2005-09-14 02:50:38 +07:00
|
|
|
/*
|
2008-06-10 23:20:58 +07:00
|
|
|
* zfcp device driver
|
2005-09-14 02:50:38 +07:00
|
|
|
*
|
2008-06-10 23:20:58 +07:00
|
|
|
* Debug traces for zfcp.
|
2005-09-14 02:50:38 +07:00
|
|
|
*
|
2016-08-10 23:30:53 +07:00
|
|
|
* Copyright IBM Corp. 2002, 2016
|
2005-09-14 02:50:38 +07:00
|
|
|
*/
|
|
|
|
|
2008-12-25 19:39:53 +07:00
|
|
|
#define KMSG_COMPONENT "zfcp"
|
|
|
|
#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
|
|
|
|
|
2011-07-30 14:25:15 +07:00
|
|
|
#include <linux/module.h>
|
2005-09-14 02:50:38 +07:00
|
|
|
#include <linux/ctype.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/slab.h>
|
2007-10-12 21:11:35 +07:00
|
|
|
#include <asm/debug.h>
|
2009-08-18 20:43:07 +07:00
|
|
|
#include "zfcp_dbf.h"
|
2005-09-14 02:50:38 +07:00
|
|
|
#include "zfcp_ext.h"
|
2009-11-24 22:54:11 +07:00
|
|
|
#include "zfcp_fc.h"
|
2005-09-14 02:50:38 +07:00
|
|
|
|
|
|
|
static u32 dbfsize = 4;
|
|
|
|
|
|
|
|
module_param(dbfsize, uint, 0400);
|
|
|
|
MODULE_PARM_DESC(dbfsize,
|
|
|
|
"number of pages for each debug feature area (default 4)");
|
|
|
|
|
2013-04-26 21:13:53 +07:00
|
|
|
static u32 dbflevel = 3;
|
|
|
|
|
|
|
|
module_param(dbflevel, uint, 0400);
|
|
|
|
MODULE_PARM_DESC(dbflevel,
|
|
|
|
"log level for each debug feature area "
|
|
|
|
"(default 3, range 0..6)");
|
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
static inline unsigned int zfcp_dbf_plen(unsigned int offset)
|
2008-03-27 20:21:55 +07:00
|
|
|
{
|
2010-12-02 21:16:14 +07:00
|
|
|
return sizeof(struct zfcp_dbf_pay) + offset - ZFCP_DBF_PAY_MAX_REC;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline
|
|
|
|
void zfcp_dbf_pl_write(struct zfcp_dbf *dbf, void *data, u16 length, char *area,
|
|
|
|
u64 req_id)
|
|
|
|
{
|
|
|
|
struct zfcp_dbf_pay *pl = &dbf->pay_buf;
|
|
|
|
u16 offset = 0, rec_length;
|
|
|
|
|
|
|
|
spin_lock(&dbf->pay_lock);
|
|
|
|
memset(pl, 0, sizeof(*pl));
|
|
|
|
pl->fsf_req_id = req_id;
|
|
|
|
memcpy(pl->area, area, ZFCP_DBF_TAG_LEN);
|
|
|
|
|
|
|
|
while (offset < length) {
|
|
|
|
rec_length = min((u16) ZFCP_DBF_PAY_MAX_REC,
|
|
|
|
(u16) (length - offset));
|
|
|
|
memcpy(pl->data, data + offset, rec_length);
|
|
|
|
debug_event(dbf->pay, 1, pl, zfcp_dbf_plen(rec_length));
|
|
|
|
|
|
|
|
offset += rec_length;
|
|
|
|
pl->counter++;
|
2008-03-27 20:21:55 +07:00
|
|
|
}
|
2010-12-02 21:16:14 +07:00
|
|
|
|
|
|
|
spin_unlock(&dbf->pay_lock);
|
2008-03-27 20:21:55 +07:00
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_hba_fsf_res - trace event for fsf responses
|
|
|
|
* @tag: tag indicating which kind of unsolicited status has been received
|
|
|
|
* @req: request for which a response was received
|
|
|
|
*/
|
2016-08-10 23:30:47 +07:00
|
|
|
void zfcp_dbf_hba_fsf_res(char *tag, int level, struct zfcp_fsf_req *req)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:14 +07:00
|
|
|
struct zfcp_dbf *dbf = req->adapter->dbf;
|
|
|
|
struct fsf_qtcb_prefix *q_pref = &req->qtcb->prefix;
|
|
|
|
struct fsf_qtcb_header *q_head = &req->qtcb->header;
|
|
|
|
struct zfcp_dbf_hba *rec = &dbf->hba_buf;
|
2005-09-14 02:50:38 +07:00
|
|
|
unsigned long flags;
|
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->hba_lock, flags);
|
2008-03-31 16:15:29 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->id = ZFCP_DBF_HBA_RES;
|
|
|
|
rec->fsf_req_id = req->req_id;
|
|
|
|
rec->fsf_req_status = req->status;
|
|
|
|
rec->fsf_cmd = req->fsf_command;
|
|
|
|
rec->fsf_seq_no = req->seq_no;
|
|
|
|
rec->u.res.req_issued = req->issued;
|
|
|
|
rec->u.res.prot_status = q_pref->prot_status;
|
|
|
|
rec->u.res.fsf_status = q_head->fsf_status;
|
2016-08-10 23:30:50 +07:00
|
|
|
rec->u.res.port_handle = q_head->port_handle;
|
|
|
|
rec->u.res.lun_handle = q_head->lun_handle;
|
2010-12-02 21:16:14 +07:00
|
|
|
|
|
|
|
memcpy(rec->u.res.prot_status_qual, &q_pref->prot_status_qual,
|
|
|
|
FSF_PROT_STATUS_QUAL_SIZE);
|
|
|
|
memcpy(rec->u.res.fsf_status_qual, &q_head->fsf_status_qual,
|
|
|
|
FSF_STATUS_QUALIFIER_SIZE);
|
|
|
|
|
|
|
|
if (req->fsf_command != FSF_QTCB_FCP_CMND) {
|
|
|
|
rec->pl_len = q_head->log_length;
|
|
|
|
zfcp_dbf_pl_write(dbf, (char *)q_pref + q_head->log_start,
|
|
|
|
rec->pl_len, "fsf_res", req->req_id);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2016-08-10 23:30:47 +07:00
|
|
|
debug_event(dbf->hba, level, rec, sizeof(*rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_unlock_irqrestore(&dbf->hba_lock, flags);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2008-03-31 16:15:31 +07:00
|
|
|
/**
|
2010-12-02 21:16:14 +07:00
|
|
|
* zfcp_dbf_hba_fsf_uss - trace event for an unsolicited status buffer
|
|
|
|
* @tag: tag indicating which kind of unsolicited status has been received
|
|
|
|
* @req: request providing the unsolicited status
|
2008-03-31 16:15:31 +07:00
|
|
|
*/
|
2010-12-02 21:16:14 +07:00
|
|
|
void zfcp_dbf_hba_fsf_uss(char *tag, struct zfcp_fsf_req *req)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:14 +07:00
|
|
|
struct zfcp_dbf *dbf = req->adapter->dbf;
|
|
|
|
struct fsf_status_read_buffer *srb = req->data;
|
|
|
|
struct zfcp_dbf_hba *rec = &dbf->hba_buf;
|
2005-09-14 02:50:38 +07:00
|
|
|
unsigned long flags;
|
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->hba_lock, flags);
|
2010-12-02 21:16:14 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
|
|
|
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->id = ZFCP_DBF_HBA_USS;
|
|
|
|
rec->fsf_req_id = req->req_id;
|
|
|
|
rec->fsf_req_status = req->status;
|
|
|
|
rec->fsf_cmd = req->fsf_command;
|
|
|
|
|
|
|
|
if (!srb)
|
|
|
|
goto log;
|
|
|
|
|
|
|
|
rec->u.uss.status_type = srb->status_type;
|
|
|
|
rec->u.uss.status_subtype = srb->status_subtype;
|
|
|
|
rec->u.uss.d_id = ntoh24(srb->d_id);
|
|
|
|
rec->u.uss.lun = srb->fcp_lun;
|
|
|
|
memcpy(&rec->u.uss.queue_designator, &srb->queue_designator,
|
|
|
|
sizeof(rec->u.uss.queue_designator));
|
|
|
|
|
|
|
|
/* status read buffer payload length */
|
|
|
|
rec->pl_len = (!srb->length) ? 0 : srb->length -
|
|
|
|
offsetof(struct fsf_status_read_buffer, payload);
|
|
|
|
|
|
|
|
if (rec->pl_len)
|
|
|
|
zfcp_dbf_pl_write(dbf, srb->payload.data, rec->pl_len,
|
|
|
|
"fsf_uss", req->req_id);
|
|
|
|
log:
|
|
|
|
debug_event(dbf->hba, 2, rec, sizeof(*rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_unlock_irqrestore(&dbf->hba_lock, flags);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2008-10-01 17:42:21 +07:00
|
|
|
/**
|
2010-12-02 21:16:14 +07:00
|
|
|
* zfcp_dbf_hba_bit_err - trace event for bit error conditions
|
|
|
|
* @tag: tag indicating which kind of unsolicited status has been received
|
|
|
|
* @req: request which caused the bit_error condition
|
2008-10-01 17:42:21 +07:00
|
|
|
*/
|
2010-12-02 21:16:14 +07:00
|
|
|
void zfcp_dbf_hba_bit_err(char *tag, struct zfcp_fsf_req *req)
|
2008-10-01 17:42:21 +07:00
|
|
|
{
|
2010-12-02 21:16:14 +07:00
|
|
|
struct zfcp_dbf *dbf = req->adapter->dbf;
|
|
|
|
struct zfcp_dbf_hba *rec = &dbf->hba_buf;
|
2008-10-01 17:42:21 +07:00
|
|
|
struct fsf_status_read_buffer *sr_buf = req->data;
|
|
|
|
unsigned long flags;
|
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->hba_lock, flags);
|
2010-12-02 21:16:14 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->id = ZFCP_DBF_HBA_BIT;
|
|
|
|
rec->fsf_req_id = req->req_id;
|
|
|
|
rec->fsf_req_status = req->status;
|
|
|
|
rec->fsf_cmd = req->fsf_command;
|
|
|
|
memcpy(&rec->u.be, &sr_buf->payload.bit_error,
|
|
|
|
sizeof(struct fsf_bit_error_payload));
|
|
|
|
|
|
|
|
debug_event(dbf->hba, 1, rec, sizeof(*rec));
|
|
|
|
spin_unlock_irqrestore(&dbf->hba_lock, flags);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2011-08-15 19:40:32 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_hba_def_err - trace event for deferred error messages
|
|
|
|
* @adapter: pointer to struct zfcp_adapter
|
|
|
|
* @req_id: request id which caused the deferred error message
|
|
|
|
* @scount: number of sbals incl. the signaling sbal
|
|
|
|
* @pl: array of all involved sbals
|
|
|
|
*/
|
|
|
|
void zfcp_dbf_hba_def_err(struct zfcp_adapter *adapter, u64 req_id, u16 scount,
|
|
|
|
void **pl)
|
|
|
|
{
|
|
|
|
struct zfcp_dbf *dbf = adapter->dbf;
|
|
|
|
struct zfcp_dbf_pay *payload = &dbf->pay_buf;
|
|
|
|
unsigned long flags;
|
|
|
|
u16 length;
|
|
|
|
|
|
|
|
if (!pl)
|
|
|
|
return;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&dbf->pay_lock, flags);
|
|
|
|
memset(payload, 0, sizeof(*payload));
|
|
|
|
|
|
|
|
memcpy(payload->area, "def_err", 7);
|
|
|
|
payload->fsf_req_id = req_id;
|
|
|
|
payload->counter = 0;
|
|
|
|
length = min((u16)sizeof(struct qdio_buffer),
|
|
|
|
(u16)ZFCP_DBF_PAY_MAX_REC);
|
|
|
|
|
2012-09-04 20:23:31 +07:00
|
|
|
while (payload->counter < scount && (char *)pl[payload->counter]) {
|
2011-08-15 19:40:32 +07:00
|
|
|
memcpy(payload->data, (char *)pl[payload->counter], length);
|
|
|
|
debug_event(dbf->pay, 1, payload, zfcp_dbf_plen(length));
|
|
|
|
payload->counter++;
|
|
|
|
}
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&dbf->pay_lock, flags);
|
|
|
|
}
|
|
|
|
|
[SCSI] zfcp: Do not wakeup while suspended
If the mapping of FCP device bus ID and corresponding subchannel
is modified while the Linux image is suspended, the resume of FCP
devices can fail. During resume, zfcp gets callbacks from cio regarding
the modified subchannels but they can be arbitrarily mixed with the
restore/resume callback. Since the cio callbacks would trigger
adapter recovery, zfcp could wakeup before the resume callback.
Therefore, ignore the cio callbacks regarding subchannels while
being suspended. We can safely do so, since zfcp does not deal itself
with subchannels. For problem determination purposes, we still trace the
ignored callback events.
The following kernel messages could be seen on resume:
kernel: <WWPN>: parent <FCP device bus ID> should not be sleeping
As part of adapter reopen recovery, zfcp performs auto port scanning
which can erroneously try to register new remote ports with
scsi_transport_fc and the device core code complains about the parent
(adapter) still sleeping.
kernel: zfcp.3dff9c: <FCP device bus ID>:\
Setting up the QDIO connection to the FCP adapter failed
<last kernel message repeated 3 more times>
kernel: zfcp.574d43: <FCP device bus ID>:\
ERP cannot recover an error on the FCP device
In such cases, the adapter gave up recovery and remained blocked along
with its child objects: remote ports and LUNs/scsi devices. Even the
adapter shutdown as part of giving up recovery failed because the ccw
device state remained disconnected. Later, the corresponding remote
ports ran into dev_loss_tmo. As a result, the LUNs were erroneously
not available again after resume.
Even a manually triggered adapter recovery (e.g. sysfs attribute
failed, or device offline/online via sysfs) could not recover the
adapter due to the remaining disconnected state of the corresponding
ccw device.
Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com>
Cc: <stable@vger.kernel.org> #2.6.32+
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2012-09-04 20:23:32 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_hba_basic - trace event for basic adapter events
|
|
|
|
* @adapter: pointer to struct zfcp_adapter
|
|
|
|
*/
|
|
|
|
void zfcp_dbf_hba_basic(char *tag, struct zfcp_adapter *adapter)
|
|
|
|
{
|
|
|
|
struct zfcp_dbf *dbf = adapter->dbf;
|
|
|
|
struct zfcp_dbf_hba *rec = &dbf->hba_buf;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&dbf->hba_lock, flags);
|
|
|
|
memset(rec, 0, sizeof(*rec));
|
|
|
|
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->id = ZFCP_DBF_HBA_BASIC;
|
|
|
|
|
|
|
|
debug_event(dbf->hba, 1, rec, sizeof(*rec));
|
|
|
|
spin_unlock_irqrestore(&dbf->hba_lock, flags);
|
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
static void zfcp_dbf_set_common(struct zfcp_dbf_rec *rec,
|
|
|
|
struct zfcp_adapter *adapter,
|
|
|
|
struct zfcp_port *port,
|
|
|
|
struct scsi_device *sdev)
|
2008-03-27 20:22:00 +07:00
|
|
|
{
|
2010-12-02 21:16:12 +07:00
|
|
|
rec->adapter_status = atomic_read(&adapter->status);
|
|
|
|
if (port) {
|
|
|
|
rec->port_status = atomic_read(&port->status);
|
|
|
|
rec->wwpn = port->wwpn;
|
|
|
|
rec->d_id = port->d_id;
|
|
|
|
}
|
|
|
|
if (sdev) {
|
|
|
|
rec->lun_status = atomic_read(&sdev_to_zfcp(sdev)->status);
|
|
|
|
rec->lun = zfcp_scsi_dev_lun(sdev);
|
2016-08-10 23:30:48 +07:00
|
|
|
} else
|
|
|
|
rec->lun = ZFCP_DBF_INVALID_LUN;
|
2008-03-27 20:22:00 +07:00
|
|
|
}
|
|
|
|
|
2008-03-27 20:22:01 +07:00
|
|
|
/**
|
2010-12-02 21:16:12 +07:00
|
|
|
* zfcp_dbf_rec_trig - trace event related to triggered recovery
|
|
|
|
* @tag: identifier for event
|
|
|
|
* @adapter: adapter on which the erp_action should run
|
|
|
|
* @port: remote port involved in the erp_action
|
|
|
|
* @sdev: scsi device involved in the erp_action
|
|
|
|
* @want: wanted erp_action
|
|
|
|
* @need: required erp_action
|
|
|
|
*
|
|
|
|
* The adapter->erp_lock has to be held.
|
2008-03-27 20:22:01 +07:00
|
|
|
*/
|
2010-12-02 21:16:12 +07:00
|
|
|
void zfcp_dbf_rec_trig(char *tag, struct zfcp_adapter *adapter,
|
|
|
|
struct zfcp_port *port, struct scsi_device *sdev,
|
|
|
|
u8 want, u8 need)
|
2008-03-27 20:22:01 +07:00
|
|
|
{
|
2010-12-02 21:16:12 +07:00
|
|
|
struct zfcp_dbf *dbf = adapter->dbf;
|
|
|
|
struct zfcp_dbf_rec *rec = &dbf->rec_buf;
|
2008-03-27 20:22:01 +07:00
|
|
|
struct list_head *entry;
|
2008-03-27 20:22:02 +07:00
|
|
|
unsigned long flags;
|
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->rec_lock, flags);
|
2010-12-02 21:16:12 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
rec->id = ZFCP_DBF_REC_TRIG;
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
zfcp_dbf_set_common(rec, adapter, port, sdev);
|
2008-03-27 20:22:02 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
list_for_each(entry, &adapter->erp_ready_head)
|
|
|
|
rec->u.trig.ready++;
|
2008-03-27 20:22:02 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
list_for_each(entry, &adapter->erp_running_head)
|
|
|
|
rec->u.trig.running++;
|
2008-03-27 20:22:02 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
rec->u.trig.want = want;
|
|
|
|
rec->u.trig.need = need;
|
2008-03-27 20:22:02 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
debug_event(dbf->rec, 1, rec, sizeof(*rec));
|
|
|
|
spin_unlock_irqrestore(&dbf->rec_lock, flags);
|
2008-03-27 20:22:02 +07:00
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
|
2008-03-27 20:22:03 +07:00
|
|
|
/**
|
scsi: zfcp: fix rport unblock race with LUN recovery
It is unavoidable that zfcp_scsi_queuecommand() has to finish requests
with DID_IMM_RETRY (like fc_remote_port_chkready()) during the time
window when zfcp detected an unavailable rport but
fc_remote_port_delete(), which is asynchronous via
zfcp_scsi_schedule_rport_block(), has not yet blocked the rport.
However, for the case when the rport becomes available again, we should
prevent unblocking the rport too early. In contrast to other FCP LLDDs,
zfcp has to open each LUN with the FCP channel hardware before it can
send I/O to a LUN. So if a port already has LUNs attached and we
unblock the rport just after port recovery, recoveries of LUNs behind
this port can still be pending which in turn force
zfcp_scsi_queuecommand() to unnecessarily finish requests with
DID_IMM_RETRY.
This also opens a time window with unblocked rport (until the followup
LUN reopen recovery has finished). If a scsi_cmnd timeout occurs during
this time window fc_timed_out() cannot work as desired and such command
would indeed time out and trigger scsi_eh. This prevents a clean and
timely path failover. This should not happen if the path issue can be
recovered on FC transport layer such as path issues involving RSCNs.
Fix this by only calling zfcp_scsi_schedule_rport_register(), to
asynchronously trigger fc_remote_port_add(), after all LUN recoveries as
children of the rport have finished and no new recoveries of equal or
higher order were triggered meanwhile. Finished intentionally includes
any recovery result no matter if successful or failed (still unblock
rport so other successful LUNs work). For simplicity, we check after
each finished LUN recovery if there is another LUN recovery pending on
the same port and then do nothing. We handle the special case of a
successful recovery of a port without LUN children the same way without
changing this case's semantics.
For debugging we introduce 2 new trace records written if the rport
unblock attempt was aborted due to still unfinished or freshly triggered
recovery. The records are only written above the default trace level.
Benjamin noticed the important special case of new recovery that can be
triggered between having given up the erp_lock and before calling
zfcp_erp_action_cleanup() within zfcp_erp_strategy(). We must avoid the
following sequence:
ERP thread rport_work other context
------------------------- -------------- --------------------------------
port is unblocked, rport still blocked,
due to pending/running ERP action,
so ((port->status & ...UNBLOCK) != 0)
and (port->rport == NULL)
unlock ERP
zfcp_erp_action_cleanup()
case ZFCP_ERP_ACTION_REOPEN_LUN:
zfcp_erp_try_rport_unblock()
((status & ...UNBLOCK) != 0) [OLD!]
zfcp_erp_port_reopen()
lock ERP
zfcp_erp_port_block()
port->status clear ...UNBLOCK
unlock ERP
zfcp_scsi_schedule_rport_block()
port->rport_task = RPORT_DEL
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task != RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_block()
if (!port->rport) return
zfcp_scsi_schedule_rport_register()
port->rport_task = RPORT_ADD
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task == RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_register()
(port->rport == NULL)
rport = fc_remote_port_add()
port->rport = rport;
Now the rport was erroneously unblocked while the zfcp_port is blocked.
This is another situation we want to avoid due to scsi_eh
potential. This state would at least remain until the new recovery from
the other context finished successfully, or potentially forever if it
failed. In order to close this race, we take the erp_lock inside
zfcp_erp_try_rport_unblock() when checking the status of zfcp_port or
LUN. With that, the possible corresponding rport state sequences would
be: (unblock[ERP thread],block[other context]) if the ERP thread gets
erp_lock first and still sees ((port->status & ...UNBLOCK) != 0),
(block[other context],NOP[ERP thread]) if the ERP thread gets erp_lock
after the other context has already cleard ...UNBLOCK from port->status.
Since checking fields of struct erp_action is unsafe because they could
have been overwritten (re-used for new recovery) meanwhile, we only
check status of zfcp_port and LUN since these are only changed under
erp_lock elsewhere. Regarding the check of the proper status flags (port
or port_forced are similar to the shown adapter recovery):
[zfcp_erp_adapter_shutdown()]
zfcp_erp_adapter_reopen()
zfcp_erp_adapter_block()
* clear UNBLOCK ---------------------------------------+
zfcp_scsi_schedule_rports_block() |
write_lock_irqsave(&adapter->erp_lock, flags);-------+ |
zfcp_erp_action_enqueue() | |
zfcp_erp_setup_act() | |
* set ERP_INUSE -----------------------------------|--|--+
write_unlock_irqrestore(&adapter->erp_lock, flags);--+ | |
.context-switch. | |
zfcp_erp_thread() | |
zfcp_erp_strategy() | |
write_lock_irqsave(&adapter->erp_lock, flags);------+ | |
... | | |
zfcp_erp_strategy_check_target() | | |
zfcp_erp_strategy_check_adapter() | | |
zfcp_erp_adapter_unblock() | | |
* set UNBLOCK -----------------------------------|--+ |
zfcp_erp_action_dequeue() | |
* clear ERP_INUSE ---------------------------------|-----+
... |
write_unlock_irqrestore(&adapter->erp_lock, flags);-+
Hence, we should check for both UNBLOCK and ERP_INUSE because they are
interleaved. Also we need to explicitly check ERP_FAILED for the link
down case which currently does not clear the UNBLOCK flag in
zfcp_fsf_link_down_info_eval().
Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com>
Fixes: 8830271c4819 ("[SCSI] zfcp: Dont fail SCSI commands when transitioning to blocked fc_rport")
Fixes: a2fa0aede07c ("[SCSI] zfcp: Block FC transport rports early on errors")
Fixes: 5f852be9e11d ("[SCSI] zfcp: Fix deadlock between zfcp ERP and SCSI")
Fixes: 338151e06608 ("[SCSI] zfcp: make use of fc_remote_port_delete when target port is unavailable")
Fixes: 3859f6a248cb ("[PATCH] zfcp: add rports to enable scsi_add_device to work again")
Cc: <stable@vger.kernel.org> #2.6.32+
Reviewed-by: Benjamin Block <bblock@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-12-09 23:16:33 +07:00
|
|
|
* zfcp_dbf_rec_run_lvl - trace event related to running recovery
|
|
|
|
* @level: trace level to be used for event
|
2010-12-02 21:16:12 +07:00
|
|
|
* @tag: identifier for event
|
|
|
|
* @erp: erp_action running
|
2008-03-27 20:22:03 +07:00
|
|
|
*/
|
scsi: zfcp: fix rport unblock race with LUN recovery
It is unavoidable that zfcp_scsi_queuecommand() has to finish requests
with DID_IMM_RETRY (like fc_remote_port_chkready()) during the time
window when zfcp detected an unavailable rport but
fc_remote_port_delete(), which is asynchronous via
zfcp_scsi_schedule_rport_block(), has not yet blocked the rport.
However, for the case when the rport becomes available again, we should
prevent unblocking the rport too early. In contrast to other FCP LLDDs,
zfcp has to open each LUN with the FCP channel hardware before it can
send I/O to a LUN. So if a port already has LUNs attached and we
unblock the rport just after port recovery, recoveries of LUNs behind
this port can still be pending which in turn force
zfcp_scsi_queuecommand() to unnecessarily finish requests with
DID_IMM_RETRY.
This also opens a time window with unblocked rport (until the followup
LUN reopen recovery has finished). If a scsi_cmnd timeout occurs during
this time window fc_timed_out() cannot work as desired and such command
would indeed time out and trigger scsi_eh. This prevents a clean and
timely path failover. This should not happen if the path issue can be
recovered on FC transport layer such as path issues involving RSCNs.
Fix this by only calling zfcp_scsi_schedule_rport_register(), to
asynchronously trigger fc_remote_port_add(), after all LUN recoveries as
children of the rport have finished and no new recoveries of equal or
higher order were triggered meanwhile. Finished intentionally includes
any recovery result no matter if successful or failed (still unblock
rport so other successful LUNs work). For simplicity, we check after
each finished LUN recovery if there is another LUN recovery pending on
the same port and then do nothing. We handle the special case of a
successful recovery of a port without LUN children the same way without
changing this case's semantics.
For debugging we introduce 2 new trace records written if the rport
unblock attempt was aborted due to still unfinished or freshly triggered
recovery. The records are only written above the default trace level.
Benjamin noticed the important special case of new recovery that can be
triggered between having given up the erp_lock and before calling
zfcp_erp_action_cleanup() within zfcp_erp_strategy(). We must avoid the
following sequence:
ERP thread rport_work other context
------------------------- -------------- --------------------------------
port is unblocked, rport still blocked,
due to pending/running ERP action,
so ((port->status & ...UNBLOCK) != 0)
and (port->rport == NULL)
unlock ERP
zfcp_erp_action_cleanup()
case ZFCP_ERP_ACTION_REOPEN_LUN:
zfcp_erp_try_rport_unblock()
((status & ...UNBLOCK) != 0) [OLD!]
zfcp_erp_port_reopen()
lock ERP
zfcp_erp_port_block()
port->status clear ...UNBLOCK
unlock ERP
zfcp_scsi_schedule_rport_block()
port->rport_task = RPORT_DEL
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task != RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_block()
if (!port->rport) return
zfcp_scsi_schedule_rport_register()
port->rport_task = RPORT_ADD
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task == RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_register()
(port->rport == NULL)
rport = fc_remote_port_add()
port->rport = rport;
Now the rport was erroneously unblocked while the zfcp_port is blocked.
This is another situation we want to avoid due to scsi_eh
potential. This state would at least remain until the new recovery from
the other context finished successfully, or potentially forever if it
failed. In order to close this race, we take the erp_lock inside
zfcp_erp_try_rport_unblock() when checking the status of zfcp_port or
LUN. With that, the possible corresponding rport state sequences would
be: (unblock[ERP thread],block[other context]) if the ERP thread gets
erp_lock first and still sees ((port->status & ...UNBLOCK) != 0),
(block[other context],NOP[ERP thread]) if the ERP thread gets erp_lock
after the other context has already cleard ...UNBLOCK from port->status.
Since checking fields of struct erp_action is unsafe because they could
have been overwritten (re-used for new recovery) meanwhile, we only
check status of zfcp_port and LUN since these are only changed under
erp_lock elsewhere. Regarding the check of the proper status flags (port
or port_forced are similar to the shown adapter recovery):
[zfcp_erp_adapter_shutdown()]
zfcp_erp_adapter_reopen()
zfcp_erp_adapter_block()
* clear UNBLOCK ---------------------------------------+
zfcp_scsi_schedule_rports_block() |
write_lock_irqsave(&adapter->erp_lock, flags);-------+ |
zfcp_erp_action_enqueue() | |
zfcp_erp_setup_act() | |
* set ERP_INUSE -----------------------------------|--|--+
write_unlock_irqrestore(&adapter->erp_lock, flags);--+ | |
.context-switch. | |
zfcp_erp_thread() | |
zfcp_erp_strategy() | |
write_lock_irqsave(&adapter->erp_lock, flags);------+ | |
... | | |
zfcp_erp_strategy_check_target() | | |
zfcp_erp_strategy_check_adapter() | | |
zfcp_erp_adapter_unblock() | | |
* set UNBLOCK -----------------------------------|--+ |
zfcp_erp_action_dequeue() | |
* clear ERP_INUSE ---------------------------------|-----+
... |
write_unlock_irqrestore(&adapter->erp_lock, flags);-+
Hence, we should check for both UNBLOCK and ERP_INUSE because they are
interleaved. Also we need to explicitly check ERP_FAILED for the link
down case which currently does not clear the UNBLOCK flag in
zfcp_fsf_link_down_info_eval().
Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com>
Fixes: 8830271c4819 ("[SCSI] zfcp: Dont fail SCSI commands when transitioning to blocked fc_rport")
Fixes: a2fa0aede07c ("[SCSI] zfcp: Block FC transport rports early on errors")
Fixes: 5f852be9e11d ("[SCSI] zfcp: Fix deadlock between zfcp ERP and SCSI")
Fixes: 338151e06608 ("[SCSI] zfcp: make use of fc_remote_port_delete when target port is unavailable")
Fixes: 3859f6a248cb ("[PATCH] zfcp: add rports to enable scsi_add_device to work again")
Cc: <stable@vger.kernel.org> #2.6.32+
Reviewed-by: Benjamin Block <bblock@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-12-09 23:16:33 +07:00
|
|
|
void zfcp_dbf_rec_run_lvl(int level, char *tag, struct zfcp_erp_action *erp)
|
2008-03-27 20:22:03 +07:00
|
|
|
{
|
2010-12-02 21:16:12 +07:00
|
|
|
struct zfcp_dbf *dbf = erp->adapter->dbf;
|
|
|
|
struct zfcp_dbf_rec *rec = &dbf->rec_buf;
|
2008-03-27 20:22:03 +07:00
|
|
|
unsigned long flags;
|
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->rec_lock, flags);
|
2010-12-02 21:16:12 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2008-03-27 20:22:03 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
rec->id = ZFCP_DBF_REC_RUN;
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
zfcp_dbf_set_common(rec, erp->adapter, erp->port, erp->sdev);
|
2008-03-27 20:22:04 +07:00
|
|
|
|
2010-12-02 21:16:12 +07:00
|
|
|
rec->u.run.fsf_req_id = erp->fsf_req_id;
|
|
|
|
rec->u.run.rec_status = erp->status;
|
|
|
|
rec->u.run.rec_step = erp->step;
|
|
|
|
rec->u.run.rec_action = erp->action;
|
|
|
|
|
|
|
|
if (erp->sdev)
|
|
|
|
rec->u.run.rec_count =
|
|
|
|
atomic_read(&sdev_to_zfcp(erp->sdev)->erp_counter);
|
|
|
|
else if (erp->port)
|
|
|
|
rec->u.run.rec_count = atomic_read(&erp->port->erp_counter);
|
|
|
|
else
|
|
|
|
rec->u.run.rec_count = atomic_read(&erp->adapter->erp_counter);
|
|
|
|
|
scsi: zfcp: fix rport unblock race with LUN recovery
It is unavoidable that zfcp_scsi_queuecommand() has to finish requests
with DID_IMM_RETRY (like fc_remote_port_chkready()) during the time
window when zfcp detected an unavailable rport but
fc_remote_port_delete(), which is asynchronous via
zfcp_scsi_schedule_rport_block(), has not yet blocked the rport.
However, for the case when the rport becomes available again, we should
prevent unblocking the rport too early. In contrast to other FCP LLDDs,
zfcp has to open each LUN with the FCP channel hardware before it can
send I/O to a LUN. So if a port already has LUNs attached and we
unblock the rport just after port recovery, recoveries of LUNs behind
this port can still be pending which in turn force
zfcp_scsi_queuecommand() to unnecessarily finish requests with
DID_IMM_RETRY.
This also opens a time window with unblocked rport (until the followup
LUN reopen recovery has finished). If a scsi_cmnd timeout occurs during
this time window fc_timed_out() cannot work as desired and such command
would indeed time out and trigger scsi_eh. This prevents a clean and
timely path failover. This should not happen if the path issue can be
recovered on FC transport layer such as path issues involving RSCNs.
Fix this by only calling zfcp_scsi_schedule_rport_register(), to
asynchronously trigger fc_remote_port_add(), after all LUN recoveries as
children of the rport have finished and no new recoveries of equal or
higher order were triggered meanwhile. Finished intentionally includes
any recovery result no matter if successful or failed (still unblock
rport so other successful LUNs work). For simplicity, we check after
each finished LUN recovery if there is another LUN recovery pending on
the same port and then do nothing. We handle the special case of a
successful recovery of a port without LUN children the same way without
changing this case's semantics.
For debugging we introduce 2 new trace records written if the rport
unblock attempt was aborted due to still unfinished or freshly triggered
recovery. The records are only written above the default trace level.
Benjamin noticed the important special case of new recovery that can be
triggered between having given up the erp_lock and before calling
zfcp_erp_action_cleanup() within zfcp_erp_strategy(). We must avoid the
following sequence:
ERP thread rport_work other context
------------------------- -------------- --------------------------------
port is unblocked, rport still blocked,
due to pending/running ERP action,
so ((port->status & ...UNBLOCK) != 0)
and (port->rport == NULL)
unlock ERP
zfcp_erp_action_cleanup()
case ZFCP_ERP_ACTION_REOPEN_LUN:
zfcp_erp_try_rport_unblock()
((status & ...UNBLOCK) != 0) [OLD!]
zfcp_erp_port_reopen()
lock ERP
zfcp_erp_port_block()
port->status clear ...UNBLOCK
unlock ERP
zfcp_scsi_schedule_rport_block()
port->rport_task = RPORT_DEL
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task != RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_block()
if (!port->rport) return
zfcp_scsi_schedule_rport_register()
port->rport_task = RPORT_ADD
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task == RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_register()
(port->rport == NULL)
rport = fc_remote_port_add()
port->rport = rport;
Now the rport was erroneously unblocked while the zfcp_port is blocked.
This is another situation we want to avoid due to scsi_eh
potential. This state would at least remain until the new recovery from
the other context finished successfully, or potentially forever if it
failed. In order to close this race, we take the erp_lock inside
zfcp_erp_try_rport_unblock() when checking the status of zfcp_port or
LUN. With that, the possible corresponding rport state sequences would
be: (unblock[ERP thread],block[other context]) if the ERP thread gets
erp_lock first and still sees ((port->status & ...UNBLOCK) != 0),
(block[other context],NOP[ERP thread]) if the ERP thread gets erp_lock
after the other context has already cleard ...UNBLOCK from port->status.
Since checking fields of struct erp_action is unsafe because they could
have been overwritten (re-used for new recovery) meanwhile, we only
check status of zfcp_port and LUN since these are only changed under
erp_lock elsewhere. Regarding the check of the proper status flags (port
or port_forced are similar to the shown adapter recovery):
[zfcp_erp_adapter_shutdown()]
zfcp_erp_adapter_reopen()
zfcp_erp_adapter_block()
* clear UNBLOCK ---------------------------------------+
zfcp_scsi_schedule_rports_block() |
write_lock_irqsave(&adapter->erp_lock, flags);-------+ |
zfcp_erp_action_enqueue() | |
zfcp_erp_setup_act() | |
* set ERP_INUSE -----------------------------------|--|--+
write_unlock_irqrestore(&adapter->erp_lock, flags);--+ | |
.context-switch. | |
zfcp_erp_thread() | |
zfcp_erp_strategy() | |
write_lock_irqsave(&adapter->erp_lock, flags);------+ | |
... | | |
zfcp_erp_strategy_check_target() | | |
zfcp_erp_strategy_check_adapter() | | |
zfcp_erp_adapter_unblock() | | |
* set UNBLOCK -----------------------------------|--+ |
zfcp_erp_action_dequeue() | |
* clear ERP_INUSE ---------------------------------|-----+
... |
write_unlock_irqrestore(&adapter->erp_lock, flags);-+
Hence, we should check for both UNBLOCK and ERP_INUSE because they are
interleaved. Also we need to explicitly check ERP_FAILED for the link
down case which currently does not clear the UNBLOCK flag in
zfcp_fsf_link_down_info_eval().
Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com>
Fixes: 8830271c4819 ("[SCSI] zfcp: Dont fail SCSI commands when transitioning to blocked fc_rport")
Fixes: a2fa0aede07c ("[SCSI] zfcp: Block FC transport rports early on errors")
Fixes: 5f852be9e11d ("[SCSI] zfcp: Fix deadlock between zfcp ERP and SCSI")
Fixes: 338151e06608 ("[SCSI] zfcp: make use of fc_remote_port_delete when target port is unavailable")
Fixes: 3859f6a248cb ("[PATCH] zfcp: add rports to enable scsi_add_device to work again")
Cc: <stable@vger.kernel.org> #2.6.32+
Reviewed-by: Benjamin Block <bblock@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-12-09 23:16:33 +07:00
|
|
|
debug_event(dbf->rec, level, rec, sizeof(*rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_unlock_irqrestore(&dbf->rec_lock, flags);
|
2008-03-27 20:22:04 +07:00
|
|
|
}
|
|
|
|
|
scsi: zfcp: fix rport unblock race with LUN recovery
It is unavoidable that zfcp_scsi_queuecommand() has to finish requests
with DID_IMM_RETRY (like fc_remote_port_chkready()) during the time
window when zfcp detected an unavailable rport but
fc_remote_port_delete(), which is asynchronous via
zfcp_scsi_schedule_rport_block(), has not yet blocked the rport.
However, for the case when the rport becomes available again, we should
prevent unblocking the rport too early. In contrast to other FCP LLDDs,
zfcp has to open each LUN with the FCP channel hardware before it can
send I/O to a LUN. So if a port already has LUNs attached and we
unblock the rport just after port recovery, recoveries of LUNs behind
this port can still be pending which in turn force
zfcp_scsi_queuecommand() to unnecessarily finish requests with
DID_IMM_RETRY.
This also opens a time window with unblocked rport (until the followup
LUN reopen recovery has finished). If a scsi_cmnd timeout occurs during
this time window fc_timed_out() cannot work as desired and such command
would indeed time out and trigger scsi_eh. This prevents a clean and
timely path failover. This should not happen if the path issue can be
recovered on FC transport layer such as path issues involving RSCNs.
Fix this by only calling zfcp_scsi_schedule_rport_register(), to
asynchronously trigger fc_remote_port_add(), after all LUN recoveries as
children of the rport have finished and no new recoveries of equal or
higher order were triggered meanwhile. Finished intentionally includes
any recovery result no matter if successful or failed (still unblock
rport so other successful LUNs work). For simplicity, we check after
each finished LUN recovery if there is another LUN recovery pending on
the same port and then do nothing. We handle the special case of a
successful recovery of a port without LUN children the same way without
changing this case's semantics.
For debugging we introduce 2 new trace records written if the rport
unblock attempt was aborted due to still unfinished or freshly triggered
recovery. The records are only written above the default trace level.
Benjamin noticed the important special case of new recovery that can be
triggered between having given up the erp_lock and before calling
zfcp_erp_action_cleanup() within zfcp_erp_strategy(). We must avoid the
following sequence:
ERP thread rport_work other context
------------------------- -------------- --------------------------------
port is unblocked, rport still blocked,
due to pending/running ERP action,
so ((port->status & ...UNBLOCK) != 0)
and (port->rport == NULL)
unlock ERP
zfcp_erp_action_cleanup()
case ZFCP_ERP_ACTION_REOPEN_LUN:
zfcp_erp_try_rport_unblock()
((status & ...UNBLOCK) != 0) [OLD!]
zfcp_erp_port_reopen()
lock ERP
zfcp_erp_port_block()
port->status clear ...UNBLOCK
unlock ERP
zfcp_scsi_schedule_rport_block()
port->rport_task = RPORT_DEL
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task != RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_block()
if (!port->rport) return
zfcp_scsi_schedule_rport_register()
port->rport_task = RPORT_ADD
queue_work(rport_work)
zfcp_scsi_rport_work()
(port->rport_task == RPORT_ADD)
port->rport_task = RPORT_NONE
zfcp_scsi_rport_register()
(port->rport == NULL)
rport = fc_remote_port_add()
port->rport = rport;
Now the rport was erroneously unblocked while the zfcp_port is blocked.
This is another situation we want to avoid due to scsi_eh
potential. This state would at least remain until the new recovery from
the other context finished successfully, or potentially forever if it
failed. In order to close this race, we take the erp_lock inside
zfcp_erp_try_rport_unblock() when checking the status of zfcp_port or
LUN. With that, the possible corresponding rport state sequences would
be: (unblock[ERP thread],block[other context]) if the ERP thread gets
erp_lock first and still sees ((port->status & ...UNBLOCK) != 0),
(block[other context],NOP[ERP thread]) if the ERP thread gets erp_lock
after the other context has already cleard ...UNBLOCK from port->status.
Since checking fields of struct erp_action is unsafe because they could
have been overwritten (re-used for new recovery) meanwhile, we only
check status of zfcp_port and LUN since these are only changed under
erp_lock elsewhere. Regarding the check of the proper status flags (port
or port_forced are similar to the shown adapter recovery):
[zfcp_erp_adapter_shutdown()]
zfcp_erp_adapter_reopen()
zfcp_erp_adapter_block()
* clear UNBLOCK ---------------------------------------+
zfcp_scsi_schedule_rports_block() |
write_lock_irqsave(&adapter->erp_lock, flags);-------+ |
zfcp_erp_action_enqueue() | |
zfcp_erp_setup_act() | |
* set ERP_INUSE -----------------------------------|--|--+
write_unlock_irqrestore(&adapter->erp_lock, flags);--+ | |
.context-switch. | |
zfcp_erp_thread() | |
zfcp_erp_strategy() | |
write_lock_irqsave(&adapter->erp_lock, flags);------+ | |
... | | |
zfcp_erp_strategy_check_target() | | |
zfcp_erp_strategy_check_adapter() | | |
zfcp_erp_adapter_unblock() | | |
* set UNBLOCK -----------------------------------|--+ |
zfcp_erp_action_dequeue() | |
* clear ERP_INUSE ---------------------------------|-----+
... |
write_unlock_irqrestore(&adapter->erp_lock, flags);-+
Hence, we should check for both UNBLOCK and ERP_INUSE because they are
interleaved. Also we need to explicitly check ERP_FAILED for the link
down case which currently does not clear the UNBLOCK flag in
zfcp_fsf_link_down_info_eval().
Signed-off-by: Steffen Maier <maier@linux.vnet.ibm.com>
Fixes: 8830271c4819 ("[SCSI] zfcp: Dont fail SCSI commands when transitioning to blocked fc_rport")
Fixes: a2fa0aede07c ("[SCSI] zfcp: Block FC transport rports early on errors")
Fixes: 5f852be9e11d ("[SCSI] zfcp: Fix deadlock between zfcp ERP and SCSI")
Fixes: 338151e06608 ("[SCSI] zfcp: make use of fc_remote_port_delete when target port is unavailable")
Fixes: 3859f6a248cb ("[PATCH] zfcp: add rports to enable scsi_add_device to work again")
Cc: <stable@vger.kernel.org> #2.6.32+
Reviewed-by: Benjamin Block <bblock@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-12-09 23:16:33 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_rec_run - trace event related to running recovery
|
|
|
|
* @tag: identifier for event
|
|
|
|
* @erp: erp_action running
|
|
|
|
*/
|
|
|
|
void zfcp_dbf_rec_run(char *tag, struct zfcp_erp_action *erp)
|
|
|
|
{
|
|
|
|
zfcp_dbf_rec_run_lvl(1, tag, erp);
|
|
|
|
}
|
|
|
|
|
2016-08-10 23:30:49 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_rec_run_wka - trace wka port event with info like running recovery
|
|
|
|
* @tag: identifier for event
|
|
|
|
* @wka_port: well known address port
|
|
|
|
* @req_id: request ID to correlate with potential HBA trace record
|
|
|
|
*/
|
|
|
|
void zfcp_dbf_rec_run_wka(char *tag, struct zfcp_fc_wka_port *wka_port,
|
|
|
|
u64 req_id)
|
|
|
|
{
|
|
|
|
struct zfcp_dbf *dbf = wka_port->adapter->dbf;
|
|
|
|
struct zfcp_dbf_rec *rec = &dbf->rec_buf;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&dbf->rec_lock, flags);
|
|
|
|
memset(rec, 0, sizeof(*rec));
|
|
|
|
|
|
|
|
rec->id = ZFCP_DBF_REC_RUN;
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->port_status = wka_port->status;
|
|
|
|
rec->d_id = wka_port->d_id;
|
|
|
|
rec->lun = ZFCP_DBF_INVALID_LUN;
|
|
|
|
|
|
|
|
rec->u.run.fsf_req_id = req_id;
|
|
|
|
rec->u.run.rec_status = ~0;
|
|
|
|
rec->u.run.rec_step = ~0;
|
|
|
|
rec->u.run.rec_action = ~0;
|
|
|
|
rec->u.run.rec_count = ~0;
|
|
|
|
|
|
|
|
debug_event(dbf->rec, 1, rec, sizeof(*rec));
|
|
|
|
spin_unlock_irqrestore(&dbf->rec_lock, flags);
|
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:13 +07:00
|
|
|
static inline
|
2016-08-10 23:30:53 +07:00
|
|
|
void zfcp_dbf_san(char *tag, struct zfcp_dbf *dbf,
|
|
|
|
char *paytag, struct scatterlist *sg, u8 id, u16 len,
|
|
|
|
u64 req_id, u32 d_id, u16 cap_len)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:13 +07:00
|
|
|
struct zfcp_dbf_san *rec = &dbf->san_buf;
|
|
|
|
u16 rec_len;
|
2008-03-31 16:15:29 +07:00
|
|
|
unsigned long flags;
|
2016-08-10 23:30:53 +07:00
|
|
|
struct zfcp_dbf_pay *payload = &dbf->pay_buf;
|
|
|
|
u16 pay_sum = 0;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->san_lock, flags);
|
2010-12-02 21:16:13 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:13 +07:00
|
|
|
rec->id = id;
|
|
|
|
rec->fsf_req_id = req_id;
|
|
|
|
rec->d_id = d_id;
|
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
2016-08-10 23:30:53 +07:00
|
|
|
rec->pl_len = len; /* full length even if we cap pay below */
|
|
|
|
if (!sg)
|
|
|
|
goto out;
|
|
|
|
rec_len = min_t(unsigned int, sg->length, ZFCP_DBF_SAN_MAX_PAYLOAD);
|
|
|
|
memcpy(rec->payload, sg_virt(sg), rec_len); /* part of 1st sg entry */
|
|
|
|
if (len <= rec_len)
|
|
|
|
goto out; /* skip pay record if full content in rec->payload */
|
|
|
|
|
|
|
|
/* if (len > rec_len):
|
|
|
|
* dump data up to cap_len ignoring small duplicate in rec->payload
|
|
|
|
*/
|
2016-10-15 03:18:39 +07:00
|
|
|
spin_lock(&dbf->pay_lock);
|
2016-08-10 23:30:53 +07:00
|
|
|
memset(payload, 0, sizeof(*payload));
|
|
|
|
memcpy(payload->area, paytag, ZFCP_DBF_TAG_LEN);
|
|
|
|
payload->fsf_req_id = req_id;
|
|
|
|
payload->counter = 0;
|
|
|
|
for (; sg && pay_sum < cap_len; sg = sg_next(sg)) {
|
|
|
|
u16 pay_len, offset = 0;
|
|
|
|
|
|
|
|
while (offset < sg->length && pay_sum < cap_len) {
|
|
|
|
pay_len = min((u16)ZFCP_DBF_PAY_MAX_REC,
|
|
|
|
(u16)(sg->length - offset));
|
|
|
|
/* cap_len <= pay_sum < cap_len+ZFCP_DBF_PAY_MAX_REC */
|
|
|
|
memcpy(payload->data, sg_virt(sg) + offset, pay_len);
|
|
|
|
debug_event(dbf->pay, 1, payload,
|
|
|
|
zfcp_dbf_plen(pay_len));
|
|
|
|
payload->counter++;
|
|
|
|
offset += pay_len;
|
|
|
|
pay_sum += pay_len;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
spin_unlock(&dbf->pay_lock);
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2016-08-10 23:30:53 +07:00
|
|
|
out:
|
2010-12-02 21:16:13 +07:00
|
|
|
debug_event(dbf->san, 1, rec, sizeof(*rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_unlock_irqrestore(&dbf->san_lock, flags);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2008-03-31 16:15:31 +07:00
|
|
|
/**
|
2010-12-02 21:16:13 +07:00
|
|
|
* zfcp_dbf_san_req - trace event for issued SAN request
|
2013-10-12 06:29:46 +07:00
|
|
|
* @tag: identifier for event
|
2010-12-02 21:16:13 +07:00
|
|
|
* @fsf_req: request containing issued CT data
|
|
|
|
* d_id: destination ID
|
2008-03-31 16:15:31 +07:00
|
|
|
*/
|
2010-12-02 21:16:13 +07:00
|
|
|
void zfcp_dbf_san_req(char *tag, struct zfcp_fsf_req *fsf, u32 d_id)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:13 +07:00
|
|
|
struct zfcp_dbf *dbf = fsf->adapter->dbf;
|
|
|
|
struct zfcp_fsf_ct_els *ct_els = fsf->data;
|
|
|
|
u16 length;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2016-08-10 23:30:53 +07:00
|
|
|
length = (u16)zfcp_qdio_real_bytes(ct_els->req);
|
|
|
|
zfcp_dbf_san(tag, dbf, "san_req", ct_els->req, ZFCP_DBF_SAN_REQ,
|
|
|
|
length, fsf->req_id, d_id, length);
|
|
|
|
}
|
|
|
|
|
|
|
|
static u16 zfcp_dbf_san_res_cap_len_if_gpn_ft(char *tag,
|
|
|
|
struct zfcp_fsf_req *fsf,
|
|
|
|
u16 len)
|
|
|
|
{
|
|
|
|
struct zfcp_fsf_ct_els *ct_els = fsf->data;
|
|
|
|
struct fc_ct_hdr *reqh = sg_virt(ct_els->req);
|
|
|
|
struct fc_ns_gid_ft *reqn = (struct fc_ns_gid_ft *)(reqh + 1);
|
|
|
|
struct scatterlist *resp_entry = ct_els->resp;
|
|
|
|
struct fc_gpn_ft_resp *acc;
|
|
|
|
int max_entries, x, last = 0;
|
|
|
|
|
|
|
|
if (!(memcmp(tag, "fsscth2", 7) == 0
|
|
|
|
&& ct_els->d_id == FC_FID_DIR_SERV
|
|
|
|
&& reqh->ct_rev == FC_CT_REV
|
|
|
|
&& reqh->ct_in_id[0] == 0
|
|
|
|
&& reqh->ct_in_id[1] == 0
|
|
|
|
&& reqh->ct_in_id[2] == 0
|
|
|
|
&& reqh->ct_fs_type == FC_FST_DIR
|
|
|
|
&& reqh->ct_fs_subtype == FC_NS_SUBTYPE
|
|
|
|
&& reqh->ct_options == 0
|
|
|
|
&& reqh->_ct_resvd1 == 0
|
|
|
|
&& reqh->ct_cmd == FC_NS_GPN_FT
|
|
|
|
/* reqh->ct_mr_size can vary so do not match but read below */
|
|
|
|
&& reqh->_ct_resvd2 == 0
|
|
|
|
&& reqh->ct_reason == 0
|
|
|
|
&& reqh->ct_explan == 0
|
|
|
|
&& reqh->ct_vendor == 0
|
|
|
|
&& reqn->fn_resvd == 0
|
|
|
|
&& reqn->fn_domain_id_scope == 0
|
|
|
|
&& reqn->fn_area_id_scope == 0
|
|
|
|
&& reqn->fn_fc4_type == FC_TYPE_FCP))
|
|
|
|
return len; /* not GPN_FT response so do not cap */
|
|
|
|
|
|
|
|
acc = sg_virt(resp_entry);
|
|
|
|
max_entries = (reqh->ct_mr_size * 4 / sizeof(struct fc_gpn_ft_resp))
|
|
|
|
+ 1 /* zfcp_fc_scan_ports: bytes correct, entries off-by-one
|
|
|
|
* to account for header as 1st pseudo "entry" */;
|
|
|
|
|
|
|
|
/* the basic CT_IU preamble is the same size as one entry in the GPN_FT
|
|
|
|
* response, allowing us to skip special handling for it - just skip it
|
|
|
|
*/
|
|
|
|
for (x = 1; x < max_entries && !last; x++) {
|
|
|
|
if (x % (ZFCP_FC_GPN_FT_ENT_PAGE + 1))
|
|
|
|
acc++;
|
|
|
|
else
|
|
|
|
acc = sg_virt(++resp_entry);
|
|
|
|
|
|
|
|
last = acc->fp_flags & FC_NS_FID_LAST;
|
|
|
|
}
|
|
|
|
len = min(len, (u16)(x * sizeof(struct fc_gpn_ft_resp)));
|
|
|
|
return len; /* cap after last entry */
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2008-03-31 16:15:31 +07:00
|
|
|
/**
|
2010-12-02 21:16:13 +07:00
|
|
|
* zfcp_dbf_san_res - trace event for received SAN request
|
2013-10-12 06:29:46 +07:00
|
|
|
* @tag: identifier for event
|
2010-12-02 21:16:13 +07:00
|
|
|
* @fsf_req: request containing issued CT data
|
2008-03-31 16:15:31 +07:00
|
|
|
*/
|
2010-12-02 21:16:13 +07:00
|
|
|
void zfcp_dbf_san_res(char *tag, struct zfcp_fsf_req *fsf)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:13 +07:00
|
|
|
struct zfcp_dbf *dbf = fsf->adapter->dbf;
|
|
|
|
struct zfcp_fsf_ct_els *ct_els = fsf->data;
|
|
|
|
u16 length;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2016-08-10 23:30:53 +07:00
|
|
|
length = (u16)zfcp_qdio_real_bytes(ct_els->resp);
|
|
|
|
zfcp_dbf_san(tag, dbf, "san_res", ct_els->resp, ZFCP_DBF_SAN_RES,
|
|
|
|
length, fsf->req_id, ct_els->d_id,
|
|
|
|
zfcp_dbf_san_res_cap_len_if_gpn_ft(tag, fsf, length));
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2008-03-31 16:15:31 +07:00
|
|
|
/**
|
2010-12-02 21:16:13 +07:00
|
|
|
* zfcp_dbf_san_in_els - trace event for incoming ELS
|
2013-10-12 06:29:46 +07:00
|
|
|
* @tag: identifier for event
|
2010-12-02 21:16:13 +07:00
|
|
|
* @fsf_req: request containing issued CT data
|
2008-03-31 16:15:31 +07:00
|
|
|
*/
|
2010-12-02 21:16:13 +07:00
|
|
|
void zfcp_dbf_san_in_els(char *tag, struct zfcp_fsf_req *fsf)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:13 +07:00
|
|
|
struct zfcp_dbf *dbf = fsf->adapter->dbf;
|
|
|
|
struct fsf_status_read_buffer *srb =
|
|
|
|
(struct fsf_status_read_buffer *) fsf->data;
|
|
|
|
u16 length;
|
2016-08-10 23:30:53 +07:00
|
|
|
struct scatterlist sg;
|
2010-12-02 21:16:13 +07:00
|
|
|
|
|
|
|
length = (u16)(srb->length -
|
|
|
|
offsetof(struct fsf_status_read_buffer, payload));
|
2016-08-10 23:30:53 +07:00
|
|
|
sg_init_one(&sg, srb->payload.data, length);
|
|
|
|
zfcp_dbf_san(tag, dbf, "san_els", &sg, ZFCP_DBF_SAN_ELS, length,
|
|
|
|
fsf->req_id, ntoh24(srb->d_id), length);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:15 +07:00
|
|
|
/**
|
|
|
|
* zfcp_dbf_scsi - trace event for scsi commands
|
|
|
|
* @tag: identifier for event
|
|
|
|
* @sc: pointer to struct scsi_cmnd
|
|
|
|
* @fsf: pointer to struct zfcp_fsf_req
|
|
|
|
*/
|
2016-08-10 23:30:47 +07:00
|
|
|
void zfcp_dbf_scsi(char *tag, int level, struct scsi_cmnd *sc,
|
|
|
|
struct zfcp_fsf_req *fsf)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:15 +07:00
|
|
|
struct zfcp_adapter *adapter =
|
|
|
|
(struct zfcp_adapter *) sc->device->host->hostdata[0];
|
|
|
|
struct zfcp_dbf *dbf = adapter->dbf;
|
|
|
|
struct zfcp_dbf_scsi *rec = &dbf->scsi_buf;
|
2009-11-24 22:54:08 +07:00
|
|
|
struct fcp_resp_with_ext *fcp_rsp;
|
2010-12-02 21:16:15 +07:00
|
|
|
struct fcp_resp_rsp_info *fcp_rsp_info;
|
|
|
|
unsigned long flags;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_irqsave(&dbf->scsi_lock, flags);
|
2010-12-02 21:16:15 +07:00
|
|
|
memset(rec, 0, sizeof(*rec));
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:15 +07:00
|
|
|
memcpy(rec->tag, tag, ZFCP_DBF_TAG_LEN);
|
|
|
|
rec->id = ZFCP_DBF_SCSI_CMND;
|
|
|
|
rec->scsi_result = sc->result;
|
|
|
|
rec->scsi_retries = sc->retries;
|
|
|
|
rec->scsi_allowed = sc->allowed;
|
|
|
|
rec->scsi_id = sc->device->id;
|
2014-06-25 20:27:36 +07:00
|
|
|
/* struct zfcp_dbf_scsi needs to be updated to handle 64bit LUNs */
|
|
|
|
rec->scsi_lun = (u32)sc->device->lun;
|
2010-12-02 21:16:15 +07:00
|
|
|
rec->host_scribble = (unsigned long)sc->host_scribble;
|
|
|
|
|
|
|
|
memcpy(rec->scsi_opcode, sc->cmnd,
|
|
|
|
min((int)sc->cmd_len, ZFCP_DBF_SCSI_OPCODE));
|
|
|
|
|
|
|
|
if (fsf) {
|
|
|
|
rec->fsf_req_id = fsf->req_id;
|
|
|
|
fcp_rsp = (struct fcp_resp_with_ext *)
|
|
|
|
&(fsf->qtcb->bottom.io.fcp_rsp);
|
|
|
|
memcpy(&rec->fcp_rsp, fcp_rsp, FCP_RESP_WITH_EXT);
|
|
|
|
if (fcp_rsp->resp.fr_flags & FCP_RSP_LEN_VAL) {
|
|
|
|
fcp_rsp_info = (struct fcp_resp_rsp_info *) &fcp_rsp[1];
|
|
|
|
rec->fcp_rsp_info = fcp_rsp_info->rsp_code;
|
|
|
|
}
|
|
|
|
if (fcp_rsp->resp.fr_flags & FCP_SNS_LEN_VAL) {
|
|
|
|
rec->pl_len = min((u16)SCSI_SENSE_BUFFERSIZE,
|
|
|
|
(u16)ZFCP_DBF_PAY_MAX_REC);
|
|
|
|
zfcp_dbf_pl_write(dbf, sc->sense_buffer, rec->pl_len,
|
|
|
|
"fcp_sns", fsf->req_id);
|
|
|
|
}
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
|
|
|
|
2016-08-10 23:30:47 +07:00
|
|
|
debug_event(dbf->scsi, level, rec, sizeof(*rec));
|
2010-12-02 21:16:15 +07:00
|
|
|
spin_unlock_irqrestore(&dbf->scsi_lock, flags);
|
|
|
|
}
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:16 +07:00
|
|
|
static debug_info_t *zfcp_dbf_reg(const char *name, int size, int rec_size)
|
2009-08-18 20:43:07 +07:00
|
|
|
{
|
|
|
|
struct debug_info *d;
|
|
|
|
|
2010-12-02 21:16:16 +07:00
|
|
|
d = debug_register(name, size, 1, rec_size);
|
2009-08-18 20:43:07 +07:00
|
|
|
if (!d)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
debug_register_view(d, &debug_hex_ascii_view);
|
2013-04-26 21:13:53 +07:00
|
|
|
debug_set_level(d, dbflevel);
|
2009-08-18 20:43:07 +07:00
|
|
|
|
|
|
|
return d;
|
|
|
|
}
|
|
|
|
|
2010-12-02 21:16:16 +07:00
|
|
|
static void zfcp_dbf_unregister(struct zfcp_dbf *dbf)
|
|
|
|
{
|
|
|
|
if (!dbf)
|
|
|
|
return;
|
|
|
|
|
|
|
|
debug_unregister(dbf->scsi);
|
|
|
|
debug_unregister(dbf->san);
|
|
|
|
debug_unregister(dbf->hba);
|
|
|
|
debug_unregister(dbf->pay);
|
|
|
|
debug_unregister(dbf->rec);
|
|
|
|
kfree(dbf);
|
|
|
|
}
|
|
|
|
|
2005-09-14 02:50:38 +07:00
|
|
|
/**
|
|
|
|
* zfcp_adapter_debug_register - registers debug feature for an adapter
|
|
|
|
* @adapter: pointer to adapter for which debug features should be registered
|
|
|
|
* return: -ENOMEM on error, 0 otherwise
|
|
|
|
*/
|
2009-08-18 20:43:21 +07:00
|
|
|
int zfcp_dbf_adapter_register(struct zfcp_adapter *adapter)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:16 +07:00
|
|
|
char name[DEBUG_MAX_NAME_LEN];
|
2009-08-18 20:43:07 +07:00
|
|
|
struct zfcp_dbf *dbf;
|
|
|
|
|
2010-07-16 20:37:40 +07:00
|
|
|
dbf = kzalloc(sizeof(struct zfcp_dbf), GFP_KERNEL);
|
2009-08-18 20:43:07 +07:00
|
|
|
if (!dbf)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
spin_lock_init(&dbf->pay_lock);
|
2009-08-18 20:43:21 +07:00
|
|
|
spin_lock_init(&dbf->hba_lock);
|
|
|
|
spin_lock_init(&dbf->san_lock);
|
|
|
|
spin_lock_init(&dbf->scsi_lock);
|
|
|
|
spin_lock_init(&dbf->rec_lock);
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2008-03-27 20:22:00 +07:00
|
|
|
/* debug feature area which records recovery activity */
|
2010-12-02 21:16:16 +07:00
|
|
|
sprintf(name, "zfcp_%s_rec", dev_name(&adapter->ccw_device->dev));
|
|
|
|
dbf->rec = zfcp_dbf_reg(name, dbfsize, sizeof(struct zfcp_dbf_rec));
|
2009-08-18 20:43:21 +07:00
|
|
|
if (!dbf->rec)
|
|
|
|
goto err_out;
|
2008-03-27 20:22:00 +07:00
|
|
|
|
2005-09-14 02:50:38 +07:00
|
|
|
/* debug feature area which records HBA (FSF and QDIO) conditions */
|
2010-12-02 21:16:16 +07:00
|
|
|
sprintf(name, "zfcp_%s_hba", dev_name(&adapter->ccw_device->dev));
|
|
|
|
dbf->hba = zfcp_dbf_reg(name, dbfsize, sizeof(struct zfcp_dbf_hba));
|
2009-08-18 20:43:21 +07:00
|
|
|
if (!dbf->hba)
|
|
|
|
goto err_out;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:14 +07:00
|
|
|
/* debug feature area which records payload info */
|
2010-12-02 21:16:16 +07:00
|
|
|
sprintf(name, "zfcp_%s_pay", dev_name(&adapter->ccw_device->dev));
|
|
|
|
dbf->pay = zfcp_dbf_reg(name, dbfsize * 2, sizeof(struct zfcp_dbf_pay));
|
2010-12-02 21:16:14 +07:00
|
|
|
if (!dbf->pay)
|
|
|
|
goto err_out;
|
|
|
|
|
2005-09-14 02:50:38 +07:00
|
|
|
/* debug feature area which records SAN command failures and recovery */
|
2010-12-02 21:16:16 +07:00
|
|
|
sprintf(name, "zfcp_%s_san", dev_name(&adapter->ccw_device->dev));
|
|
|
|
dbf->san = zfcp_dbf_reg(name, dbfsize, sizeof(struct zfcp_dbf_san));
|
2009-08-18 20:43:21 +07:00
|
|
|
if (!dbf->san)
|
|
|
|
goto err_out;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
|
|
|
/* debug feature area which records SCSI command failures and recovery */
|
2010-12-02 21:16:16 +07:00
|
|
|
sprintf(name, "zfcp_%s_scsi", dev_name(&adapter->ccw_device->dev));
|
|
|
|
dbf->scsi = zfcp_dbf_reg(name, dbfsize, sizeof(struct zfcp_dbf_scsi));
|
2009-08-18 20:43:21 +07:00
|
|
|
if (!dbf->scsi)
|
|
|
|
goto err_out;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2009-08-18 20:43:07 +07:00
|
|
|
adapter->dbf = dbf;
|
2005-09-14 02:50:38 +07:00
|
|
|
|
2010-12-02 21:16:16 +07:00
|
|
|
return 0;
|
2009-08-18 20:43:21 +07:00
|
|
|
err_out:
|
2010-12-02 21:16:16 +07:00
|
|
|
zfcp_dbf_unregister(dbf);
|
2005-09-14 02:50:38 +07:00
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* zfcp_adapter_debug_unregister - unregisters debug feature for an adapter
|
2010-12-02 21:16:16 +07:00
|
|
|
* @adapter: pointer to adapter for which debug features should be unregistered
|
2005-09-14 02:50:38 +07:00
|
|
|
*/
|
2010-12-02 21:16:16 +07:00
|
|
|
void zfcp_dbf_adapter_unregister(struct zfcp_adapter *adapter)
|
2005-09-14 02:50:38 +07:00
|
|
|
{
|
2010-12-02 21:16:16 +07:00
|
|
|
struct zfcp_dbf *dbf = adapter->dbf;
|
|
|
|
|
|
|
|
adapter->dbf = NULL;
|
|
|
|
zfcp_dbf_unregister(dbf);
|
2005-09-14 02:50:38 +07:00
|
|
|
}
|
2009-08-18 20:43:21 +07:00
|
|
|
|