iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

scsi: mpt3sas: Handle firmware faults during second half of IOC init

Browse Source 
If a firmware fault occurs while scanning the devices during IOC
initialization then the driver issues the hard reset operation to recover
the IOC. However, the driver is not issuing a Port enable request
message as part of hard reset operation during IOC initialization.  Due to
this, the driver will not receive get any device discovery-related events
and hence devices will not be accessible.

Teach the driver to gracefully handle firmware faults while scanning for
target devices during IOC initialization. Make the driver issue a port
enable request message as part of hard reset operation. This permits
receiving device discovery-related events from the firmware after the hard
reset operation completes.

Link: https://lore.kernel.org/r/20210518051625.1596742-4-suganath-prabu.subramani@broadcom.com
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This commit is contained in:
Suganath Prabu S 2021-05-18 10:46:25 +05:30 committed by Martin K. Petersen
parent 19a622c39a
commit a0815c45c8
3 changed files with 145 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
drivers/scsi/mpt3sas/mpt3sas_base.c
View File

@ -7205,7 +7205,7 @@ mpt3sas_port_enable_done(struct MPT3SAS_ADAPTER *ioc, u16 smid, u8 msix_index,
if (ioc_status != MPI2_IOCSTATUS_SUCCESS)
ioc->port_enable_failed = 1;
if (ioc->is_driver_loading) {
if (ioc->port_enable_cmds.status & MPT3_CMD_COMPLETE_ASYNC) {
if (ioc_status == MPI2_IOCSTATUS_SUCCESS) {
mpt3sas_port_enable_complete(ioc);
return 1;
@ -7214,6 +7214,7 @@ mpt3sas_port_enable_done(struct MPT3SAS_ADAPTER *ioc, u16 smid, u8 msix_index,
ioc->start_scan = 0;
return 1;
}
ioc->port_enable_cmds.status &= ~MPT3_CMD_COMPLETE_ASYNC;
}
complete(&ioc->port_enable_cmds.done);
return 1;
@ -7308,6 +7309,7 @@ mpt3sas_port_enable(struct MPT3SAS_ADAPTER *ioc)
}
ioc->drv_internal_flags |= MPT_DRV_INTERNAL_FIRST_PE_ISSUED;
ioc->port_enable_cmds.status = MPT3_CMD_PENDING;
ioc->port_enable_cmds.status |= MPT3_CMD_COMPLETE_ASYNC;
mpi_request = mpt3sas_base_get_msg_frame(ioc, smid);
ioc->port_enable_cmds.smid = smid;
memset(mpi_request, 0, sizeof(Mpi2PortEnableRequest_t));
@ -7856,7 +7858,7 @@ _base_make_ioc_operational(struct MPT3SAS_ADAPTER *ioc)
if (r)
return r;
if (ioc->is_driver_loading) {
if (!ioc->shost_recovery) {
if (ioc->is_warpdrive && ioc->manu_pg10.OEMIdentifier
== 0x80) {
@ -8276,8 +8278,6 @@ _base_clear_outstanding_mpt_commands(struct MPT3SAS_ADAPTER *ioc)
ioc->start_scan_failed =
MPI2_IOCSTATUS_INTERNAL_ERROR;
ioc->start_scan = 0;
ioc->port_enable_cmds.status =
MPT3_CMD_NOT_USED;
} else {
complete(&ioc->port_enable_cmds.done);
}

1
drivers/scsi/mpt3sas/mpt3sas_base.h
View File

@ -500,6 +500,7 @@ struct MPT3SAS_DEVICE {
#define MPT3_CMD_PENDING 0x0002 /* pending */
#define MPT3_CMD_REPLY_VALID 0x0004 /* reply is valid */
#define MPT3_CMD_RESET 0x0008 /* host reset dropped the command */
#define MPT3_CMD_COMPLETE_ASYNC 0x0010 /* tells whether cmd completes in same thread or not */
/**
* struct _internal_cmd - internal commands struct

152
drivers/scsi/mpt3sas/mpt3sas_scsih.c
View File

@ -78,6 +78,7 @@ static void _scsih_pcie_device_remove_from_sml(struct MPT3SAS_ADAPTER *ioc,
static void
_scsih_pcie_check_device(struct MPT3SAS_ADAPTER *ioc, u16 handle);
static u8 _scsih_check_for_pending_tm(struct MPT3SAS_ADAPTER *ioc, u16 smid);
static void _scsih_complete_devices_scanning(struct MPT3SAS_ADAPTER *ioc);
/* global parameters */
LIST_HEAD(mpt3sas_ioc_list);
@ -3631,8 +3632,6 @@ _scsih_error_recovery_delete_devices(struct MPT3SAS_ADAPTER *ioc)
{
struct fw_event_work *fw_event;
if (ioc->is_driver_loading)
return;
fw_event = alloc_fw_event_work(0);
if (!fw_event)
return;
@ -3693,6 +3692,14 @@ _scsih_fw_event_cleanup_queue(struct MPT3SAS_ADAPTER *ioc)
if ((list_empty(&ioc->fw_event_list) && !ioc->current_event) ||
!ioc->firmware_event_thread)
return;
/*
* Set current running event as ignore, so that
* current running event will exit quickly.
* As diag reset has occurred it is of no use
* to process remaining stale event data entries.
*/
if (ioc->shost_recovery && ioc->current_event)
ioc->current_event->ignore = 1;
ioc->fw_events_cleanup = 1;
while ((fw_event = dequeue_next_fw_event(ioc)) ||
@ -3719,6 +3726,19 @@ _scsih_fw_event_cleanup_queue(struct MPT3SAS_ADAPTER *ioc)
continue;
}
/*
* Driver has to clear ioc->start_scan flag when
* it is cleaning up MPT3SAS_PORT_ENABLE_COMPLETE,
* otherwise scsi_scan_host() API waits for the
* 5 minute timer to expire. If we exit from
* scsi_scan_host() early then we can issue the
* new port enable request as part of current diag reset.
*/
if (fw_event->event == MPT3SAS_PORT_ENABLE_COMPLETE) {
ioc->port_enable_cmds.status |= MPT3_CMD_RESET;
ioc->start_scan = 0;
}
/*
* Wait on the fw_event to complete. If this returns 1, then
* the event was never executed, and we need a put for the
@ -10140,6 +10160,17 @@ _scsih_remove_unresponding_devices(struct MPT3SAS_ADAPTER *ioc)
* owner for the reference the list had on any object we prune.
*/
spin_lock_irqsave(&ioc->sas_device_lock, flags);
/*
* Clean up the sas_device_init_list list as
* driver goes for fresh scan as part of diag reset.
*/
list_for_each_entry_safe(sas_device, sas_device_next,
&ioc->sas_device_init_list, list) {
list_del_init(&sas_device->list);
sas_device_put(sas_device);
}
list_for_each_entry_safe(sas_device, sas_device_next,
&ioc->sas_device_list, list) {
if (!sas_device->responding)
@ -10161,6 +10192,16 @@ _scsih_remove_unresponding_devices(struct MPT3SAS_ADAPTER *ioc)
ioc_info(ioc, "Removing unresponding devices: pcie end-devices\n");
INIT_LIST_HEAD(&head);
spin_lock_irqsave(&ioc->pcie_device_lock, flags);
/*
* Clean up the pcie_device_init_list list as
* driver goes for fresh scan as part of diag reset.
*/
list_for_each_entry_safe(pcie_device, pcie_device_next,
&ioc->pcie_device_init_list, list) {
list_del_init(&pcie_device->list);
pcie_device_put(pcie_device);
}
list_for_each_entry_safe(pcie_device, pcie_device_next,
&ioc->pcie_device_list, list) {
if (!pcie_device->responding)
@ -10563,8 +10604,7 @@ void
mpt3sas_scsih_reset_done_handler(struct MPT3SAS_ADAPTER *ioc)
{
dtmprintk(ioc, ioc_info(ioc, "%s: MPT3_IOC_DONE_RESET\n", __func__));
if ((!ioc->is_driver_loading) && !(disable_discovery > 0 &&
!ioc->sas_hba.num_phys)) {
if (!(disable_discovery > 0 && !ioc->sas_hba.num_phys)) {
if (ioc->multipath_on_hba) {
_scsih_sas_port_refresh(ioc);
_scsih_update_vphys_after_reset(ioc);
@ -10619,6 +10659,18 @@ _mpt3sas_fw_work(struct MPT3SAS_ADAPTER *ioc, struct fw_event_work *fw_event)
_scsih_del_dirty_vphy(ioc);
_scsih_del_dirty_port_entries(ioc);
_scsih_scan_for_devices_after_reset(ioc);
/*
* If diag reset has occurred during the driver load
* then driver has to complete the driver load operation
* by executing the following items:
*- Register the devices from sas_device_init_list to SML
*- clear is_driver_loading flag,
*- start the watchdog thread.
* In happy driver load path, above things are taken care of when
* driver executes scsih_scan_finished().
*/
if (ioc->is_driver_loading)
_scsih_complete_devices_scanning(ioc);
_scsih_set_nvme_max_shutdown_latency(ioc);
break;
case MPT3SAS_PORT_ENABLE_COMPLETE:
@ -10764,11 +10816,23 @@ mpt3sas_scsih_event_callback(struct MPT3SAS_ADAPTER *ioc, u8 msix_index,
_scsih_check_topo_delete_events(ioc,
(Mpi2EventDataSasTopologyChangeList_t *)
mpi_reply->EventData);
/*
* No need to add the topology change list
* event to fw event work queue when
* diag reset is going on. Since during diag
* reset driver scan the devices by reading
* sas device page0's not by processing the
* events.
*/
if (ioc->shost_recovery)
return 1;
break;
case MPI2_EVENT_PCIE_TOPOLOGY_CHANGE_LIST:
_scsih_check_pcie_topo_remove_events(ioc,
(Mpi26EventDataPCIeTopologyChangeList_t *)
mpi_reply->EventData);
if (ioc->shost_recovery)
return 1;
break;
case MPI2_EVENT_IR_CONFIGURATION_CHANGE_LIST:
_scsih_check_ir_config_unhide_events(ioc,
@ -11284,13 +11348,27 @@ _scsih_probe_boot_devices(struct MPT3SAS_ADAPTER *ioc)
if (channel == RAID_CHANNEL) {
raid_device = device;
/*
* If this boot vd is already registered with SML then
* no need to register it again as part of device scanning
* after diag reset during driver load operation.
*/
if (raid_device->starget)
return;
rc = scsi_add_device(ioc->shost, RAID_CHANNEL,
raid_device->id, 0);
if (rc)
_scsih_raid_device_remove(ioc, raid_device);
} else if (channel == PCIE_CHANNEL) {
spin_lock_irqsave(&ioc->pcie_device_lock, flags);
pcie_device = device;
/*
* If this boot NVMe device is already registered with SML then
* no need to register it again as part of device scanning
* after diag reset during driver load operation.
*/
if (pcie_device->starget)
return;
spin_lock_irqsave(&ioc->pcie_device_lock, flags);
tid = pcie_device->id;
list_move_tail(&pcie_device->list, &ioc->pcie_device_list);
spin_unlock_irqrestore(&ioc->pcie_device_lock, flags);
@ -11298,8 +11376,15 @@ _scsih_probe_boot_devices(struct MPT3SAS_ADAPTER *ioc)
if (rc)
_scsih_pcie_device_remove(ioc, pcie_device);
} else {
spin_lock_irqsave(&ioc->sas_device_lock, flags);
sas_device = device;
/*
* If this boot sas/sata device is already registered with SML
* then no need to register it again as part of device scanning
* after diag reset during driver load operation.
*/
if (sas_device->starget)
return;
spin_lock_irqsave(&ioc->sas_device_lock, flags);
handle = sas_device->handle;
sas_address_parent = sas_device->sas_address_parent;
sas_address = sas_device->sas_address;
@ -11597,6 +11682,25 @@ scsih_scan_start(struct Scsi_Host *shost)
ioc_info(ioc, "port enable: FAILED\n");
}
/**
* _scsih_complete_devices_scanning - add the devices to sml and
* complete ioc initialization.
* @ioc: per adapter object
*
* Return nothing.
*/
static void _scsih_complete_devices_scanning(struct MPT3SAS_ADAPTER *ioc)
{
if (ioc->wait_for_discovery_to_complete) {
ioc->wait_for_discovery_to_complete = 0;
_scsih_probe_devices(ioc);
}
mpt3sas_base_start_watchdog(ioc);
ioc->is_driver_loading = 0;
}
/**
* scsih_scan_finished - scsi lld callback for .scan_finished
* @shost: SCSI host pointer
@ -11610,6 +11714,8 @@ static int
scsih_scan_finished(struct Scsi_Host *shost, unsigned long time)
{
struct MPT3SAS_ADAPTER *ioc = shost_priv(shost);
u32 ioc_state;
int issue_hard_reset = 0;
if (disable_discovery > 0) {
ioc->is_driver_loading = 0;
@ -11624,9 +11730,30 @@ scsih_scan_finished(struct Scsi_Host *shost, unsigned long time)
return 1;
}
if (ioc->start_scan)
if (ioc->start_scan) {
ioc_state = mpt3sas_base_get_iocstate(ioc, 0);
if ((ioc_state & MPI2_IOC_STATE_MASK) == MPI2_IOC_STATE_FAULT) {
mpt3sas_print_fault_code(ioc, ioc_state &
MPI2_DOORBELL_DATA_MASK);
issue_hard_reset = 1;
goto out;
} else if ((ioc_state & MPI2_IOC_STATE_MASK) ==
MPI2_IOC_STATE_COREDUMP) {
mpt3sas_base_coredump_info(ioc, ioc_state &
MPI2_DOORBELL_DATA_MASK);
mpt3sas_base_wait_for_coredump_completion(ioc, __func__);
issue_hard_reset = 1;
goto out;
}
return 0;
}
if (ioc->port_enable_cmds.status & MPT3_CMD_RESET) {
ioc_info(ioc,
"port enable: aborted due to diag reset\n");
ioc->port_enable_cmds.status = MPT3_CMD_NOT_USED;
goto out;
}
if (ioc->start_scan_failed) {
ioc_info(ioc, "port enable: FAILED with (ioc_status=0x%08x)\n",
ioc->start_scan_failed);
@ -11638,13 +11765,14 @@ scsih_scan_finished(struct Scsi_Host *shost, unsigned long time)
ioc_info(ioc, "port enable: SUCCESS\n");
ioc->port_enable_cmds.status = MPT3_CMD_NOT_USED;
_scsih_complete_devices_scanning(ioc);
if (ioc->wait_for_discovery_to_complete) {
ioc->wait_for_discovery_to_complete = 0;
_scsih_probe_devices(ioc);
out:
if (issue_hard_reset) {
ioc->port_enable_cmds.status = MPT3_CMD_NOT_USED;
if (mpt3sas_base_hard_reset_handler(ioc, SOFT_RESET))
ioc->is_driver_loading = 0;
}
mpt3sas_base_start_watchdog(ioc);
ioc->is_driver_loading = 0;
return 1;
}