2012-01-26 18:17:21 +00:00
|
|
|
/*-
|
2015-02-24 22:07:42 +00:00
|
|
|
* Copyright (c) 2011-2015 LSI Corp.
|
|
|
|
* Copyright (c) 2013-2015 Avago Technologies
|
2012-01-26 18:17:21 +00:00
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
* are met:
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*
|
2015-02-24 22:07:42 +00:00
|
|
|
* Avago Technologies (LSI) MPT-Fusion Host Adapter FreeBSD
|
2012-01-26 18:17:21 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
2015-02-24 22:07:42 +00:00
|
|
|
/* Communications core for Avago Technologies (LSI) MPT2 */
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
/* TODO Move headers to mpsvar */
|
|
|
|
#include <sys/types.h>
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/selinfo.h>
|
|
|
|
#include <sys/module.h>
|
|
|
|
#include <sys/bus.h>
|
|
|
|
#include <sys/conf.h>
|
|
|
|
#include <sys/bio.h>
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
#include <sys/uio.h>
|
|
|
|
#include <sys/sysctl.h>
|
|
|
|
#include <sys/endian.h>
|
|
|
|
#include <sys/queue.h>
|
|
|
|
#include <sys/kthread.h>
|
|
|
|
#include <sys/taskqueue.h>
|
|
|
|
#include <sys/sbuf.h>
|
|
|
|
|
|
|
|
#include <machine/bus.h>
|
|
|
|
#include <machine/resource.h>
|
|
|
|
#include <sys/rman.h>
|
|
|
|
|
|
|
|
#include <machine/stdarg.h>
|
|
|
|
|
|
|
|
#include <cam/cam.h>
|
|
|
|
#include <cam/cam_ccb.h>
|
|
|
|
#include <cam/cam_debug.h>
|
|
|
|
#include <cam/cam_sim.h>
|
|
|
|
#include <cam/cam_xpt_sim.h>
|
|
|
|
#include <cam/cam_xpt_periph.h>
|
|
|
|
#include <cam/cam_periph.h>
|
|
|
|
#include <cam/scsi/scsi_all.h>
|
|
|
|
#include <cam/scsi/scsi_message.h>
|
|
|
|
|
|
|
|
#include <dev/mps/mpi/mpi2_type.h>
|
|
|
|
#include <dev/mps/mpi/mpi2.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_ioc.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_sas.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_cnfg.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_init.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_raid.h>
|
|
|
|
#include <dev/mps/mpi/mpi2_tool.h>
|
|
|
|
#include <dev/mps/mps_ioctl.h>
|
|
|
|
#include <dev/mps/mpsvar.h>
|
|
|
|
#include <dev/mps/mps_table.h>
|
|
|
|
#include <dev/mps/mps_sas.h>
|
|
|
|
|
|
|
|
/* For Hashed SAS Address creation for SATA Drives */
|
|
|
|
#define MPT2SAS_SN_LEN 20
|
|
|
|
#define MPT2SAS_MN_LEN 40
|
|
|
|
|
|
|
|
struct mps_fw_event_work {
|
|
|
|
u16 event;
|
|
|
|
void *event_data;
|
|
|
|
TAILQ_ENTRY(mps_fw_event_work) ev_link;
|
|
|
|
};
|
|
|
|
|
|
|
|
union _sata_sas_address {
|
|
|
|
u8 wwid[8];
|
|
|
|
struct {
|
|
|
|
u32 high;
|
|
|
|
u32 low;
|
|
|
|
} word;
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* define the IDENTIFY DEVICE structure
|
|
|
|
*/
|
|
|
|
struct _ata_identify_device_data {
|
|
|
|
u16 reserved1[10]; /* 0-9 */
|
|
|
|
u16 serial_number[10]; /* 10-19 */
|
|
|
|
u16 reserved2[7]; /* 20-26 */
|
|
|
|
u16 model_number[20]; /* 27-46*/
|
2015-02-24 22:07:42 +00:00
|
|
|
u16 reserved3[170]; /* 47-216 */
|
|
|
|
u16 rotational_speed; /* 217 */
|
|
|
|
u16 reserved4[38]; /* 218-255 */
|
2012-01-26 18:17:21 +00:00
|
|
|
};
|
2012-02-09 00:16:12 +00:00
|
|
|
static u32 event_count;
|
2012-01-26 18:17:21 +00:00
|
|
|
static void mpssas_fw_work(struct mps_softc *sc,
|
|
|
|
struct mps_fw_event_work *fw_event);
|
|
|
|
static void mpssas_fw_event_free(struct mps_softc *,
|
|
|
|
struct mps_fw_event_work *);
|
|
|
|
static int mpssas_add_device(struct mps_softc *sc, u16 handle, u8 linkrate);
|
|
|
|
static int mpssas_get_sata_identify(struct mps_softc *sc, u16 handle,
|
|
|
|
Mpi2SataPassthroughReply_t *mpi_reply, char *id_buffer, int sz,
|
|
|
|
u32 devinfo);
|
2015-02-24 22:07:42 +00:00
|
|
|
static void mpssas_ata_id_timeout(void *data);
|
2012-01-26 18:17:21 +00:00
|
|
|
int mpssas_get_sas_address_for_sata_disk(struct mps_softc *sc,
|
2015-02-24 22:07:42 +00:00
|
|
|
u64 *sas_address, u16 handle, u32 device_info, u8 *is_SATA_SSD);
|
2012-01-26 18:17:21 +00:00
|
|
|
static int mpssas_volume_add(struct mps_softc *sc,
|
2012-02-09 00:16:12 +00:00
|
|
|
u16 handle);
|
2014-07-30 17:58:17 +00:00
|
|
|
static void mpssas_SSU_to_SATA_devices(struct mps_softc *sc);
|
|
|
|
static void mpssas_stop_unit_done(struct cam_periph *periph,
|
|
|
|
union ccb *done_ccb);
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
void
|
|
|
|
mpssas_evt_handler(struct mps_softc *sc, uintptr_t data,
|
|
|
|
MPI2_EVENT_NOTIFICATION_REPLY *event)
|
|
|
|
{
|
|
|
|
struct mps_fw_event_work *fw_event;
|
|
|
|
u16 sz;
|
|
|
|
|
|
|
|
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
|
|
|
|
mps_print_evt_sas(sc, event);
|
|
|
|
mpssas_record_event(sc, event);
|
|
|
|
|
|
|
|
fw_event = malloc(sizeof(struct mps_fw_event_work), M_MPT2,
|
|
|
|
M_ZERO|M_NOWAIT);
|
|
|
|
if (!fw_event) {
|
|
|
|
printf("%s: allocate failed for fw_event\n", __func__);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
sz = le16toh(event->EventDataLength) * 4;
|
|
|
|
fw_event->event_data = malloc(sz, M_MPT2, M_ZERO|M_NOWAIT);
|
|
|
|
if (!fw_event->event_data) {
|
|
|
|
printf("%s: allocate failed for event_data\n", __func__);
|
|
|
|
free(fw_event, M_MPT2);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
bcopy(event->EventData, fw_event->event_data, sz);
|
|
|
|
fw_event->event = event->Event;
|
|
|
|
if ((event->Event == MPI2_EVENT_SAS_TOPOLOGY_CHANGE_LIST ||
|
|
|
|
event->Event == MPI2_EVENT_SAS_ENCL_DEVICE_STATUS_CHANGE ||
|
|
|
|
event->Event == MPI2_EVENT_IR_CONFIGURATION_CHANGE_LIST) &&
|
|
|
|
sc->track_mapping_events)
|
|
|
|
sc->pending_map_events++;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When wait_for_port_enable flag is set, make sure that all the events
|
|
|
|
* are processed. Increment the startup_refcount and decrement it after
|
|
|
|
* events are processed.
|
|
|
|
*/
|
|
|
|
if ((event->Event == MPI2_EVENT_SAS_TOPOLOGY_CHANGE_LIST ||
|
|
|
|
event->Event == MPI2_EVENT_IR_CONFIGURATION_CHANGE_LIST) &&
|
|
|
|
sc->wait_for_port_enable)
|
|
|
|
mpssas_startup_increment(sc->sassc);
|
|
|
|
|
|
|
|
TAILQ_INSERT_TAIL(&sc->sassc->ev_queue, fw_event, ev_link);
|
|
|
|
taskqueue_enqueue(sc->sassc->ev_tq, &sc->sassc->ev_task);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
mpssas_fw_event_free(struct mps_softc *sc, struct mps_fw_event_work *fw_event)
|
|
|
|
{
|
|
|
|
|
|
|
|
free(fw_event->event_data, M_MPT2);
|
|
|
|
free(fw_event, M_MPT2);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* _mps_fw_work - delayed task for processing firmware events
|
|
|
|
* @sc: per adapter object
|
|
|
|
* @fw_event: The fw_event_work object
|
|
|
|
* Context: user.
|
|
|
|
*
|
|
|
|
* Return nothing.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
mpssas_fw_work(struct mps_softc *sc, struct mps_fw_event_work *fw_event)
|
|
|
|
{
|
|
|
|
struct mpssas_softc *sassc;
|
|
|
|
sassc = sc->sassc;
|
|
|
|
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, "(%d)->(%s) Working on Event: [%x]\n",
|
2012-02-09 00:16:12 +00:00
|
|
|
event_count++,__func__,fw_event->event);
|
2012-01-26 18:17:21 +00:00
|
|
|
switch (fw_event->event) {
|
|
|
|
case MPI2_EVENT_SAS_TOPOLOGY_CHANGE_LIST:
|
|
|
|
{
|
|
|
|
MPI2_EVENT_DATA_SAS_TOPOLOGY_CHANGE_LIST *data;
|
|
|
|
MPI2_EVENT_SAS_TOPO_PHY_ENTRY *phy;
|
|
|
|
int i;
|
|
|
|
|
|
|
|
data = (MPI2_EVENT_DATA_SAS_TOPOLOGY_CHANGE_LIST *)
|
|
|
|
fw_event->event_data;
|
|
|
|
|
|
|
|
mps_mapping_topology_change_event(sc, fw_event->event_data);
|
|
|
|
|
|
|
|
for (i = 0; i < data->NumEntries; i++) {
|
|
|
|
phy = &data->PHY[i];
|
|
|
|
switch (phy->PhyStatus & MPI2_EVENT_SAS_TOPO_RC_MASK) {
|
|
|
|
case MPI2_EVENT_SAS_TOPO_RC_TARG_ADDED:
|
|
|
|
if (mpssas_add_device(sc,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(phy->AttachedDevHandle), phy->LinkRate)){
|
2012-01-26 18:17:21 +00:00
|
|
|
printf("%s: failed to add device with "
|
|
|
|
"handle 0x%x\n", __func__,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(phy->AttachedDevHandle));
|
|
|
|
mpssas_prepare_remove(sassc, le16toh(
|
|
|
|
phy->AttachedDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
case MPI2_EVENT_SAS_TOPO_RC_TARG_NOT_RESPONDING:
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
mpssas_prepare_remove(sassc,le16toh(
|
|
|
|
phy->AttachedDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_SAS_TOPO_RC_PHY_CHANGED:
|
|
|
|
case MPI2_EVENT_SAS_TOPO_RC_NO_CHANGE:
|
|
|
|
case MPI2_EVENT_SAS_TOPO_RC_DELAY_NOT_RESPONDING:
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* refcount was incremented for this event in
|
|
|
|
* mpssas_evt_handler. Decrement it here because the event has
|
|
|
|
* been processed.
|
|
|
|
*/
|
|
|
|
mpssas_startup_decrement(sassc);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_SAS_DISCOVERY:
|
|
|
|
{
|
|
|
|
MPI2_EVENT_DATA_SAS_DISCOVERY *data;
|
|
|
|
|
|
|
|
data = (MPI2_EVENT_DATA_SAS_DISCOVERY *)fw_event->event_data;
|
|
|
|
|
|
|
|
if (data->ReasonCode & MPI2_EVENT_SAS_DISC_RC_STARTED)
|
|
|
|
mps_dprint(sc, MPS_TRACE,"SAS discovery start event\n");
|
|
|
|
if (data->ReasonCode & MPI2_EVENT_SAS_DISC_RC_COMPLETED) {
|
|
|
|
mps_dprint(sc, MPS_TRACE,"SAS discovery stop event\n");
|
|
|
|
sassc->flags &= ~MPSSAS_IN_DISCOVERY;
|
|
|
|
mpssas_discovery_end(sassc);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_SAS_ENCL_DEVICE_STATUS_CHANGE:
|
|
|
|
{
|
|
|
|
Mpi2EventDataSasEnclDevStatusChange_t *data;
|
|
|
|
data = (Mpi2EventDataSasEnclDevStatusChange_t *)
|
|
|
|
fw_event->event_data;
|
|
|
|
mps_mapping_enclosure_dev_status_change_event(sc,
|
|
|
|
fw_event->event_data);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_IR_CONFIGURATION_CHANGE_LIST:
|
|
|
|
{
|
|
|
|
Mpi2EventIrConfigElement_t *element;
|
|
|
|
int i;
|
|
|
|
u8 foreign_config;
|
|
|
|
Mpi2EventDataIrConfigChangeList_t *event_data;
|
|
|
|
struct mpssas_target *targ;
|
|
|
|
unsigned int id;
|
|
|
|
|
|
|
|
event_data = fw_event->event_data;
|
|
|
|
foreign_config = (le32toh(event_data->Flags) &
|
|
|
|
MPI2_EVENT_IR_CHANGE_FLAGS_FOREIGN_CONFIG) ? 1 : 0;
|
|
|
|
|
|
|
|
element =
|
|
|
|
(Mpi2EventIrConfigElement_t *)&event_data->ConfigElement[0];
|
|
|
|
id = mps_mapping_get_raid_id_from_handle
|
|
|
|
(sc, element->VolDevHandle);
|
|
|
|
|
|
|
|
mps_mapping_ir_config_change_event(sc, event_data);
|
|
|
|
|
|
|
|
for (i = 0; i < event_data->NumElements; i++, element++) {
|
|
|
|
switch (element->ReasonCode) {
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_VOLUME_CREATED:
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_ADDED:
|
|
|
|
if (!foreign_config) {
|
2012-02-09 00:16:12 +00:00
|
|
|
if (mpssas_volume_add(sc, le16toh(element->VolDevHandle))){
|
2012-01-26 18:17:21 +00:00
|
|
|
printf("%s: failed to add RAID "
|
|
|
|
"volume with handle 0x%x\n",
|
|
|
|
__func__, le16toh(element->
|
|
|
|
VolDevHandle));
|
|
|
|
}
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_VOLUME_DELETED:
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_REMOVED:
|
|
|
|
/*
|
|
|
|
* Rescan after volume is deleted or removed.
|
|
|
|
*/
|
|
|
|
if (!foreign_config) {
|
|
|
|
if (id == MPS_MAP_BAD_ID) {
|
|
|
|
printf("%s: could not get ID "
|
|
|
|
"for volume with handle "
|
|
|
|
"0x%04x\n", __func__,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(element->VolDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
targ = &sassc->targets[id];
|
|
|
|
targ->handle = 0x0;
|
|
|
|
targ->encl_slot = 0x0;
|
|
|
|
targ->encl_handle = 0x0;
|
|
|
|
targ->exp_dev_handle = 0x0;
|
|
|
|
targ->phy_num = 0x0;
|
|
|
|
targ->linkrate = 0x0;
|
|
|
|
mpssas_rescan_target(sc, targ);
|
|
|
|
printf("RAID target id 0x%x removed\n",
|
|
|
|
targ->tid);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_PD_CREATED:
|
2012-02-09 00:16:12 +00:00
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_HIDE:
|
2012-01-26 18:17:21 +00:00
|
|
|
/*
|
|
|
|
* Phys Disk of a volume has been created. Hide
|
|
|
|
* it from the OS.
|
|
|
|
*/
|
2012-02-09 00:16:12 +00:00
|
|
|
targ = mpssas_find_target_by_handle(sassc, 0, element->PhysDiskDevHandle);
|
|
|
|
if (targ == NULL)
|
|
|
|
break;
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
|
|
|
|
/* Set raid component flags only if it is not WD.
|
|
|
|
* OR WrapDrive with WD_HIDE_ALWAYS/WD_HIDE_IF_VOLUME is set in NVRAM
|
|
|
|
*/
|
|
|
|
if((!sc->WD_available) ||
|
|
|
|
((sc->WD_available &&
|
|
|
|
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
|
|
|
|
(sc->WD_valid_config && (sc->WD_hide_expose ==
|
|
|
|
MPS_WD_HIDE_IF_VOLUME)))) {
|
|
|
|
targ->flags |= MPS_TARGET_FLAGS_RAID_COMPONENT;
|
|
|
|
}
|
2012-02-09 00:16:12 +00:00
|
|
|
mpssas_rescan_target(sc, targ);
|
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_CHANGE_RC_PD_DELETED:
|
|
|
|
/*
|
|
|
|
* Phys Disk of a volume has been deleted.
|
|
|
|
* Expose it to the OS.
|
|
|
|
*/
|
|
|
|
if (mpssas_add_device(sc,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(element->PhysDiskDevHandle), 0)){
|
2012-01-26 18:17:21 +00:00
|
|
|
printf("%s: failed to add device with "
|
|
|
|
"handle 0x%x\n", __func__,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(element->PhysDiskDevHandle));
|
|
|
|
mpssas_prepare_remove(sassc, le16toh(element->
|
|
|
|
PhysDiskDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* refcount was incremented for this event in
|
|
|
|
* mpssas_evt_handler. Decrement it here because the event has
|
|
|
|
* been processed.
|
|
|
|
*/
|
|
|
|
mpssas_startup_decrement(sassc);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_IR_VOLUME:
|
|
|
|
{
|
|
|
|
Mpi2EventDataIrVolume_t *event_data = fw_event->event_data;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Informational only.
|
|
|
|
*/
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, "Received IR Volume event:\n");
|
2012-01-26 18:17:21 +00:00
|
|
|
switch (event_data->ReasonCode) {
|
|
|
|
case MPI2_EVENT_IR_VOLUME_RC_SETTINGS_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Volume Settings "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"changed from 0x%x to 0x%x for Volome with "
|
|
|
|
"handle 0x%x", le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue),
|
|
|
|
le16toh(event_data->VolDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_VOLUME_RC_STATUS_FLAGS_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Volume Status "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"changed from 0x%x to 0x%x for Volome with "
|
|
|
|
"handle 0x%x", le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue),
|
|
|
|
le16toh(event_data->VolDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_VOLUME_RC_STATE_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Volume State "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"changed from 0x%x to 0x%x for Volome with "
|
|
|
|
"handle 0x%x", le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue),
|
|
|
|
le16toh(event_data->VolDevHandle));
|
2012-02-09 00:16:12 +00:00
|
|
|
u32 state;
|
|
|
|
struct mpssas_target *targ;
|
|
|
|
state = le32toh(event_data->NewValue);
|
|
|
|
switch (state) {
|
|
|
|
case MPI2_RAID_VOL_STATE_MISSING:
|
|
|
|
case MPI2_RAID_VOL_STATE_FAILED:
|
|
|
|
mpssas_prepare_volume_remove(sassc, event_data->
|
|
|
|
VolDevHandle);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case MPI2_RAID_VOL_STATE_ONLINE:
|
|
|
|
case MPI2_RAID_VOL_STATE_DEGRADED:
|
|
|
|
case MPI2_RAID_VOL_STATE_OPTIMAL:
|
|
|
|
targ = mpssas_find_target_by_handle(sassc, 0, event_data->VolDevHandle);
|
|
|
|
if (targ) {
|
|
|
|
printf("%s %d: Volume handle 0x%x is already added \n",
|
|
|
|
__func__, __LINE__ , event_data->VolDevHandle);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
if (mpssas_volume_add(sc, le16toh(event_data->VolDevHandle))) {
|
|
|
|
printf("%s: failed to add RAID "
|
|
|
|
"volume with handle 0x%x\n",
|
|
|
|
__func__, le16toh(event_data->
|
|
|
|
VolDevHandle));
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_IR_PHYSICAL_DISK:
|
|
|
|
{
|
|
|
|
Mpi2EventDataIrPhysicalDisk_t *event_data =
|
|
|
|
fw_event->event_data;
|
2012-02-09 00:16:12 +00:00
|
|
|
struct mpssas_target *targ;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Informational only.
|
|
|
|
*/
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, "Received IR Phys Disk event:\n");
|
2012-01-26 18:17:21 +00:00
|
|
|
switch (event_data->ReasonCode) {
|
|
|
|
case MPI2_EVENT_IR_PHYSDISK_RC_SETTINGS_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Phys Disk Settings "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"changed from 0x%x to 0x%x for Phys Disk Number "
|
|
|
|
"%d and handle 0x%x at Enclosure handle 0x%x, Slot "
|
|
|
|
"%d", le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue),
|
|
|
|
event_data->PhysDiskNum,
|
|
|
|
le16toh(event_data->PhysDiskDevHandle),
|
|
|
|
le16toh(event_data->EnclosureHandle), le16toh(event_data->Slot));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_PHYSDISK_RC_STATUS_FLAGS_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Phys Disk Status changed "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"from 0x%x to 0x%x for Phys Disk Number %d and "
|
|
|
|
"handle 0x%x at Enclosure handle 0x%x, Slot %d",
|
|
|
|
le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue), event_data->PhysDiskNum,
|
|
|
|
le16toh(event_data->PhysDiskDevHandle),
|
|
|
|
le16toh(event_data->EnclosureHandle), le16toh(event_data->Slot));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
case MPI2_EVENT_IR_PHYSDISK_RC_STATE_CHANGED:
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, " Phys Disk State changed "
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
"from 0x%x to 0x%x for Phys Disk Number %d and "
|
|
|
|
"handle 0x%x at Enclosure handle 0x%x, Slot %d",
|
|
|
|
le32toh(event_data->PreviousValue),
|
|
|
|
le32toh(event_data->NewValue), event_data->PhysDiskNum,
|
|
|
|
le16toh(event_data->PhysDiskDevHandle),
|
|
|
|
le16toh(event_data->EnclosureHandle), le16toh(event_data->Slot));
|
2012-02-09 00:16:12 +00:00
|
|
|
switch (event_data->NewValue) {
|
|
|
|
case MPI2_RAID_PD_STATE_ONLINE:
|
|
|
|
case MPI2_RAID_PD_STATE_DEGRADED:
|
|
|
|
case MPI2_RAID_PD_STATE_REBUILDING:
|
|
|
|
case MPI2_RAID_PD_STATE_OPTIMAL:
|
|
|
|
case MPI2_RAID_PD_STATE_HOT_SPARE:
|
|
|
|
targ = mpssas_find_target_by_handle(sassc, 0,
|
|
|
|
event_data->PhysDiskDevHandle);
|
|
|
|
if (targ) {
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
if(!sc->WD_available) {
|
|
|
|
targ->flags |= MPS_TARGET_FLAGS_RAID_COMPONENT;
|
|
|
|
printf("%s %d: Found Target for handle 0x%x. \n",
|
|
|
|
__func__, __LINE__ , event_data->PhysDiskDevHandle);
|
|
|
|
} else if ((sc->WD_available &&
|
|
|
|
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
|
|
|
|
(sc->WD_valid_config && (sc->WD_hide_expose ==
|
|
|
|
MPS_WD_HIDE_IF_VOLUME))) {
|
|
|
|
targ->flags |= MPS_TARGET_FLAGS_RAID_COMPONENT;
|
|
|
|
printf("%s %d: WD: Found Target for handle 0x%x. \n",
|
|
|
|
__func__, __LINE__ , event_data->PhysDiskDevHandle);
|
|
|
|
}
|
|
|
|
}
|
2012-02-09 00:16:12 +00:00
|
|
|
break;
|
|
|
|
case MPI2_RAID_PD_STATE_OFFLINE:
|
|
|
|
case MPI2_RAID_PD_STATE_NOT_CONFIGURED:
|
|
|
|
case MPI2_RAID_PD_STATE_NOT_COMPATIBLE:
|
|
|
|
default:
|
|
|
|
targ = mpssas_find_target_by_handle(sassc, 0,
|
|
|
|
event_data->PhysDiskDevHandle);
|
|
|
|
if (targ) {
|
|
|
|
targ->flags |= ~MPS_TARGET_FLAGS_RAID_COMPONENT;
|
|
|
|
printf("%s %d: Found Target for handle 0x%x. \n",
|
|
|
|
__func__, __LINE__ , event_data->PhysDiskDevHandle);
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_IR_OPERATION_STATUS:
|
|
|
|
{
|
|
|
|
Mpi2EventDataIrOperationStatus_t *event_data =
|
|
|
|
fw_event->event_data;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Informational only.
|
|
|
|
*/
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, "Received IR Op Status event:\n");
|
|
|
|
mps_dprint(sc, MPS_EVENT, " RAID Operation of %d is %d "
|
2012-01-26 18:17:21 +00:00
|
|
|
"percent complete for Volume with handle 0x%x",
|
|
|
|
event_data->RAIDOperation, event_data->PercentComplete,
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
le16toh(event_data->VolDevHandle));
|
2012-01-26 18:17:21 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_LOG_ENTRY_ADDED:
|
|
|
|
{
|
|
|
|
pMpi2EventDataLogEntryAdded_t logEntry;
|
|
|
|
uint16_t logQualifier;
|
|
|
|
uint8_t logCode;
|
|
|
|
|
|
|
|
logEntry = (pMpi2EventDataLogEntryAdded_t)fw_event->event_data;
|
|
|
|
logQualifier = logEntry->LogEntryQualifier;
|
|
|
|
|
|
|
|
if (logQualifier == MPI2_WD_LOG_ENTRY) {
|
|
|
|
logCode = logEntry->LogData[0];
|
|
|
|
|
|
|
|
switch (logCode) {
|
|
|
|
case MPI2_WD_SSD_THROTTLING:
|
|
|
|
printf("WarpDrive Warning: IO Throttling has "
|
|
|
|
"occurred in the WarpDrive subsystem. "
|
|
|
|
"Check WarpDrive documentation for "
|
|
|
|
"additional details\n");
|
|
|
|
break;
|
|
|
|
case MPI2_WD_DRIVE_LIFE_WARN:
|
|
|
|
printf("WarpDrive Warning: Program/Erase "
|
|
|
|
"Cycles for the WarpDrive subsystem in "
|
|
|
|
"degraded range. Check WarpDrive "
|
|
|
|
"documentation for additional details\n");
|
|
|
|
break;
|
|
|
|
case MPI2_WD_DRIVE_LIFE_DEAD:
|
|
|
|
printf("WarpDrive Fatal Error: There are no "
|
|
|
|
"Program/Erase Cycles for the WarpDrive "
|
|
|
|
"subsystem. The storage device will be in "
|
|
|
|
"read-only mode. Check WarpDrive "
|
|
|
|
"documentation for additional details\n");
|
|
|
|
break;
|
|
|
|
case MPI2_WD_RAIL_MON_FAIL:
|
|
|
|
printf("WarpDrive Fatal Error: The Backup Rail "
|
|
|
|
"Monitor has failed on the WarpDrive "
|
|
|
|
"subsystem. Check WarpDrive documentation "
|
|
|
|
"for additional details\n");
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
case MPI2_EVENT_SAS_DEVICE_STATUS_CHANGE:
|
|
|
|
case MPI2_EVENT_SAS_BROADCAST_PRIMITIVE:
|
|
|
|
default:
|
|
|
|
mps_dprint(sc, MPS_TRACE,"Unhandled event 0x%0X\n",
|
|
|
|
fw_event->event);
|
|
|
|
break;
|
|
|
|
|
|
|
|
}
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_EVENT, "(%d)->(%s) Event Free: [%x]\n",event_count,__func__, fw_event->event);
|
2012-01-26 18:17:21 +00:00
|
|
|
mpssas_fw_event_free(sc, fw_event);
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
|
|
|
mpssas_firmware_event_work(void *arg, int pending)
|
|
|
|
{
|
|
|
|
struct mps_fw_event_work *fw_event;
|
|
|
|
struct mps_softc *sc;
|
|
|
|
|
|
|
|
sc = (struct mps_softc *)arg;
|
|
|
|
mps_lock(sc);
|
|
|
|
while ((fw_event = TAILQ_FIRST(&sc->sassc->ev_queue)) != NULL) {
|
|
|
|
TAILQ_REMOVE(&sc->sassc->ev_queue, fw_event, ev_link);
|
|
|
|
mpssas_fw_work(sc, fw_event);
|
|
|
|
}
|
|
|
|
mps_unlock(sc);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
mpssas_add_device(struct mps_softc *sc, u16 handle, u8 linkrate){
|
|
|
|
char devstring[80];
|
|
|
|
struct mpssas_softc *sassc;
|
|
|
|
struct mpssas_target *targ;
|
|
|
|
Mpi2ConfigReply_t mpi_reply;
|
|
|
|
Mpi2SasDevicePage0_t config_page;
|
2015-02-24 22:07:42 +00:00
|
|
|
uint64_t sas_address;
|
2012-01-26 18:17:21 +00:00
|
|
|
uint64_t parent_sas_address = 0;
|
|
|
|
u32 device_info, parent_devinfo = 0;
|
|
|
|
unsigned int id;
|
2015-02-24 22:07:42 +00:00
|
|
|
int ret = 1, error = 0, i;
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
struct mpssas_lun *lun;
|
2015-02-24 22:07:42 +00:00
|
|
|
u8 is_SATA_SSD = 0;
|
|
|
|
struct mps_command *cm;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
sassc = sc->sassc;
|
|
|
|
mpssas_startup_increment(sassc);
|
|
|
|
if ((mps_config_get_sas_device_pg0(sc, &mpi_reply, &config_page,
|
|
|
|
MPI2_SAS_DEVICE_PGAD_FORM_HANDLE, handle))) {
|
|
|
|
printf("%s: error reading SAS device page0\n", __func__);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
device_info = le32toh(config_page.DeviceInfo);
|
|
|
|
|
|
|
|
if (((device_info & MPI2_SAS_DEVICE_INFO_SMP_TARGET) == 0)
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
&& (le16toh(config_page.ParentDevHandle) != 0)) {
|
2012-01-26 18:17:21 +00:00
|
|
|
Mpi2ConfigReply_t tmp_mpi_reply;
|
|
|
|
Mpi2SasDevicePage0_t parent_config_page;
|
|
|
|
|
|
|
|
if ((mps_config_get_sas_device_pg0(sc, &tmp_mpi_reply,
|
|
|
|
&parent_config_page, MPI2_SAS_DEVICE_PGAD_FORM_HANDLE,
|
|
|
|
le16toh(config_page.ParentDevHandle)))) {
|
|
|
|
printf("%s: error reading SAS device %#x page0\n",
|
|
|
|
__func__, le16toh(config_page.ParentDevHandle));
|
|
|
|
} else {
|
|
|
|
parent_sas_address = parent_config_page.SASAddress.High;
|
|
|
|
parent_sas_address = (parent_sas_address << 32) |
|
|
|
|
parent_config_page.SASAddress.Low;
|
|
|
|
parent_devinfo = le32toh(parent_config_page.DeviceInfo);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
/* TODO Check proper endianess */
|
|
|
|
sas_address = config_page.SASAddress.High;
|
2015-02-24 22:07:42 +00:00
|
|
|
sas_address = (sas_address << 32) | config_page.SASAddress.Low;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
2015-02-24 22:07:42 +00:00
|
|
|
/*
|
|
|
|
* Always get SATA Identify information because this is used to
|
|
|
|
* determine if Start/Stop Unit should be sent to the drive when the
|
|
|
|
* system is shutdown.
|
|
|
|
*/
|
|
|
|
if (device_info & MPI2_SAS_DEVICE_INFO_SATA_DEVICE) {
|
|
|
|
ret = mpssas_get_sas_address_for_sata_disk(sc, &sas_address,
|
|
|
|
handle, device_info, &is_SATA_SSD);
|
|
|
|
if (ret) {
|
|
|
|
mps_dprint(sc, MPS_INFO, "%s: failed to get disk type "
|
|
|
|
"(SSD or HDD) for SATA device with handle 0x%04x\n",
|
|
|
|
__func__, handle);
|
|
|
|
} else {
|
|
|
|
mps_dprint(sc, MPS_INFO, "SAS Address from SATA "
|
|
|
|
"device = %jx\n", sas_address);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
id = mps_mapping_get_sas_id(sc, sas_address, handle);
|
2012-01-26 18:17:21 +00:00
|
|
|
if (id == MPS_MAP_BAD_ID) {
|
|
|
|
printf("failure at %s:%d/%s()! Could not get ID for device "
|
|
|
|
"with handle 0x%04x\n", __FILE__, __LINE__, __func__,
|
|
|
|
handle);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
2013-08-09 01:09:02 +00:00
|
|
|
|
|
|
|
if (mpssas_check_id(sassc, id) != 0) {
|
|
|
|
device_printf(sc->mps_dev, "Excluding target id %d\n", id);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "SAS Address from SAS device page0 = %jx\n",
|
2012-01-26 18:17:21 +00:00
|
|
|
sas_address);
|
|
|
|
targ = &sassc->targets[id];
|
|
|
|
targ->devinfo = device_info;
|
|
|
|
targ->devname = le32toh(config_page.DeviceName.High);
|
|
|
|
targ->devname = (targ->devname << 32) |
|
|
|
|
le32toh(config_page.DeviceName.Low);
|
|
|
|
targ->encl_handle = le16toh(config_page.EnclosureHandle);
|
|
|
|
targ->encl_slot = le16toh(config_page.Slot);
|
|
|
|
targ->handle = handle;
|
|
|
|
targ->parent_handle = le16toh(config_page.ParentDevHandle);
|
|
|
|
targ->sasaddr = mps_to_u64(&config_page.SASAddress);
|
|
|
|
targ->parent_sasaddr = le64toh(parent_sas_address);
|
|
|
|
targ->parent_devinfo = parent_devinfo;
|
|
|
|
targ->tid = id;
|
|
|
|
targ->linkrate = (linkrate>>4);
|
|
|
|
targ->flags = 0;
|
2015-02-24 22:07:42 +00:00
|
|
|
if (is_SATA_SSD) {
|
|
|
|
targ->flags = MPS_TARGET_IS_SATA_SSD;
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
TAILQ_INIT(&targ->commands);
|
|
|
|
TAILQ_INIT(&targ->timedout_commands);
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
while(!SLIST_EMPTY(&targ->luns)) {
|
|
|
|
lun = SLIST_FIRST(&targ->luns);
|
|
|
|
SLIST_REMOVE_HEAD(&targ->luns, lun_link);
|
|
|
|
free(lun, M_MPT2);
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
SLIST_INIT(&targ->luns);
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
mps_describe_devinfo(targ->devinfo, devstring, 80);
|
2015-02-24 22:07:42 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "Found device <%s> <%s> <0x%04x> <%d/%d>\n",
|
|
|
|
devstring, mps_describe_table(mps_linkrate_names, targ->linkrate),
|
2012-01-26 18:17:21 +00:00
|
|
|
targ->handle, targ->encl_handle, targ->encl_slot);
|
CAM and mps(4) driver scanning changes.
Add a PIM_NOSCAN flag to the CAM path inquiry CCB. This tells CAM
not to perform a rescan on a bus when it is registered.
We now use this flag in the mps(4) driver. Since it knows what
devices it has attached, it is more efficient for it to just issue
a target rescan on the targets that are attached.
Also, remove the private rescan thread from the mps(4) driver in
favor of the rescan thread already built into CAM. Without this
change, but with the change above, the MPS scanner could run before
or during CAM's initial setup, which would cause duplicate device
reprobes and announcements.
sys/param.h:
Bump __FreeBSD_version to 1000039 for the inclusion of the
PIM_RESCAN CAM path inquiry flag.
sys/cam/cam_ccb.h:
sys/cam/cam_xpt.c:
Added a PIM_NOSCAN flag. If a SIM sets this in the path
inquiry ccb, then CAM won't rescan the bus in
xpt_bus_regsister.
sys/dev/mps/mps_sas.c
For versions of FreeBSD that have the PIM_NOSCAN path
inquiry flag, don't freeze the sim queue during scanning,
because CAM won't be scanning this bus. Instead, hold
up the boot. Don't call mpssas_rescan_target in
mpssas_startup_decrement; it's redundant and I don't
know why it was in there.
Set PIM_NOSCAN in path inquiry CCBs.
Remove methods related to the internal rescan daemon.
Always use async events to trigger a probe for EEDP support.
In older versions of FreeBSD where AC_ADVINFO_CHANGED is
not available, use AC_FOUND_DEVICE and issue the
necessary READ CAPACITY manually.
Provide a path to xpt_register_async() so that we only
receive events for our own SCSI domain.
Improve error reporting in cases where setup for EEDP
detection fails.
sys/dev/mps/mps_sas.h:
Remove softc flags and data related to the scanner thread.
sys/dev/mps/mps_sas_lsi.c:
Unconditionally rescan the target whenever a device is added.
Sponsored by: Spectra Logic
MFC after: 1 week
2013-07-22 18:37:07 +00:00
|
|
|
|
|
|
|
#if __FreeBSD_version < 1000039
|
2012-01-26 18:17:21 +00:00
|
|
|
if ((sassc->flags & MPSSAS_IN_STARTUP) == 0)
|
CAM and mps(4) driver scanning changes.
Add a PIM_NOSCAN flag to the CAM path inquiry CCB. This tells CAM
not to perform a rescan on a bus when it is registered.
We now use this flag in the mps(4) driver. Since it knows what
devices it has attached, it is more efficient for it to just issue
a target rescan on the targets that are attached.
Also, remove the private rescan thread from the mps(4) driver in
favor of the rescan thread already built into CAM. Without this
change, but with the change above, the MPS scanner could run before
or during CAM's initial setup, which would cause duplicate device
reprobes and announcements.
sys/param.h:
Bump __FreeBSD_version to 1000039 for the inclusion of the
PIM_RESCAN CAM path inquiry flag.
sys/cam/cam_ccb.h:
sys/cam/cam_xpt.c:
Added a PIM_NOSCAN flag. If a SIM sets this in the path
inquiry ccb, then CAM won't rescan the bus in
xpt_bus_regsister.
sys/dev/mps/mps_sas.c
For versions of FreeBSD that have the PIM_NOSCAN path
inquiry flag, don't freeze the sim queue during scanning,
because CAM won't be scanning this bus. Instead, hold
up the boot. Don't call mpssas_rescan_target in
mpssas_startup_decrement; it's redundant and I don't
know why it was in there.
Set PIM_NOSCAN in path inquiry CCBs.
Remove methods related to the internal rescan daemon.
Always use async events to trigger a probe for EEDP support.
In older versions of FreeBSD where AC_ADVINFO_CHANGED is
not available, use AC_FOUND_DEVICE and issue the
necessary READ CAPACITY manually.
Provide a path to xpt_register_async() so that we only
receive events for our own SCSI domain.
Improve error reporting in cases where setup for EEDP
detection fails.
sys/dev/mps/mps_sas.h:
Remove softc flags and data related to the scanner thread.
sys/dev/mps/mps_sas_lsi.c:
Unconditionally rescan the target whenever a device is added.
Sponsored by: Spectra Logic
MFC after: 1 week
2013-07-22 18:37:07 +00:00
|
|
|
#endif
|
2012-01-26 18:17:21 +00:00
|
|
|
mpssas_rescan_target(sc, targ);
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "Target id 0x%x added\n", targ->tid);
|
2015-02-24 22:07:42 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Check all commands to see if the SATA_ID_TIMEOUT flag has been set.
|
|
|
|
* If so, send a Target Reset TM to the target that was just created.
|
|
|
|
* An Abort Task TM should be used instead of a Target Reset, but that
|
|
|
|
* would be much more difficult because targets have not been fully
|
|
|
|
* discovered yet, and LUN's haven't been setup. So, just reset the
|
|
|
|
* target instead of the LUN.
|
|
|
|
*/
|
|
|
|
for (i = 1; i < sc->num_reqs; i++) {
|
|
|
|
cm = &sc->commands[i];
|
|
|
|
if (cm->cm_flags & MPS_CM_FLAGS_SATA_ID_TIMEOUT) {
|
|
|
|
targ->timeouts++;
|
|
|
|
cm->cm_state = MPS_CM_STATE_TIMEDOUT;
|
|
|
|
|
|
|
|
if ((targ->tm = mpssas_alloc_tm(sc)) != NULL) {
|
|
|
|
mps_dprint(sc, MPS_INFO, "%s: sending Target "
|
|
|
|
"Reset for stuck SATA identify command "
|
|
|
|
"(cm = %p)\n", __func__, cm);
|
|
|
|
targ->tm->cm_targ = targ;
|
|
|
|
mpssas_send_reset(sc, targ->tm,
|
|
|
|
MPI2_SCSITASKMGMT_TASKTYPE_TARGET_RESET);
|
|
|
|
} else {
|
|
|
|
mps_dprint(sc, MPS_ERROR, "Failed to allocate "
|
|
|
|
"tm for Target Reset after SATA ID command "
|
|
|
|
"timed out (cm %p)\n", cm);
|
|
|
|
}
|
|
|
|
/*
|
|
|
|
* No need to check for more since the target is
|
|
|
|
* already being reset.
|
|
|
|
*/
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
out:
|
2015-02-24 22:07:42 +00:00
|
|
|
/*
|
|
|
|
* Free the commands that may not have been freed from the SATA ID call
|
|
|
|
*/
|
|
|
|
for (i = 1; i < sc->num_reqs; i++) {
|
|
|
|
cm = &sc->commands[i];
|
|
|
|
if (cm->cm_flags & MPS_CM_FLAGS_SATA_ID_TIMEOUT) {
|
|
|
|
mps_free_command(sc, cm);
|
|
|
|
}
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
mpssas_startup_decrement(sassc);
|
|
|
|
return (error);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
|
|
|
mpssas_get_sas_address_for_sata_disk(struct mps_softc *sc,
|
2015-02-24 22:07:42 +00:00
|
|
|
u64 *sas_address, u16 handle, u32 device_info, u8 *is_SATA_SSD)
|
2012-01-26 18:17:21 +00:00
|
|
|
{
|
|
|
|
Mpi2SataPassthroughReply_t mpi_reply;
|
|
|
|
int i, rc, try_count;
|
|
|
|
u32 *bufferptr;
|
|
|
|
union _sata_sas_address hash_address;
|
|
|
|
struct _ata_identify_device_data ata_identify;
|
|
|
|
u8 buffer[MPT2SAS_MN_LEN + MPT2SAS_SN_LEN];
|
|
|
|
u32 ioc_status;
|
|
|
|
u8 sas_status;
|
|
|
|
|
|
|
|
memset(&ata_identify, 0, sizeof(ata_identify));
|
|
|
|
try_count = 0;
|
|
|
|
do {
|
|
|
|
rc = mpssas_get_sata_identify(sc, handle, &mpi_reply,
|
|
|
|
(char *)&ata_identify, sizeof(ata_identify), device_info);
|
|
|
|
try_count++;
|
|
|
|
ioc_status = le16toh(mpi_reply.IOCStatus)
|
|
|
|
& MPI2_IOCSTATUS_MASK;
|
|
|
|
sas_status = mpi_reply.SASStatus;
|
2015-02-24 22:07:42 +00:00
|
|
|
if (ioc_status != MPI2_IOCSTATUS_SUCCESS) {
|
|
|
|
if (sc->spinup_wait_time > 0) {
|
|
|
|
mps_dprint(sc, MPS_INFO, "Sleeping %d seconds "
|
|
|
|
"after SATA ID error to wait for spinup\n",
|
|
|
|
sc->spinup_wait_time);
|
|
|
|
msleep(&sc->msleep_fake_chan, &sc->mps_mtx, 0,
|
|
|
|
"mpsid", sc->spinup_wait_time * hz);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} while (((rc && (rc != EWOULDBLOCK)) || ioc_status || sas_status) &&
|
2012-01-26 18:17:21 +00:00
|
|
|
(try_count < 5));
|
|
|
|
|
|
|
|
if (rc == 0 && !ioc_status && !sas_status) {
|
2015-02-24 22:07:42 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "%s: got SATA identify "
|
|
|
|
"successfully for handle = 0x%x with try_count = %d\n",
|
|
|
|
__func__, handle, try_count);
|
2012-01-26 18:17:21 +00:00
|
|
|
} else {
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "%s: handle = 0x%x failed\n",
|
2015-02-24 22:07:42 +00:00
|
|
|
__func__, handle);
|
2012-01-26 18:17:21 +00:00
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
/* Copy & byteswap the 40 byte model number to a buffer */
|
|
|
|
for (i = 0; i < MPT2SAS_MN_LEN; i += 2) {
|
|
|
|
buffer[i] = ((u8 *)ata_identify.model_number)[i + 1];
|
|
|
|
buffer[i + 1] = ((u8 *)ata_identify.model_number)[i];
|
|
|
|
}
|
|
|
|
/* Copy & byteswap the 20 byte serial number to a buffer */
|
|
|
|
for (i = 0; i < MPT2SAS_SN_LEN; i += 2) {
|
|
|
|
buffer[MPT2SAS_MN_LEN + i] =
|
2015-02-24 22:07:42 +00:00
|
|
|
((u8 *)ata_identify.serial_number)[i + 1];
|
2012-01-26 18:17:21 +00:00
|
|
|
buffer[MPT2SAS_MN_LEN + i + 1] =
|
2015-02-24 22:07:42 +00:00
|
|
|
((u8 *)ata_identify.serial_number)[i];
|
2012-01-26 18:17:21 +00:00
|
|
|
}
|
|
|
|
bufferptr = (u32 *)buffer;
|
|
|
|
/* There are 60 bytes to hash down to 8. 60 isn't divisible by 8,
|
|
|
|
* so loop through the first 56 bytes (7*8),
|
|
|
|
* and then add in the last dword.
|
|
|
|
*/
|
|
|
|
hash_address.word.low = 0;
|
|
|
|
hash_address.word.high = 0;
|
|
|
|
for (i = 0; (i < ((MPT2SAS_MN_LEN+MPT2SAS_SN_LEN)/8)); i++) {
|
|
|
|
hash_address.word.low += *bufferptr;
|
|
|
|
bufferptr++;
|
|
|
|
hash_address.word.high += *bufferptr;
|
|
|
|
bufferptr++;
|
|
|
|
}
|
|
|
|
/* Add the last dword */
|
|
|
|
hash_address.word.low += *bufferptr;
|
|
|
|
/* Make sure the hash doesn't start with 5, because it could clash
|
|
|
|
* with a SAS address. Change 5 to a D.
|
|
|
|
*/
|
|
|
|
if ((hash_address.word.high & 0x000000F0) == (0x00000050))
|
|
|
|
hash_address.word.high |= 0x00000080;
|
|
|
|
*sas_address = (u64)hash_address.wwid[0] << 56 |
|
|
|
|
(u64)hash_address.wwid[1] << 48 | (u64)hash_address.wwid[2] << 40 |
|
|
|
|
(u64)hash_address.wwid[3] << 32 | (u64)hash_address.wwid[4] << 24 |
|
|
|
|
(u64)hash_address.wwid[5] << 16 | (u64)hash_address.wwid[6] << 8 |
|
|
|
|
(u64)hash_address.wwid[7];
|
2015-02-24 22:07:42 +00:00
|
|
|
if (ata_identify.rotational_speed == 1) {
|
|
|
|
*is_SATA_SSD = 1;
|
|
|
|
}
|
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
mpssas_get_sata_identify(struct mps_softc *sc, u16 handle,
|
|
|
|
Mpi2SataPassthroughReply_t *mpi_reply, char *id_buffer, int sz, u32 devinfo)
|
|
|
|
{
|
|
|
|
Mpi2SataPassthroughRequest_t *mpi_request;
|
|
|
|
Mpi2SataPassthroughReply_t *reply;
|
|
|
|
struct mps_command *cm;
|
|
|
|
char *buffer;
|
|
|
|
int error = 0;
|
|
|
|
|
|
|
|
buffer = malloc( sz, M_MPT2, M_NOWAIT | M_ZERO);
|
|
|
|
if (!buffer)
|
|
|
|
return ENOMEM;
|
|
|
|
|
2012-07-18 09:06:07 +00:00
|
|
|
if ((cm = mps_alloc_command(sc)) == NULL) {
|
|
|
|
free(buffer, M_MPT2);
|
2012-01-26 18:17:21 +00:00
|
|
|
return (EBUSY);
|
2012-07-18 09:06:07 +00:00
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
mpi_request = (MPI2_SATA_PASSTHROUGH_REQUEST *)cm->cm_req;
|
|
|
|
bzero(mpi_request,sizeof(MPI2_SATA_PASSTHROUGH_REQUEST));
|
|
|
|
mpi_request->Function = MPI2_FUNCTION_SATA_PASSTHROUGH;
|
|
|
|
mpi_request->VF_ID = 0;
|
|
|
|
mpi_request->DevHandle = htole16(handle);
|
|
|
|
mpi_request->PassthroughFlags = (MPI2_SATA_PT_REQ_PT_FLAGS_PIO |
|
|
|
|
MPI2_SATA_PT_REQ_PT_FLAGS_READ);
|
|
|
|
mpi_request->DataLength = htole32(sz);
|
|
|
|
mpi_request->CommandFIS[0] = 0x27;
|
|
|
|
mpi_request->CommandFIS[1] = 0x80;
|
|
|
|
mpi_request->CommandFIS[2] = (devinfo &
|
|
|
|
MPI2_SAS_DEVICE_INFO_ATAPI_DEVICE) ? 0xA1 : 0xEC;
|
|
|
|
cm->cm_sge = &mpi_request->SGL;
|
|
|
|
cm->cm_sglsize = sizeof(MPI2_SGE_IO_UNION);
|
|
|
|
cm->cm_flags = MPS_CM_FLAGS_SGE_SIMPLE | MPS_CM_FLAGS_DATAIN;
|
|
|
|
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
|
|
|
|
cm->cm_data = buffer;
|
|
|
|
cm->cm_length = htole32(sz);
|
2015-02-24 22:07:42 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Start a timeout counter specifically for the SATA ID command. This
|
|
|
|
* is used to fix a problem where the FW does not send a reply sometimes
|
|
|
|
* when a bad disk is in the topology. So, this is used to timeout the
|
|
|
|
* command so that processing can continue normally.
|
|
|
|
*/
|
|
|
|
mps_dprint(sc, MPS_XINFO, "%s start timeout counter for SATA ID "
|
|
|
|
"command\n", __func__);
|
|
|
|
callout_reset(&cm->cm_callout, MPS_ATA_ID_TIMEOUT * hz,
|
|
|
|
mpssas_ata_id_timeout, cm);
|
|
|
|
error = mps_wait_command(sc, cm, 60, CAN_SLEEP);
|
|
|
|
mps_dprint(sc, MPS_XINFO, "%s stop timeout counter for SATA ID "
|
|
|
|
"command\n", __func__);
|
|
|
|
callout_stop(&cm->cm_callout);
|
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
reply = (Mpi2SataPassthroughReply_t *)cm->cm_reply;
|
|
|
|
if (error || (reply == NULL)) {
|
|
|
|
/* FIXME */
|
Merge in phase 14+ -> 16 mps driver fixes from LSI:
---------------------------------------------------------------
System panics during a Port reset with ouststanding I/O
---------------------------------------------------------------
It is possible to call mps_mapping_free_memory after this
memory is already freed, causing a panic. Removed this extra
call to mps_mappiing_free_memory and call mps_mapping_exit
in place of the mps_mapping_free_memory call so that any
outstanding mapping items can be flushed before memory is
freed.
---------------------------------------------------------------
Correct memory leak during a Port reset with ouststanding I/O
---------------------------------------------------------------
In mps_reinit function, the mapping memory was not being
freed before being re-allocated. Added line to call the
memory free function for mapping memory.
---------------------------------------------------------------
Use CAM_SIM_QUEUED flag in Driver IO path.
---------------------------------------------------------------
This flag informs the XPT that successful abort of a CCB
requires an abort ccb to be issued to the SIM. While
processing SCSI IO's, set the CAM_SIM_QUEUED flag in the
status for the IO. When the command completes, clear this
flag.
---------------------------------------------------------------
Check for CAM_REQ_INPROG in I/O path.
---------------------------------------------------------------
Added a check in mpssas_action_scsiio for the In Progress
status for the IO. If this flag is set, the IO has already
been aborted by the upper layer (before CAM_SIM_QUEUED was
set) and there is no need to send the IO. The request will
be completed without error.
---------------------------------------------------------------
Improve "doorbell handshake method" for mps_get_iocfacts
---------------------------------------------------------------
Removed call to get Port Facts since this information is
not used currently.
Added mps_iocfacts_allocate function to allocate memory
that is based on IOC Facts data. Added mps_iocfacts_free
function to free memory that is based on IOC Facts data.
Both of the functions are used when a Diag Reset is performed
or when the driver is attached/detached. This is needed in
case IOC Facts changes after a Diag Reset, which could
happen if FW is upgraded.
Moved call of mps_bases_static_config_pages from the attach
routine to after the IOC is ready to process accesses based
on the new memory allocations (instead of polling through
the Doorbell).
---------------------------------------------------------------
Set TimeStamp in INIT message in millisecond format Set the IOC
---------------------------------------------------------------
---------------------------------------------------------------
Prefer mps_wait_command to mps_request_polled
---------------------------------------------------------------
Instead of using mps_request_polled, call mps_wait_command
whenever possible. Change the mps_wait_command function to
check the current context and either use interrupt context
or poll if required by using the pause or DELAY function.
Added a check after waiting 50mSecs to see if the command
has timed out. This is only done if polliing, the msleep
command will automatically timeout if the command has taken
too long to complete.
---------------------------------------------------------------
Integrated RAID: Volume Activation Failed error message is
displayed though the volume has been activated.
---------------------------------------------------------------
Instead of failing an IOCTL request that does not have a
large enough buffer to hold the complete reply, copy as
much data from the reply as possible into the user's buffer
and log a message saying that the user's buffer was smaller
than the returned data.
---------------------------------------------------------------
mapping_add_new_device failure due to persistent table FULL
---------------------------------------------------------------
When a new device is added, if it is determined that the
device persistent table is being used and is full, instead
of displaying a message for this condition every time, only
log a message if the MPS_INFO bit is set in the debug_flags.
Submitted by: LSI
MFC after: 1 week
2013-07-22 18:41:53 +00:00
|
|
|
/*
|
|
|
|
* If the request returns an error then we need to do a diag
|
|
|
|
* reset
|
|
|
|
*/
|
|
|
|
printf("%s: request for page completed with error %d",
|
2012-01-26 18:17:21 +00:00
|
|
|
__func__, error);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
bcopy(buffer, id_buffer, sz);
|
|
|
|
bcopy(reply, mpi_reply, sizeof(Mpi2SataPassthroughReply_t));
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
if ((le16toh(reply->IOCStatus) & MPI2_IOCSTATUS_MASK) !=
|
2012-01-26 18:17:21 +00:00
|
|
|
MPI2_IOCSTATUS_SUCCESS) {
|
|
|
|
printf("%s: error reading SATA PASSTHRU; iocstatus = 0x%x\n",
|
|
|
|
__func__, reply->IOCStatus);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
out:
|
2015-02-24 22:07:42 +00:00
|
|
|
/*
|
|
|
|
* If the SATA_ID_TIMEOUT flag has been set for this command, don't free
|
|
|
|
* it. The command will be freed after sending a target reset TM. If
|
|
|
|
* the command did timeout, use EWOULDBLOCK.
|
|
|
|
*/
|
|
|
|
if ((cm->cm_flags & MPS_CM_FLAGS_SATA_ID_TIMEOUT) == 0)
|
|
|
|
mps_free_command(sc, cm);
|
|
|
|
else if (error == 0)
|
|
|
|
error = EWOULDBLOCK;
|
|
|
|
free(buffer, M_MPT2);
|
2012-01-26 18:17:21 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
2015-02-24 22:07:42 +00:00
|
|
|
static void
|
|
|
|
mpssas_ata_id_timeout(void *data)
|
|
|
|
{
|
|
|
|
struct mps_softc *sc;
|
|
|
|
struct mps_command *cm;
|
|
|
|
|
|
|
|
cm = (struct mps_command *)data;
|
|
|
|
sc = cm->cm_sc;
|
|
|
|
mtx_assert(&sc->mps_mtx, MA_OWNED);
|
|
|
|
|
|
|
|
mps_dprint(sc, MPS_INFO, "%s checking ATA ID command %p sc %p\n",
|
|
|
|
__func__, cm, sc);
|
|
|
|
if ((callout_pending(&cm->cm_callout)) ||
|
|
|
|
(!callout_active(&cm->cm_callout))) {
|
|
|
|
mps_dprint(sc, MPS_INFO, "%s ATA ID command almost timed out\n",
|
|
|
|
__func__);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
callout_deactivate(&cm->cm_callout);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Run the interrupt handler to make sure it's not pending. This
|
|
|
|
* isn't perfect because the command could have already completed
|
|
|
|
* and been re-used, though this is unlikely.
|
|
|
|
*/
|
|
|
|
mps_intr_locked(sc);
|
|
|
|
if (cm->cm_state == MPS_CM_STATE_FREE) {
|
|
|
|
mps_dprint(sc, MPS_INFO, "%s ATA ID command almost timed out\n",
|
|
|
|
__func__);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
mps_dprint(sc, MPS_INFO, "ATA ID command timeout cm %p\n", cm);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Send wakeup() to the sleeping thread that issued this ATA ID command.
|
|
|
|
* wakeup() will cause msleep to return a 0 (not EWOULDBLOCK), and this
|
|
|
|
* will keep reinit() from being called. This way, an Abort Task TM can
|
|
|
|
* be issued so that the timed out command can be cleared. The Abort
|
|
|
|
* Task cannot be sent from here because the driver has not completed
|
|
|
|
* setting up targets. Instead, the command is flagged so that special
|
|
|
|
* handling will be used to send the abort.
|
|
|
|
*/
|
|
|
|
cm->cm_flags |= MPS_CM_FLAGS_SATA_ID_TIMEOUT;
|
|
|
|
wakeup(cm);
|
|
|
|
}
|
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
static int
|
2012-02-09 00:16:12 +00:00
|
|
|
mpssas_volume_add(struct mps_softc *sc, u16 handle)
|
2012-01-26 18:17:21 +00:00
|
|
|
{
|
|
|
|
struct mpssas_softc *sassc;
|
|
|
|
struct mpssas_target *targ;
|
|
|
|
u64 wwid;
|
|
|
|
unsigned int id;
|
|
|
|
int error = 0;
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
struct mpssas_lun *lun;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
sassc = sc->sassc;
|
|
|
|
mpssas_startup_increment(sassc);
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
/* wwid is endian safe */
|
2012-01-26 18:17:21 +00:00
|
|
|
mps_config_get_volume_wwid(sc, handle, &wwid);
|
|
|
|
if (!wwid) {
|
|
|
|
printf("%s: invalid WWID; cannot add volume to mapping table\n",
|
|
|
|
__func__);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
id = mps_mapping_get_raid_id(sc, wwid, handle);
|
|
|
|
if (id == MPS_MAP_BAD_ID) {
|
|
|
|
printf("%s: could not get ID for volume with handle 0x%04x and "
|
|
|
|
"WWID 0x%016llx\n", __func__, handle,
|
|
|
|
(unsigned long long)wwid);
|
|
|
|
error = ENXIO;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
targ = &sassc->targets[id];
|
|
|
|
targ->tid = id;
|
|
|
|
targ->handle = handle;
|
|
|
|
targ->devname = wwid;
|
|
|
|
TAILQ_INIT(&targ->commands);
|
|
|
|
TAILQ_INIT(&targ->timedout_commands);
|
Bring in LSI's latest mps(4) 6Gb SAS and WarpDrive driver, version
14.00.00.01-fbsd.
Their description of the changes is as follows:
1. Copyright contents has been changed in all respective .c
and .h files
2. Support for WRITE12 and READ12 for direct-io (warpdrive only)
has been added.
3. Driver has added checks to see if Drive has READ_CAP_16
support before sending it down to the device.
If SPC3_SID_PROTECT flag is set in the inquiry data, the
device supports protection information, and must support
the 16 byte read capacity command, otherwise continue without
sending read cap 16. This will optimize driver performance,
since it will not send READ_CAP_16 to the drive which does
not have support of READ_CAP_16.
4. With new approach, "MPTIOCTL_RESET_ADAPTER" IOCTL will not
use DELAY() which is busy loop implementation.
It will use <msleep> (Better way to sleep without busy
loop). Also from the HBA reset code path and some other
places, DELAY() is replaced with msleep() or "pause()",
which is based on sleep/wakeup style calls. Driver use
msleep()/pause() instead of DELAY based on CAN_SLEEP/NO_SLEEP
flags to avoid busy loop which is not required all the
time.e.a
a. While driver is getting loaded, driver calls most of the
commands with NO_SLEEP.
b. When Driver is functional and it needs Reinit of HBA,
CAN_SLEEP flag is used.
5. <mpslsi> driver is not Endian safe. It will not work on Big
Endian machines like Sparc and PowerPC platforms because it
assumes it is running on a Little Endian machine.
Driver code is modified such way that it does not assume CPU
arch is Little Endian.
a. All places where Driver interacts from HBA to Host, it
converts Little Endian format to CPU format.
b. All places where Driver interacts from Host to HBA, it
converts CPU format to Little Endian.
6. Findout memory leaks in FreeBSD Driver and resolve those,
such as memory leak in targ's luns creation/deletion.
Also added additional checks to see memory allocation
success/fail.
7. Add loginfo prints as debug message, i.e. When FW sends any
loginfo, Driver should print those as debug message.
This will help for debugging purpose.
8. There is possibility to get config request timeout. Current
driver is able to detect config request timetout, but it does
not do anything on config_request timeout. Driver should
call mps_reinit() if any request_poll (which is called as
part of config_request) is time out.
9. cdb length check is required for 32 byte CDB. Add correct mpi
control value for 32 bit CDB as below while submitting SCSI IO
Request to controller.
mpi_control |= 4 << MPI2_SCSIIO_CONTROL_ADDCDBLEN_SHIFT;
10. Check the actual status of Message unit reset
(mps_message_unit_reset).Previously FreeBSD Driver just writes
MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET and never check the ack
(it just wait for 50 millisecond). So, Driver now check the
status of "MPI2_FUNCTION_IOC_MESSAGE_UNIT_RESET" after writing
it to the FW.
Now it also checking for whether doorbell ack uses msleep with
proper sleep flags, instead of <DELAY>.
11. Previously CAM does not detect Multi-Lun Devices. In order to
detect Multi-Lun Devices by CAM the driver needs following change
set:
a. There is "max_lun" field which Driver need to set based on
hw/fw support. Currently LSI released driver does not set
this field.
b. Default of "max_lun" should not be 0 in OS, but it is
currently set to 0 in CAM layer.
c. Export max_lun capacity to 255
12. Driver will not reset target info after port enable complete and
also do Device removal when Device remove from FW. The detail
description is as follows
a. When Driver receive WD PD add events, it will add all
information in driver local data structure.
b. Only for WD, we have below checks after port enable
completes, where driver clear off all information retrieved
at #1.
if ((sc->WD_available &&
(sc->WD_hide_expose == MPS_WD_HIDE_ALWAYS)) ||
(sc->WD_valid_config && (sc->WD_hide_expose ==
MPS_WD_HIDE_IF_VOLUME)) {
// clear off target data structure.
}
It is mainly not to attach PDs to OS.
FreeBSD does bus rescan as older Parallel scsi style. So Driver
needs to handle which Drive is visible to OS. That is a reason
we have to clear off targ information for PDs.
Again, above logic was implemented long time ago. Similar concept
we have for non-wd also. For that, LSI have introduced different
logic to hide PDs.
Eventually, because of above gap, when Phy goes offline, we
observe below failure. That is what Driver is not doing complete
removal of device with FW. (which was pointed by Scott)
Apr 5 02:39:24 Freebsd7 kernel: mpslsi0: mpssas_prepare_remove
Apr 5 02:39:24 Freebsd7 kernel: mpssas_prepare_remove 497 : invalid handle 0xe
Now Driver will not reset target info after port enable complete
and also will do Device removal when Device remove from FW.
13. Returning "CAM_SEL_TIMEOUT" instead of "CAM_TID_INVALID"
error code on request to the Target IDs that have no devices
conected at that moment. As if "CAM_TID_INVALID" error code
is returned to the CAM Layaer then it results in a huge chain
of errors in verbose kernel messages on boot and every
hot-plug event.
Submitted by: Sreekanth Reddy <Sreekanth.Reddy@lsi.com>
MFC after: 3 days
2012-06-28 03:48:54 +00:00
|
|
|
while(!SLIST_EMPTY(&targ->luns)) {
|
|
|
|
lun = SLIST_FIRST(&targ->luns);
|
|
|
|
SLIST_REMOVE_HEAD(&targ->luns, lun_link);
|
|
|
|
free(lun, M_MPT2);
|
|
|
|
}
|
2012-01-26 18:17:21 +00:00
|
|
|
SLIST_INIT(&targ->luns);
|
2013-09-12 22:06:12 +00:00
|
|
|
#if __FreeBSD_version < 1000039
|
2012-01-26 18:17:21 +00:00
|
|
|
if ((sassc->flags & MPSSAS_IN_STARTUP) == 0)
|
2013-09-12 22:06:12 +00:00
|
|
|
#endif
|
2012-01-26 18:17:21 +00:00
|
|
|
mpssas_rescan_target(sc, targ);
|
2013-07-19 00:12:41 +00:00
|
|
|
mps_dprint(sc, MPS_MAPPING, "RAID target id %d added (WWID = 0x%jx)\n",
|
2012-01-26 18:17:21 +00:00
|
|
|
targ->tid, wwid);
|
|
|
|
out:
|
|
|
|
mpssas_startup_decrement(sassc);
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
2014-07-30 17:58:17 +00:00
|
|
|
/**
|
|
|
|
* mpssas_SSU_to_SATA_devices
|
|
|
|
* @sc: per adapter object
|
|
|
|
*
|
|
|
|
* Looks through the target list and issues a StartStopUnit SCSI command to each
|
|
|
|
* SATA direct-access device. This helps to ensure that data corruption is
|
|
|
|
* avoided when the system is being shut down. This must be called after the IR
|
|
|
|
* System Shutdown RAID Action is sent if in IR mode.
|
|
|
|
*
|
|
|
|
* Return nothing.
|
|
|
|
*/
|
|
|
|
static void
|
|
|
|
mpssas_SSU_to_SATA_devices(struct mps_softc *sc)
|
|
|
|
{
|
|
|
|
struct mpssas_softc *sassc = sc->sassc;
|
|
|
|
union ccb *ccb;
|
|
|
|
path_id_t pathid = cam_sim_path(sassc->sim);
|
|
|
|
target_id_t targetid;
|
|
|
|
struct mpssas_target *target;
|
|
|
|
char path_str[64];
|
|
|
|
struct timeval cur_time, start_time;
|
|
|
|
|
|
|
|
/*
|
2015-02-24 22:07:42 +00:00
|
|
|
* For each target, issue a StartStopUnit command to stop the device.
|
2014-07-30 17:58:17 +00:00
|
|
|
*/
|
|
|
|
sc->SSU_started = TRUE;
|
|
|
|
sc->SSU_refcount = 0;
|
|
|
|
for (targetid = 0; targetid < sc->facts->MaxTargets; targetid++) {
|
|
|
|
target = &sassc->targets[targetid];
|
|
|
|
if (target->handle == 0x0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2015-02-24 22:07:42 +00:00
|
|
|
ccb = xpt_alloc_ccb_nowait();
|
|
|
|
if (ccb == NULL) {
|
|
|
|
mps_dprint(sc, MPS_FAULT, "Unable to alloc CCB to stop "
|
|
|
|
"unit.\n");
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The stop_at_shutdown flag will be set if this device is
|
|
|
|
* a SATA direct-access end device.
|
|
|
|
*/
|
|
|
|
if (target->stop_at_shutdown) {
|
|
|
|
if (xpt_create_path(&ccb->ccb_h.path,
|
|
|
|
xpt_periph, pathid, targetid,
|
|
|
|
CAM_LUN_WILDCARD) != CAM_REQ_CMP) {
|
|
|
|
mps_dprint(sc, MPS_FAULT, "Unable to create "
|
|
|
|
"LUN path to stop unit.\n");
|
|
|
|
xpt_free_ccb(ccb);
|
2014-07-30 17:58:17 +00:00
|
|
|
return;
|
|
|
|
}
|
2015-02-24 22:07:42 +00:00
|
|
|
xpt_path_string(ccb->ccb_h.path, path_str,
|
|
|
|
sizeof(path_str));
|
2014-07-30 17:58:17 +00:00
|
|
|
|
2015-02-24 22:07:42 +00:00
|
|
|
mps_dprint(sc, MPS_INFO, "Sending StopUnit: path %s "
|
|
|
|
"handle %d\n", path_str, target->handle);
|
|
|
|
|
2014-07-30 17:58:17 +00:00
|
|
|
/*
|
2015-02-24 22:07:42 +00:00
|
|
|
* Issue a START STOP UNIT command for the target.
|
|
|
|
* Increment the SSU counter to be used to count the
|
|
|
|
* number of required replies.
|
2014-07-30 17:58:17 +00:00
|
|
|
*/
|
2015-02-24 22:07:42 +00:00
|
|
|
mps_dprint(sc, MPS_INFO, "Incrementing SSU count\n");
|
|
|
|
sc->SSU_refcount++;
|
|
|
|
ccb->ccb_h.target_id =
|
|
|
|
xpt_path_target_id(ccb->ccb_h.path);
|
|
|
|
ccb->ccb_h.ppriv_ptr1 = sassc;
|
|
|
|
scsi_start_stop(&ccb->csio,
|
|
|
|
/*retries*/0,
|
|
|
|
mpssas_stop_unit_done,
|
|
|
|
MSG_SIMPLE_Q_TAG,
|
|
|
|
/*start*/FALSE,
|
|
|
|
/*load/eject*/0,
|
|
|
|
/*immediate*/FALSE,
|
|
|
|
MPS_SENSE_LEN,
|
|
|
|
/*timeout*/10000);
|
|
|
|
xpt_action(ccb);
|
2014-07-30 17:58:17 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Wait until all of the SSU commands have completed or time has
|
2015-02-24 22:07:42 +00:00
|
|
|
* expired (60 seconds). Pause for 100ms each time through. If any
|
2014-07-30 17:58:17 +00:00
|
|
|
* command times out, the target will be reset in the SCSI command
|
|
|
|
* timeout routine.
|
|
|
|
*/
|
|
|
|
getmicrotime(&start_time);
|
|
|
|
while (sc->SSU_refcount) {
|
|
|
|
pause("mpswait", hz/10);
|
|
|
|
|
|
|
|
getmicrotime(&cur_time);
|
|
|
|
if ((cur_time.tv_sec - start_time.tv_sec) > 60) {
|
|
|
|
mps_dprint(sc, MPS_FAULT, "Time has expired waiting "
|
|
|
|
"for SSU commands to complete.\n");
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
mpssas_stop_unit_done(struct cam_periph *periph, union ccb *done_ccb)
|
|
|
|
{
|
|
|
|
struct mpssas_softc *sassc;
|
|
|
|
char path_str[64];
|
|
|
|
|
|
|
|
sassc = (struct mpssas_softc *)done_ccb->ccb_h.ppriv_ptr1;
|
|
|
|
|
|
|
|
xpt_path_string(done_ccb->ccb_h.path, path_str, sizeof(path_str));
|
|
|
|
mps_dprint(sassc->sc, MPS_INFO, "Completing stop unit for %s\n",
|
|
|
|
path_str);
|
|
|
|
|
|
|
|
if (done_ccb == NULL)
|
|
|
|
return;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Nothing more to do except free the CCB and path. If the command
|
|
|
|
* timed out, an abort reset, then target reset will be issued during
|
|
|
|
* the SCSI Command process.
|
|
|
|
*/
|
|
|
|
xpt_free_path(done_ccb->ccb_h.path);
|
|
|
|
xpt_free_ccb(done_ccb);
|
|
|
|
}
|
|
|
|
|
2012-01-26 18:17:21 +00:00
|
|
|
/**
|
|
|
|
* mpssas_ir_shutdown - IR shutdown notification
|
|
|
|
* @sc: per adapter object
|
|
|
|
*
|
|
|
|
* Sending RAID Action to alert the Integrated RAID subsystem of the IOC that
|
|
|
|
* the host system is shutting down.
|
|
|
|
*
|
|
|
|
* Return nothing.
|
|
|
|
*/
|
|
|
|
void
|
|
|
|
mpssas_ir_shutdown(struct mps_softc *sc)
|
|
|
|
{
|
|
|
|
u16 volume_mapping_flags;
|
|
|
|
u16 ioc_pg8_flags = le16toh(sc->ioc_pg8.Flags);
|
|
|
|
struct dev_mapping_table *mt_entry;
|
|
|
|
u32 start_idx, end_idx;
|
|
|
|
unsigned int id, found_volume = 0;
|
|
|
|
struct mps_command *cm;
|
|
|
|
Mpi2RaidActionRequest_t *action;
|
2015-02-24 22:07:42 +00:00
|
|
|
target_id_t targetid;
|
|
|
|
struct mpssas_target *target;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
|
|
|
|
|
|
|
|
/* is IR firmware build loaded? */
|
|
|
|
if (!sc->ir_firmware)
|
2014-07-30 17:58:17 +00:00
|
|
|
goto out;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
/* are there any volumes? Look at IR target IDs. */
|
|
|
|
// TODO-later, this should be looked up in the RAID config structure
|
|
|
|
// when it is implemented.
|
|
|
|
volume_mapping_flags = le16toh(sc->ioc_pg8.IRVolumeMappingFlags) &
|
|
|
|
MPI2_IOCPAGE8_IRFLAGS_MASK_VOLUME_MAPPING_MODE;
|
|
|
|
if (volume_mapping_flags == MPI2_IOCPAGE8_IRFLAGS_LOW_VOLUME_MAPPING) {
|
|
|
|
start_idx = 0;
|
|
|
|
if (ioc_pg8_flags & MPI2_IOCPAGE8_FLAGS_RESERVED_TARGETID_0)
|
|
|
|
start_idx = 1;
|
|
|
|
} else
|
|
|
|
start_idx = sc->max_devices - sc->max_volumes;
|
|
|
|
end_idx = start_idx + sc->max_volumes - 1;
|
|
|
|
|
|
|
|
for (id = start_idx; id < end_idx; id++) {
|
|
|
|
mt_entry = &sc->mapping_table[id];
|
|
|
|
if ((mt_entry->physical_id != 0) &&
|
|
|
|
(mt_entry->missing_count == 0)) {
|
|
|
|
found_volume = 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!found_volume)
|
2014-07-30 17:58:17 +00:00
|
|
|
goto out;
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
if ((cm = mps_alloc_command(sc)) == NULL) {
|
|
|
|
printf("%s: command alloc failed\n", __func__);
|
2014-07-30 17:58:17 +00:00
|
|
|
goto out;
|
2012-01-26 18:17:21 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
action = (MPI2_RAID_ACTION_REQUEST *)cm->cm_req;
|
|
|
|
action->Function = MPI2_FUNCTION_RAID_ACTION;
|
|
|
|
action->Action = MPI2_RAID_ACTION_SYSTEM_SHUTDOWN_INITIATED;
|
|
|
|
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
|
2012-02-09 00:16:12 +00:00
|
|
|
mps_lock(sc);
|
2015-02-24 22:07:42 +00:00
|
|
|
mps_wait_command(sc, cm, 5, CAN_SLEEP);
|
2012-02-09 00:16:12 +00:00
|
|
|
mps_unlock(sc);
|
2012-01-26 18:17:21 +00:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Don't check for reply, just leave.
|
|
|
|
*/
|
|
|
|
if (cm)
|
|
|
|
mps_free_command(sc, cm);
|
2014-07-30 17:58:17 +00:00
|
|
|
|
|
|
|
out:
|
2015-02-24 22:07:42 +00:00
|
|
|
/*
|
|
|
|
* All of the targets must have the correct value set for
|
|
|
|
* 'stop_at_shutdown' for the current 'enable_ssu' sysctl variable.
|
|
|
|
*
|
|
|
|
* The possible values for the 'enable_ssu' variable are:
|
|
|
|
* 0: disable to SSD and HDD
|
|
|
|
* 1: disable only to HDD (default)
|
|
|
|
* 2: disable only to SSD
|
|
|
|
* 3: enable to SSD and HDD
|
|
|
|
* anything else will default to 1.
|
|
|
|
*/
|
|
|
|
for (targetid = 0; targetid < sc->facts->MaxTargets; targetid++) {
|
|
|
|
target = &sc->sassc->targets[targetid];
|
|
|
|
if (target->handle == 0x0) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (target->supports_SSU) {
|
|
|
|
switch (sc->enable_ssu) {
|
|
|
|
case MPS_SSU_DISABLE_SSD_DISABLE_HDD:
|
|
|
|
target->stop_at_shutdown = FALSE;
|
|
|
|
break;
|
|
|
|
case MPS_SSU_DISABLE_SSD_ENABLE_HDD:
|
|
|
|
target->stop_at_shutdown = TRUE;
|
|
|
|
if (target->flags & MPS_TARGET_IS_SATA_SSD) {
|
|
|
|
target->stop_at_shutdown = FALSE;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
case MPS_SSU_ENABLE_SSD_ENABLE_HDD:
|
|
|
|
target->stop_at_shutdown = TRUE;
|
|
|
|
break;
|
|
|
|
case MPS_SSU_ENABLE_SSD_DISABLE_HDD:
|
|
|
|
default:
|
|
|
|
target->stop_at_shutdown = TRUE;
|
|
|
|
if ((target->flags &
|
|
|
|
MPS_TARGET_IS_SATA_SSD) == 0) {
|
|
|
|
target->stop_at_shutdown = FALSE;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2014-07-30 17:58:17 +00:00
|
|
|
mpssas_SSU_to_SATA_devices(sc);
|
2012-01-26 18:17:21 +00:00
|
|
|
}
|