Disable HBA heartbeat timer..
# 0 = HBA Heartbeat disabled
# 1 = HBA Heartbeat enabled (default)
# Value range is [0,1]. Default value is 1.
*/
LPFC_ATTR_R(enable_hba_heartbeat, 0, 0, 1, "Enable HBA Heartbeat.");
/*
# lpfc_EnableXLane: Enable Express Lane Feature
# 0x0 Express Lane Feature disabled
# 0x1 Express Lane Feature enabled
# Value range is [0,1]. Default value is 0.
*/
LPFC_ATTR_R(EnableXLane, 0, 0, 1, "Enable Express Lane Feature.");
/*
# lpfc_XLanePriority: Define CS_CTL priority for Express Lane Feature
# 0x0 - 0x7f = CS_CTL field in FC header (high 7 bits)
# Value range is [0x0,0x7f]. Default value is 0
*/
LPFC_ATTR_RW(XLanePriority, 0, 0x0, 0x7f, "CS_CTL for Express Lane Feature.");
/*
# lpfc_enable_bg: Enable BlockGuard (Emulex's Implementation of T10-DIF)
# 0 = BlockGuard disabled (default)
# 1 = BlockGuard enabled
# Value range is [0,1]. Default value is 0.
*/
LPFC_ATTR_R(enable_bg, 0, 0, 1, "Enable BlockGuard Support");
/*
# lpfc_fcp_look_ahead: Look ahead for completions in FCP start routine
# 0 = disabled (default)
# 1 = enabled
# Value range is [0,1]. Default value is 0.
#
# This feature in under investigation and may be supported in the future.
*/
unsigned int lpfc_fcp_look_ahead = LPFC_LOOK_AHEAD_OFF;
/*
# lpfc_prot_mask: i
# - Bit mask of host protection capabilities used to register with the
# SCSI mid-layer
# - Only meaningful if BG is turned on (lpfc_enable_bg=1).
# - Allows you to ultimately specify which profiles to use
# - Default will result in registering capabilities for all profiles.
# - SHOST_DIF_TYPE1_PROTECTION 1
# HBA supports T10 DIF Type 1: HBA to Target Type 1 Protection
# - SHOST_DIX_TYPE0_PROTECTION 8
# HBA supports DIX Type 0: Host to HBA protection only
# - SHOST_DIX_TYPE1_PROTECTION 16
# HBA supports DIX Type 1: Host to HBA Type 1 protection
#
*/
LPFC_ATTR(prot_mask,
(SHOST_DIF_TYPE1_PROTECTION |
SHOST_DIX_TYPE0_PROTECTION |
SHOST_DIX_TYPE1_PROTECTION),
0,
(SHOST_DIF_TYPE1_PROTECTION |
SHOST_DIX_TYPE0_PROTECTION |
SHOST_DIX_TYPE1_PROTECTION),
"T10-DIF host protection capabilities mask");
/*
# lpfc_prot_guard: i
# - Bit mask of protection guard types to register with the SCSI mid-layer
# - Guard types are currently either 1) T10-DIF CRC 2) IP checksum
# - Allows you to ultimately specify which profiles to use
# - Default will result in registering capabilities for all guard types
#
*/
LPFC_ATTR(prot_guard,
SHOST_DIX_GUARD_IP, SHOST_DIX_GUARD_CRC, SHOST_DIX_GUARD_IP,
"T10-DIF host protection guard type");
/*
* Delay initial NPort discovery when Clean Address bit is cleared in
* FLOGI/FDISC accept and FCID/Fabric name/Fabric portname is changed.
* This parameter can have value 0 or 1.
* When this parameter is set to 0, no delay is added to the initial
* discovery.
* When this parameter is set to non-zero value, initial Nport discovery is
* delayed by ra_tov seconds when Clean Address bit is cleared in FLOGI/FDISC
* accept and FCID/Fabric name/Fabric portname is changed.
* Driver always delay Nport discovery for subsequent FLOGI/FDISC completion
* when Clean Address bit is cleared in FLOGI/FDISC
* accept and FCID/Fabric name/Fabric portname is changed.
* Default value is 0.
*/
LPFC_ATTR(delay_discovery, 0, 0, 1,
"Delay NPort discovery when Clean Address bit is cleared.");
/*
* lpfc_sg_seg_cnt - Initial Maximum DMA Segment Count
* This value can be set to values between 64 and 4096. The default value is
* 64, but may be increased to allow for larger Max I/O sizes. The scsi layer
* will be allowed to request I/Os of sizes up to (MAX_SEG_COUNT * SEG_SIZE).
* Because of the additional overhead involved in setting up T10-DIF,
* this parameter will be limited to 128 if BlockGuard is enabled under SLI4
* and will be limited to 512 if BlockGuard is enabled under SLI3.
*/
LPFC_ATTR_R(sg_seg_cnt, LPFC_DEFAULT_SG_SEG_CNT, LPFC_DEFAULT_SG_SEG_CNT,
LPFC_MAX_SG_SEG_CNT, "Max Scatter Gather Segment Count");
/*
* lpfc_enable_mds_diags: Enable MDS Diagnostics
* 0 = MDS Diagnostics disabled (default)
* 1 = MDS Diagnostics enabled
* Value range is [0,1]. Default value is 0.
*/
LPFC_ATTR_R(enable_mds_diags, 0, 0, 1, "Enable MDS Diagnostics");
struct device_attribute *lpfc_hba_attrs[] = {
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
&dev_attr_nvme_info,
&dev_attr_bg_info,
&dev_attr_bg_guard_err,
&dev_attr_bg_apptag_err,
&dev_attr_bg_reftag_err,
&dev_attr_info,
&dev_attr_serialnum,
&dev_attr_modeldesc,
&dev_attr_modelname,
&dev_attr_programtype,
&dev_attr_portnum,
&dev_attr_fwrev,
&dev_attr_hdw,
&dev_attr_option_rom_version,
&dev_attr_link_state,
&dev_attr_num_discovered_ports,
&dev_attr_menlo_mgmt_mode,
&dev_attr_lpfc_drvr_version,
&dev_attr_lpfc_enable_fip,
&dev_attr_lpfc_temp_sensor,
&dev_attr_lpfc_log_verbose,
&dev_attr_lpfc_lun_queue_depth,
&dev_attr_lpfc_tgt_queue_depth,
&dev_attr_lpfc_hba_queue_depth,
&dev_attr_lpfc_peer_port_login,
&dev_attr_lpfc_nodev_tmo,
&dev_attr_lpfc_devloss_tmo,
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
&dev_attr_lpfc_enable_fc4_type,
&dev_attr_lpfc_xri_split,
&dev_attr_lpfc_fcp_class,
&dev_attr_lpfc_use_adisc,
&dev_attr_lpfc_first_burst_size,
&dev_attr_lpfc_ack0,
&dev_attr_lpfc_topology,
&dev_attr_lpfc_scan_down,
&dev_attr_lpfc_link_speed,
&dev_attr_lpfc_fcp_io_sched,
&dev_attr_lpfc_fcp2_no_tgt_reset,
&dev_attr_lpfc_cr_delay,
&dev_attr_lpfc_cr_count,
&dev_attr_lpfc_multi_ring_support,
&dev_attr_lpfc_multi_ring_rctl,
&dev_attr_lpfc_multi_ring_type,
&dev_attr_lpfc_fdmi_on,
&dev_attr_lpfc_enable_SmartSAN,
&dev_attr_lpfc_max_luns,
&dev_attr_lpfc_enable_npiv,
&dev_attr_lpfc_fcf_failover_policy,
&dev_attr_lpfc_enable_rrq,
&dev_attr_nport_evt_cnt,
&dev_attr_board_mode,
&dev_attr_max_vpi,
&dev_attr_used_vpi,
&dev_attr_max_rpi,
&dev_attr_used_rpi,
&dev_attr_max_xri,
&dev_attr_used_xri,
&dev_attr_npiv_info,
&dev_attr_issue_reset,
&dev_attr_lpfc_poll,
&dev_attr_lpfc_poll_tmo,
&dev_attr_lpfc_task_mgmt_tmo,
&dev_attr_lpfc_use_msi,
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
&dev_attr_lpfc_nvme_oas,
&dev_attr_lpfc_fcp_imax,
&dev_attr_lpfc_fcp_cpu_map,
&dev_attr_lpfc_fcp_io_channel,
&dev_attr_lpfc_suppress_rsp,
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
&dev_attr_lpfc_nvme_io_channel,
&dev_attr_lpfc_nvmet_mrq,
&dev_attr_lpfc_nvmet_mrq_post,
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
&dev_attr_lpfc_nvme_enable_fb,
&dev_attr_lpfc_nvmet_fb_size,
&dev_attr_lpfc_enable_bg,
&dev_attr_lpfc_soft_wwnn,
&dev_attr_lpfc_soft_wwpn,
&dev_attr_lpfc_soft_wwn_enable,
&dev_attr_lpfc_enable_hba_reset,
&dev_attr_lpfc_enable_hba_heartbeat,
&dev_attr_lpfc_EnableXLane,
&dev_attr_lpfc_XLanePriority,
&dev_attr_lpfc_xlane_lun,
&dev_attr_lpfc_xlane_tgt,
&dev_attr_lpfc_xlane_vpt,
&dev_attr_lpfc_xlane_lun_state,
&dev_attr_lpfc_xlane_lun_status,
&dev_attr_lpfc_xlane_priority,
&dev_attr_lpfc_sg_seg_cnt,
&dev_attr_lpfc_max_scsicmpl_time,
&dev_attr_lpfc_stat_data_ctrl,
&dev_attr_lpfc_aer_support,
&dev_attr_lpfc_aer_state_cleanup,
&dev_attr_lpfc_sriov_nr_virtfn,
&dev_attr_lpfc_req_fw_upgrade,
&dev_attr_lpfc_suppress_link_up,
&dev_attr_lpfc_iocb_cnt,
&dev_attr_iocb_hw,
&dev_attr_txq_hw,
&dev_attr_txcmplq_hw,
&dev_attr_lpfc_fips_level,
&dev_attr_lpfc_fips_rev,
&dev_attr_lpfc_dss,
&dev_attr_lpfc_sriov_hw_max_virtfn,
&dev_attr_protocol,
&dev_attr_lpfc_xlane_supported,
&dev_attr_lpfc_enable_mds_diags,
NULL,
};
struct device_attribute *lpfc_vport_attrs[] = {
&dev_attr_info,
&dev_attr_link_state,
&dev_attr_num_discovered_ports,
&dev_attr_lpfc_drvr_version,
&dev_attr_lpfc_log_verbose,
&dev_attr_lpfc_lun_queue_depth,
&dev_attr_lpfc_tgt_queue_depth,
&dev_attr_lpfc_nodev_tmo,
&dev_attr_lpfc_devloss_tmo,
&dev_attr_lpfc_hba_queue_depth,
&dev_attr_lpfc_peer_port_login,
&dev_attr_lpfc_restrict_login,
&dev_attr_lpfc_fcp_class,
&dev_attr_lpfc_use_adisc,
&dev_attr_lpfc_first_burst_size,
&dev_attr_lpfc_max_luns,
&dev_attr_nport_evt_cnt,
&dev_attr_npiv_info,
&dev_attr_lpfc_enable_da_id,
&dev_attr_lpfc_max_scsicmpl_time,
&dev_attr_lpfc_stat_data_ctrl,
&dev_attr_lpfc_static_vport,
&dev_attr_lpfc_fips_level,
&dev_attr_lpfc_fips_rev,
NULL,
};
/**
* sysfs_ctlreg_write - Write method for writing to ctlreg
* @filp: open sysfs file
* @kobj: kernel kobject that contains the kernel class device.
* @bin_attr: kernel attributes passed to us.
* @buf: contains the data to be written to the adapter IOREG space.
* @off: offset into buffer to beginning of data.
* @count: bytes to transfer.
*
* Description:
* Accessed via /sys/class/scsi_host/hostxxx/ctlreg.
* Uses the adapter io control registers to send buf contents to the adapter.
*
* Returns:
* -ERANGE off and count combo out of range
* -EINVAL off, count or buff address invalid
* -EPERM adapter is offline
* value of count, buf contents written
**/
static ssize_t
sysfs_ctlreg_write(struct file *filp, struct kobject *kobj,
struct bin_attribute *bin_attr,
sysfs: add parameter "struct bin_attribute *" in .read/.write methods for sysfs binary attributes Well, first of all, I don't want to change so many files either. What I do: Adding a new parameter "struct bin_attribute *" in the .read/.write methods for the sysfs binary attributes. In fact, only the four lines change in fs/sysfs/bin.c and include/linux/sysfs.h do the real work. But I have to update all the files that use binary attributes to make them compatible with the new .read and .write methods. I'm not sure if I missed any. :( Why I do this: For a sysfs attribute, we can get a pointer pointing to the struct attribute in the .show/.store method, while we can't do this for the binary attributes. I don't know why this is different, but this does make it not so handy to use the binary attributes as the regular ones. So I think this patch is reasonable. :) Who benefits from it: The patch that exposes ACPI tables in sysfs requires such an improvement. All the table binary attributes share the same .read method. Parameter "struct bin_attribute *" is used to get the table signature and instance number which are used to distinguish different ACPI table binary attributes. Without this parameter, we need to offer different .read methods for different ACPI table binary attributes. This is impossible as there are various ACPI tables on different platforms, and we don't know what they are until they are loaded. Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2007-06-09 12:57:22 +07:00
char *buf, loff_t off, size_t count)
{
size_t buf_off;
struct device *dev = container_of(kobj, struct device, kobj);
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (phba->sli_rev >= LPFC_SLI_REV4)
return -EPERM;
if ((off + count) > FF_REG_AREA_SIZE)
return -ERANGE;
if (count <= LPFC_REG_WRITE_KEY_SIZE)
return 0;
if (off % 4 || count % 4 || (unsigned long)buf % 4)
return -EINVAL;
/* This is to protect HBA registers from accidental writes. */
if (memcmp(buf, LPFC_REG_WRITE_KEY, LPFC_REG_WRITE_KEY_SIZE))
return -EINVAL;
if (!(vport->fc_flag & FC_OFFLINE_MODE))
return -EPERM;
spin_lock_irq(&phba->hbalock);
for (buf_off = 0; buf_off < count - LPFC_REG_WRITE_KEY_SIZE;
buf_off += sizeof(uint32_t))
writel(*((uint32_t *)(buf + buf_off + LPFC_REG_WRITE_KEY_SIZE)),
phba->ctrl_regs_memmap_p + off + buf_off);
spin_unlock_irq(&phba->hbalock);
return count;
}
/**
* sysfs_ctlreg_read - Read method for reading from ctlreg
* @filp: open sysfs file
* @kobj: kernel kobject that contains the kernel class device.
* @bin_attr: kernel attributes passed to us.
* @buf: if successful contains the data from the adapter IOREG space.
* @off: offset into buffer to beginning of data.
* @count: bytes to transfer.
*
* Description:
* Accessed via /sys/class/scsi_host/hostxxx/ctlreg.
* Uses the adapter io control registers to read data into buf.
*
* Returns:
* -ERANGE off and count combo out of range
* -EINVAL off, count or buff address invalid
* value of count, buf contents read
**/
static ssize_t
sysfs_ctlreg_read(struct file *filp, struct kobject *kobj,
struct bin_attribute *bin_attr,
sysfs: add parameter "struct bin_attribute *" in .read/.write methods for sysfs binary attributes Well, first of all, I don't want to change so many files either. What I do: Adding a new parameter "struct bin_attribute *" in the .read/.write methods for the sysfs binary attributes. In fact, only the four lines change in fs/sysfs/bin.c and include/linux/sysfs.h do the real work. But I have to update all the files that use binary attributes to make them compatible with the new .read and .write methods. I'm not sure if I missed any. :( Why I do this: For a sysfs attribute, we can get a pointer pointing to the struct attribute in the .show/.store method, while we can't do this for the binary attributes. I don't know why this is different, but this does make it not so handy to use the binary attributes as the regular ones. So I think this patch is reasonable. :) Who benefits from it: The patch that exposes ACPI tables in sysfs requires such an improvement. All the table binary attributes share the same .read method. Parameter "struct bin_attribute *" is used to get the table signature and instance number which are used to distinguish different ACPI table binary attributes. Without this parameter, we need to offer different .read methods for different ACPI table binary attributes. This is impossible as there are various ACPI tables on different platforms, and we don't know what they are until they are loaded. Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2007-06-09 12:57:22 +07:00
char *buf, loff_t off, size_t count)
{
size_t buf_off;
uint32_t * tmp_ptr;
struct device *dev = container_of(kobj, struct device, kobj);
struct Scsi_Host *shost = class_to_shost(dev);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
if (phba->sli_rev >= LPFC_SLI_REV4)
return -EPERM;
if (off > FF_REG_AREA_SIZE)
return -ERANGE;
if ((off + count) > FF_REG_AREA_SIZE)
count = FF_REG_AREA_SIZE - off;
if (count == 0) return 0;
if (off % 4 || count % 4 || (unsigned long)buf % 4)
return -EINVAL;
spin_lock_irq(&phba->hbalock);
for (buf_off = 0; buf_off < count; buf_off += sizeof(uint32_t)) {
tmp_ptr = (uint32_t *)(buf + buf_off);
*tmp_ptr = readl(phba->ctrl_regs_memmap_p + off + buf_off);
}
spin_unlock_irq(&phba->hbalock);
return count;
}
static struct bin_attribute sysfs_ctlreg_attr = {
.attr = {
.name = "ctlreg",
.mode = S_IRUSR | S_IWUSR,
},
.size = 256,
.read = sysfs_ctlreg_read,
.write = sysfs_ctlreg_write,
};
/**
* sysfs_mbox_write - Write method for writing information via mbox
* @filp: open sysfs file
* @kobj: kernel kobject that contains the kernel class device.
* @bin_attr: kernel attributes passed to us.
* @buf: contains the data to be written to sysfs mbox.
* @off: offset into buffer to beginning of data.
* @count: bytes to transfer.
*
* Description:
* Deprecated function. All mailbox access from user space is performed via the
* bsg interface.
*
* Returns:
* -EPERM operation not permitted
**/
static ssize_t
sysfs_mbox_write(struct file *filp, struct kobject *kobj,
struct bin_attribute *bin_attr,
sysfs: add parameter "struct bin_attribute *" in .read/.write methods for sysfs binary attributes Well, first of all, I don't want to change so many files either. What I do: Adding a new parameter "struct bin_attribute *" in the .read/.write methods for the sysfs binary attributes. In fact, only the four lines change in fs/sysfs/bin.c and include/linux/sysfs.h do the real work. But I have to update all the files that use binary attributes to make them compatible with the new .read and .write methods. I'm not sure if I missed any. :( Why I do this: For a sysfs attribute, we can get a pointer pointing to the struct attribute in the .show/.store method, while we can't do this for the binary attributes. I don't know why this is different, but this does make it not so handy to use the binary attributes as the regular ones. So I think this patch is reasonable. :) Who benefits from it: The patch that exposes ACPI tables in sysfs requires such an improvement. All the table binary attributes share the same .read method. Parameter "struct bin_attribute *" is used to get the table signature and instance number which are used to distinguish different ACPI table binary attributes. Without this parameter, we need to offer different .read methods for different ACPI table binary attributes. This is impossible as there are various ACPI tables on different platforms, and we don't know what they are until they are loaded. Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2007-06-09 12:57:22 +07:00
char *buf, loff_t off, size_t count)
{
return -EPERM;
}
/**
* sysfs_mbox_read - Read method for reading information via mbox
* @filp: open sysfs file
* @kobj: kernel kobject that contains the kernel class device.
* @bin_attr: kernel attributes passed to us.
* @buf: contains the data to be read from sysfs mbox.
* @off: offset into buffer to beginning of data.
* @count: bytes to transfer.
*
* Description:
* Deprecated function. All mailbox access from user space is performed via the
* bsg interface.
*
* Returns:
* -EPERM operation not permitted
**/
static ssize_t
sysfs_mbox_read(struct file *filp, struct kobject *kobj,
struct bin_attribute *bin_attr,
sysfs: add parameter "struct bin_attribute *" in .read/.write methods for sysfs binary attributes Well, first of all, I don't want to change so many files either. What I do: Adding a new parameter "struct bin_attribute *" in the .read/.write methods for the sysfs binary attributes. In fact, only the four lines change in fs/sysfs/bin.c and include/linux/sysfs.h do the real work. But I have to update all the files that use binary attributes to make them compatible with the new .read and .write methods. I'm not sure if I missed any. :( Why I do this: For a sysfs attribute, we can get a pointer pointing to the struct attribute in the .show/.store method, while we can't do this for the binary attributes. I don't know why this is different, but this does make it not so handy to use the binary attributes as the regular ones. So I think this patch is reasonable. :) Who benefits from it: The patch that exposes ACPI tables in sysfs requires such an improvement. All the table binary attributes share the same .read method. Parameter "struct bin_attribute *" is used to get the table signature and instance number which are used to distinguish different ACPI table binary attributes. Without this parameter, we need to offer different .read methods for different ACPI table binary attributes. This is impossible as there are various ACPI tables on different platforms, and we don't know what they are until they are loaded. Signed-off-by: Zhang Rui <rui.zhang@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2007-06-09 12:57:22 +07:00
char *buf, loff_t off, size_t count)
{
return -EPERM;
}
static struct bin_attribute sysfs_mbox_attr = {
.attr = {
.name = "mbox",
.mode = S_IRUSR | S_IWUSR,
},
.size = MAILBOX_SYSFS_MAX,
.read = sysfs_mbox_read,
.write = sysfs_mbox_write,
};
/**
* lpfc_alloc_sysfs_attr - Creates the ctlreg and mbox entries
* @vport: address of lpfc vport structure.
*
* Return codes:
* zero on success
* error return code from sysfs_create_bin_file()
**/
int
lpfc_alloc_sysfs_attr(struct lpfc_vport *vport)
{
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
int error;
error = sysfs_create_bin_file(&shost->shost_dev.kobj,
&sysfs_drvr_stat_data_attr);
/* Virtual ports do not need ctrl_reg and mbox */
if (error || vport->port_type == LPFC_NPIV_PORT)
goto out;
error = sysfs_create_bin_file(&shost->shost_dev.kobj,
&sysfs_ctlreg_attr);
if (error)
goto out_remove_stat_attr;
error = sysfs_create_bin_file(&shost->shost_dev.kobj,
&sysfs_mbox_attr);
if (error)
goto out_remove_ctlreg_attr;
return 0;
out_remove_ctlreg_attr:
sysfs_remove_bin_file(&shost->shost_dev.kobj, &sysfs_ctlreg_attr);
out_remove_stat_attr:
sysfs_remove_bin_file(&shost->shost_dev.kobj,
&sysfs_drvr_stat_data_attr);
out:
return error;
}
/**
* lpfc_free_sysfs_attr - Removes the ctlreg and mbox entries
* @vport: address of lpfc vport structure.
**/
void
lpfc_free_sysfs_attr(struct lpfc_vport *vport)
{
struct Scsi_Host *shost = lpfc_shost_from_vport(vport);
sysfs_remove_bin_file(&shost->shost_dev.kobj,
&sysfs_drvr_stat_data_attr);
/* Virtual ports do not need ctrl_reg and mbox */
if (vport->port_type == LPFC_NPIV_PORT)
return;
sysfs_remove_bin_file(&shost->shost_dev.kobj, &sysfs_mbox_attr);
sysfs_remove_bin_file(&shost->shost_dev.kobj, &sysfs_ctlreg_attr);
}
/*
* Dynamic FC Host Attributes Support
*/
/**
* lpfc_get_host_symbolic_name - Copy symbolic name into the scsi host
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_symbolic_name(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *)shost->hostdata;
lpfc_vport_symbolic_node_name(vport, fc_host_symbolic_name(shost),
sizeof fc_host_symbolic_name(shost));
}
/**
* lpfc_get_host_port_id - Copy the vport DID into the scsi host port id
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_port_id(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
/* note: fc_myDID already in cpu endianness */
fc_host_port_id(shost) = vport->fc_myDID;
}
/**
* lpfc_get_host_port_type - Set the value of the scsi host port type
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_port_type(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
spin_lock_irq(shost->host_lock);
if (vport->port_type == LPFC_NPIV_PORT) {
fc_host_port_type(shost) = FC_PORTTYPE_NPIV;
} else if (lpfc_is_link_up(phba)) {
2010-11-21 11:11:48 +07:00
if (phba->fc_topology == LPFC_TOPOLOGY_LOOP) {
if (vport->fc_flag & FC_PUBLIC_LOOP)
fc_host_port_type(shost) = FC_PORTTYPE_NLPORT;
else
fc_host_port_type(shost) = FC_PORTTYPE_LPORT;
} else {
if (vport->fc_flag & FC_FABRIC)
fc_host_port_type(shost) = FC_PORTTYPE_NPORT;
else
fc_host_port_type(shost) = FC_PORTTYPE_PTP;
}
} else
fc_host_port_type(shost) = FC_PORTTYPE_UNKNOWN;
spin_unlock_irq(shost->host_lock);
}
/**
* lpfc_get_host_port_state - Set the value of the scsi host port state
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_port_state(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
spin_lock_irq(shost->host_lock);
if (vport->fc_flag & FC_OFFLINE_MODE)
fc_host_port_state(shost) = FC_PORTSTATE_OFFLINE;
else {
switch (phba->link_state) {
case LPFC_LINK_UNKNOWN:
case LPFC_LINK_DOWN:
fc_host_port_state(shost) = FC_PORTSTATE_LINKDOWN;
break;
case LPFC_LINK_UP:
case LPFC_CLEAR_LA:
case LPFC_HBA_READY:
/* Links up, reports port state accordingly */
if (vport->port_state < LPFC_VPORT_READY)
fc_host_port_state(shost) =
FC_PORTSTATE_BYPASSED;
else
fc_host_port_state(shost) =
FC_PORTSTATE_ONLINE;
break;
case LPFC_HBA_ERROR:
fc_host_port_state(shost) = FC_PORTSTATE_ERROR;
break;
default:
fc_host_port_state(shost) = FC_PORTSTATE_UNKNOWN;
break;
}
}
spin_unlock_irq(shost->host_lock);
}
/**
* lpfc_get_host_speed - Set the value of the scsi host speed
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_speed(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
spin_lock_irq(shost->host_lock);
if ((lpfc_is_link_up(phba)) && (!(phba->hba_flag & HBA_FCOE_MODE))) {
switch(phba->fc_linkspeed) {
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_1GHZ:
fc_host_speed(shost) = FC_PORTSPEED_1GBIT;
break;
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_2GHZ:
fc_host_speed(shost) = FC_PORTSPEED_2GBIT;
break;
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_4GHZ:
fc_host_speed(shost) = FC_PORTSPEED_4GBIT;
break;
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_8GHZ:
fc_host_speed(shost) = FC_PORTSPEED_8GBIT;
break;
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_10GHZ:
fc_host_speed(shost) = FC_PORTSPEED_10GBIT;
break;
2010-11-21 11:11:48 +07:00
case LPFC_LINK_SPEED_16GHZ:
fc_host_speed(shost) = FC_PORTSPEED_16GBIT;
break;
case LPFC_LINK_SPEED_32GHZ:
fc_host_speed(shost) = FC_PORTSPEED_32GBIT;
break;
2010-11-21 11:11:48 +07:00
default:
fc_host_speed(shost) = FC_PORTSPEED_UNKNOWN;
break;
}
[SCSI] lpfc 8.2.4 : Miscellaneous Fixes Miscellaneous Fixes: - Fix a couple of sparse complaints - Reset the FCP recovery flag when the node is not a FCP2 device. - Speed up offline prep delays - Fixed a memory leak in lpfc_mem_alloc failure path - Fixed external loopback test. - Fixed error code returned from the driver when HBA is over heated. - Correct Max NPIV vport to limits read from adapter - Add missing locks around fc_flag and FC_NEEDS_REG_VPI - Add missing hba ids for device identification - Added support for SET_VARIABLE and MBX_WRITE_WWN mailbox commands - Changed all temperature event messages from warning to error - Fix reporting of link speed when link is down - Added support for MBX_WRITE_WWN mailbox command - Change del_timer_sync() in ISR to del_timer() in interrupt handler - Correct instances of beXX_to_cpu() that should be cpu_to_beXX() - Perform target flush before releasing node references on module unload - Avoid bogus devloss_tmo messages when driver unloads - Fix panic when HBA generates ERATT interupt - Fix mbox race condition and a workaround on back-to-back mailbox commands - Force NPIV off for pt2pt mode between 2 NPorts - Stop worker thread before removing fc_host. - Fix up discovery timeout error case due to missing clear_la - Tighten mailbox polling code to speed up detection of fast completions - Only allow DUMP_MEMORY if adapter offline due to overtemp errors - Added extended error information to the log messages in chip init. Signed-off-by: James Smart <James.Smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2008-01-11 13:52:54 +07:00
} else
fc_host_speed(shost) = FC_PORTSPEED_UNKNOWN;
spin_unlock_irq(shost->host_lock);
}
/**
* lpfc_get_host_fabric_name - Set the value of the scsi host fabric name
* @shost: kernel scsi host pointer.
**/
static void
lpfc_get_host_fabric_name (struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
u64 node_name;
spin_lock_irq(shost->host_lock);
2011-10-11 08:32:10 +07:00
if ((vport->port_state > LPFC_FLOGI) &&
((vport->fc_flag & FC_FABRIC) ||
((phba->fc_topology == LPFC_TOPOLOGY_LOOP) &&
(vport->fc_flag & FC_PUBLIC_LOOP))))
node_name = wwn_to_u64(phba->fc_fabparam.nodeName.u.wwn);
else
/* fabric is local port if there is no F/FL_Port */
[SCSI] lpfc 8.2.4 : Miscellaneous Fixes Miscellaneous Fixes: - Fix a couple of sparse complaints - Reset the FCP recovery flag when the node is not a FCP2 device. - Speed up offline prep delays - Fixed a memory leak in lpfc_mem_alloc failure path - Fixed external loopback test. - Fixed error code returned from the driver when HBA is over heated. - Correct Max NPIV vport to limits read from adapter - Add missing locks around fc_flag and FC_NEEDS_REG_VPI - Add missing hba ids for device identification - Added support for SET_VARIABLE and MBX_WRITE_WWN mailbox commands - Changed all temperature event messages from warning to error - Fix reporting of link speed when link is down - Added support for MBX_WRITE_WWN mailbox command - Change del_timer_sync() in ISR to del_timer() in interrupt handler - Correct instances of beXX_to_cpu() that should be cpu_to_beXX() - Perform target flush before releasing node references on module unload - Avoid bogus devloss_tmo messages when driver unloads - Fix panic when HBA generates ERATT interupt - Fix mbox race condition and a workaround on back-to-back mailbox commands - Force NPIV off for pt2pt mode between 2 NPorts - Stop worker thread before removing fc_host. - Fix up discovery timeout error case due to missing clear_la - Tighten mailbox polling code to speed up detection of fast completions - Only allow DUMP_MEMORY if adapter offline due to overtemp errors - Added extended error information to the log messages in chip init. Signed-off-by: James Smart <James.Smart@emulex.com> Signed-off-by: James Bottomley <James.Bottomley@HansenPartnership.com>
2008-01-11 13:52:54 +07:00
node_name = 0;
spin_unlock_irq(shost->host_lock);
fc_host_fabric_name(shost) = node_name;
}
/**
* lpfc_get_stats - Return statistical information about the adapter
* @shost: kernel scsi host pointer.
*
* Notes:
* NULL on error for link down, no mbox pool, sli2 active,
* management not allowed, memory allocation error, or mbox error.
*
* Returns:
* NULL for error
* address of the adapter host statistics
**/
static struct fc_host_statistics *
lpfc_get_stats(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct lpfc_sli *psli = &phba->sli;
struct fc_host_statistics *hs = &phba->link_stats;
struct lpfc_lnk_stat * lso = &psli->lnk_stat_offsets;
LPFC_MBOXQ_t *pmboxq;
MAILBOX_t *pmb;
unsigned long seconds;
int rc = 0;
/*
* prevent udev from issuing mailbox commands until the port is
* configured.
*/
if (phba->link_state < LPFC_LINK_DOWN ||
!phba->mbox_mem_pool ||
(phba->sli.sli_flag & LPFC_SLI_ACTIVE) == 0)
return NULL;
if (phba->sli.sli_flag & LPFC_BLOCK_MGMT_IO)
return NULL;
pmboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
if (!pmboxq)
return NULL;
memset(pmboxq, 0, sizeof (LPFC_MBOXQ_t));
pmb = &pmboxq->u.mb;
pmb->mbxCommand = MBX_READ_STATUS;
pmb->mbxOwner = OWN_HOST;
pmboxq->context1 = NULL;
pmboxq->vport = vport;
if (vport->fc_flag & FC_OFFLINE_MODE)
rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_POLL);
else
rc = lpfc_sli_issue_mbox_wait(phba, pmboxq, phba->fc_ratov * 2);
if (rc != MBX_SUCCESS) {
if (rc != MBX_TIMEOUT)
mempool_free(pmboxq, phba->mbox_mem_pool);
return NULL;
}
memset(hs, 0, sizeof (struct fc_host_statistics));
hs->tx_frames = pmb->un.varRdStatus.xmitFrameCnt;
2011-10-11 08:32:10 +07:00
/*
* The MBX_READ_STATUS returns tx_k_bytes which has to
* converted to words
*/
hs->tx_words = (uint64_t)
((uint64_t)pmb->un.varRdStatus.xmitByteCnt
* (uint64_t)256);
hs->rx_frames = pmb->un.varRdStatus.rcvFrameCnt;
2011-10-11 08:32:10 +07:00
hs->rx_words = (uint64_t)
((uint64_t)pmb->un.varRdStatus.rcvByteCnt
* (uint64_t)256);
memset(pmboxq, 0, sizeof (LPFC_MBOXQ_t));
pmb->mbxCommand = MBX_READ_LNK_STAT;
pmb->mbxOwner = OWN_HOST;
pmboxq->context1 = NULL;
pmboxq->vport = vport;
if (vport->fc_flag & FC_OFFLINE_MODE)
rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_POLL);
else
rc = lpfc_sli_issue_mbox_wait(phba, pmboxq, phba->fc_ratov * 2);
if (rc != MBX_SUCCESS) {
if (rc != MBX_TIMEOUT)
mempool_free(pmboxq, phba->mbox_mem_pool);
return NULL;
}
hs->link_failure_count = pmb->un.varRdLnk.linkFailureCnt;
hs->loss_of_sync_count = pmb->un.varRdLnk.lossSyncCnt;
hs->loss_of_signal_count = pmb->un.varRdLnk.lossSignalCnt;
hs->prim_seq_protocol_err_count = pmb->un.varRdLnk.primSeqErrCnt;
hs->invalid_tx_word_count = pmb->un.varRdLnk.invalidXmitWord;
hs->invalid_crc_count = pmb->un.varRdLnk.crcCnt;
hs->error_frames = pmb->un.varRdLnk.crcCnt;
hs->link_failure_count -= lso->link_failure_count;
hs->loss_of_sync_count -= lso->loss_of_sync_count;
hs->loss_of_signal_count -= lso->loss_of_signal_count;
hs->prim_seq_protocol_err_count -= lso->prim_seq_protocol_err_count;
hs->invalid_tx_word_count -= lso->invalid_tx_word_count;
hs->invalid_crc_count -= lso->invalid_crc_count;
hs->error_frames -= lso->error_frames;
2010-11-21 11:11:48 +07:00
if (phba->hba_flag & HBA_FCOE_MODE) {
hs->lip_count = -1;
hs->nos_count = (phba->link_events >> 1);
hs->nos_count -= lso->link_events;
2010-11-21 11:11:48 +07:00
} else if (phba->fc_topology == LPFC_TOPOLOGY_LOOP) {
hs->lip_count = (phba->fc_eventTag >> 1);
hs->lip_count -= lso->link_events;
hs->nos_count = -1;
} else {
hs->lip_count = -1;
hs->nos_count = (phba->fc_eventTag >> 1);
hs->nos_count -= lso->link_events;
}
hs->dumped_frames = -1;
seconds = get_seconds();
if (seconds < psli->stats_start)
hs->seconds_since_last_reset = seconds +
((unsigned long)-1 - psli->stats_start);
else
hs->seconds_since_last_reset = seconds - psli->stats_start;
mempool_free(pmboxq, phba->mbox_mem_pool);
return hs;
}
/**
* lpfc_reset_stats - Copy the adapter link stats information
* @shost: kernel scsi host pointer.
**/
static void
lpfc_reset_stats(struct Scsi_Host *shost)
{
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_hba *phba = vport->phba;
struct lpfc_sli *psli = &phba->sli;
struct lpfc_lnk_stat *lso = &psli->lnk_stat_offsets;
LPFC_MBOXQ_t *pmboxq;
MAILBOX_t *pmb;
int rc = 0;
if (phba->sli.sli_flag & LPFC_BLOCK_MGMT_IO)
return;
pmboxq = mempool_alloc(phba->mbox_mem_pool, GFP_KERNEL);
if (!pmboxq)
return;
memset(pmboxq, 0, sizeof(LPFC_MBOXQ_t));
pmb = &pmboxq->u.mb;
pmb->mbxCommand = MBX_READ_STATUS;
pmb->mbxOwner = OWN_HOST;
pmb->un.varWords[0] = 0x1; /* reset request */
pmboxq->context1 = NULL;
pmboxq->vport = vport;
if ((vport->fc_flag & FC_OFFLINE_MODE) ||
(!(psli->sli_flag & LPFC_SLI_ACTIVE)))
rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_POLL);
else
rc = lpfc_sli_issue_mbox_wait(phba, pmboxq, phba->fc_ratov * 2);
if (rc != MBX_SUCCESS) {
if (rc != MBX_TIMEOUT)
mempool_free(pmboxq, phba->mbox_mem_pool);
return;
}
memset(pmboxq, 0, sizeof(LPFC_MBOXQ_t));
pmb->mbxCommand = MBX_READ_LNK_STAT;
pmb->mbxOwner = OWN_HOST;
pmboxq->context1 = NULL;
pmboxq->vport = vport;
if ((vport->fc_flag & FC_OFFLINE_MODE) ||
(!(psli->sli_flag & LPFC_SLI_ACTIVE)))
rc = lpfc_sli_issue_mbox(phba, pmboxq, MBX_POLL);
else
rc = lpfc_sli_issue_mbox_wait(phba, pmboxq, phba->fc_ratov * 2);
if (rc != MBX_SUCCESS) {
if (rc != MBX_TIMEOUT)
mempool_free( pmboxq, phba->mbox_mem_pool);
return;
}
lso->link_failure_count = pmb->un.varRdLnk.linkFailureCnt;
lso->loss_of_sync_count = pmb->un.varRdLnk.lossSyncCnt;
lso->loss_of_signal_count = pmb->un.varRdLnk.lossSignalCnt;
lso->prim_seq_protocol_err_count = pmb->un.varRdLnk.primSeqErrCnt;
lso->invalid_tx_word_count = pmb->un.varRdLnk.invalidXmitWord;
lso->invalid_crc_count = pmb->un.varRdLnk.crcCnt;
lso->error_frames = pmb->un.varRdLnk.crcCnt;
2010-11-21 11:11:48 +07:00
if (phba->hba_flag & HBA_FCOE_MODE)
lso->link_events = (phba->link_events >> 1);
else
lso->link_events = (phba->fc_eventTag >> 1);
psli->stats_start = get_seconds();
mempool_free(pmboxq, phba->mbox_mem_pool);
return;
}
/*
* The LPFC driver treats linkdown handling as target loss events so there
* are no sysfs handlers for link_down_tmo.
*/
/**
* lpfc_get_node_by_target - Return the nodelist for a target
* @starget: kernel scsi target pointer.
*
* Returns:
* address of the node list if found
* NULL target not found
**/
static struct lpfc_nodelist *
lpfc_get_node_by_target(struct scsi_target *starget)
{
struct Scsi_Host *shost = dev_to_shost(starget->dev.parent);
struct lpfc_vport *vport = (struct lpfc_vport *) shost->hostdata;
struct lpfc_nodelist *ndlp;
spin_lock_irq(shost->host_lock);
/* Search for this, mapped, target ID */
list_for_each_entry(ndlp, &vport->fc_nodes, nlp_listp) {
if (NLP_CHK_NODE_ACT(ndlp) &&
ndlp->nlp_state == NLP_STE_MAPPED_NODE &&
starget->id == ndlp->nlp_sid) {
spin_unlock_irq(shost->host_lock);
return ndlp;
}
}
spin_unlock_irq(shost->host_lock);
return NULL;
}
/**
* lpfc_get_starget_port_id - Set the target port id to the ndlp DID or -1
* @starget: kernel scsi target pointer.
**/
static void
lpfc_get_starget_port_id(struct scsi_target *starget)
{
struct lpfc_nodelist *ndlp = lpfc_get_node_by_target(starget);
fc_starget_port_id(starget) = ndlp ? ndlp->nlp_DID : -1;
}
/**
* lpfc_get_starget_node_name - Set the target node name
* @starget: kernel scsi target pointer.
*
* Description: Set the target node name to the ndlp node name wwn or zero.
**/
static void
lpfc_get_starget_node_name(struct scsi_target *starget)
{
struct lpfc_nodelist *ndlp = lpfc_get_node_by_target(starget);
fc_starget_node_name(starget) =
ndlp ? wwn_to_u64(ndlp->nlp_nodename.u.wwn) : 0;
}
/**
* lpfc_get_starget_port_name - Set the target port name
* @starget: kernel scsi target pointer.
*
* Description: set the target port name to the ndlp port name wwn or zero.
**/
static void
lpfc_get_starget_port_name(struct scsi_target *starget)
{
struct lpfc_nodelist *ndlp = lpfc_get_node_by_target(starget);
fc_starget_port_name(starget) =
ndlp ? wwn_to_u64(ndlp->nlp_portname.u.wwn) : 0;
}
/**
* lpfc_set_rport_loss_tmo - Set the rport dev loss tmo
* @rport: fc rport address.
* @timeout: new value for dev loss tmo.
*
* Description:
* If timeout is non zero set the dev_loss_tmo to timeout, else set
* dev_loss_tmo to one.
**/
static void
lpfc_set_rport_loss_tmo(struct fc_rport *rport, uint32_t timeout)
{
if (timeout)
rport->dev_loss_tmo = timeout;
else
rport->dev_loss_tmo = 1;
}
/**
* lpfc_rport_show_function - Return rport target information
*
* Description:
* Macro that uses field to generate a function with the name lpfc_show_rport_
*
* lpfc_show_rport_##field: returns the bytes formatted in buf
* @cdev: class converted to an fc_rport.
* @buf: on return contains the target_field or zero.
*
* Returns: size of formatted string.
**/
#define lpfc_rport_show_function(field, format_string, sz, cast) \
static ssize_t \
lpfc_show_rport_##field (struct device *dev, \
struct device_attribute *attr, \
char *buf) \
{ \
struct fc_rport *rport = transport_class_to_rport(dev); \
struct lpfc_rport_data *rdata = rport->hostdata; \
return snprintf(buf, sz, format_string, \
(rdata->target) ? cast rdata->target->field : 0); \
}
#define lpfc_rport_rd_attr(field, format_string, sz) \
lpfc_rport_show_function(field, format_string, sz, ) \
static FC_RPORT_ATTR(field, S_IRUGO, lpfc_show_rport_##field, NULL)
/**
* lpfc_set_vport_symbolic_name - Set the vport's symbolic name
* @fc_vport: The fc_vport who's symbolic name has been changed.
*
* Description:
* This function is called by the transport after the @fc_vport's symbolic name
* has been changed. This function re-registers the symbolic name with the
* switch to propagate the change into the fabric if the vport is active.
**/
static void
lpfc_set_vport_symbolic_name(struct fc_vport *fc_vport)
{
struct lpfc_vport *vport = *(struct lpfc_vport **)fc_vport->dd_data;
if (vport->port_state == LPFC_VPORT_READY)
lpfc_ns_cmd(vport, SLI_CTNS_RSPN_ID, 0, 0);
}
/**
* lpfc_hba_log_verbose_init - Set hba's log verbose level
* @phba: Pointer to lpfc_hba struct.
*
* This function is called by the lpfc_get_cfgparam() routine to set the
* module lpfc_log_verbose into the @phba cfg_log_verbose for use with
* log message according to the module's lpfc_log_verbose parameter setting
* before hba port or vport created.
**/
static void
lpfc_hba_log_verbose_init(struct lpfc_hba *phba, uint32_t verbose)
{
phba->cfg_log_verbose = verbose;
}
struct fc_function_template lpfc_transport_functions = {
/* fixed attributes the driver supports */
.show_host_node_name = 1,
.show_host_port_name = 1,
.show_host_supported_classes = 1,
.show_host_supported_fc4s = 1,
.show_host_supported_speeds = 1,
.show_host_maxframe_size = 1,
.get_host_symbolic_name = lpfc_get_host_symbolic_name,
.show_host_symbolic_name = 1,
/* dynamic attributes the driver supports */
.get_host_port_id = lpfc_get_host_port_id,
.show_host_port_id = 1,
.get_host_port_type = lpfc_get_host_port_type,
.show_host_port_type = 1,
.get_host_port_state = lpfc_get_host_port_state,
.show_host_port_state = 1,
/* active_fc4s is shown but doesn't change (thus no get function) */
.show_host_active_fc4s = 1,
.get_host_speed = lpfc_get_host_speed,
.show_host_speed = 1,
.get_host_fabric_name = lpfc_get_host_fabric_name,
.show_host_fabric_name = 1,
/*
* The LPFC driver treats linkdown handling as target loss events
* so there are no sysfs handlers for link_down_tmo.
*/
.get_fc_host_stats = lpfc_get_stats,
.reset_fc_host_stats = lpfc_reset_stats,
.dd_fcrport_size = sizeof(struct lpfc_rport_data),
.show_rport_maxframe_size = 1,
.show_rport_supported_classes = 1,
.set_rport_dev_loss_tmo = lpfc_set_rport_loss_tmo,
.show_rport_dev_loss_tmo = 1,
.get_starget_port_id = lpfc_get_starget_port_id,
.show_starget_port_id = 1,
.get_starget_node_name = lpfc_get_starget_node_name,
.show_starget_node_name = 1,
.get_starget_port_name = lpfc_get_starget_port_name,
.show_starget_port_name = 1,
.issue_fc_host_lip = lpfc_issue_lip,
.dev_loss_tmo_callbk = lpfc_dev_loss_tmo_callbk,
.terminate_rport_io = lpfc_terminate_rport_io,
.dd_fcvport_size = sizeof(struct lpfc_vport *),
.vport_disable = lpfc_vport_disable,
.set_vport_symbolic_name = lpfc_set_vport_symbolic_name,
.bsg_request = lpfc_bsg_request,
.bsg_timeout = lpfc_bsg_timeout,
};
struct fc_function_template lpfc_vport_transport_functions = {
/* fixed attributes the driver supports */
.show_host_node_name = 1,
.show_host_port_name = 1,
.show_host_supported_classes = 1,
.show_host_supported_fc4s = 1,
.show_host_supported_speeds = 1,
.show_host_maxframe_size = 1,
.get_host_symbolic_name = lpfc_get_host_symbolic_name,
.show_host_symbolic_name = 1,
/* dynamic attributes the driver supports */
.get_host_port_id = lpfc_get_host_port_id,
.show_host_port_id = 1,
.get_host_port_type = lpfc_get_host_port_type,
.show_host_port_type = 1,
.get_host_port_state = lpfc_get_host_port_state,
.show_host_port_state = 1,
/* active_fc4s is shown but doesn't change (thus no get function) */
.show_host_active_fc4s = 1,
.get_host_speed = lpfc_get_host_speed,
.show_host_speed = 1,
.get_host_fabric_name = lpfc_get_host_fabric_name,
.show_host_fabric_name = 1,
/*
* The LPFC driver treats linkdown handling as target loss events
* so there are no sysfs handlers for link_down_tmo.
*/
.get_fc_host_stats = lpfc_get_stats,
.reset_fc_host_stats = lpfc_reset_stats,
.dd_fcrport_size = sizeof(struct lpfc_rport_data),
.show_rport_maxframe_size = 1,
.show_rport_supported_classes = 1,
.set_rport_dev_loss_tmo = lpfc_set_rport_loss_tmo,
.show_rport_dev_loss_tmo = 1,
.get_starget_port_id = lpfc_get_starget_port_id,
.show_starget_port_id = 1,
.get_starget_node_name = lpfc_get_starget_node_name,
.show_starget_node_name = 1,
.get_starget_port_name = lpfc_get_starget_port_name,
.show_starget_port_name = 1,
.dev_loss_tmo_callbk = lpfc_dev_loss_tmo_callbk,
.terminate_rport_io = lpfc_terminate_rport_io,
.vport_disable = lpfc_vport_disable,
.set_vport_symbolic_name = lpfc_set_vport_symbolic_name,
};
/**
* lpfc_get_cfgparam - Used during probe_one to init the adapter structure
* @phba: lpfc_hba pointer.
**/
void
lpfc_get_cfgparam(struct lpfc_hba *phba)
{
lpfc_fcp_io_sched_init(phba, lpfc_fcp_io_sched);
lpfc_fcp2_no_tgt_reset_init(phba, lpfc_fcp2_no_tgt_reset);
lpfc_cr_delay_init(phba, lpfc_cr_delay);
lpfc_cr_count_init(phba, lpfc_cr_count);
lpfc_multi_ring_support_init(phba, lpfc_multi_ring_support);
lpfc_multi_ring_rctl_init(phba, lpfc_multi_ring_rctl);
lpfc_multi_ring_type_init(phba, lpfc_multi_ring_type);
lpfc_ack0_init(phba, lpfc_ack0);
lpfc_topology_init(phba, lpfc_topology);
lpfc_link_speed_init(phba, lpfc_link_speed);
[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 04:32:13 +07:00
lpfc_poll_tmo_init(phba, lpfc_poll_tmo);
lpfc_task_mgmt_tmo_init(phba, lpfc_task_mgmt_tmo);
lpfc_enable_npiv_init(phba, lpfc_enable_npiv);
lpfc_fcf_failover_policy_init(phba, lpfc_fcf_failover_policy);
lpfc_enable_rrq_init(phba, lpfc_enable_rrq);
lpfc_fdmi_on_init(phba, lpfc_fdmi_on);
lpfc_enable_SmartSAN_init(phba, lpfc_enable_SmartSAN);
lpfc_use_msi_init(phba, lpfc_use_msi);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
lpfc_nvme_oas_init(phba, lpfc_nvme_oas);
lpfc_fcp_imax_init(phba, lpfc_fcp_imax);
lpfc_fcp_cpu_map_init(phba, lpfc_fcp_cpu_map);
lpfc_enable_hba_reset_init(phba, lpfc_enable_hba_reset);
lpfc_enable_hba_heartbeat_init(phba, lpfc_enable_hba_heartbeat);
lpfc_EnableXLane_init(phba, lpfc_EnableXLane);
if (phba->sli_rev != LPFC_SLI_REV4)
phba->cfg_EnableXLane = 0;
lpfc_XLanePriority_init(phba, lpfc_XLanePriority);
memset(phba->cfg_oas_tgt_wwpn, 0, (8 * sizeof(uint8_t)));
memset(phba->cfg_oas_vpt_wwpn, 0, (8 * sizeof(uint8_t)));
phba->cfg_oas_lun_state = 0;
phba->cfg_oas_lun_status = 0;
phba->cfg_oas_flags = 0;
phba->cfg_oas_priority = 0;
lpfc_enable_bg_init(phba, lpfc_enable_bg);
lpfc_prot_mask_init(phba, lpfc_prot_mask);
lpfc_prot_guard_init(phba, lpfc_prot_guard);
if (phba->sli_rev == LPFC_SLI_REV4)
phba->cfg_poll = 0;
else
phba->cfg_poll = lpfc_poll;
lpfc_suppress_rsp_init(phba, lpfc_suppress_rsp);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
lpfc_enable_fc4_type_init(phba, lpfc_enable_fc4_type);
lpfc_nvmet_mrq_init(phba, lpfc_nvmet_mrq);
lpfc_nvmet_mrq_post_init(phba, lpfc_nvmet_mrq_post);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
/* Initialize first burst. Target vs Initiator are different. */
lpfc_nvme_enable_fb_init(phba, lpfc_nvme_enable_fb);
lpfc_nvmet_fb_size_init(phba, lpfc_nvmet_fb_size);
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
lpfc_fcp_io_channel_init(phba, lpfc_fcp_io_channel);
lpfc_nvme_io_channel_init(phba, lpfc_nvme_io_channel);
if (phba->sli_rev != LPFC_SLI_REV4) {
/* NVME only supported on SLI4 */
phba->nvmet_support = 0;
phba->cfg_enable_fc4_type = LPFC_ENABLE_FCP;
} else {
/* We MUST have FCP support */
if (!(phba->cfg_enable_fc4_type & LPFC_ENABLE_FCP))
phba->cfg_enable_fc4_type |= LPFC_ENABLE_FCP;
}
/* A value of 0 means use the number of CPUs found in the system */
if (phba->cfg_fcp_io_channel == 0)
phba->cfg_fcp_io_channel = phba->sli4_hba.num_present_cpu;
if (phba->cfg_nvme_io_channel == 0)
phba->cfg_nvme_io_channel = phba->sli4_hba.num_present_cpu;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (phba->cfg_enable_fc4_type == LPFC_ENABLE_NVME)
phba->cfg_fcp_io_channel = 0;
if (phba->cfg_enable_fc4_type == LPFC_ENABLE_FCP)
phba->cfg_nvme_io_channel = 0;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (phba->cfg_fcp_io_channel > phba->cfg_nvme_io_channel)
phba->io_channel_irqs = phba->cfg_fcp_io_channel;
else
phba->io_channel_irqs = phba->cfg_nvme_io_channel;
phba->cfg_soft_wwnn = 0L;
phba->cfg_soft_wwpn = 0L;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
lpfc_xri_split_init(phba, lpfc_xri_split);
lpfc_sg_seg_cnt_init(phba, lpfc_sg_seg_cnt);
lpfc_hba_queue_depth_init(phba, lpfc_hba_queue_depth);
lpfc_hba_log_verbose_init(phba, lpfc_log_verbose);
lpfc_aer_support_init(phba, lpfc_aer_support);
lpfc_sriov_nr_virtfn_init(phba, lpfc_sriov_nr_virtfn);
lpfc_request_firmware_upgrade_init(phba, lpfc_req_fw_upgrade);
lpfc_suppress_link_up_init(phba, lpfc_suppress_link_up);
lpfc_iocb_cnt_init(phba, lpfc_iocb_cnt);
lpfc_delay_discovery_init(phba, lpfc_delay_discovery);
lpfc_sli_mode_init(phba, lpfc_sli_mode);
phba->cfg_enable_dss = 1;
lpfc_enable_mds_diags_init(phba, lpfc_enable_mds_diags);
return;
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
/**
* lpfc_nvme_mod_param_dep - Adjust module parameter value based on
* dependencies between protocols and roles.
* @phba: lpfc_hba pointer.
**/
void
lpfc_nvme_mod_param_dep(struct lpfc_hba *phba)
{
if (phba->cfg_nvme_io_channel > phba->sli4_hba.num_present_cpu)
phba->cfg_nvme_io_channel = phba->sli4_hba.num_present_cpu;
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (phba->cfg_fcp_io_channel > phba->sli4_hba.num_present_cpu)
phba->cfg_fcp_io_channel = phba->sli4_hba.num_present_cpu;
if (phba->cfg_enable_fc4_type & LPFC_ENABLE_NVME &&
phba->nvmet_support) {
phba->cfg_enable_fc4_type &= ~LPFC_ENABLE_FCP;
phba->cfg_fcp_io_channel = 0;
lpfc_printf_log(phba, KERN_INFO, LOG_NVME_DISC,
"6013 %s x%x fb_size x%x, fb_max x%x\n",
"NVME Target PRLI ACC enable_fb ",
phba->cfg_nvme_enable_fb,
phba->cfg_nvmet_fb_size,
LPFC_NVMET_FB_SZ_MAX);
if (phba->cfg_nvme_enable_fb == 0)
phba->cfg_nvmet_fb_size = 0;
else {
if (phba->cfg_nvmet_fb_size > LPFC_NVMET_FB_SZ_MAX)
phba->cfg_nvmet_fb_size = LPFC_NVMET_FB_SZ_MAX;
}
/* Adjust lpfc_nvmet_mrq to avoid running out of WQE slots */
if (phba->cfg_nvmet_mrq > phba->cfg_nvme_io_channel) {
phba->cfg_nvmet_mrq = phba->cfg_nvme_io_channel;
lpfc_printf_log(phba, KERN_ERR, LOG_NVME_DISC,
"6018 Adjust lpfc_nvmet_mrq to %d\n",
phba->cfg_nvmet_mrq);
}
} else {
/* Not NVME Target mode. Turn off Target parameters. */
phba->nvmet_support = 0;
phba->cfg_nvmet_mrq = 0;
phba->cfg_nvmet_mrq_post = 0;
phba->cfg_nvmet_fb_size = 0;
}
scsi: lpfc: NVME Initiator: Base modifications NVME Initiator: Base modifications This patch adds base modifications for NVME initiator support. The base modifications consist of: - Formal split of SLI3 rings from SLI-4 WQs (sometimes referred to as rings as well) as implementation now widely varies between the two. - Addition of configuration modes: SCSI initiator only; NVME initiator only; NVME target only; and SCSI and NVME initiator. The configuration mode drives overall adapter configuration, offloads enabled, and resource splits. NVME support is only available on SLI-4 devices and newer fw. - Implements the following based on configuration mode: - Exchange resources are split by protocol; Obviously, if only 1 mode, then no split occurs. Default is 50/50. module attribute allows tuning. - Pools and config parameters are separated per-protocol - Each protocol has it's own set of queues, but share interrupt vectors. SCSI: SLI3 devices have few queues and the original style of queue allocation remains. SLI4 devices piggy back on an "io-channel" concept that eventually needs to merge with scsi-mq/blk-mq support (it is underway). For now, the paradigm continues as it existed prior. io channel allocates N msix and N WQs (N=4 default) and either round robins or uses cpu # modulo N for scheduling. A bunch of module parameters allow the configuration to be tuned. NVME (initiator): Allocates an msix per cpu (or whatever pci_alloc_irq_vectors gets) Allocates a WQ per cpu, and maps the WQs to msix on a WQ # modulo msix vector count basis. Module parameters exist to cap/control the config if desired. - Each protocol has its own buffer and dma pools. I apologize for the size of the patch. Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com> Signed-off-by: James Smart <james.smart@broadcom.com> ---- Reviewed-by: Hannes Reinecke <hare@suse.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2017-02-13 04:52:30 +07:00
if (phba->cfg_fcp_io_channel > phba->cfg_nvme_io_channel)
phba->io_channel_irqs = phba->cfg_fcp_io_channel;
else
phba->io_channel_irqs = phba->cfg_nvme_io_channel;
}
/**
* lpfc_get_vport_cfgparam - Used during port create, init the vport structure
* @vport: lpfc_vport pointer.
**/
void
lpfc_get_vport_cfgparam(struct lpfc_vport *vport)
{
lpfc_log_verbose_init(vport, lpfc_log_verbose);
lpfc_lun_queue_depth_init(vport, lpfc_lun_queue_depth);
lpfc_tgt_queue_depth_init(vport, lpfc_tgt_queue_depth);
lpfc_devloss_tmo_init(vport, lpfc_devloss_tmo);
lpfc_nodev_tmo_init(vport, lpfc_nodev_tmo);
lpfc_peer_port_login_init(vport, lpfc_peer_port_login);
lpfc_restrict_login_init(vport, lpfc_restrict_login);
lpfc_fcp_class_init(vport, lpfc_fcp_class);
lpfc_use_adisc_init(vport, lpfc_use_adisc);
lpfc_first_burst_size_init(vport, lpfc_first_burst_size);
lpfc_max_scsicmpl_time_init(vport, lpfc_max_scsicmpl_time);
lpfc_discovery_threads_init(vport, lpfc_discovery_threads);
lpfc_max_luns_init(vport, lpfc_max_luns);
lpfc_scan_down_init(vport, lpfc_scan_down);
lpfc_enable_da_id_init(vport, lpfc_enable_da_id);
return;
}