freebsd-skq/sys/dev/mps/mps_sas_lsi.c

1291 lines
39 KiB
C
Raw Normal View History

/*-
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
* Copyright (c) 2011-2015 LSI Corp.
* Copyright (c) 2013-2015 Avago Technologies
* 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.
*
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
* Avago Technologies (LSI) MPT-Fusion Host Adapter FreeBSD
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
/* Communications core for Avago Technologies (LSI) MPT2 */
/* 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*/
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
u16 reserved3[170]; /* 47-216 */
u16 rotational_speed; /* 217 */
u16 reserved4[38]; /* 218-255 */
};
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
static u32 event_count;
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);
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
static void mpssas_ata_id_timeout(void *data);
int mpssas_get_sas_address_for_sata_disk(struct mps_softc *sc,
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
u64 *sas_address, u16 handle, u32 device_info, u8 *is_SATA_SSD);
static int mpssas_volume_add(struct mps_softc *sc,
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
u16 handle);
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);
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;
mps_dprint(sc, MPS_EVENT, "(%d)->(%s) Working on Event: [%x]\n",
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
event_count++,__func__,fw_event->event);
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)){
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));
}
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));
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) {
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
if (mpssas_volume_add(sc, le16toh(element->VolDevHandle))){
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));
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:
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
case MPI2_EVENT_IR_CHANGE_RC_HIDE:
/*
* Phys Disk of a volume has been created. Hide
* it from the OS.
*/
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
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;
}
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
mpssas_rescan_target(sc, targ);
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)){
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));
}
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.
*/
mps_dprint(sc, MPS_EVENT, "Received IR Volume event:\n");
switch (event_data->ReasonCode) {
case MPI2_EVENT_IR_VOLUME_RC_SETTINGS_CHANGED:
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));
break;
case MPI2_EVENT_IR_VOLUME_RC_STATUS_FLAGS_CHANGED:
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));
break;
case MPI2_EVENT_IR_VOLUME_RC_STATE_CHANGED:
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));
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
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;
}
break;
default:
break;
}
break;
}
case MPI2_EVENT_IR_PHYSICAL_DISK:
{
Mpi2EventDataIrPhysicalDisk_t *event_data =
fw_event->event_data;
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
struct mpssas_target *targ;
/*
* Informational only.
*/
mps_dprint(sc, MPS_EVENT, "Received IR Phys Disk event:\n");
switch (event_data->ReasonCode) {
case MPI2_EVENT_IR_PHYSDISK_RC_SETTINGS_CHANGED:
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));
break;
case MPI2_EVENT_IR_PHYSDISK_RC_STATUS_FLAGS_CHANGED:
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));
break;
case MPI2_EVENT_IR_PHYSDISK_RC_STATE_CHANGED:
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));
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
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);
}
}
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
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;
}
default:
break;
}
break;
}
case MPI2_EVENT_IR_OPERATION_STATUS:
{
Mpi2EventDataIrOperationStatus_t *event_data =
fw_event->event_data;
/*
* Informational only.
*/
mps_dprint(sc, MPS_EVENT, "Received IR Op Status event:\n");
mps_dprint(sc, MPS_EVENT, " RAID Operation of %d is %d "
"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));
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;
}
mps_dprint(sc, MPS_EVENT, "(%d)->(%s) Event Free: [%x]\n",event_count,__func__, fw_event->event);
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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
uint64_t sas_address;
uint64_t parent_sas_address = 0;
u32 device_info, parent_devinfo = 0;
unsigned int id;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
u8 is_SATA_SSD = 0;
struct mps_command *cm;
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)) {
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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
sas_address = (sas_address << 32) | config_page.SASAddress.Low;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
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;
}
if (mpssas_check_id(sassc, id) != 0) {
device_printf(sc->mps_dev, "Excluding target id %d\n", id);
error = ENXIO;
goto out;
}
mps_dprint(sc, MPS_MAPPING, "SAS Address from SAS device page0 = %jx\n",
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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
if (is_SATA_SSD) {
targ->flags = MPS_TARGET_IS_SATA_SSD;
}
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);
}
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
mps_describe_devinfo(targ->devinfo, devstring, 80);
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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),
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
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
mpssas_rescan_target(sc, targ);
mps_dprint(sc, MPS_MAPPING, "Target id 0x%x added\n", targ->tid);
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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;
}
}
out:
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
}
}
mpssas_startup_decrement(sassc);
return (error);
}
int
mpssas_get_sas_address_for_sata_disk(struct mps_softc *sc,
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
u64 *sas_address, u16 handle, u32 device_info, u8 *is_SATA_SSD)
{
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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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) &&
(try_count < 5));
if (rc == 0 && !ioc_status && !sas_status) {
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
} else {
mps_dprint(sc, MPS_MAPPING, "%s: handle = 0x%x failed\n",
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
__func__, handle);
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] =
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
((u8 *)ata_identify.serial_number)[i + 1];
buffer[MPT2SAS_MN_LEN + i + 1] =
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
((u8 *)ata_identify.serial_number)[i];
}
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];
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
if (ata_identify.rotational_speed == 1) {
*is_SATA_SSD = 1;
}
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;
if ((cm = mps_alloc_command(sc)) == NULL) {
free(buffer, M_MPT2);
return (EBUSY);
}
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);
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
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",
__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) !=
MPI2_IOCSTATUS_SUCCESS) {
printf("%s: error reading SATA PASSTHRU; iocstatus = 0x%x\n",
__func__, reply->IOCStatus);
error = ENXIO;
goto out;
}
out:
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
return (error);
}
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
}
static int
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
mpssas_volume_add(struct mps_softc *sc, u16 handle)
{
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;
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 */
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);
}
SLIST_INIT(&targ->luns);
#if __FreeBSD_version < 1000039
if ((sassc->flags & MPSSAS_IN_STARTUP) == 0)
#endif
mpssas_rescan_target(sc, targ);
mps_dprint(sc, MPS_MAPPING, "RAID target id %d added (WWID = 0x%jx)\n",
targ->tid, wwid);
out:
mpssas_startup_decrement(sassc);
return (error);
}
/**
* 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;
/*
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
* For each target, issue a StartStopUnit command to stop the device.
*/
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;
}
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
return;
}
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
xpt_path_string(ccb->ccb_h.path, path_str,
sizeof(path_str));
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
mps_dprint(sc, MPS_INFO, "Sending StopUnit: path %s "
"handle %d\n", path_str, target->handle);
/*
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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.
*/
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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);
}
}
/*
* Wait until all of the SSU commands have completed or time has
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
* expired (60 seconds). Pause for 100ms each time through. If any
* 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);
}
/**
* 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;
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
target_id_t targetid;
struct mpssas_target *target;
mps_dprint(sc, MPS_TRACE, "%s\n", __func__);
/* is IR firmware build loaded? */
if (!sc->ir_firmware)
goto out;
/* 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)
goto out;
if ((cm = mps_alloc_command(sc)) == NULL) {
printf("%s: command alloc failed\n", __func__);
goto out;
}
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;
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
mps_lock(sc);
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
2015-02-24 22:07:42 +00:00
mps_wait_command(sc, cm, 5, CAN_SLEEP);
Bring in a number of mps(4) driver fixes from LSI: 1. Fixed timeout specification for the msleep in mps_wait_command(). Added 30 second timeout for mps_wait_command() calls in mps_user.c. 2. Make sure we call mps_detach_user() from the kldunload path. 3. Raid Hotplug behavior change. The driver now removes a volume when it goes to a failed state, so we also need to add volume back to the OS when it goes to opitimal/degraded/online from failed/missing. Handle raid volume add and remove from the IR_Volume event. 4. Added some more debugging information. 5. Replace xpt_async(AC_LOST_DEVICE, path, NULL) with mpssas_rescan_target(). This is to work around a panic in CAM that shows up when adding a drive with a rescan and removing another device from the driver thread with an AC_LOST_DEVICE async notification. This problem was encountered in testing with the LSI sas2ircu utility, which was used to create a RAID volume from physical disks. The driver has to create the RAID volume target and remove the physical disk targets, and triggered a panic in the process. The CAM issue needs to be fully diagnosed and fixed, but this works around the issue for now. 6. Fix some memory initialization issues in mps_free_command(). 7. Resolve the "devq freeze forever" issue. This was caused by the internal read capacity command issued in the non-head version of the driver. When the command completed with an error, the driver wasn't unfreezing thd device queue. The version in head uses the CAM infrastructure for getting the read capacity information, and therefore doesn't have the same issue. 8. Bump the version to 13.00.00.00-fbsd. (this is very close to LSI's internal stable driver 13.00.00.00) Submitted by: Kashyap Desai <Kashyap.Desai@lsi.com> MFC after: 3 days
2012-02-09 00:16:12 +00:00
mps_unlock(sc);
/*
* Don't check for reply, just leave.
*/
if (cm)
mps_free_command(sc, cm);
out:
- Updated all files with 2015 Avago copyright, and updated LSI's copyright dates. - Changed all of the PCI device strings from LSI to Avago Technologies (LSI). - Added a sysctl variable to control how StartStopUnit behavior works. User can select to spin down disks based on if disk is SSD or HDD. - Inquiry data is required to tell if a disk will support SSU at shutdown or not. Due to the addition of mpssas_async, which gets Advanced Info but not Inquiry data, the setting of supports_SSU was moved to the mpssas_scsiio_complete function, which snoops for any Inquiry commands. And, since disks are shutdown as a target and not a LUN, this process was simplified by basing it on targets and not LUNs. - Added a sysctl variable that sets the amount of time to retry after sending a failed SATA ID command. This helps with some bad disks and large disks that require a lot of time to spin up. Part of this change was to add a callout to handle timeouts with the SATA ID command. The callout function is called mpssas_ata_id_timeout(). (Fixes PR 191348) - Changed the way resets work by allowing I/O to continue to devices that are not currently under a reset condition. This uses devq's instead of simq's and makes use of the MPSSAS_TARGET_INRESET flag. This change also adds a function called mpssas_prepare_tm(). - Some changes were made to reduce code duplication when getting a SAS address for a SATA disk. - Fixed some formatting and whitespace. - Bump version of mps driver to 20.00.00.00-fbsd PR: 191348 Reviewed by: ken, scottl Approved by: ken, scottl MFC after: 2 weeks
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;
}
}
}
mpssas_SSU_to_SATA_devices(sc);
}