linux/drivers/scsi/fnic/fnic_scsi.c

2769 lines
78 KiB
C
Raw Normal View History

/*
* Copyright 2008 Cisco Systems, Inc. All rights reserved.
* Copyright 2007 Nuova Systems, Inc. All rights reserved.
*
* This program is free software; you may redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; version 2 of the License.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
* NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
* BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
* ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
* CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
* SOFTWARE.
*/
#include <linux/mempool.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/workqueue.h>
#include <linux/pci.h>
#include <linux/scatterlist.h>
#include <linux/skbuff.h>
#include <linux/spinlock.h>
#include <linux/if_ether.h>
#include <linux/if_vlan.h>
#include <linux/delay.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 16:04:11 +08:00
#include <linux/gfp.h>
#include <scsi/scsi.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_device.h>
#include <scsi/scsi_cmnd.h>
#include <scsi/scsi_tcq.h>
#include <scsi/fc/fc_els.h>
#include <scsi/fc/fc_fcoe.h>
#include <scsi/libfc.h>
#include <scsi/fc_frame.h>
#include "fnic_io.h"
#include "fnic.h"
const char *fnic_state_str[] = {
[FNIC_IN_FC_MODE] = "FNIC_IN_FC_MODE",
[FNIC_IN_FC_TRANS_ETH_MODE] = "FNIC_IN_FC_TRANS_ETH_MODE",
[FNIC_IN_ETH_MODE] = "FNIC_IN_ETH_MODE",
[FNIC_IN_ETH_TRANS_FC_MODE] = "FNIC_IN_ETH_TRANS_FC_MODE",
};
static const char *fnic_ioreq_state_str[] = {
[FNIC_IOREQ_NOT_INITED] = "FNIC_IOREQ_NOT_INITED",
[FNIC_IOREQ_CMD_PENDING] = "FNIC_IOREQ_CMD_PENDING",
[FNIC_IOREQ_ABTS_PENDING] = "FNIC_IOREQ_ABTS_PENDING",
[FNIC_IOREQ_ABTS_COMPLETE] = "FNIC_IOREQ_ABTS_COMPLETE",
[FNIC_IOREQ_CMD_COMPLETE] = "FNIC_IOREQ_CMD_COMPLETE",
};
static const char *fcpio_status_str[] = {
[FCPIO_SUCCESS] = "FCPIO_SUCCESS", /*0x0*/
[FCPIO_INVALID_HEADER] = "FCPIO_INVALID_HEADER",
[FCPIO_OUT_OF_RESOURCE] = "FCPIO_OUT_OF_RESOURCE",
[FCPIO_INVALID_PARAM] = "FCPIO_INVALID_PARAM]",
[FCPIO_REQ_NOT_SUPPORTED] = "FCPIO_REQ_NOT_SUPPORTED",
[FCPIO_IO_NOT_FOUND] = "FCPIO_IO_NOT_FOUND",
[FCPIO_ABORTED] = "FCPIO_ABORTED", /*0x41*/
[FCPIO_TIMEOUT] = "FCPIO_TIMEOUT",
[FCPIO_SGL_INVALID] = "FCPIO_SGL_INVALID",
[FCPIO_MSS_INVALID] = "FCPIO_MSS_INVALID",
[FCPIO_DATA_CNT_MISMATCH] = "FCPIO_DATA_CNT_MISMATCH",
[FCPIO_FW_ERR] = "FCPIO_FW_ERR",
[FCPIO_ITMF_REJECTED] = "FCPIO_ITMF_REJECTED",
[FCPIO_ITMF_FAILED] = "FCPIO_ITMF_FAILED",
[FCPIO_ITMF_INCORRECT_LUN] = "FCPIO_ITMF_INCORRECT_LUN",
[FCPIO_CMND_REJECTED] = "FCPIO_CMND_REJECTED",
[FCPIO_NO_PATH_AVAIL] = "FCPIO_NO_PATH_AVAIL",
[FCPIO_PATH_FAILED] = "FCPIO_PATH_FAILED",
[FCPIO_LUNMAP_CHNG_PEND] = "FCPIO_LUNHMAP_CHNG_PEND",
};
const char *fnic_state_to_str(unsigned int state)
{
if (state >= ARRAY_SIZE(fnic_state_str) || !fnic_state_str[state])
return "unknown";
return fnic_state_str[state];
}
static const char *fnic_ioreq_state_to_str(unsigned int state)
{
if (state >= ARRAY_SIZE(fnic_ioreq_state_str) ||
!fnic_ioreq_state_str[state])
return "unknown";
return fnic_ioreq_state_str[state];
}
static const char *fnic_fcpio_status_to_str(unsigned int status)
{
if (status >= ARRAY_SIZE(fcpio_status_str) || !fcpio_status_str[status])
return "unknown";
return fcpio_status_str[status];
}
static void fnic_cleanup_io(struct fnic *fnic, int exclude_id);
static inline spinlock_t *fnic_io_lock_hash(struct fnic *fnic,
struct scsi_cmnd *sc)
{
u32 hash = sc->request->tag & (FNIC_IO_LOCKS - 1);
return &fnic->io_req_lock[hash];
}
static inline spinlock_t *fnic_io_lock_tag(struct fnic *fnic,
int tag)
{
return &fnic->io_req_lock[tag & (FNIC_IO_LOCKS - 1)];
}
/*
* Unmap the data buffer and sense buffer for an io_req,
* also unmap and free the device-private scatter/gather list.
*/
static void fnic_release_ioreq_buf(struct fnic *fnic,
struct fnic_io_req *io_req,
struct scsi_cmnd *sc)
{
if (io_req->sgl_list_pa)
pci_unmap_single(fnic->pdev, io_req->sgl_list_pa,
sizeof(io_req->sgl_list[0]) * io_req->sgl_cnt,
PCI_DMA_TODEVICE);
scsi_dma_unmap(sc);
if (io_req->sgl_cnt)
mempool_free(io_req->sgl_list_alloc,
fnic->io_sgl_pool[io_req->sgl_type]);
if (io_req->sense_buf_pa)
pci_unmap_single(fnic->pdev, io_req->sense_buf_pa,
SCSI_SENSE_BUFFERSIZE, PCI_DMA_FROMDEVICE);
}
/* Free up Copy Wq descriptors. Called with copy_wq lock held */
static int free_wq_copy_descs(struct fnic *fnic, struct vnic_wq_copy *wq)
{
/* if no Ack received from firmware, then nothing to clean */
if (!fnic->fw_ack_recd[0])
return 1;
/*
* Update desc_available count based on number of freed descriptors
* Account for wraparound
*/
if (wq->to_clean_index <= fnic->fw_ack_index[0])
wq->ring.desc_avail += (fnic->fw_ack_index[0]
- wq->to_clean_index + 1);
else
wq->ring.desc_avail += (wq->ring.desc_count
- wq->to_clean_index
+ fnic->fw_ack_index[0] + 1);
/*
* just bump clean index to ack_index+1 accounting for wraparound
* this will essentially free up all descriptors between
* to_clean_index and fw_ack_index, both inclusive
*/
wq->to_clean_index =
(fnic->fw_ack_index[0] + 1) % wq->ring.desc_count;
/* we have processed the acks received so far */
fnic->fw_ack_recd[0] = 0;
return 0;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/**
* __fnic_set_state_flags
* Sets/Clears bits in fnic's state_flags
**/
void
__fnic_set_state_flags(struct fnic *fnic, unsigned long st_flags,
unsigned long clearbits)
{
struct Scsi_Host *host = fnic->lport->host;
int sh_locked = spin_is_locked(host->host_lock);
unsigned long flags = 0;
if (!sh_locked)
spin_lock_irqsave(host->host_lock, flags);
if (clearbits)
fnic->state_flags &= ~st_flags;
else
fnic->state_flags |= st_flags;
if (!sh_locked)
spin_unlock_irqrestore(host->host_lock, flags);
return;
}
/*
* fnic_fw_reset_handler
* Routine to send reset msg to fw
*/
int fnic_fw_reset_handler(struct fnic *fnic)
{
struct vnic_wq_copy *wq = &fnic->wq_copy[0];
int ret = 0;
unsigned long flags;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/* indicate fwreset to io path */
fnic_set_state_flags(fnic, FNIC_FLAGS_FWRESET);
skb_queue_purge(&fnic->frame_queue);
skb_queue_purge(&fnic->tx_queue);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/* wait for io cmpl */
while (atomic_read(&fnic->in_flight))
schedule_timeout(msecs_to_jiffies(1));
spin_lock_irqsave(&fnic->wq_copy_lock[0], flags);
if (vnic_wq_copy_desc_avail(wq) <= fnic->wq_copy_desc_low[0])
free_wq_copy_descs(fnic, wq);
if (!vnic_wq_copy_desc_avail(wq))
ret = -EAGAIN;
else {
fnic_queue_wq_copy_desc_fw_reset(wq, SCSI_NO_TAG);
atomic64_inc(&fnic->fnic_stats.fw_stats.active_fw_reqs);
if (atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs) >
atomic64_read(&fnic->fnic_stats.fw_stats.max_fw_reqs))
atomic64_set(&fnic->fnic_stats.fw_stats.max_fw_reqs,
atomic64_read(
&fnic->fnic_stats.fw_stats.active_fw_reqs));
}
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], flags);
if (!ret) {
atomic64_inc(&fnic->fnic_stats.reset_stats.fw_resets);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Issued fw reset\n");
} else {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
fnic_clear_state_flags(fnic, FNIC_FLAGS_FWRESET);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Failed to issue fw reset\n");
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
}
return ret;
}
/*
* fnic_flogi_reg_handler
* Routine to send flogi register msg to fw
*/
int fnic_flogi_reg_handler(struct fnic *fnic, u32 fc_id)
{
struct vnic_wq_copy *wq = &fnic->wq_copy[0];
enum fcpio_flogi_reg_format_type format;
struct fc_lport *lp = fnic->lport;
u8 gw_mac[ETH_ALEN];
int ret = 0;
unsigned long flags;
spin_lock_irqsave(&fnic->wq_copy_lock[0], flags);
if (vnic_wq_copy_desc_avail(wq) <= fnic->wq_copy_desc_low[0])
free_wq_copy_descs(fnic, wq);
if (!vnic_wq_copy_desc_avail(wq)) {
ret = -EAGAIN;
goto flogi_reg_ioreq_end;
}
if (fnic->ctlr.map_dest) {
memset(gw_mac, 0xff, ETH_ALEN);
format = FCPIO_FLOGI_REG_DEF_DEST;
} else {
memcpy(gw_mac, fnic->ctlr.dest_addr, ETH_ALEN);
format = FCPIO_FLOGI_REG_GW_DEST;
}
if ((fnic->config.flags & VFCF_FIP_CAPABLE) && !fnic->ctlr.map_dest) {
fnic_queue_wq_copy_desc_fip_reg(wq, SCSI_NO_TAG,
fc_id, gw_mac,
fnic->data_src_addr,
lp->r_a_tov, lp->e_d_tov);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"FLOGI FIP reg issued fcid %x src %pM dest %pM\n",
fc_id, fnic->data_src_addr, gw_mac);
} else {
fnic_queue_wq_copy_desc_flogi_reg(wq, SCSI_NO_TAG,
format, fc_id, gw_mac);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"FLOGI reg issued fcid %x map %d dest %pM\n",
fc_id, fnic->ctlr.map_dest, gw_mac);
}
atomic64_inc(&fnic->fnic_stats.fw_stats.active_fw_reqs);
if (atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs) >
atomic64_read(&fnic->fnic_stats.fw_stats.max_fw_reqs))
atomic64_set(&fnic->fnic_stats.fw_stats.max_fw_reqs,
atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs));
flogi_reg_ioreq_end:
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], flags);
return ret;
}
/*
* fnic_queue_wq_copy_desc
* Routine to enqueue a wq copy desc
*/
static inline int fnic_queue_wq_copy_desc(struct fnic *fnic,
struct vnic_wq_copy *wq,
struct fnic_io_req *io_req,
struct scsi_cmnd *sc,
int sg_count)
{
struct scatterlist *sg;
struct fc_rport *rport = starget_to_rport(scsi_target(sc->device));
struct fc_rport_libfc_priv *rp = rport->dd_data;
struct host_sg_desc *desc;
struct misc_stats *misc_stats = &fnic->fnic_stats.misc_stats;
unsigned int i;
unsigned long intr_flags;
int flags;
u8 exch_flags;
struct scsi_lun fc_lun;
int r;
if (sg_count) {
/* For each SGE, create a device desc entry */
desc = io_req->sgl_list;
for_each_sg(scsi_sglist(sc), sg, sg_count, i) {
desc->addr = cpu_to_le64(sg_dma_address(sg));
desc->len = cpu_to_le32(sg_dma_len(sg));
desc->_resvd = 0;
desc++;
}
io_req->sgl_list_pa = pci_map_single
(fnic->pdev,
io_req->sgl_list,
sizeof(io_req->sgl_list[0]) * sg_count,
PCI_DMA_TODEVICE);
r = pci_dma_mapping_error(fnic->pdev, io_req->sgl_list_pa);
if (r) {
printk(KERN_ERR "PCI mapping failed with error %d\n", r);
return SCSI_MLQUEUE_HOST_BUSY;
}
}
io_req->sense_buf_pa = pci_map_single(fnic->pdev,
sc->sense_buffer,
SCSI_SENSE_BUFFERSIZE,
PCI_DMA_FROMDEVICE);
r = pci_dma_mapping_error(fnic->pdev, io_req->sense_buf_pa);
if (r) {
pci_unmap_single(fnic->pdev, io_req->sgl_list_pa,
sizeof(io_req->sgl_list[0]) * sg_count,
PCI_DMA_TODEVICE);
printk(KERN_ERR "PCI mapping failed with error %d\n", r);
return SCSI_MLQUEUE_HOST_BUSY;
}
int_to_scsilun(sc->device->lun, &fc_lun);
/* Enqueue the descriptor in the Copy WQ */
spin_lock_irqsave(&fnic->wq_copy_lock[0], intr_flags);
if (vnic_wq_copy_desc_avail(wq) <= fnic->wq_copy_desc_low[0])
free_wq_copy_descs(fnic, wq);
if (unlikely(!vnic_wq_copy_desc_avail(wq))) {
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], intr_flags);
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"fnic_queue_wq_copy_desc failure - no descriptors\n");
atomic64_inc(&misc_stats->io_cpwq_alloc_failures);
return SCSI_MLQUEUE_HOST_BUSY;
}
flags = 0;
if (sc->sc_data_direction == DMA_FROM_DEVICE)
flags = FCPIO_ICMND_RDDATA;
else if (sc->sc_data_direction == DMA_TO_DEVICE)
flags = FCPIO_ICMND_WRDATA;
exch_flags = 0;
if ((fnic->config.flags & VFCF_FCP_SEQ_LVL_ERR) &&
(rp->flags & FC_RP_FLAGS_RETRY))
exch_flags |= FCPIO_ICMND_SRFLAG_RETRY;
fnic_queue_wq_copy_desc_icmnd_16(wq, sc->request->tag,
0, exch_flags, io_req->sgl_cnt,
SCSI_SENSE_BUFFERSIZE,
io_req->sgl_list_pa,
io_req->sense_buf_pa,
0, /* scsi cmd ref, always 0 */
FCPIO_ICMND_PTA_SIMPLE,
/* scsi pri and tag */
flags, /* command flags */
sc->cmnd, sc->cmd_len,
scsi_bufflen(sc),
fc_lun.scsi_lun, io_req->port_id,
rport->maxframe_size, rp->r_a_tov,
rp->e_d_tov);
atomic64_inc(&fnic->fnic_stats.fw_stats.active_fw_reqs);
if (atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs) >
atomic64_read(&fnic->fnic_stats.fw_stats.max_fw_reqs))
atomic64_set(&fnic->fnic_stats.fw_stats.max_fw_reqs,
atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs));
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], intr_flags);
return 0;
}
/*
* fnic_queuecommand
* Routine to send a scsi cdb
* Called with host_lock held and interrupts disabled.
*/
static int fnic_queuecommand_lck(struct scsi_cmnd *sc, void (*done)(struct scsi_cmnd *))
{
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
struct fc_lport *lp = shost_priv(sc->device->host);
struct fc_rport *rport;
struct fnic_io_req *io_req = NULL;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
struct fnic *fnic = lport_priv(lp);
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
struct vnic_wq_copy *wq;
int ret;
u64 cmd_trace;
int sg_count = 0;
unsigned long flags = 0;
unsigned long ptr;
spinlock_t *io_lock = NULL;
int io_lock_acquired = 0;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (unlikely(fnic_chk_state_flags_locked(fnic, FNIC_FLAGS_IO_BLOCKED)))
return SCSI_MLQUEUE_HOST_BUSY;
rport = starget_to_rport(scsi_target(sc->device));
ret = fc_remote_port_chkready(rport);
if (ret) {
atomic64_inc(&fnic_stats->misc_stats.rport_not_ready);
sc->result = ret;
done(sc);
return 0;
}
if (rport) {
struct fc_rport_libfc_priv *rp = rport->dd_data;
if (!rp || rp->rp_state != RPORT_ST_READY) {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"returning DID_NO_CONNECT for IO as rport is removed\n");
atomic64_inc(&fnic_stats->misc_stats.rport_not_ready);
sc->result = DID_NO_CONNECT<<16;
done(sc);
return 0;
}
}
if (lp->state != LPORT_ST_READY || !(lp->link_up))
return SCSI_MLQUEUE_HOST_BUSY;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
atomic_inc(&fnic->in_flight);
/*
* Release host lock, use driver resource specific locks from here.
* Don't re-enable interrupts in case they were disabled prior to the
* caller disabling them.
*/
spin_unlock(lp->host->host_lock);
CMD_STATE(sc) = FNIC_IOREQ_NOT_INITED;
CMD_FLAGS(sc) = FNIC_NO_FLAGS;
/* Get a new io_req for this SCSI IO */
io_req = mempool_alloc(fnic->io_req_pool, GFP_ATOMIC);
if (!io_req) {
atomic64_inc(&fnic_stats->io_stats.alloc_failures);
ret = SCSI_MLQUEUE_HOST_BUSY;
goto out;
}
memset(io_req, 0, sizeof(*io_req));
/* Map the data buffer */
sg_count = scsi_dma_map(sc);
if (sg_count < 0) {
FNIC_TRACE(fnic_queuecommand, sc->device->host->host_no,
sc->request->tag, sc, 0, sc->cmnd[0],
sg_count, CMD_STATE(sc));
mempool_free(io_req, fnic->io_req_pool);
goto out;
}
/* Determine the type of scatter/gather list we need */
io_req->sgl_cnt = sg_count;
io_req->sgl_type = FNIC_SGL_CACHE_DFLT;
if (sg_count > FNIC_DFLT_SG_DESC_CNT)
io_req->sgl_type = FNIC_SGL_CACHE_MAX;
if (sg_count) {
io_req->sgl_list =
mempool_alloc(fnic->io_sgl_pool[io_req->sgl_type],
GFP_ATOMIC);
if (!io_req->sgl_list) {
atomic64_inc(&fnic_stats->io_stats.alloc_failures);
ret = SCSI_MLQUEUE_HOST_BUSY;
scsi_dma_unmap(sc);
mempool_free(io_req, fnic->io_req_pool);
goto out;
}
/* Cache sgl list allocated address before alignment */
io_req->sgl_list_alloc = io_req->sgl_list;
ptr = (unsigned long) io_req->sgl_list;
if (ptr % FNIC_SG_DESC_ALIGN) {
io_req->sgl_list = (struct host_sg_desc *)
(((unsigned long) ptr
+ FNIC_SG_DESC_ALIGN - 1)
& ~(FNIC_SG_DESC_ALIGN - 1));
}
}
/*
* Will acquire lock defore setting to IO initialized.
*/
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
/* initialize rest of io_req */
io_lock_acquired = 1;
io_req->port_id = rport->port_id;
io_req->start_time = jiffies;
CMD_STATE(sc) = FNIC_IOREQ_CMD_PENDING;
CMD_SP(sc) = (char *)io_req;
CMD_FLAGS(sc) |= FNIC_IO_INITIALIZED;
sc->scsi_done = done;
/* create copy wq desc and enqueue it */
wq = &fnic->wq_copy[0];
ret = fnic_queue_wq_copy_desc(fnic, wq, io_req, sc, sg_count);
if (ret) {
/*
* In case another thread cancelled the request,
* refetch the pointer under the lock.
*/
FNIC_TRACE(fnic_queuecommand, sc->device->host->host_no,
sc->request->tag, sc, 0, 0, 0,
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
io_req = (struct fnic_io_req *)CMD_SP(sc);
CMD_SP(sc) = NULL;
CMD_STATE(sc) = FNIC_IOREQ_CMD_COMPLETE;
spin_unlock_irqrestore(io_lock, flags);
if (io_req) {
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
}
atomic_dec(&fnic->in_flight);
/* acquire host lock before returning to SCSI */
spin_lock(lp->host->host_lock);
return ret;
} else {
atomic64_inc(&fnic_stats->io_stats.active_ios);
atomic64_inc(&fnic_stats->io_stats.num_ios);
if (atomic64_read(&fnic_stats->io_stats.active_ios) >
atomic64_read(&fnic_stats->io_stats.max_active_ios))
atomic64_set(&fnic_stats->io_stats.max_active_ios,
atomic64_read(&fnic_stats->io_stats.active_ios));
/* REVISIT: Use per IO lock in the final code */
CMD_FLAGS(sc) |= FNIC_IO_ISSUED;
}
out:
cmd_trace = ((u64)sc->cmnd[0] << 56 | (u64)sc->cmnd[7] << 40 |
(u64)sc->cmnd[8] << 32 | (u64)sc->cmnd[2] << 24 |
(u64)sc->cmnd[3] << 16 | (u64)sc->cmnd[4] << 8 |
sc->cmnd[5]);
FNIC_TRACE(fnic_queuecommand, sc->device->host->host_no,
sc->request->tag, sc, io_req,
sg_count, cmd_trace,
(((u64)CMD_FLAGS(sc) >> 32) | CMD_STATE(sc)));
/* if only we issued IO, will we have the io lock */
if (io_lock_acquired)
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
atomic_dec(&fnic->in_flight);
/* acquire host lock before returning to SCSI */
spin_lock(lp->host->host_lock);
return ret;
}
DEF_SCSI_QCMD(fnic_queuecommand)
/*
* fnic_fcpio_fw_reset_cmpl_handler
* Routine to handle fw reset completion
*/
static int fnic_fcpio_fw_reset_cmpl_handler(struct fnic *fnic,
struct fcpio_fw_req *desc)
{
u8 type;
u8 hdr_status;
struct fcpio_tag tag;
int ret = 0;
unsigned long flags;
struct reset_stats *reset_stats = &fnic->fnic_stats.reset_stats;
fcpio_header_dec(&desc->hdr, &type, &hdr_status, &tag);
atomic64_inc(&reset_stats->fw_reset_completions);
/* Clean up all outstanding io requests */
fnic_cleanup_io(fnic, SCSI_NO_TAG);
atomic64_set(&fnic->fnic_stats.fw_stats.active_fw_reqs, 0);
atomic64_set(&fnic->fnic_stats.io_stats.active_ios, 0);
spin_lock_irqsave(&fnic->fnic_lock, flags);
/* fnic should be in FC_TRANS_ETH_MODE */
if (fnic->state == FNIC_IN_FC_TRANS_ETH_MODE) {
/* Check status of reset completion */
if (!hdr_status) {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"reset cmpl success\n");
/* Ready to send flogi out */
fnic->state = FNIC_IN_ETH_MODE;
} else {
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"fnic fw_reset : failed %s\n",
fnic_fcpio_status_to_str(hdr_status));
/*
* Unable to change to eth mode, cannot send out flogi
* Change state to fc mode, so that subsequent Flogi
* requests from libFC will cause more attempts to
* reset the firmware. Free the cached flogi
*/
fnic->state = FNIC_IN_FC_MODE;
atomic64_inc(&reset_stats->fw_reset_failures);
ret = -1;
}
} else {
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"Unexpected state %s while processing"
" reset cmpl\n", fnic_state_to_str(fnic->state));
atomic64_inc(&reset_stats->fw_reset_failures);
ret = -1;
}
/* Thread removing device blocks till firmware reset is complete */
if (fnic->remove_wait)
complete(fnic->remove_wait);
/*
* If fnic is being removed, or fw reset failed
* free the flogi frame. Else, send it out
*/
if (fnic->remove_wait || ret) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
skb_queue_purge(&fnic->tx_queue);
goto reset_cmpl_handler_end;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_flush_tx(fnic);
reset_cmpl_handler_end:
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
fnic_clear_state_flags(fnic, FNIC_FLAGS_FWRESET);
return ret;
}
/*
* fnic_fcpio_flogi_reg_cmpl_handler
* Routine to handle flogi register completion
*/
static int fnic_fcpio_flogi_reg_cmpl_handler(struct fnic *fnic,
struct fcpio_fw_req *desc)
{
u8 type;
u8 hdr_status;
struct fcpio_tag tag;
int ret = 0;
unsigned long flags;
fcpio_header_dec(&desc->hdr, &type, &hdr_status, &tag);
/* Update fnic state based on status of flogi reg completion */
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->state == FNIC_IN_ETH_TRANS_FC_MODE) {
/* Check flogi registration completion status */
if (!hdr_status) {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"flog reg succeeded\n");
fnic->state = FNIC_IN_FC_MODE;
} else {
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"fnic flogi reg :failed %s\n",
fnic_fcpio_status_to_str(hdr_status));
fnic->state = FNIC_IN_ETH_MODE;
ret = -1;
}
} else {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Unexpected fnic state %s while"
" processing flogi reg completion\n",
fnic_state_to_str(fnic->state));
ret = -1;
}
if (!ret) {
if (fnic->stop_rx_link_events) {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
goto reg_cmpl_handler_end;
}
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
fnic_flush_tx(fnic);
queue_work(fnic_event_queue, &fnic->frame_work);
} else {
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
reg_cmpl_handler_end:
return ret;
}
static inline int is_ack_index_in_range(struct vnic_wq_copy *wq,
u16 request_out)
{
if (wq->to_clean_index <= wq->to_use_index) {
/* out of range, stale request_out index */
if (request_out < wq->to_clean_index ||
request_out >= wq->to_use_index)
return 0;
} else {
/* out of range, stale request_out index */
if (request_out < wq->to_clean_index &&
request_out >= wq->to_use_index)
return 0;
}
/* request_out index is in range */
return 1;
}
/*
* Mark that ack received and store the Ack index. If there are multiple
* acks received before Tx thread cleans it up, the latest value will be
* used which is correct behavior. This state should be in the copy Wq
* instead of in the fnic
*/
static inline void fnic_fcpio_ack_handler(struct fnic *fnic,
unsigned int cq_index,
struct fcpio_fw_req *desc)
{
struct vnic_wq_copy *wq;
u16 request_out = desc->u.ack.request_out;
unsigned long flags;
u64 *ox_id_tag = (u64 *)(void *)desc;
/* mark the ack state */
wq = &fnic->wq_copy[cq_index - fnic->raw_wq_count - fnic->rq_count];
spin_lock_irqsave(&fnic->wq_copy_lock[0], flags);
fnic->fnic_stats.misc_stats.last_ack_time = jiffies;
if (is_ack_index_in_range(wq, request_out)) {
fnic->fw_ack_index[0] = request_out;
fnic->fw_ack_recd[0] = 1;
} else
atomic64_inc(
&fnic->fnic_stats.misc_stats.ack_index_out_of_range);
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], flags);
FNIC_TRACE(fnic_fcpio_ack_handler,
fnic->lport->host->host_no, 0, 0, ox_id_tag[2], ox_id_tag[3],
ox_id_tag[4], ox_id_tag[5]);
}
/*
* fnic_fcpio_icmnd_cmpl_handler
* Routine to handle icmnd completions
*/
static void fnic_fcpio_icmnd_cmpl_handler(struct fnic *fnic,
struct fcpio_fw_req *desc)
{
u8 type;
u8 hdr_status;
struct fcpio_tag tag;
u32 id;
u64 xfer_len = 0;
struct fcpio_icmnd_cmpl *icmnd_cmpl;
struct fnic_io_req *io_req;
struct scsi_cmnd *sc;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
unsigned long flags;
spinlock_t *io_lock;
u64 cmd_trace;
unsigned long start_time;
/* Decode the cmpl description to get the io_req id */
fcpio_header_dec(&desc->hdr, &type, &hdr_status, &tag);
fcpio_tag_id_dec(&tag, &id);
icmnd_cmpl = &desc->u.icmnd_cmpl;
if (id >= fnic->fnic_max_tag_id) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
shost_printk(KERN_ERR, fnic->lport->host,
"Tag out of range tag %x hdr status = %s\n",
id, fnic_fcpio_status_to_str(hdr_status));
return;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
}
sc = scsi_host_find_tag(fnic->lport->host, id);
WARN_ON_ONCE(!sc);
if (!sc) {
atomic64_inc(&fnic_stats->io_stats.sc_null);
shost_printk(KERN_ERR, fnic->lport->host,
"icmnd_cmpl sc is null - "
"hdr status = %s tag = 0x%x desc = 0x%p\n",
fnic_fcpio_status_to_str(hdr_status), id, desc);
FNIC_TRACE(fnic_fcpio_icmnd_cmpl_handler,
fnic->lport->host->host_no, id,
((u64)icmnd_cmpl->_resvd0[1] << 16 |
(u64)icmnd_cmpl->_resvd0[0]),
((u64)hdr_status << 16 |
(u64)icmnd_cmpl->scsi_status << 8 |
(u64)icmnd_cmpl->flags), desc,
(u64)icmnd_cmpl->residual, 0);
return;
}
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
WARN_ON_ONCE(!io_req);
if (!io_req) {
atomic64_inc(&fnic_stats->io_stats.ioreq_null);
CMD_FLAGS(sc) |= FNIC_IO_REQ_NULL;
spin_unlock_irqrestore(io_lock, flags);
shost_printk(KERN_ERR, fnic->lport->host,
"icmnd_cmpl io_req is null - "
"hdr status = %s tag = 0x%x sc 0x%p\n",
fnic_fcpio_status_to_str(hdr_status), id, sc);
return;
}
start_time = io_req->start_time;
/* firmware completed the io */
io_req->io_completed = 1;
/*
* if SCSI-ML has already issued abort on this command,
* ignore completion of the IO. The abts path will clean it up
*/
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_IO_ABTS_PENDING;
switch (hdr_status) {
case FCPIO_SUCCESS:
CMD_FLAGS(sc) |= FNIC_IO_DONE;
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"icmnd_cmpl ABTS pending hdr status = %s "
"sc 0x%p scsi_status %x residual %d\n",
fnic_fcpio_status_to_str(hdr_status), sc,
icmnd_cmpl->scsi_status,
icmnd_cmpl->residual);
break;
case FCPIO_ABORTED:
CMD_FLAGS(sc) |= FNIC_IO_ABORTED;
break;
default:
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"icmnd_cmpl abts pending "
"hdr status = %s tag = 0x%x sc = 0x%p\n",
fnic_fcpio_status_to_str(hdr_status),
id, sc);
break;
}
return;
}
/* Mark the IO as complete */
CMD_STATE(sc) = FNIC_IOREQ_CMD_COMPLETE;
icmnd_cmpl = &desc->u.icmnd_cmpl;
switch (hdr_status) {
case FCPIO_SUCCESS:
sc->result = (DID_OK << 16) | icmnd_cmpl->scsi_status;
xfer_len = scsi_bufflen(sc);
scsi_set_resid(sc, icmnd_cmpl->residual);
if (icmnd_cmpl->flags & FCPIO_ICMND_CMPL_RESID_UNDER)
xfer_len -= icmnd_cmpl->residual;
if (icmnd_cmpl->scsi_status == SAM_STAT_TASK_SET_FULL)
atomic64_inc(&fnic_stats->misc_stats.queue_fulls);
break;
case FCPIO_TIMEOUT: /* request was timed out */
atomic64_inc(&fnic_stats->misc_stats.fcpio_timeout);
sc->result = (DID_TIME_OUT << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_ABORTED: /* request was aborted */
atomic64_inc(&fnic_stats->misc_stats.fcpio_aborted);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_DATA_CNT_MISMATCH: /* recv/sent more/less data than exp. */
atomic64_inc(&fnic_stats->misc_stats.data_count_mismatch);
scsi_set_resid(sc, icmnd_cmpl->residual);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_OUT_OF_RESOURCE: /* out of resources to complete request */
atomic64_inc(&fnic_stats->fw_stats.fw_out_of_resources);
sc->result = (DID_REQUEUE << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_IO_NOT_FOUND: /* requested I/O was not found */
atomic64_inc(&fnic_stats->io_stats.io_not_found);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_SGL_INVALID: /* request was aborted due to sgl error */
atomic64_inc(&fnic_stats->misc_stats.sgl_invalid);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_FW_ERR: /* request was terminated due fw error */
atomic64_inc(&fnic_stats->fw_stats.io_fw_errs);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_MSS_INVALID: /* request was aborted due to mss error */
atomic64_inc(&fnic_stats->misc_stats.mss_invalid);
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
case FCPIO_INVALID_HEADER: /* header contains invalid data */
case FCPIO_INVALID_PARAM: /* some parameter in request invalid */
case FCPIO_REQ_NOT_SUPPORTED:/* request type is not supported */
default:
sc->result = (DID_ERROR << 16) | icmnd_cmpl->scsi_status;
break;
}
/* Break link with the SCSI command */
CMD_SP(sc) = NULL;
CMD_FLAGS(sc) |= FNIC_IO_DONE;
spin_unlock_irqrestore(io_lock, flags);
fnic: move printk()s outside of the critical code section. This patch moves a printk() outside of the code section where interrupt are disabled. In some cases a flood of error messages may cause a kernel panic. It also removes one of the printk()s because the same error message was printed twice. [709686.317197] Kernel panic - not syncing: Watchdog detected hard LOCKUP on cpu 12 [709686.317200] CPU: 12 PID: 1963 Comm: systemd-journal Tainted: GF O-------------- 3.10.0-229.el7.x86_64 #1 [709686.317201] Hardware name: Cisco Systems Inc UCSB-B200-M3/UCSB-B200-M3, BIOS B200M3.2.2.3.6.030620151309 03/06/2015 [709686.317206] ffffffff8182b2e8 00000000392722ba ffff88046fcc5c48 ffffffff81603f36 [709686.317209] ffff88046fcc5cc8 ffffffff815fd7da 0000000000000010 ffff88046fcc5cd8 [709686.317211] ffff88046fcc5c78 00000000392722ba ffff88046fcc5c88 000000000000000c [709686.317212] Call Trace: [709686.317221] <NMI> [<ffffffff81603f36>] dump_stack+0x19/0x1b [709686.317223] [<ffffffff815fd7da>] panic+0xd8/0x1e7 [709686.317227] [<ffffffff8110a760>] ? watchdog_enable_all_cpus.part.2+0x40/0x40 [709686.317229] [<ffffffff8110a822>] watchdog_overflow_callback+0xc2/0xd0 [709686.317233] [<ffffffff8114c901>] __perf_event_overflow+0xa1/0x250 [709686.317235] [<ffffffff8114d404>] perf_event_overflow+0x14/0x20 [709686.317239] [<ffffffff810301fd>] intel_pmu_handle_irq+0x1fd/0x410 [709686.317242] [<ffffffff811908d1>] ? unmap_kernel_range_noflush+0x11/0x20 [709686.317246] [<ffffffff81373574>] ? ghes_copy_tofrom_phys+0x124/0x210 [709686.317249] [<ffffffff8160cfcb>] perf_event_nmi_handler+0x2b/0x50 [709686.317251] [<ffffffff8160c719>] nmi_handle.isra.0+0x69/0xb0 [709686.317252] [<ffffffff8160c830>] do_nmi+0xd0/0x340 [709686.317256] [<ffffffff8160bb71>] end_repeat_nmi+0x1e/0x2e [709686.317260] [<ffffffff812e24fd>] ? memcpy+0xd/0x110 [709686.317263] [<ffffffff812e24fd>] ? memcpy+0xd/0x110 [709686.317265] [<ffffffff812e24fd>] ? memcpy+0xd/0x110 [709686.317269] <<EOE>> [<ffffffff8132c297>] ? vgacon_scroll+0x2d7/0x330 [709686.317273] [<ffffffff813a086c>] scrup+0xfc/0x110 [709686.317275] [<ffffffff813a0920>] lf+0xa0/0xb0 [709686.317278] [<ffffffff813a1b32>] vt_console_print+0x2d2/0x420 [709686.317283] [<ffffffff8106f4a1>] call_console_drivers.constprop.15+0x91/0xf0 [709686.317287] [<ffffffff8107069f>] console_unlock+0x3bf/0x400 [709686.317291] [<ffffffff81070996>] vprintk_emit+0x2b6/0x530 [709686.317294] [<ffffffff815fd961>] printk_emit+0x44/0x5b [709686.317297] [<ffffffff81070d98>] devkmsg_writev+0x158/0x1d0 [709686.317303] [<ffffffff811c5ef9>] do_sync_readv_writev+0x79/0xd0 [709686.317307] [<ffffffff811c73ee>] do_readv_writev+0xce/0x260 [709686.317310] [<ffffffff811c8d18>] ? __sb_start_write+0x58/0x110 [709686.317314] [<ffffffff811c7615>] vfs_writev+0x35/0x60 [709686.317318] [<ffffffff811c776c>] SyS_writev+0x5c/0xd0 [709686.317322] [<ffffffff81613da9>] system_call_fastpath+0x16/0x1b Signed-off-by: Maurizio Lombardi <mlombard@redhat.com> Reviewed-by: Laurence Oberman <loberman@redhat.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-03-16 21:44:08 +08:00
if (hdr_status != FCPIO_SUCCESS) {
atomic64_inc(&fnic_stats->io_stats.io_failures);
shost_printk(KERN_ERR, fnic->lport->host, "hdr status = %s\n",
fnic_fcpio_status_to_str(hdr_status));
}
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
cmd_trace = ((u64)hdr_status << 56) |
(u64)icmnd_cmpl->scsi_status << 48 |
(u64)icmnd_cmpl->flags << 40 | (u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 | (u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5];
FNIC_TRACE(fnic_fcpio_icmnd_cmpl_handler,
sc->device->host->host_no, id, sc,
((u64)icmnd_cmpl->_resvd0[1] << 56 |
(u64)icmnd_cmpl->_resvd0[0] << 48 |
jiffies_to_msecs(jiffies - start_time)),
desc, cmd_trace,
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
if (sc->sc_data_direction == DMA_FROM_DEVICE) {
fnic->lport->host_stats.fcp_input_requests++;
fnic->fcp_input_bytes += xfer_len;
} else if (sc->sc_data_direction == DMA_TO_DEVICE) {
fnic->lport->host_stats.fcp_output_requests++;
fnic->fcp_output_bytes += xfer_len;
} else
fnic->lport->host_stats.fcp_control_requests++;
atomic64_dec(&fnic_stats->io_stats.active_ios);
if (atomic64_read(&fnic->io_cmpl_skip))
atomic64_dec(&fnic->io_cmpl_skip);
else
atomic64_inc(&fnic_stats->io_stats.io_completions);
/* Call SCSI completion function to complete the IO */
if (sc->scsi_done)
sc->scsi_done(sc);
}
/* fnic_fcpio_itmf_cmpl_handler
* Routine to handle itmf completions
*/
static void fnic_fcpio_itmf_cmpl_handler(struct fnic *fnic,
struct fcpio_fw_req *desc)
{
u8 type;
u8 hdr_status;
struct fcpio_tag tag;
u32 id;
struct scsi_cmnd *sc;
struct fnic_io_req *io_req;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
struct abort_stats *abts_stats = &fnic->fnic_stats.abts_stats;
struct terminate_stats *term_stats = &fnic->fnic_stats.term_stats;
struct misc_stats *misc_stats = &fnic->fnic_stats.misc_stats;
unsigned long flags;
spinlock_t *io_lock;
unsigned long start_time;
fcpio_header_dec(&desc->hdr, &type, &hdr_status, &tag);
fcpio_tag_id_dec(&tag, &id);
if ((id & FNIC_TAG_MASK) >= fnic->fnic_max_tag_id) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
shost_printk(KERN_ERR, fnic->lport->host,
"Tag out of range tag %x hdr status = %s\n",
id, fnic_fcpio_status_to_str(hdr_status));
return;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
}
sc = scsi_host_find_tag(fnic->lport->host, id & FNIC_TAG_MASK);
WARN_ON_ONCE(!sc);
if (!sc) {
atomic64_inc(&fnic_stats->io_stats.sc_null);
shost_printk(KERN_ERR, fnic->lport->host,
"itmf_cmpl sc is null - hdr status = %s tag = 0x%x\n",
fnic_fcpio_status_to_str(hdr_status), id);
return;
}
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
WARN_ON_ONCE(!io_req);
if (!io_req) {
atomic64_inc(&fnic_stats->io_stats.ioreq_null);
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_REQ_NULL;
shost_printk(KERN_ERR, fnic->lport->host,
"itmf_cmpl io_req is null - "
"hdr status = %s tag = 0x%x sc 0x%p\n",
fnic_fcpio_status_to_str(hdr_status), id, sc);
return;
}
start_time = io_req->start_time;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if ((id & FNIC_TAG_ABORT) && (id & FNIC_TAG_DEV_RST)) {
/* Abort and terminate completion of device reset req */
/* REVISIT : Add asserts about various flags */
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"dev reset abts cmpl recd. id %x status %s\n",
id, fnic_fcpio_status_to_str(hdr_status));
CMD_STATE(sc) = FNIC_IOREQ_ABTS_COMPLETE;
CMD_ABTS_STATUS(sc) = hdr_status;
CMD_FLAGS(sc) |= FNIC_DEV_RST_DONE;
if (io_req->abts_done)
complete(io_req->abts_done);
spin_unlock_irqrestore(io_lock, flags);
} else if (id & FNIC_TAG_ABORT) {
/* Completion of abort cmd */
switch (hdr_status) {
case FCPIO_SUCCESS:
break;
case FCPIO_TIMEOUT:
if (CMD_FLAGS(sc) & FNIC_IO_ABTS_ISSUED)
atomic64_inc(&abts_stats->abort_fw_timeouts);
else
atomic64_inc(
&term_stats->terminate_fw_timeouts);
break;
case FCPIO_ITMF_REJECTED:
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"abort reject recd. id %d\n",
(int)(id & FNIC_TAG_MASK));
break;
case FCPIO_IO_NOT_FOUND:
if (CMD_FLAGS(sc) & FNIC_IO_ABTS_ISSUED)
atomic64_inc(&abts_stats->abort_io_not_found);
else
atomic64_inc(
&term_stats->terminate_io_not_found);
break;
default:
if (CMD_FLAGS(sc) & FNIC_IO_ABTS_ISSUED)
atomic64_inc(&abts_stats->abort_failures);
else
atomic64_inc(
&term_stats->terminate_failures);
break;
}
if (CMD_STATE(sc) != FNIC_IOREQ_ABTS_PENDING) {
/* This is a late completion. Ignore it */
spin_unlock_irqrestore(io_lock, flags);
return;
}
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_DONE;
/* If the status is IO not found consider it as success */
if (hdr_status == FCPIO_IO_NOT_FOUND)
CMD_ABTS_STATUS(sc) = FCPIO_SUCCESS;
else
CMD_ABTS_STATUS(sc) = hdr_status;
atomic64_dec(&fnic_stats->io_stats.active_ios);
if (atomic64_read(&fnic->io_cmpl_skip))
atomic64_dec(&fnic->io_cmpl_skip);
else
atomic64_inc(&fnic_stats->io_stats.io_completions);
if (!(CMD_FLAGS(sc) & (FNIC_IO_ABORTED | FNIC_IO_DONE)))
atomic64_inc(&misc_stats->no_icmnd_itmf_cmpls);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"abts cmpl recd. id %d status %s\n",
(int)(id & FNIC_TAG_MASK),
fnic_fcpio_status_to_str(hdr_status));
/*
* If scsi_eh thread is blocked waiting for abts to complete,
* signal completion to it. IO will be cleaned in the thread
* else clean it in this context
*/
if (io_req->abts_done) {
complete(io_req->abts_done);
spin_unlock_irqrestore(io_lock, flags);
} else {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"abts cmpl, completing IO\n");
CMD_SP(sc) = NULL;
sc->result = (DID_ERROR << 16);
spin_unlock_irqrestore(io_lock, flags);
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
if (sc->scsi_done) {
FNIC_TRACE(fnic_fcpio_itmf_cmpl_handler,
sc->device->host->host_no, id,
sc,
jiffies_to_msecs(jiffies - start_time),
desc,
(((u64)hdr_status << 40) |
(u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 |
(u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5]),
(((u64)CMD_FLAGS(sc) << 32) |
CMD_STATE(sc)));
sc->scsi_done(sc);
}
}
} else if (id & FNIC_TAG_DEV_RST) {
/* Completion of device reset */
CMD_LR_STATUS(sc) = hdr_status;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_DEV_RST_ABTS_PENDING;
FNIC_TRACE(fnic_fcpio_itmf_cmpl_handler,
sc->device->host->host_no, id, sc,
jiffies_to_msecs(jiffies - start_time),
desc, 0,
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Terminate pending "
"dev reset cmpl recd. id %d status %s\n",
(int)(id & FNIC_TAG_MASK),
fnic_fcpio_status_to_str(hdr_status));
return;
}
if (CMD_FLAGS(sc) & FNIC_DEV_RST_TIMED_OUT) {
/* Need to wait for terminate completion */
spin_unlock_irqrestore(io_lock, flags);
FNIC_TRACE(fnic_fcpio_itmf_cmpl_handler,
sc->device->host->host_no, id, sc,
jiffies_to_msecs(jiffies - start_time),
desc, 0,
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"dev reset cmpl recd after time out. "
"id %d status %s\n",
(int)(id & FNIC_TAG_MASK),
fnic_fcpio_status_to_str(hdr_status));
return;
}
CMD_STATE(sc) = FNIC_IOREQ_CMD_COMPLETE;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
CMD_FLAGS(sc) |= FNIC_DEV_RST_DONE;
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"dev reset cmpl recd. id %d status %s\n",
(int)(id & FNIC_TAG_MASK),
fnic_fcpio_status_to_str(hdr_status));
if (io_req->dr_done)
complete(io_req->dr_done);
spin_unlock_irqrestore(io_lock, flags);
} else {
shost_printk(KERN_ERR, fnic->lport->host,
"Unexpected itmf io state %s tag %x\n",
fnic_ioreq_state_to_str(CMD_STATE(sc)), id);
spin_unlock_irqrestore(io_lock, flags);
}
}
/*
* fnic_fcpio_cmpl_handler
* Routine to service the cq for wq_copy
*/
static int fnic_fcpio_cmpl_handler(struct vnic_dev *vdev,
unsigned int cq_index,
struct fcpio_fw_req *desc)
{
struct fnic *fnic = vnic_dev_priv(vdev);
switch (desc->hdr.type) {
case FCPIO_ICMND_CMPL: /* fw completed a command */
case FCPIO_ITMF_CMPL: /* fw completed itmf (abort cmd, lun reset)*/
case FCPIO_FLOGI_REG_CMPL: /* fw completed flogi_reg */
case FCPIO_FLOGI_FIP_REG_CMPL: /* fw completed flogi_fip_reg */
case FCPIO_RESET_CMPL: /* fw completed reset */
atomic64_dec(&fnic->fnic_stats.fw_stats.active_fw_reqs);
break;
default:
break;
}
switch (desc->hdr.type) {
case FCPIO_ACK: /* fw copied copy wq desc to its queue */
fnic_fcpio_ack_handler(fnic, cq_index, desc);
break;
case FCPIO_ICMND_CMPL: /* fw completed a command */
fnic_fcpio_icmnd_cmpl_handler(fnic, desc);
break;
case FCPIO_ITMF_CMPL: /* fw completed itmf (abort cmd, lun reset)*/
fnic_fcpio_itmf_cmpl_handler(fnic, desc);
break;
case FCPIO_FLOGI_REG_CMPL: /* fw completed flogi_reg */
case FCPIO_FLOGI_FIP_REG_CMPL: /* fw completed flogi_fip_reg */
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
fnic_fcpio_flogi_reg_cmpl_handler(fnic, desc);
break;
case FCPIO_RESET_CMPL: /* fw completed reset */
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
fnic_fcpio_fw_reset_cmpl_handler(fnic, desc);
break;
default:
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"firmware completion type %d\n",
desc->hdr.type);
break;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
return 0;
}
/*
* fnic_wq_copy_cmpl_handler
* Routine to process wq copy
*/
int fnic_wq_copy_cmpl_handler(struct fnic *fnic, int copy_work_to_do)
{
unsigned int wq_work_done = 0;
unsigned int i, cq_index;
unsigned int cur_work_done;
for (i = 0; i < fnic->wq_copy_count; i++) {
cq_index = i + fnic->raw_wq_count + fnic->rq_count;
cur_work_done = vnic_cq_copy_service(&fnic->cq[cq_index],
fnic_fcpio_cmpl_handler,
copy_work_to_do);
wq_work_done += cur_work_done;
}
return wq_work_done;
}
static void fnic_cleanup_io(struct fnic *fnic, int exclude_id)
{
int i;
struct fnic_io_req *io_req;
unsigned long flags = 0;
struct scsi_cmnd *sc;
spinlock_t *io_lock;
unsigned long start_time = 0;
struct fnic_stats *fnic_stats = &fnic->fnic_stats;
for (i = 0; i < fnic->fnic_max_tag_id; i++) {
if (i == exclude_id)
continue;
io_lock = fnic_io_lock_tag(fnic, i);
spin_lock_irqsave(io_lock, flags);
sc = scsi_host_find_tag(fnic->lport->host, i);
if (!sc) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
io_req = (struct fnic_io_req *)CMD_SP(sc);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if ((CMD_FLAGS(sc) & FNIC_DEVICE_RESET) &&
!(CMD_FLAGS(sc) & FNIC_DEV_RST_DONE)) {
/*
* We will be here only when FW completes reset
* without sending completions for outstanding ios.
*/
CMD_FLAGS(sc) |= FNIC_DEV_RST_DONE;
if (io_req && io_req->dr_done)
complete(io_req->dr_done);
else if (io_req && io_req->abts_done)
complete(io_req->abts_done);
spin_unlock_irqrestore(io_lock, flags);
continue;
} else if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
goto cleanup_scsi_cmd;
}
CMD_SP(sc) = NULL;
spin_unlock_irqrestore(io_lock, flags);
/*
* If there is a scsi_cmnd associated with this io_req, then
* free the corresponding state
*/
start_time = io_req->start_time;
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
cleanup_scsi_cmd:
sc->result = DID_TRANSPORT_DISRUPTED << 16;
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"%s: sc duration = %lu DID_TRANSPORT_DISRUPTED\n",
__func__, (jiffies - start_time));
if (atomic64_read(&fnic->io_cmpl_skip))
atomic64_dec(&fnic->io_cmpl_skip);
else
atomic64_inc(&fnic_stats->io_stats.io_completions);
/* Complete the command to SCSI */
if (sc->scsi_done) {
FNIC_TRACE(fnic_cleanup_io,
sc->device->host->host_no, i, sc,
jiffies_to_msecs(jiffies - start_time),
0, ((u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 |
(u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5]),
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
sc->scsi_done(sc);
}
}
}
void fnic_wq_copy_cleanup_handler(struct vnic_wq_copy *wq,
struct fcpio_host_req *desc)
{
u32 id;
struct fnic *fnic = vnic_dev_priv(wq->vdev);
struct fnic_io_req *io_req;
struct scsi_cmnd *sc;
unsigned long flags;
spinlock_t *io_lock;
unsigned long start_time = 0;
/* get the tag reference */
fcpio_tag_id_dec(&desc->hdr.tag, &id);
id &= FNIC_TAG_MASK;
if (id >= fnic->fnic_max_tag_id)
return;
sc = scsi_host_find_tag(fnic->lport->host, id);
if (!sc)
return;
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
/* Get the IO context which this desc refers to */
io_req = (struct fnic_io_req *)CMD_SP(sc);
/* fnic interrupts are turned off by now */
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
goto wq_copy_cleanup_scsi_cmd;
}
CMD_SP(sc) = NULL;
spin_unlock_irqrestore(io_lock, flags);
start_time = io_req->start_time;
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
wq_copy_cleanup_scsi_cmd:
sc->result = DID_NO_CONNECT << 16;
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host, "wq_copy_cleanup_handler:"
" DID_NO_CONNECT\n");
if (sc->scsi_done) {
FNIC_TRACE(fnic_wq_copy_cleanup_handler,
sc->device->host->host_no, id, sc,
jiffies_to_msecs(jiffies - start_time),
0, ((u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 | (u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5]),
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
sc->scsi_done(sc);
}
}
static inline int fnic_queue_abort_io_req(struct fnic *fnic, int tag,
u32 task_req, u8 *fc_lun,
struct fnic_io_req *io_req)
{
struct vnic_wq_copy *wq = &fnic->wq_copy[0];
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
struct Scsi_Host *host = fnic->lport->host;
struct misc_stats *misc_stats = &fnic->fnic_stats.misc_stats;
unsigned long flags;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_lock_irqsave(host->host_lock, flags);
if (unlikely(fnic_chk_state_flags_locked(fnic,
FNIC_FLAGS_IO_BLOCKED))) {
spin_unlock_irqrestore(host->host_lock, flags);
return 1;
} else
atomic_inc(&fnic->in_flight);
spin_unlock_irqrestore(host->host_lock, flags);
spin_lock_irqsave(&fnic->wq_copy_lock[0], flags);
if (vnic_wq_copy_desc_avail(wq) <= fnic->wq_copy_desc_low[0])
free_wq_copy_descs(fnic, wq);
if (!vnic_wq_copy_desc_avail(wq)) {
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
atomic_dec(&fnic->in_flight);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
"fnic_queue_abort_io_req: failure: no descriptors\n");
atomic64_inc(&misc_stats->abts_cpwq_alloc_failures);
return 1;
}
fnic_queue_wq_copy_desc_itmf(wq, tag | FNIC_TAG_ABORT,
0, task_req, tag, fc_lun, io_req->port_id,
fnic->config.ra_tov, fnic->config.ed_tov);
atomic64_inc(&fnic->fnic_stats.fw_stats.active_fw_reqs);
if (atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs) >
atomic64_read(&fnic->fnic_stats.fw_stats.max_fw_reqs))
atomic64_set(&fnic->fnic_stats.fw_stats.max_fw_reqs,
atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs));
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
atomic_dec(&fnic->in_flight);
return 0;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
static void fnic_rport_exch_reset(struct fnic *fnic, u32 port_id)
{
int tag;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
int abt_tag;
int term_cnt = 0;
struct fnic_io_req *io_req;
spinlock_t *io_lock;
unsigned long flags;
struct scsi_cmnd *sc;
struct reset_stats *reset_stats = &fnic->fnic_stats.reset_stats;
struct terminate_stats *term_stats = &fnic->fnic_stats.term_stats;
struct scsi_lun fc_lun;
enum fnic_ioreq_state old_ioreq_state;
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
"fnic_rport_exch_reset called portid 0x%06x\n",
port_id);
if (fnic->in_remove)
return;
for (tag = 0; tag < fnic->fnic_max_tag_id; tag++) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
abt_tag = tag;
io_lock = fnic_io_lock_tag(fnic, tag);
spin_lock_irqsave(io_lock, flags);
sc = scsi_host_find_tag(fnic->lport->host, tag);
if (!sc) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req || io_req->port_id != port_id) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if ((CMD_FLAGS(sc) & FNIC_DEVICE_RESET) &&
(!(CMD_FLAGS(sc) & FNIC_DEV_RST_ISSUED))) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_rport_exch_reset dev rst not pending sc 0x%p\n",
sc);
spin_unlock_irqrestore(io_lock, flags);
continue;
}
/*
* Found IO that is still pending with firmware and
* belongs to rport that went away
*/
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (io_req->abts_done) {
shost_printk(KERN_ERR, fnic->lport->host,
"fnic_rport_exch_reset: io_req->abts_done is set "
"state is %s\n",
fnic_ioreq_state_to_str(CMD_STATE(sc)));
}
if (!(CMD_FLAGS(sc) & FNIC_IO_ISSUED)) {
shost_printk(KERN_ERR, fnic->lport->host,
"rport_exch_reset "
"IO not yet issued %p tag 0x%x flags "
"%x state %d\n",
sc, tag, CMD_FLAGS(sc), CMD_STATE(sc));
}
old_ioreq_state = CMD_STATE(sc);
CMD_STATE(sc) = FNIC_IOREQ_ABTS_PENDING;
CMD_ABTS_STATUS(sc) = FCPIO_INVALID_CODE;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) {
atomic64_inc(&reset_stats->device_reset_terminates);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
abt_tag = (tag | FNIC_TAG_DEV_RST);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_rport_exch_reset dev rst sc 0x%p\n",
sc);
}
BUG_ON(io_req->abts_done);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_rport_reset_exch: Issuing abts\n");
spin_unlock_irqrestore(io_lock, flags);
/* Now queue the abort command to firmware */
int_to_scsilun(sc->device->lun, &fc_lun);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (fnic_queue_abort_io_req(fnic, abt_tag,
FCPIO_ITMF_ABT_TASK_TERM,
fc_lun.scsi_lun, io_req)) {
/*
* Revert the cmd state back to old state, if
* it hasn't changed in between. This cmd will get
* aborted later by scsi_eh, or cleaned up during
* lun reset
*/
spin_lock_irqsave(io_lock, flags);
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING)
CMD_STATE(sc) = old_ioreq_state;
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
} else {
spin_lock_irqsave(io_lock, flags);
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET)
CMD_FLAGS(sc) |= FNIC_DEV_RST_TERM_ISSUED;
else
CMD_FLAGS(sc) |= FNIC_IO_INTERNAL_TERM_ISSUED;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_unlock_irqrestore(io_lock, flags);
atomic64_inc(&term_stats->terminates);
term_cnt++;
}
}
if (term_cnt > atomic64_read(&term_stats->max_terminates))
atomic64_set(&term_stats->max_terminates, term_cnt);
}
void fnic_terminate_rport_io(struct fc_rport *rport)
{
int tag;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
int abt_tag;
int term_cnt = 0;
struct fnic_io_req *io_req;
spinlock_t *io_lock;
unsigned long flags;
struct scsi_cmnd *sc;
struct scsi_lun fc_lun;
struct fc_rport_libfc_priv *rdata;
struct fc_lport *lport;
struct fnic *fnic;
struct fc_rport *cmd_rport;
struct reset_stats *reset_stats;
struct terminate_stats *term_stats;
enum fnic_ioreq_state old_ioreq_state;
if (!rport) {
printk(KERN_ERR "fnic_terminate_rport_io: rport is NULL\n");
return;
}
rdata = rport->dd_data;
if (!rdata) {
printk(KERN_ERR "fnic_terminate_rport_io: rdata is NULL\n");
return;
}
lport = rdata->local_port;
if (!lport) {
printk(KERN_ERR "fnic_terminate_rport_io: lport is NULL\n");
return;
}
fnic = lport_priv(lport);
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host, "fnic_terminate_rport_io called"
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
" wwpn 0x%llx, wwnn0x%llx, rport 0x%p, portid 0x%06x\n",
rport->port_name, rport->node_name, rport,
rport->port_id);
if (fnic->in_remove)
return;
reset_stats = &fnic->fnic_stats.reset_stats;
term_stats = &fnic->fnic_stats.term_stats;
for (tag = 0; tag < fnic->fnic_max_tag_id; tag++) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
abt_tag = tag;
io_lock = fnic_io_lock_tag(fnic, tag);
spin_lock_irqsave(io_lock, flags);
sc = scsi_host_find_tag(fnic->lport->host, tag);
if (!sc) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
cmd_rport = starget_to_rport(scsi_target(sc->device));
if (rport != cmd_rport) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req || rport != cmd_rport) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if ((CMD_FLAGS(sc) & FNIC_DEVICE_RESET) &&
(!(CMD_FLAGS(sc) & FNIC_DEV_RST_ISSUED))) {
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_terminate_rport_io dev rst not pending sc 0x%p\n",
sc);
spin_unlock_irqrestore(io_lock, flags);
continue;
}
/*
* Found IO that is still pending with firmware and
* belongs to rport that went away
*/
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (io_req->abts_done) {
shost_printk(KERN_ERR, fnic->lport->host,
"fnic_terminate_rport_io: io_req->abts_done is set "
"state is %s\n",
fnic_ioreq_state_to_str(CMD_STATE(sc)));
}
if (!(CMD_FLAGS(sc) & FNIC_IO_ISSUED)) {
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"fnic_terminate_rport_io "
"IO not yet issued %p tag 0x%x flags "
"%x state %d\n",
sc, tag, CMD_FLAGS(sc), CMD_STATE(sc));
}
old_ioreq_state = CMD_STATE(sc);
CMD_STATE(sc) = FNIC_IOREQ_ABTS_PENDING;
CMD_ABTS_STATUS(sc) = FCPIO_INVALID_CODE;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) {
atomic64_inc(&reset_stats->device_reset_terminates);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
abt_tag = (tag | FNIC_TAG_DEV_RST);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_terminate_rport_io dev rst sc 0x%p\n", sc);
}
BUG_ON(io_req->abts_done);
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"fnic_terminate_rport_io: Issuing abts\n");
spin_unlock_irqrestore(io_lock, flags);
/* Now queue the abort command to firmware */
int_to_scsilun(sc->device->lun, &fc_lun);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (fnic_queue_abort_io_req(fnic, abt_tag,
FCPIO_ITMF_ABT_TASK_TERM,
fc_lun.scsi_lun, io_req)) {
/*
* Revert the cmd state back to old state, if
* it hasn't changed in between. This cmd will get
* aborted later by scsi_eh, or cleaned up during
* lun reset
*/
spin_lock_irqsave(io_lock, flags);
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING)
CMD_STATE(sc) = old_ioreq_state;
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
} else {
spin_lock_irqsave(io_lock, flags);
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET)
CMD_FLAGS(sc) |= FNIC_DEV_RST_TERM_ISSUED;
else
CMD_FLAGS(sc) |= FNIC_IO_INTERNAL_TERM_ISSUED;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_unlock_irqrestore(io_lock, flags);
atomic64_inc(&term_stats->terminates);
term_cnt++;
}
}
if (term_cnt > atomic64_read(&term_stats->max_terminates))
atomic64_set(&term_stats->max_terminates, term_cnt);
}
/*
* This function is exported to SCSI for sending abort cmnds.
* A SCSI IO is represented by a io_req in the driver.
* The ioreq is linked to the SCSI Cmd, thus a link with the ULP's IO.
*/
int fnic_abort_cmd(struct scsi_cmnd *sc)
{
struct fc_lport *lp;
struct fnic *fnic;
struct fnic_io_req *io_req = NULL;
struct fc_rport *rport;
spinlock_t *io_lock;
unsigned long flags;
unsigned long start_time = 0;
int ret = SUCCESS;
u32 task_req = 0;
struct scsi_lun fc_lun;
struct fnic_stats *fnic_stats;
struct abort_stats *abts_stats;
struct terminate_stats *term_stats;
enum fnic_ioreq_state old_ioreq_state;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
int tag;
DECLARE_COMPLETION_ONSTACK(tm_done);
/* Wait for rport to unblock */
fc_block_scsi_eh(sc);
/* Get local-port, check ready and link up */
lp = shost_priv(sc->device->host);
fnic = lport_priv(lp);
fnic_stats = &fnic->fnic_stats;
abts_stats = &fnic->fnic_stats.abts_stats;
term_stats = &fnic->fnic_stats.term_stats;
rport = starget_to_rport(scsi_target(sc->device));
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
tag = sc->request->tag;
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"Abort Cmd called FCID 0x%x, LUN 0x%llx TAG %x flags %x\n",
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
rport->port_id, sc->device->lun, tag, CMD_FLAGS(sc));
CMD_FLAGS(sc) = FNIC_NO_FLAGS;
if (lp->state != LPORT_ST_READY || !(lp->link_up)) {
ret = FAILED;
goto fnic_abort_cmd_end;
}
/*
* Avoid a race between SCSI issuing the abort and the device
* completing the command.
*
* If the command is already completed by the fw cmpl code,
* we just return SUCCESS from here. This means that the abort
* succeeded. In the SCSI ML, since the timeout for command has
* happened, the completion wont actually complete the command
* and it will be considered as an aborted command
*
* The CMD_SP will not be cleared except while holding io_req_lock.
*/
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
goto fnic_abort_cmd_end;
}
io_req->abts_done = &tm_done;
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
goto wait_pending;
}
/*
* Command is still pending, need to abort it
* If the firmware completes the command after this point,
* the completion wont be done till mid-layer, since abort
* has already started.
*/
old_ioreq_state = CMD_STATE(sc);
CMD_STATE(sc) = FNIC_IOREQ_ABTS_PENDING;
CMD_ABTS_STATUS(sc) = FCPIO_INVALID_CODE;
spin_unlock_irqrestore(io_lock, flags);
/*
* Check readiness of the remote port. If the path to remote
* port is up, then send abts to the remote port to terminate
* the IO. Else, just locally terminate the IO in the firmware
*/
if (fc_remote_port_chkready(rport) == 0)
task_req = FCPIO_ITMF_ABT_TASK;
else {
atomic64_inc(&fnic_stats->misc_stats.rport_not_ready);
task_req = FCPIO_ITMF_ABT_TASK_TERM;
}
/* Now queue the abort command to firmware */
int_to_scsilun(sc->device->lun, &fc_lun);
if (fnic_queue_abort_io_req(fnic, sc->request->tag, task_req,
fc_lun.scsi_lun, io_req)) {
spin_lock_irqsave(io_lock, flags);
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING)
CMD_STATE(sc) = old_ioreq_state;
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (io_req)
io_req->abts_done = NULL;
spin_unlock_irqrestore(io_lock, flags);
ret = FAILED;
goto fnic_abort_cmd_end;
}
if (task_req == FCPIO_ITMF_ABT_TASK) {
CMD_FLAGS(sc) |= FNIC_IO_ABTS_ISSUED;
atomic64_inc(&fnic_stats->abts_stats.aborts);
} else {
CMD_FLAGS(sc) |= FNIC_IO_TERM_ISSUED;
atomic64_inc(&fnic_stats->term_stats.terminates);
}
/*
* We queued an abort IO, wait for its completion.
* Once the firmware completes the abort command, it will
* wake up this thread.
*/
wait_pending:
wait_for_completion_timeout(&tm_done,
msecs_to_jiffies
(2 * fnic->config.ra_tov +
fnic->config.ed_tov));
/* Check the abort status */
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req) {
atomic64_inc(&fnic_stats->io_stats.ioreq_null);
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_REQ_NULL;
ret = FAILED;
goto fnic_abort_cmd_end;
}
io_req->abts_done = NULL;
/* fw did not complete abort, timed out */
if (CMD_ABTS_STATUS(sc) == FCPIO_INVALID_CODE) {
spin_unlock_irqrestore(io_lock, flags);
if (task_req == FCPIO_ITMF_ABT_TASK) {
atomic64_inc(&abts_stats->abort_drv_timeouts);
} else {
atomic64_inc(&term_stats->terminate_drv_timeouts);
}
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_TIMED_OUT;
ret = FAILED;
goto fnic_abort_cmd_end;
}
/* IO out of order */
if (!(CMD_FLAGS(sc) & (FNIC_IO_ABORTED | FNIC_IO_DONE))) {
spin_unlock_irqrestore(io_lock, flags);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Issuing Host reset due to out of order IO\n");
if (fnic_host_reset(sc) == FAILED) {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_host_reset failed.\n");
}
ret = FAILED;
goto fnic_abort_cmd_end;
}
CMD_STATE(sc) = FNIC_IOREQ_ABTS_COMPLETE;
start_time = io_req->start_time;
/*
* firmware completed the abort, check the status,
* free the io_req if successful. If abort fails,
* Device reset will clean the I/O.
*/
if (CMD_ABTS_STATUS(sc) == FCPIO_SUCCESS)
CMD_SP(sc) = NULL;
else {
ret = FAILED;
spin_unlock_irqrestore(io_lock, flags);
goto fnic_abort_cmd_end;
}
spin_unlock_irqrestore(io_lock, flags);
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
if (sc->scsi_done) {
/* Call SCSI completion function to complete the IO */
sc->result = (DID_ABORT << 16);
sc->scsi_done(sc);
}
fnic_abort_cmd_end:
FNIC_TRACE(fnic_abort_cmd, sc->device->host->host_no,
sc->request->tag, sc,
jiffies_to_msecs(jiffies - start_time),
0, ((u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 | (u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5]),
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Returning from abort cmd type %x %s\n", task_req,
(ret == SUCCESS) ?
"SUCCESS" : "FAILED");
return ret;
}
static inline int fnic_queue_dr_io_req(struct fnic *fnic,
struct scsi_cmnd *sc,
struct fnic_io_req *io_req)
{
struct vnic_wq_copy *wq = &fnic->wq_copy[0];
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
struct Scsi_Host *host = fnic->lport->host;
struct misc_stats *misc_stats = &fnic->fnic_stats.misc_stats;
struct scsi_lun fc_lun;
int ret = 0;
unsigned long intr_flags;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_lock_irqsave(host->host_lock, intr_flags);
if (unlikely(fnic_chk_state_flags_locked(fnic,
FNIC_FLAGS_IO_BLOCKED))) {
spin_unlock_irqrestore(host->host_lock, intr_flags);
return FAILED;
} else
atomic_inc(&fnic->in_flight);
spin_unlock_irqrestore(host->host_lock, intr_flags);
spin_lock_irqsave(&fnic->wq_copy_lock[0], intr_flags);
if (vnic_wq_copy_desc_avail(wq) <= fnic->wq_copy_desc_low[0])
free_wq_copy_descs(fnic, wq);
if (!vnic_wq_copy_desc_avail(wq)) {
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"queue_dr_io_req failure - no descriptors\n");
atomic64_inc(&misc_stats->devrst_cpwq_alloc_failures);
ret = -EAGAIN;
goto lr_io_req_end;
}
/* fill in the lun info */
int_to_scsilun(sc->device->lun, &fc_lun);
fnic_queue_wq_copy_desc_itmf(wq, sc->request->tag | FNIC_TAG_DEV_RST,
0, FCPIO_ITMF_LUN_RESET, SCSI_NO_TAG,
fc_lun.scsi_lun, io_req->port_id,
fnic->config.ra_tov, fnic->config.ed_tov);
atomic64_inc(&fnic->fnic_stats.fw_stats.active_fw_reqs);
if (atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs) >
atomic64_read(&fnic->fnic_stats.fw_stats.max_fw_reqs))
atomic64_set(&fnic->fnic_stats.fw_stats.max_fw_reqs,
atomic64_read(&fnic->fnic_stats.fw_stats.active_fw_reqs));
lr_io_req_end:
spin_unlock_irqrestore(&fnic->wq_copy_lock[0], intr_flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
atomic_dec(&fnic->in_flight);
return ret;
}
/*
* Clean up any pending aborts on the lun
* For each outstanding IO on this lun, whose abort is not completed by fw,
* issue a local abort. Wait for abort to complete. Return 0 if all commands
* successfully aborted, 1 otherwise
*/
static int fnic_clean_pending_aborts(struct fnic *fnic,
struct scsi_cmnd *lr_sc,
bool new_sc)
{
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
int tag, abt_tag;
struct fnic_io_req *io_req;
spinlock_t *io_lock;
unsigned long flags;
int ret = 0;
struct scsi_cmnd *sc;
struct scsi_lun fc_lun;
struct scsi_device *lun_dev = lr_sc->device;
DECLARE_COMPLETION_ONSTACK(tm_done);
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
enum fnic_ioreq_state old_ioreq_state;
for (tag = 0; tag < fnic->fnic_max_tag_id; tag++) {
io_lock = fnic_io_lock_tag(fnic, tag);
spin_lock_irqsave(io_lock, flags);
sc = scsi_host_find_tag(fnic->lport->host, tag);
/*
* ignore this lun reset cmd if issued using new SC
* or cmds that do not belong to this lun
*/
if (!sc || ((sc == lr_sc) && new_sc) || sc->device != lun_dev) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req || sc->device != lun_dev) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
/*
* Found IO that is still pending with firmware and
* belongs to the LUN that we are resetting
*/
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Found IO in %s on lun\n",
fnic_ioreq_state_to_str(CMD_STATE(sc)));
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
if ((CMD_FLAGS(sc) & FNIC_DEVICE_RESET) &&
(!(CMD_FLAGS(sc) & FNIC_DEV_RST_ISSUED))) {
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"%s dev rst not pending sc 0x%p\n", __func__,
sc);
spin_unlock_irqrestore(io_lock, flags);
continue;
}
if (io_req->abts_done)
shost_printk(KERN_ERR, fnic->lport->host,
"%s: io_req->abts_done is set state is %s\n",
__func__, fnic_ioreq_state_to_str(CMD_STATE(sc)));
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
old_ioreq_state = CMD_STATE(sc);
/*
* Any pending IO issued prior to reset is expected to be
* in abts pending state, if not we need to set
* FNIC_IOREQ_ABTS_PENDING to indicate the IO is abort pending.
* When IO is completed, the IO will be handed over and
* handled in this function.
*/
CMD_STATE(sc) = FNIC_IOREQ_ABTS_PENDING;
BUG_ON(io_req->abts_done);
abt_tag = tag;
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) {
abt_tag |= FNIC_TAG_DEV_RST;
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"%s: dev rst sc 0x%p\n", __func__, sc);
}
CMD_ABTS_STATUS(sc) = FCPIO_INVALID_CODE;
io_req->abts_done = &tm_done;
spin_unlock_irqrestore(io_lock, flags);
/* Now queue the abort command to firmware */
int_to_scsilun(sc->device->lun, &fc_lun);
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
if (fnic_queue_abort_io_req(fnic, abt_tag,
FCPIO_ITMF_ABT_TASK_TERM,
fc_lun.scsi_lun, io_req)) {
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (io_req)
io_req->abts_done = NULL;
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING)
CMD_STATE(sc) = old_ioreq_state;
spin_unlock_irqrestore(io_lock, flags);
ret = 1;
goto clean_pending_aborts_end;
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
} else {
spin_lock_irqsave(io_lock, flags);
if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET)
CMD_FLAGS(sc) |= FNIC_DEV_RST_TERM_ISSUED;
spin_unlock_irqrestore(io_lock, flags);
}
CMD_FLAGS(sc) |= FNIC_IO_INTERNAL_TERM_ISSUED;
wait_for_completion_timeout(&tm_done,
msecs_to_jiffies
(fnic->config.ed_tov));
/* Recheck cmd state to check if it is now aborted */
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_REQ_NULL;
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
continue;
}
io_req->abts_done = NULL;
/* if abort is still pending with fw, fail */
if (CMD_ABTS_STATUS(sc) == FCPIO_INVALID_CODE) {
spin_unlock_irqrestore(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_IO_ABT_TERM_DONE;
ret = 1;
goto clean_pending_aborts_end;
}
CMD_STATE(sc) = FNIC_IOREQ_ABTS_COMPLETE;
/* original sc used for lr is handled by dev reset code */
if (sc != lr_sc)
CMD_SP(sc) = NULL;
spin_unlock_irqrestore(io_lock, flags);
/* original sc used for lr is handled by dev reset code */
if (sc != lr_sc) {
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
}
/*
* Any IO is returned during reset, it needs to call scsi_done
* to return the scsi_cmnd to upper layer.
*/
if (sc->scsi_done) {
/* Set result to let upper SCSI layer retry */
sc->result = DID_RESET << 16;
sc->scsi_done(sc);
}
}
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
schedule_timeout(msecs_to_jiffies(2 * fnic->config.ed_tov));
/* walk again to check, if IOs are still pending in fw */
if (fnic_is_abts_pending(fnic, lr_sc))
ret = FAILED;
clean_pending_aborts_end:
return ret;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/**
* fnic_scsi_host_start_tag
* Allocates tagid from host's tag list
**/
static inline int
fnic_scsi_host_start_tag(struct fnic *fnic, struct scsi_cmnd *sc)
{
struct blk_queue_tag *bqt = fnic->lport->host->bqt;
int tag, ret = SCSI_NO_TAG;
BUG_ON(!bqt);
if (!bqt) {
pr_err("Tags are not supported\n");
goto end;
}
do {
tag = find_next_zero_bit(bqt->tag_map, bqt->max_depth, 1);
if (tag >= bqt->max_depth) {
pr_err("Tag allocation failure\n");
goto end;
}
} while (test_and_set_bit(tag, bqt->tag_map));
bqt->tag_index[tag] = sc->request;
sc->request->tag = tag;
sc->tag = tag;
if (!sc->request->special)
sc->request->special = sc;
ret = tag;
end:
return ret;
}
/**
* fnic_scsi_host_end_tag
* frees tag allocated by fnic_scsi_host_start_tag.
**/
static inline void
fnic_scsi_host_end_tag(struct fnic *fnic, struct scsi_cmnd *sc)
{
struct blk_queue_tag *bqt = fnic->lport->host->bqt;
int tag = sc->request->tag;
if (tag == SCSI_NO_TAG)
return;
BUG_ON(!bqt || !bqt->tag_index[tag]);
if (!bqt)
return;
bqt->tag_index[tag] = NULL;
clear_bit(tag, bqt->tag_map);
return;
}
/*
* SCSI Eh thread issues a Lun Reset when one or more commands on a LUN
* fail to get aborted. It calls driver's eh_device_reset with a SCSI command
* on the LUN.
*/
int fnic_device_reset(struct scsi_cmnd *sc)
{
struct fc_lport *lp;
struct fnic *fnic;
struct fnic_io_req *io_req = NULL;
struct fc_rport *rport;
int status;
int ret = FAILED;
spinlock_t *io_lock;
unsigned long flags;
unsigned long start_time = 0;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
struct scsi_lun fc_lun;
struct fnic_stats *fnic_stats;
struct reset_stats *reset_stats;
int tag = 0;
DECLARE_COMPLETION_ONSTACK(tm_done);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
int tag_gen_flag = 0; /*to track tags allocated by fnic driver*/
bool new_sc = 0;
/* Wait for rport to unblock */
fc_block_scsi_eh(sc);
/* Get local-port, check ready and link up */
lp = shost_priv(sc->device->host);
fnic = lport_priv(lp);
fnic_stats = &fnic->fnic_stats;
reset_stats = &fnic->fnic_stats.reset_stats;
atomic64_inc(&reset_stats->device_resets);
rport = starget_to_rport(scsi_target(sc->device));
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Device reset called FCID 0x%x, LUN 0x%llx sc 0x%p\n",
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
rport->port_id, sc->device->lun, sc);
if (lp->state != LPORT_ST_READY || !(lp->link_up))
goto fnic_device_reset_end;
/* Check if remote port up */
if (fc_remote_port_chkready(rport)) {
atomic64_inc(&fnic_stats->misc_stats.rport_not_ready);
goto fnic_device_reset_end;
}
CMD_FLAGS(sc) = FNIC_DEVICE_RESET;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/* Allocate tag if not present */
tag = sc->request->tag;
if (unlikely(tag < 0)) {
/*
* XXX(hch): current the midlayer fakes up a struct
* request for the explicit reset ioctls, and those
* don't have a tag allocated to them. The below
* code pokes into midlayer structures to paper over
* this design issue, but that won't work for blk-mq.
*
* Either someone who can actually test the hardware
* will have to come up with a similar hack for the
* blk-mq case, or we'll have to bite the bullet and
* fix the way the EH ioctls work for real, but until
* that happens we fail these explicit requests here.
*/
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
tag = fnic_scsi_host_start_tag(fnic, sc);
if (unlikely(tag == SCSI_NO_TAG))
goto fnic_device_reset_end;
tag_gen_flag = 1;
new_sc = 1;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
}
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
/*
* If there is a io_req attached to this command, then use it,
* else allocate a new one.
*/
if (!io_req) {
io_req = mempool_alloc(fnic->io_req_pool, GFP_ATOMIC);
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
goto fnic_device_reset_end;
}
memset(io_req, 0, sizeof(*io_req));
io_req->port_id = rport->port_id;
CMD_SP(sc) = (char *)io_req;
}
io_req->dr_done = &tm_done;
CMD_STATE(sc) = FNIC_IOREQ_CMD_PENDING;
CMD_LR_STATUS(sc) = FCPIO_INVALID_CODE;
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host, "TAG %x\n", tag);
/*
* issue the device reset, if enqueue failed, clean up the ioreq
* and break assoc with scsi cmd
*/
if (fnic_queue_dr_io_req(fnic, sc, io_req)) {
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (io_req)
io_req->dr_done = NULL;
goto fnic_device_reset_clean;
}
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_lock_irqsave(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_DEV_RST_ISSUED;
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
spin_unlock_irqrestore(io_lock, flags);
/*
* Wait on the local completion for LUN reset. The io_req may be
* freed while we wait since we hold no lock.
*/
wait_for_completion_timeout(&tm_done,
msecs_to_jiffies(FNIC_LUN_RESET_TIMEOUT));
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req) {
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"io_req is null tag 0x%x sc 0x%p\n", tag, sc);
goto fnic_device_reset_end;
}
io_req->dr_done = NULL;
status = CMD_LR_STATUS(sc);
/*
* If lun reset not completed, bail out with failed. io_req
* gets cleaned up during higher levels of EH
*/
if (status == FCPIO_INVALID_CODE) {
atomic64_inc(&reset_stats->device_reset_timeouts);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Device reset timed out\n");
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
CMD_FLAGS(sc) |= FNIC_DEV_RST_TIMED_OUT;
spin_unlock_irqrestore(io_lock, flags);
int_to_scsilun(sc->device->lun, &fc_lun);
/*
* Issue abort and terminate on device reset request.
* If q'ing of terminate fails, retry it after a delay.
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
*/
while (1) {
spin_lock_irqsave(io_lock, flags);
if (CMD_FLAGS(sc) & FNIC_DEV_RST_TERM_ISSUED) {
spin_unlock_irqrestore(io_lock, flags);
break;
}
spin_unlock_irqrestore(io_lock, flags);
if (fnic_queue_abort_io_req(fnic,
tag | FNIC_TAG_DEV_RST,
FCPIO_ITMF_ABT_TASK_TERM,
fc_lun.scsi_lun, io_req)) {
wait_for_completion_timeout(&tm_done,
msecs_to_jiffies(FNIC_ABT_TERM_DELAY_TIMEOUT));
} else {
spin_lock_irqsave(io_lock, flags);
CMD_FLAGS(sc) |= FNIC_DEV_RST_TERM_ISSUED;
CMD_STATE(sc) = FNIC_IOREQ_ABTS_PENDING;
io_req->abts_done = &tm_done;
spin_unlock_irqrestore(io_lock, flags);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Abort and terminate issued on Device reset "
"tag 0x%x sc 0x%p\n", tag, sc);
break;
}
}
while (1) {
spin_lock_irqsave(io_lock, flags);
if (!(CMD_FLAGS(sc) & FNIC_DEV_RST_DONE)) {
spin_unlock_irqrestore(io_lock, flags);
wait_for_completion_timeout(&tm_done,
msecs_to_jiffies(FNIC_LUN_RESET_TIMEOUT));
break;
} else {
io_req = (struct fnic_io_req *)CMD_SP(sc);
io_req->abts_done = NULL;
goto fnic_device_reset_clean;
}
}
} else {
spin_unlock_irqrestore(io_lock, flags);
}
/* Completed, but not successful, clean up the io_req, return fail */
if (status != FCPIO_SUCCESS) {
spin_lock_irqsave(io_lock, flags);
FNIC_SCSI_DBG(KERN_DEBUG,
fnic->lport->host,
"Device reset completed - failed\n");
io_req = (struct fnic_io_req *)CMD_SP(sc);
goto fnic_device_reset_clean;
}
/*
* Clean up any aborts on this lun that have still not
* completed. If any of these fail, then LUN reset fails.
* clean_pending_aborts cleans all cmds on this lun except
* the lun reset cmd. If all cmds get cleaned, the lun reset
* succeeds
*/
if (fnic_clean_pending_aborts(fnic, sc, new_sc)) {
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Device reset failed"
" since could not abort all IOs\n");
goto fnic_device_reset_clean;
}
/* Clean lun reset command */
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (io_req)
/* Completed, and successful */
ret = SUCCESS;
fnic_device_reset_clean:
if (io_req)
CMD_SP(sc) = NULL;
spin_unlock_irqrestore(io_lock, flags);
if (io_req) {
start_time = io_req->start_time;
fnic_release_ioreq_buf(fnic, io_req, sc);
mempool_free(io_req, fnic->io_req_pool);
}
fnic_device_reset_end:
FNIC_TRACE(fnic_device_reset, sc->device->host->host_no,
sc->request->tag, sc,
jiffies_to_msecs(jiffies - start_time),
0, ((u64)sc->cmnd[0] << 32 |
(u64)sc->cmnd[2] << 24 | (u64)sc->cmnd[3] << 16 |
(u64)sc->cmnd[4] << 8 | sc->cmnd[5]),
(((u64)CMD_FLAGS(sc) << 32) | CMD_STATE(sc)));
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
/* free tag if it is allocated */
if (unlikely(tag_gen_flag))
fnic_scsi_host_end_tag(fnic, sc);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Returning from device reset %s\n",
(ret == SUCCESS) ?
"SUCCESS" : "FAILED");
if (ret == FAILED)
atomic64_inc(&reset_stats->device_reset_failures);
return ret;
}
/* Clean up all IOs, clean up libFC local port */
int fnic_reset(struct Scsi_Host *shost)
{
struct fc_lport *lp;
struct fnic *fnic;
int ret = 0;
struct reset_stats *reset_stats;
lp = shost_priv(shost);
fnic = lport_priv(lp);
reset_stats = &fnic->fnic_stats.reset_stats;
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_reset called\n");
atomic64_inc(&reset_stats->fnic_resets);
/*
* Reset local port, this will clean up libFC exchanges,
* reset remote port sessions, and if link is up, begin flogi
*/
ret = fc_lport_reset(lp);
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"Returning from fnic reset %s\n",
(ret == 0) ?
"SUCCESS" : "FAILED");
if (ret == 0)
atomic64_inc(&reset_stats->fnic_reset_completions);
else
atomic64_inc(&reset_stats->fnic_reset_failures);
return ret;
}
/*
* SCSI Error handling calls driver's eh_host_reset if all prior
* error handling levels return FAILED. If host reset completes
* successfully, and if link is up, then Fabric login begins.
*
* Host Reset is the highest level of error recovery. If this fails, then
* host is offlined by SCSI.
*
*/
int fnic_host_reset(struct scsi_cmnd *sc)
{
int ret;
unsigned long wait_host_tmo;
struct Scsi_Host *shost = sc->device->host;
struct fc_lport *lp = shost_priv(shost);
/*
* If fnic_reset is successful, wait for fabric login to complete
* scsi-ml tries to send a TUR to every device if host reset is
* successful, so before returning to scsi, fabric should be up
*/
ret = (fnic_reset(shost) == 0) ? SUCCESS : FAILED;
if (ret == SUCCESS) {
wait_host_tmo = jiffies + FNIC_HOST_RESET_SETTLE_TIME * HZ;
ret = FAILED;
while (time_before(jiffies, wait_host_tmo)) {
if ((lp->state == LPORT_ST_READY) &&
(lp->link_up)) {
ret = SUCCESS;
break;
}
ssleep(1);
}
}
return ret;
}
/*
* This fxn is called from libFC when host is removed
*/
void fnic_scsi_abort_io(struct fc_lport *lp)
{
int err = 0;
unsigned long flags;
enum fnic_state old_state;
struct fnic *fnic = lport_priv(lp);
DECLARE_COMPLETION_ONSTACK(remove_wait);
/* Issue firmware reset for fnic, wait for reset to complete */
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
retry_fw_reset:
spin_lock_irqsave(&fnic->fnic_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (unlikely(fnic->state == FNIC_IN_FC_TRANS_ETH_MODE)) {
/* fw reset is in progress, poll for its completion */
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
schedule_timeout(msecs_to_jiffies(100));
goto retry_fw_reset;
}
fnic->remove_wait = &remove_wait;
old_state = fnic->state;
fnic->state = FNIC_IN_FC_TRANS_ETH_MODE;
fnic_update_mac_locked(fnic, fnic->ctlr.ctl_src_addr);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
err = fnic_fw_reset_handler(fnic);
if (err) {
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->state == FNIC_IN_FC_TRANS_ETH_MODE)
fnic->state = old_state;
fnic->remove_wait = NULL;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
return;
}
/* Wait for firmware reset to complete */
wait_for_completion_timeout(&remove_wait,
msecs_to_jiffies(FNIC_RMDEVICE_TIMEOUT));
spin_lock_irqsave(&fnic->fnic_lock, flags);
fnic->remove_wait = NULL;
FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host,
"fnic_scsi_abort_io %s\n",
(fnic->state == FNIC_IN_ETH_MODE) ?
"SUCCESS" : "FAILED");
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
/*
* This fxn called from libFC to clean up driver IO state on link down
*/
void fnic_scsi_cleanup(struct fc_lport *lp)
{
unsigned long flags;
enum fnic_state old_state;
struct fnic *fnic = lport_priv(lp);
/* issue fw reset */
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
retry_fw_reset:
spin_lock_irqsave(&fnic->fnic_lock, flags);
[SCSI] fnic: fixing issues in device and firmware reset code 1. Handling overlapped firmware resets This fix serialize multiple firmware resets to avoid situation where fnic device fails to come up for link up event, when firmware resets are issued back to back. If there are overlapped firmware resets are issued, the firmware reset operation checks whether there is any firmware reset in progress, if so it polls for its completion in a loop with 100ms delay. 2. Handling device reset timeout fnic_device_reset code has been modified to handle Device reset timeout: - Issue terminate on device reset timeout. - Introduced flags field (one of the scratch fields in scsi_cmnd). With this, device reset request would have DEVICE_RESET flag set for other routines to determine the type of the request. Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion routines to handle SCSI commands with DEVICE_RESET flag. 3. LUN/Device Reset hangs when issued through IOCTL using utilities like sg_reset. Each SCSI command is associated with a valid tag, fnic uses this tag to retrieve associated scsi command on completion. the LUN/Device Reset issued through IOCTL resulting into a SCSI command that is not associated with a valid tag. So fnic fails to retrieve associated scsi command on completion, which causes hang. This fix allocates tag, associates it with the scsi command and frees the tag, when the operation completed. 4. Preventing IOs during firmware reset. Current fnic implementation allows IO submissions during firmware reset. This fix synchronizes IO submissions and firmware reset operations. It ensures that IOs issued to fnic prior to reset will be issued to the firmware before firmware reset. Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:00:58 +08:00
if (unlikely(fnic->state == FNIC_IN_FC_TRANS_ETH_MODE)) {
/* fw reset is in progress, poll for its completion */
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
schedule_timeout(msecs_to_jiffies(100));
goto retry_fw_reset;
}
old_state = fnic->state;
fnic->state = FNIC_IN_FC_TRANS_ETH_MODE;
fnic_update_mac_locked(fnic, fnic->ctlr.ctl_src_addr);
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
if (fnic_fw_reset_handler(fnic)) {
spin_lock_irqsave(&fnic->fnic_lock, flags);
if (fnic->state == FNIC_IN_FC_TRANS_ETH_MODE)
fnic->state = old_state;
spin_unlock_irqrestore(&fnic->fnic_lock, flags);
}
}
void fnic_empty_scsi_cleanup(struct fc_lport *lp)
{
}
void fnic_exch_mgr_reset(struct fc_lport *lp, u32 sid, u32 did)
{
struct fnic *fnic = lport_priv(lp);
/* Non-zero sid, nothing to do */
if (sid)
goto call_fc_exch_mgr_reset;
if (did) {
fnic_rport_exch_reset(fnic, did);
goto call_fc_exch_mgr_reset;
}
/*
* sid = 0, did = 0
* link down or device being removed
*/
if (!fnic->in_remove)
fnic_scsi_cleanup(lp);
else
fnic_scsi_abort_io(lp);
/* call libFC exch mgr reset to reset its exchanges */
call_fc_exch_mgr_reset:
fc_exch_mgr_reset(lp, sid, did);
}
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
/*
* fnic_is_abts_pending() is a helper function that
* walks through tag map to check if there is any IOs pending,if there is one,
* then it returns 1 (true), otherwise 0 (false)
* if @lr_sc is non NULL, then it checks IOs specific to particular LUN,
* otherwise, it checks for all IOs.
*/
int fnic_is_abts_pending(struct fnic *fnic, struct scsi_cmnd *lr_sc)
{
int tag;
struct fnic_io_req *io_req;
spinlock_t *io_lock;
unsigned long flags;
int ret = 0;
struct scsi_cmnd *sc;
struct scsi_device *lun_dev = NULL;
if (lr_sc)
lun_dev = lr_sc->device;
/* walk again to check, if IOs are still pending in fw */
for (tag = 0; tag < fnic->fnic_max_tag_id; tag++) {
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
sc = scsi_host_find_tag(fnic->lport->host, tag);
/*
* ignore this lun reset cmd or cmds that do not belong to
* this lun
*/
if (!sc || (lr_sc && (sc->device != lun_dev || sc == lr_sc)))
continue;
io_lock = fnic_io_lock_hash(fnic, sc);
spin_lock_irqsave(io_lock, flags);
io_req = (struct fnic_io_req *)CMD_SP(sc);
if (!io_req || sc->device != lun_dev) {
spin_unlock_irqrestore(io_lock, flags);
continue;
}
/*
* Found IO that is still pending with firmware and
* belongs to the LUN that we are resetting
*/
FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host,
"Found IO in %s on lun\n",
fnic_ioreq_state_to_str(CMD_STATE(sc)));
if (CMD_STATE(sc) == FNIC_IOREQ_ABTS_PENDING)
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
ret = 1;
spin_unlock_irqrestore(io_lock, flags);
[SCSI] fnic: fnic driver may hit BUG_ON on device reset The issue was observed when LUN Reset is issued through IOCTL or sg_reset utility. fnic driver issues LUN RESET to firmware. On successful completion of device reset, driver cleans up all the pending IOs that were issued prior to device reset. These pending IOs are expected to be in ABTS_PENDING state. This works fine, when the device reset operation resulted from midlayer, but not when device reset was triggered from IOCTL path as the pending IOs were not in ABTS_PENDING state. execution path hits panic if the pending IO is not in ABTS_PENDING state. Changes: The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips if they were already in ABTS_PENDING state. An extra check is added to validate the abort status of the commands after a delay of 2 * E_D_TOV using a helper function. The helper function returns 1 if it finds any pending IO in ABTS_PENDING state, belong to the LUN on which device reset was issued else 0. With this, device reset operation returns success only if the helper funciton returns 0, otherwise it returns failure. Other changes: - Removed code in fnic_clean_pending_aborts() that returns failure if it finds io_req NULL, instead of returning failure added code to continue with next io - Added device reset flags for debugging in fnic_terminate_rport_io, fnic_rport_exch_reset, and fnic_clean_pending_aborts Signed-off-by: Narsimhulu Musini <nmusini@cisco.com> Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
2013-02-13 09:01:00 +08:00
}
return ret;
}