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

2528 lines
68 KiB
C
Raw Normal View History

/*-
* SPDX-License-Identifier: BSD-3-Clause
*
* Copyright (c) 2008 Yahoo!, Inc.
* All rights reserved.
* Written by: John Baldwin <jhb@FreeBSD.org>
*
* 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.
* 3. Neither the name of the author nor the names of any co-contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* 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 userland interface
*/
/*-
- 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
*
* $FreeBSD$
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/* 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/ioccom.h>
#include <sys/endian.h>
#include <sys/queue.h>
#include <sys/kthread.h>
#include <sys/taskqueue.h>
#include <sys/proc.h>
#include <sys/sysent.h>
#include <machine/bus.h>
#include <machine/resource.h>
#include <sys/rman.h>
#include <cam/cam.h>
- 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
#include <cam/cam_ccb.h>
#include <cam/scsi/scsi_all.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_cnfg.h>
#include <dev/mps/mpi/mpi2_init.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>
- 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
#include <dev/mps/mps_sas.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
static d_open_t mps_open;
static d_close_t mps_close;
static d_ioctl_t mps_ioctl_devsw;
static struct cdevsw mps_cdevsw = {
.d_version = D_VERSION,
.d_flags = 0,
.d_open = mps_open,
.d_close = mps_close,
.d_ioctl = mps_ioctl_devsw,
.d_name = "mps",
};
typedef int (mps_user_f)(struct mps_command *, struct mps_usr_command *);
static mps_user_f mpi_pre_ioc_facts;
static mps_user_f mpi_pre_port_facts;
static mps_user_f mpi_pre_fw_download;
static mps_user_f mpi_pre_fw_upload;
static mps_user_f mpi_pre_sata_passthrough;
static mps_user_f mpi_pre_smp_passthrough;
static mps_user_f mpi_pre_config;
static mps_user_f mpi_pre_sas_io_unit_control;
static int mps_user_read_cfg_header(struct mps_softc *,
struct mps_cfg_page_req *);
static int mps_user_read_cfg_page(struct mps_softc *,
struct mps_cfg_page_req *, void *);
static int mps_user_read_extcfg_header(struct mps_softc *,
struct mps_ext_cfg_page_req *);
static int mps_user_read_extcfg_page(struct mps_softc *,
struct mps_ext_cfg_page_req *, void *);
static int mps_user_write_cfg_page(struct mps_softc *,
struct mps_cfg_page_req *, void *);
static int mps_user_setup_request(struct mps_command *,
struct mps_usr_command *);
static int mps_user_command(struct mps_softc *, struct mps_usr_command *);
static int mps_user_pass_thru(struct mps_softc *sc, mps_pass_thru_t *data);
static void mps_user_get_adapter_data(struct mps_softc *sc,
mps_adapter_data_t *data);
static void mps_user_read_pci_info(struct mps_softc *sc,
mps_pci_info_t *data);
static uint8_t mps_get_fw_diag_buffer_number(struct mps_softc *sc,
uint32_t unique_id);
static int mps_post_fw_diag_buffer(struct mps_softc *sc,
mps_fw_diagnostic_buffer_t *pBuffer, uint32_t *return_code);
static int mps_release_fw_diag_buffer(struct mps_softc *sc,
mps_fw_diagnostic_buffer_t *pBuffer, uint32_t *return_code,
uint32_t diag_type);
static int mps_diag_register(struct mps_softc *sc,
mps_fw_diag_register_t *diag_register, uint32_t *return_code);
static int mps_diag_unregister(struct mps_softc *sc,
mps_fw_diag_unregister_t *diag_unregister, uint32_t *return_code);
static int mps_diag_query(struct mps_softc *sc, mps_fw_diag_query_t *diag_query,
uint32_t *return_code);
static int mps_diag_read_buffer(struct mps_softc *sc,
mps_diag_read_buffer_t *diag_read_buffer, uint8_t *ioctl_buf,
uint32_t *return_code);
static int mps_diag_release(struct mps_softc *sc,
mps_fw_diag_release_t *diag_release, uint32_t *return_code);
static int mps_do_diag_action(struct mps_softc *sc, uint32_t action,
uint8_t *diag_action, uint32_t length, uint32_t *return_code);
static int mps_user_diag_action(struct mps_softc *sc, mps_diag_action_t *data);
static void mps_user_event_query(struct mps_softc *sc, mps_event_query_t *data);
static void mps_user_event_enable(struct mps_softc *sc,
mps_event_enable_t *data);
static int mps_user_event_report(struct mps_softc *sc,
mps_event_report_t *data);
static int mps_user_reg_access(struct mps_softc *sc, mps_reg_access_t *data);
static int mps_user_btdh(struct mps_softc *sc, mps_btdh_mapping_t *data);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
MALLOC_DEFINE(M_MPSUSER, "mps_user", "Buffers for mps(4) ioctls");
/* Macros from compat/freebsd32/freebsd32.h */
#define PTRIN(v) (void *)(uintptr_t)(v)
#define PTROUT(v) (uint32_t)(uintptr_t)(v)
#define CP(src,dst,fld) do { (dst).fld = (src).fld; } while (0)
#define PTRIN_CP(src,dst,fld) \
do { (dst).fld = PTRIN((src).fld); } while (0)
#define PTROUT_CP(src,dst,fld) \
do { (dst).fld = PTROUT((src).fld); } while (0)
int
mps_attach_user(struct mps_softc *sc)
{
int unit;
unit = device_get_unit(sc->mps_dev);
sc->mps_cdev = make_dev(&mps_cdevsw, unit, UID_ROOT, GID_OPERATOR, 0640,
"mps%d", unit);
if (sc->mps_cdev == NULL) {
return (ENOMEM);
}
sc->mps_cdev->si_drv1 = sc;
return (0);
}
void
mps_detach_user(struct mps_softc *sc)
{
/* XXX: do a purge of pending requests? */
if (sc->mps_cdev != NULL)
destroy_dev(sc->mps_cdev);
}
static int
mps_open(struct cdev *dev, int flags, int fmt, struct thread *td)
{
return (0);
}
static int
mps_close(struct cdev *dev, int flags, int fmt, struct thread *td)
{
return (0);
}
static int
mps_user_read_cfg_header(struct mps_softc *sc,
struct mps_cfg_page_req *page_req)
{
MPI2_CONFIG_PAGE_HEADER *hdr;
struct mps_config_params params;
int error;
hdr = &params.hdr.Struct;
params.action = MPI2_CONFIG_ACTION_PAGE_HEADER;
params.page_address = le32toh(page_req->page_address);
hdr->PageVersion = 0;
hdr->PageLength = 0;
hdr->PageNumber = page_req->header.PageNumber;
hdr->PageType = page_req->header.PageType;
params.buffer = NULL;
params.length = 0;
params.callback = NULL;
if ((error = mps_read_config_page(sc, &params)) != 0) {
/*
* Leave the request. Without resetting the chip, it's
* still owned by it and we'll just get into trouble
* freeing it now. Mark it as abandoned so that if it
* shows up later it can be freed.
*/
mps_printf(sc, "read_cfg_header timed out\n");
return (ETIMEDOUT);
}
page_req->ioc_status = htole16(params.status);
if ((page_req->ioc_status & MPI2_IOCSTATUS_MASK) ==
MPI2_IOCSTATUS_SUCCESS) {
bcopy(hdr, &page_req->header, sizeof(page_req->header));
}
return (0);
}
static int
mps_user_read_cfg_page(struct mps_softc *sc, struct mps_cfg_page_req *page_req,
void *buf)
{
MPI2_CONFIG_PAGE_HEADER *reqhdr, *hdr;
struct mps_config_params params;
int error;
reqhdr = buf;
hdr = &params.hdr.Struct;
hdr->PageVersion = reqhdr->PageVersion;
hdr->PageLength = reqhdr->PageLength;
hdr->PageNumber = reqhdr->PageNumber;
hdr->PageType = reqhdr->PageType & MPI2_CONFIG_PAGETYPE_MASK;
params.action = MPI2_CONFIG_ACTION_PAGE_READ_CURRENT;
params.page_address = le32toh(page_req->page_address);
params.buffer = buf;
params.length = le32toh(page_req->len);
params.callback = NULL;
if ((error = mps_read_config_page(sc, &params)) != 0) {
mps_printf(sc, "mps_user_read_cfg_page timed out\n");
return (ETIMEDOUT);
}
page_req->ioc_status = htole16(params.status);
return (0);
}
static int
mps_user_read_extcfg_header(struct mps_softc *sc,
struct mps_ext_cfg_page_req *ext_page_req)
{
MPI2_CONFIG_EXTENDED_PAGE_HEADER *hdr;
struct mps_config_params params;
int error;
hdr = &params.hdr.Ext;
params.action = MPI2_CONFIG_ACTION_PAGE_HEADER;
hdr->PageVersion = ext_page_req->header.PageVersion;
hdr->PageType = MPI2_CONFIG_PAGETYPE_EXTENDED;
hdr->ExtPageLength = 0;
hdr->PageNumber = ext_page_req->header.PageNumber;
hdr->ExtPageType = ext_page_req->header.ExtPageType;
params.page_address = le32toh(ext_page_req->page_address);
params.buffer = NULL;
params.length = 0;
params.callback = NULL;
if ((error = mps_read_config_page(sc, &params)) != 0) {
/*
* Leave the request. Without resetting the chip, it's
* still owned by it and we'll just get into trouble
* freeing it now. Mark it as abandoned so that if it
* shows up later it can be freed.
*/
mps_printf(sc, "mps_user_read_extcfg_header timed out\n");
return (ETIMEDOUT);
}
ext_page_req->ioc_status = htole16(params.status);
if ((ext_page_req->ioc_status & MPI2_IOCSTATUS_MASK) ==
MPI2_IOCSTATUS_SUCCESS) {
ext_page_req->header.PageVersion = hdr->PageVersion;
ext_page_req->header.PageNumber = hdr->PageNumber;
ext_page_req->header.PageType = hdr->PageType;
ext_page_req->header.ExtPageLength = hdr->ExtPageLength;
ext_page_req->header.ExtPageType = hdr->ExtPageType;
}
return (0);
}
static int
mps_user_read_extcfg_page(struct mps_softc *sc,
struct mps_ext_cfg_page_req *ext_page_req, void *buf)
{
MPI2_CONFIG_EXTENDED_PAGE_HEADER *reqhdr, *hdr;
struct mps_config_params params;
int error;
reqhdr = buf;
hdr = &params.hdr.Ext;
params.action = MPI2_CONFIG_ACTION_PAGE_READ_CURRENT;
params.page_address = le32toh(ext_page_req->page_address);
hdr->PageVersion = reqhdr->PageVersion;
hdr->PageType = MPI2_CONFIG_PAGETYPE_EXTENDED;
hdr->PageNumber = reqhdr->PageNumber;
hdr->ExtPageType = reqhdr->ExtPageType;
hdr->ExtPageLength = reqhdr->ExtPageLength;
params.buffer = buf;
params.length = le32toh(ext_page_req->len);
params.callback = NULL;
if ((error = mps_read_config_page(sc, &params)) != 0) {
mps_printf(sc, "mps_user_read_extcfg_page timed out\n");
return (ETIMEDOUT);
}
ext_page_req->ioc_status = htole16(params.status);
return (0);
}
static int
mps_user_write_cfg_page(struct mps_softc *sc,
struct mps_cfg_page_req *page_req, void *buf)
{
MPI2_CONFIG_PAGE_HEADER *reqhdr, *hdr;
struct mps_config_params params;
u_int hdr_attr;
int error;
reqhdr = buf;
hdr = &params.hdr.Struct;
hdr_attr = reqhdr->PageType & MPI2_CONFIG_PAGEATTR_MASK;
if (hdr_attr != MPI2_CONFIG_PAGEATTR_CHANGEABLE &&
hdr_attr != MPI2_CONFIG_PAGEATTR_PERSISTENT) {
mps_printf(sc, "page type 0x%x not changeable\n",
reqhdr->PageType & MPI2_CONFIG_PAGETYPE_MASK);
return (EINVAL);
}
/*
* There isn't any point in restoring stripped out attributes
* if you then mask them going down to issue the request.
*/
hdr->PageVersion = reqhdr->PageVersion;
hdr->PageLength = reqhdr->PageLength;
hdr->PageNumber = reqhdr->PageNumber;
hdr->PageType = reqhdr->PageType;
params.action = MPI2_CONFIG_ACTION_PAGE_WRITE_CURRENT;
params.page_address = le32toh(page_req->page_address);
params.buffer = buf;
params.length = le32toh(page_req->len);
params.callback = NULL;
if ((error = mps_write_config_page(sc, &params)) != 0) {
mps_printf(sc, "mps_write_cfg_page timed out\n");
return (ETIMEDOUT);
}
page_req->ioc_status = htole16(params.status);
return (0);
}
Add Serial Management Protocol (SMP) passthrough support to CAM. This includes support in the kernel, camcontrol(8), libcam and the mps(4) driver for SMP passthrough. The CAM SCSI probe code has been modified to fetch Inquiry VPD page 0x00 to determine supported pages, and will now fetch page 0x83 in addition to page 0x80 if supported. Add two new CAM CCBs, XPT_SMP_IO, and XPT_GDEV_ADVINFO. The SMP CCB is intended for SMP requests and responses. The ADVINFO is currently used to fetch cached VPD page 0x83 data from the transport layer, but is intended to be extensible to fetch other types of device-specific data. SMP-only devices are not currently represented in the CAM topology, and so the current semantics are that the SIM will route SMP CCBs to either the addressed device, if it contains an SMP target, or its parent, if it contains an SMP target. (This is noted in cam_ccb.h, since it will change later once we have the ability to have SMP-only devices in CAM's topology.) smp_all.c, smp_all.h: New helper routines for SMP. This includes SMP request building routines, response parsing routines, error decoding routines, and structure definitions for a number of SMP commands. libcam/Makefile: Add smp_all.c to libcam, so that SMP functionality is available to userland applications. camcontrol.8, camcontrol.c: Add smp passthrough support to camcontrol. Several new subcommands are now available: 'smpcmd' functions much like 'cmd', except that it allows the user to send generic SMP commands. 'smprg' sends the SMP report general command, and displays the decoded output. It will automatically fetch extended output if it is available. 'smppc' sends the SMP phy control command, with any number of potential options. Among other things, this allows the user to reset a phy on a SAS expander, or disable a phy on an expander. 'smpmaninfo' sends the SMP report manufacturer information and displays the decoded output. 'smpphylist' displays a list of phys on an expander, and the CAM devices attached to those phys, if any. cam.h, cam.c: Add a status value for SMP errors (CAM_SMP_STATUS_ERROR). Add a missing description for CAM_SCSI_IT_NEXUS_LOST. Add support for SMP commands to cam_error_string(). cam_ccb.h: Rename the CAM_DIR_RESV flag to CAM_DIR_BOTH. SMP commands are by nature bi-directional, and we may need to support bi-directional SCSI commands later. Add the XPT_SMP_IO CCB. Since SMP commands are bi-directional, there are pointers for both the request and response. Add a fill routine for SMP CCBs. Add the XPT_GDEV_ADVINFO CCB. This is currently used to fetch cached page 0x83 data from the transport later, but is extensible to fetch many other types of data. cam_periph.c: Add support in cam_periph_mapmem() for XPT_SMP_IO and XPT_GDEV_ADVINFO CCBs. cam_xpt.c: Add support for executing XPT_SMP_IO CCBs. cam_xpt_internal.h: Add fields for VPD pages 0x00 and 0x83 in struct cam_ed. scsi_all.c: Add scsi_get_sas_addr(), a function that parses VPD page 0x83 data and pulls out a SAS address. scsi_all.h: Add VPD page 0x00 and 0x83 structures, and a prototype for scsi_get_sas_addr(). scsi_pass.c: Add support for mapping buffers in XPT_SMP_IO and XPT_GDEV_ADVINFO CCBs. scsi_xpt.c: In the SCSI probe code, first ask the device for VPD page 0x00. If any VPD pages are supported, that page is required to be implemented. Based on the response, we may probe for the serial number (page 0x80) or device id (page 0x83). Add support for the XPT_GDEV_ADVINFO CCB. sys/conf/files: Add smp_all.c. mps.c: Add support for passing in a uio in mps_map_command(), so we can map a S/G list at once. Add support for SMP passthrough commands in mps_data_cb(). SMP is a special case, because the first buffer in the S/G list is outbound and the second buffer is inbound. Add support for warning the user if the busdma code comes back with more buffers than will work for the command. This will, for example, help the user determine why an SMP command failed if busdma comes back with three buffers. mps_pci.c: Add sys/uio.h. mps_sas.c: Add the SAS address and the parent handle to the list of fields we pull from device page 0 and cache in struct mpssas_target. These are needed for SMP passthrough. Add support for the XPT_SMP_IO CCB. For now, this CCB is routed to the addressed device if it supports SMP, or to its parent if it does not and the parent does. This is necessary because CAM does not currently support SMP-only nodes in the topology. Make SMP passthrough support conditional on __FreeBSD_version >= 900026. This will make it easier to MFC this change to the driver without MFCing the CAM changes as well. mps_user.c: Un-staticize mpi_init_sge() so we can use it for the SMP passthrough code. mpsvar.h: Add a uio and iovecs into struct mps_command for SMP passthrough commands. Add a cm_max_segs field to struct mps_command so that we can warn the user if busdma comes back with too many segments. Clear the cm_reply when a command gets freed. If it is not cleared, reply frames will eventually get freed into the pool multiple times and corrupt the pool. (This fix is from scottl.) Add a prototype for mpi_init_sge(). sys/param.h: Bump __FreeBSD_version to 900026 for the for the inclusion of the XPT_GDEV_ADVINFO and XPT_SMP_IO CAM CCBs.
2010-11-30 22:39:46 +00:00
void
mpi_init_sge(struct mps_command *cm, void *req, void *sge)
{
int off, space;
space = (int)cm->cm_sc->reqframesz;
off = (uintptr_t)sge - (uintptr_t)req;
KASSERT(off < space, ("bad pointers %p %p, off %d, space %d",
req, sge, off, space));
cm->cm_sge = sge;
cm->cm_sglsize = space - off;
}
/*
* Prepare the mps_command for an IOC_FACTS request.
*/
static int
mpi_pre_ioc_facts(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_IOC_FACTS_REQUEST *req = (void *)cm->cm_req;
MPI2_IOC_FACTS_REPLY *rpl;
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
cm->cm_sge = NULL;
cm->cm_sglsize = 0;
return (0);
}
/*
* Prepare the mps_command for a PORT_FACTS request.
*/
static int
mpi_pre_port_facts(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_PORT_FACTS_REQUEST *req = (void *)cm->cm_req;
MPI2_PORT_FACTS_REPLY *rpl;
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
cm->cm_sge = NULL;
cm->cm_sglsize = 0;
return (0);
}
/*
* Prepare the mps_command for a FW_DOWNLOAD request.
*/
static int
mpi_pre_fw_download(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_FW_DOWNLOAD_REQUEST *req = (void *)cm->cm_req;
MPI2_FW_DOWNLOAD_REPLY *rpl;
2010-10-14 16:44:44 +00:00
MPI2_FW_DOWNLOAD_TCSGE tc;
int error;
/*
* This code assumes there is room in the request's SGL for
* the TransactionContext plus at least a SGL chain element.
*/
CTASSERT(sizeof req->SGL >= sizeof tc + MPS_SGC_SIZE);
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
2010-10-14 16:44:44 +00:00
if (cmd->len == 0)
return (EINVAL);
error = copyin(cmd->buf, cm->cm_data, cmd->len);
if (error != 0)
return (error);
mpi_init_sge(cm, req, &req->SGL);
2010-10-14 16:44:44 +00:00
bzero(&tc, sizeof tc);
/*
* For now, the F/W image must be provided in a single request.
*/
if ((req->MsgFlags & MPI2_FW_DOWNLOAD_MSGFLGS_LAST_SEGMENT) == 0)
return (EINVAL);
if (req->TotalImageSize != cmd->len)
return (EINVAL);
/*
* The value of the first two elements is specified in the
* Fusion-MPT Message Passing Interface document.
*/
tc.ContextSize = 0;
tc.DetailsLength = 12;
tc.ImageOffset = 0;
tc.ImageSize = cmd->len;
cm->cm_flags |= MPS_CM_FLAGS_DATAOUT;
return (mps_push_sge(cm, &tc, sizeof tc, 0));
}
/*
* Prepare the mps_command for a FW_UPLOAD request.
*/
static int
mpi_pre_fw_upload(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_FW_UPLOAD_REQUEST *req = (void *)cm->cm_req;
MPI2_FW_UPLOAD_REPLY *rpl;
MPI2_FW_UPLOAD_TCSGE tc;
/*
* This code assumes there is room in the request's SGL for
* the TransactionContext plus at least a SGL chain element.
*/
CTASSERT(sizeof req->SGL >= sizeof tc + MPS_SGC_SIZE);
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
mpi_init_sge(cm, req, &req->SGL);
bzero(&tc, sizeof tc);
/*
* The value of the first two elements is specified in the
* Fusion-MPT Message Passing Interface document.
*/
tc.ContextSize = 0;
tc.DetailsLength = 12;
/*
* XXX Is there any reason to fetch a partial image? I.e. to
* set ImageOffset to something other than 0?
*/
tc.ImageOffset = 0;
tc.ImageSize = cmd->len;
cm->cm_flags |= MPS_CM_FLAGS_DATAIN;
return (mps_push_sge(cm, &tc, sizeof tc, 0));
}
/*
* Prepare the mps_command for a SATA_PASSTHROUGH request.
*/
static int
mpi_pre_sata_passthrough(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_SATA_PASSTHROUGH_REQUEST *req = (void *)cm->cm_req;
MPI2_SATA_PASSTHROUGH_REPLY *rpl;
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
mpi_init_sge(cm, req, &req->SGL);
return (0);
}
/*
* Prepare the mps_command for a SMP_PASSTHROUGH request.
*/
static int
mpi_pre_smp_passthrough(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_SMP_PASSTHROUGH_REQUEST *req = (void *)cm->cm_req;
MPI2_SMP_PASSTHROUGH_REPLY *rpl;
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
mpi_init_sge(cm, req, &req->SGL);
return (0);
}
/*
* Prepare the mps_command for a CONFIG request.
*/
static int
mpi_pre_config(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_CONFIG_REQUEST *req = (void *)cm->cm_req;
MPI2_CONFIG_REPLY *rpl;
if (cmd->req_len != sizeof *req)
return (EINVAL);
if (cmd->rpl_len != sizeof *rpl)
return (EINVAL);
mpi_init_sge(cm, req, &req->PageBufferSGE);
return (0);
}
/*
* Prepare the mps_command for a SAS_IO_UNIT_CONTROL request.
*/
static int
mpi_pre_sas_io_unit_control(struct mps_command *cm,
struct mps_usr_command *cmd)
{
cm->cm_sge = NULL;
cm->cm_sglsize = 0;
return (0);
}
/*
* A set of functions to prepare an mps_command for the various
* supported requests.
*/
struct mps_user_func {
U8 Function;
mps_user_f *f_pre;
} mps_user_func_list[] = {
{ MPI2_FUNCTION_IOC_FACTS, mpi_pre_ioc_facts },
{ MPI2_FUNCTION_PORT_FACTS, mpi_pre_port_facts },
{ MPI2_FUNCTION_FW_DOWNLOAD, mpi_pre_fw_download },
{ MPI2_FUNCTION_FW_UPLOAD, mpi_pre_fw_upload },
{ MPI2_FUNCTION_SATA_PASSTHROUGH, mpi_pre_sata_passthrough },
{ MPI2_FUNCTION_SMP_PASSTHROUGH, mpi_pre_smp_passthrough},
{ MPI2_FUNCTION_CONFIG, mpi_pre_config},
{ MPI2_FUNCTION_SAS_IO_UNIT_CONTROL, mpi_pre_sas_io_unit_control },
{ 0xFF, NULL } /* list end */
};
static int
mps_user_setup_request(struct mps_command *cm, struct mps_usr_command *cmd)
{
MPI2_REQUEST_HEADER *hdr = (MPI2_REQUEST_HEADER *)cm->cm_req;
struct mps_user_func *f;
for (f = mps_user_func_list; f->f_pre != NULL; f++) {
if (hdr->Function == f->Function)
return (f->f_pre(cm, cmd));
}
return (EINVAL);
}
static int
mps_user_command(struct mps_softc *sc, struct mps_usr_command *cmd)
{
MPI2_REQUEST_HEADER *hdr;
MPI2_DEFAULT_REPLY *rpl;
void *buf = NULL;
struct mps_command *cm = NULL;
int err = 0;
int sz;
mps_lock(sc);
cm = mps_alloc_command(sc);
if (cm == NULL) {
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
mps_printf(sc, "%s: no mps requests\n", __func__);
err = ENOMEM;
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
goto RetFree;
}
mps_unlock(sc);
hdr = (MPI2_REQUEST_HEADER *)cm->cm_req;
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
mps_dprint(sc, MPS_USER, "%s: req %p %d rpl %p %d\n", __func__,
cmd->req, cmd->req_len, cmd->rpl, cmd->rpl_len);
if (cmd->req_len > (int)sc->reqframesz) {
err = EINVAL;
goto RetFreeUnlocked;
}
err = copyin(cmd->req, hdr, cmd->req_len);
if (err != 0)
goto RetFreeUnlocked;
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
mps_dprint(sc, MPS_USER, "%s: Function %02X MsgFlags %02X\n", __func__,
hdr->Function, hdr->MsgFlags);
if (cmd->len > 0) {
buf = malloc(cmd->len, M_MPSUSER, M_WAITOK|M_ZERO);
cm->cm_data = buf;
cm->cm_length = cmd->len;
} else {
cm->cm_data = NULL;
cm->cm_length = 0;
}
cm->cm_flags = MPS_CM_FLAGS_SGE_SIMPLE;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
err = mps_user_setup_request(cm, cmd);
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 (err == EINVAL) {
mps_printf(sc, "%s: unsupported parameter or unsupported "
"function in request (function = 0x%X)\n", __func__,
hdr->Function);
}
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 (err != 0)
goto RetFreeUnlocked;
mps_lock(sc);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
err = mps_wait_command(sc, &cm, 60, CAN_SLEEP);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
if (err || (cm == NULL)) {
mps_printf(sc, "%s: invalid request: error %d\n",
__func__, err);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
goto RetFree;
}
rpl = (MPI2_DEFAULT_REPLY *)cm->cm_reply;
if (rpl != NULL)
sz = rpl->MsgLength * 4;
else
sz = 0;
if (sz > cmd->rpl_len) {
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
mps_printf(sc, "%s: user reply buffer (%d) smaller than "
"returned buffer (%d)\n", __func__, cmd->rpl_len, sz);
sz = cmd->rpl_len;
}
mps_unlock(sc);
copyout(rpl, cmd->rpl, sz);
if (buf != NULL)
copyout(buf, cmd->buf, cmd->len);
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
mps_dprint(sc, MPS_USER, "%s: reply size %d\n", __func__, sz);
RetFreeUnlocked:
mps_lock(sc);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
RetFree:
if (cm != NULL)
mps_free_command(sc, cm);
mps_unlock(sc);
if (buf != NULL)
free(buf, M_MPSUSER);
return (err);
}
static int
mps_user_pass_thru(struct mps_softc *sc, mps_pass_thru_t *data)
{
MPI2_REQUEST_HEADER *hdr, tmphdr;
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
MPI2_DEFAULT_REPLY *rpl = NULL;
struct mps_command *cm = NULL;
int err = 0, dir = 0, sz;
uint8_t function = 0;
u_int sense_len;
- 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
struct mpssas_target *targ = NULL;
/*
* Only allow one passthru command at a time. Use the MPS_FLAGS_BUSY
* bit to denote that a passthru is being processed.
*/
mps_lock(sc);
if (sc->mps_flags & MPS_FLAGS_BUSY) {
mps_dprint(sc, MPS_USER, "%s: Only one passthru command "
"allowed at a single time.", __func__);
mps_unlock(sc);
return (EBUSY);
}
sc->mps_flags |= MPS_FLAGS_BUSY;
mps_unlock(sc);
/*
* Do some validation on data direction. Valid cases are:
* 1) DataSize is 0 and direction is NONE
* 2) DataSize is non-zero and one of:
* a) direction is READ or
* b) direction is WRITE or
* c) direction is BOTH and DataOutSize is non-zero
* If valid and the direction is BOTH, change the direction to READ.
* if valid and the direction is not BOTH, make sure DataOutSize is 0.
*/
if (((data->DataSize == 0) &&
(data->DataDirection == MPS_PASS_THRU_DIRECTION_NONE)) ||
((data->DataSize != 0) &&
((data->DataDirection == MPS_PASS_THRU_DIRECTION_READ) ||
(data->DataDirection == MPS_PASS_THRU_DIRECTION_WRITE) ||
((data->DataDirection == MPS_PASS_THRU_DIRECTION_BOTH) &&
(data->DataOutSize != 0))))) {
if (data->DataDirection == MPS_PASS_THRU_DIRECTION_BOTH)
data->DataDirection = MPS_PASS_THRU_DIRECTION_READ;
else
data->DataOutSize = 0;
} else
return (EINVAL);
mps_dprint(sc, MPS_USER, "%s: req 0x%jx %d rpl 0x%jx %d "
"data in 0x%jx %d data out 0x%jx %d data dir %d\n", __func__,
data->PtrRequest, data->RequestSize, data->PtrReply,
data->ReplySize, data->PtrData, data->DataSize,
data->PtrDataOut, data->DataOutSize, data->DataDirection);
/*
* copy in the header so we know what we're dealing with before we
* commit to allocating a command for it.
*/
err = copyin(PTRIN(data->PtrRequest), &tmphdr, data->RequestSize);
if (err != 0)
goto RetFreeUnlocked;
if (data->RequestSize > (int)sc->reqframesz) {
err = EINVAL;
goto RetFreeUnlocked;
}
function = tmphdr.Function;
mps_dprint(sc, MPS_USER, "%s: Function %02X MsgFlags %02X\n", __func__,
function, tmphdr.MsgFlags);
/*
* Handle a passthru TM request.
*/
if (function == MPI2_FUNCTION_SCSI_TASK_MGMT) {
MPI2_SCSI_TASK_MANAGE_REQUEST *task;
mps_lock(sc);
cm = mpssas_alloc_tm(sc);
if (cm == NULL) {
err = EINVAL;
goto Ret;
}
/* Copy the header in. Only a small fixup is needed. */
task = (MPI2_SCSI_TASK_MANAGE_REQUEST *)cm->cm_req;
bcopy(&tmphdr, task, data->RequestSize);
task->TaskMID = cm->cm_desc.Default.SMID;
cm->cm_data = NULL;
cm->cm_complete = NULL;
cm->cm_complete_data = NULL;
- 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
targ = mpssas_find_target_by_handle(sc->sassc, 0,
task->DevHandle);
if (targ == NULL) {
mps_dprint(sc, MPS_INFO,
"%s %d : invalid handle for requested TM 0x%x \n",
__func__, __LINE__, task->DevHandle);
err = 1;
} else {
mpssas_prepare_for_tm(sc, cm, targ, CAM_LUN_WILDCARD);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
err = mps_wait_command(sc, &cm, 30, CAN_SLEEP);
- 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 (err != 0) {
err = EIO;
mps_dprint(sc, MPS_FAULT, "%s: task management failed",
__func__);
}
/*
* Copy the reply data and sense data to user space.
*/
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
if ((cm != NULL) && (cm->cm_reply != NULL)) {
rpl = (MPI2_DEFAULT_REPLY *)cm->cm_reply;
sz = rpl->MsgLength * 4;
if (sz > data->ReplySize) {
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
mps_printf(sc, "%s: user reply buffer (%d) "
"smaller than returned buffer (%d)\n",
__func__, data->ReplySize, sz);
}
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
mps_unlock(sc);
copyout(cm->cm_reply, PTRIN(data->PtrReply),
data->ReplySize);
mps_lock(sc);
}
mpssas_free_tm(sc, cm);
goto Ret;
}
mps_lock(sc);
cm = mps_alloc_command(sc);
if (cm == NULL) {
mps_printf(sc, "%s: no mps requests\n", __func__);
err = ENOMEM;
goto Ret;
}
mps_unlock(sc);
hdr = (MPI2_REQUEST_HEADER *)cm->cm_req;
bcopy(&tmphdr, hdr, data->RequestSize);
/*
* Do some checking to make sure the IOCTL request contains a valid
* request. Then set the SGL info.
*/
mpi_init_sge(cm, hdr, (void *)((uint8_t *)hdr + data->RequestSize));
/*
* Set up for read, write or both. From check above, DataOutSize will
* be 0 if direction is READ or WRITE, but it will have some non-zero
* value if the direction is BOTH. So, just use the biggest size to get
* the cm_data buffer size. If direction is BOTH, 2 SGLs need to be set
* up; the first is for the request and the second will contain the
* response data. cm_out_len needs to be set here and this will be used
* when the SGLs are set up.
*/
cm->cm_data = NULL;
cm->cm_length = MAX(data->DataSize, data->DataOutSize);
cm->cm_out_len = data->DataOutSize;
cm->cm_flags = 0;
if (cm->cm_length != 0) {
cm->cm_data = malloc(cm->cm_length, M_MPSUSER, M_WAITOK |
M_ZERO);
cm->cm_flags = MPS_CM_FLAGS_DATAIN;
if (data->DataOutSize) {
cm->cm_flags |= MPS_CM_FLAGS_DATAOUT;
err = copyin(PTRIN(data->PtrDataOut),
cm->cm_data, data->DataOutSize);
} else if (data->DataDirection ==
MPS_PASS_THRU_DIRECTION_WRITE) {
cm->cm_flags = MPS_CM_FLAGS_DATAOUT;
err = copyin(PTRIN(data->PtrData),
cm->cm_data, data->DataSize);
}
if (err != 0)
mps_dprint(sc, MPS_FAULT, "%s: failed to copy "
"IOCTL data from user space\n", __func__);
}
cm->cm_flags |= MPS_CM_FLAGS_SGE_SIMPLE;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
/*
* Set up Sense buffer and SGL offset for IO passthru. SCSI IO request
* uses SCSI IO descriptor.
*/
if ((function == MPI2_FUNCTION_SCSI_IO_REQUEST) ||
(function == MPI2_FUNCTION_RAID_SCSI_IO_PASSTHROUGH)) {
MPI2_SCSI_IO_REQUEST *scsi_io_req;
scsi_io_req = (MPI2_SCSI_IO_REQUEST *)hdr;
/*
* Put SGE for data and data_out buffer at the end of
* scsi_io_request message header (64 bytes in total).
* Following above SGEs, the residual space will be used by
* sense data.
*/
scsi_io_req->SenseBufferLength = (uint8_t)(data->RequestSize -
64);
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
scsi_io_req->SenseBufferLowAddress = htole32(cm->cm_sense_busaddr);
/*
* Set SGLOffset0 value. This is the number of dwords that SGL
* is offset from the beginning of MPI2_SCSI_IO_REQUEST struct.
*/
scsi_io_req->SGLOffset0 = 24;
/*
* Setup descriptor info. RAID passthrough must use the
* default request descriptor which is already set, so if this
* is a SCSI IO request, change the descriptor to SCSI IO.
* Also, if this is a SCSI IO request, handle the reply in the
* mpssas_scsio_complete function.
*/
if (function == MPI2_FUNCTION_SCSI_IO_REQUEST) {
cm->cm_desc.SCSIIO.RequestFlags =
MPI2_REQ_DESCRIPT_FLAGS_SCSI_IO;
cm->cm_desc.SCSIIO.DevHandle = scsi_io_req->DevHandle;
/*
* Make sure the DevHandle is not 0 because this is a
* likely error.
*/
if (scsi_io_req->DevHandle == 0) {
err = EINVAL;
goto RetFreeUnlocked;
}
}
}
mps_lock(sc);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
err = mps_wait_command(sc, &cm, 30, CAN_SLEEP);
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
if (err || (cm == NULL)) {
mps_printf(sc, "%s: invalid request: error %d\n", __func__,
err);
mps_unlock(sc);
goto RetFreeUnlocked;
}
/*
* Sync the DMA data, if any. Then copy the data to user space.
*/
if (cm->cm_data != NULL) {
if (cm->cm_flags & MPS_CM_FLAGS_DATAIN)
dir = BUS_DMASYNC_POSTREAD;
else if (cm->cm_flags & MPS_CM_FLAGS_DATAOUT)
dir = BUS_DMASYNC_POSTWRITE;
bus_dmamap_sync(sc->buffer_dmat, cm->cm_dmamap, dir);
bus_dmamap_unload(sc->buffer_dmat, cm->cm_dmamap);
if (cm->cm_flags & MPS_CM_FLAGS_DATAIN) {
mps_unlock(sc);
err = copyout(cm->cm_data,
PTRIN(data->PtrData), data->DataSize);
mps_lock(sc);
if (err != 0)
mps_dprint(sc, MPS_FAULT, "%s: failed to copy "
"IOCTL data to user space\n", __func__);
}
}
/*
* Copy the reply data and sense data to user space.
*/
if (cm->cm_reply != NULL) {
rpl = (MPI2_DEFAULT_REPLY *)cm->cm_reply;
sz = rpl->MsgLength * 4;
if (sz > data->ReplySize) {
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
mps_printf(sc, "%s: user reply buffer (%d) smaller "
"than returned buffer (%d)\n", __func__,
data->ReplySize, sz);
}
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
mps_unlock(sc);
copyout(cm->cm_reply, PTRIN(data->PtrReply), data->ReplySize);
mps_lock(sc);
if ((function == MPI2_FUNCTION_SCSI_IO_REQUEST) ||
(function == MPI2_FUNCTION_RAID_SCSI_IO_PASSTHROUGH)) {
if (((MPI2_SCSI_IO_REPLY *)rpl)->SCSIState &
MPI2_SCSI_STATE_AUTOSENSE_VALID) {
sense_len =
MIN((le32toh(((MPI2_SCSI_IO_REPLY *)rpl)->
SenseCount)), sizeof(struct
scsi_sense_data));
mps_unlock(sc);
copyout(cm->cm_sense, (PTRIN(data->PtrReply +
sizeof(MPI2_SCSI_IO_REPLY))), sense_len);
mps_lock(sc);
}
}
}
mps_unlock(sc);
RetFreeUnlocked:
mps_lock(sc);
if (cm != NULL) {
if (cm->cm_data)
free(cm->cm_data, M_MPSUSER);
mps_free_command(sc, cm);
}
Ret:
sc->mps_flags &= ~MPS_FLAGS_BUSY;
mps_unlock(sc);
return (err);
}
static void
mps_user_get_adapter_data(struct mps_softc *sc, mps_adapter_data_t *data)
{
Mpi2ConfigReply_t mpi_reply;
Mpi2BiosPage3_t config_page;
/*
* Use the PCI interface functions to get the Bus, Device, and Function
* information.
*/
data->PciInformation.u.bits.BusNumber = pci_get_bus(sc->mps_dev);
data->PciInformation.u.bits.DeviceNumber = pci_get_slot(sc->mps_dev);
data->PciInformation.u.bits.FunctionNumber =
pci_get_function(sc->mps_dev);
/*
* Get the FW version that should already be saved in IOC Facts.
*/
data->MpiFirmwareVersion = sc->facts->FWVersion.Word;
/*
* General device info.
*/
data->AdapterType = MPSIOCTL_ADAPTER_TYPE_SAS2;
if (sc->mps_flags & MPS_FLAGS_WD_AVAILABLE)
data->AdapterType = MPSIOCTL_ADAPTER_TYPE_SAS2_SSS6200;
data->PCIDeviceHwId = pci_get_device(sc->mps_dev);
data->PCIDeviceHwRev = pci_read_config(sc->mps_dev, PCIR_REVID, 1);
data->SubSystemId = pci_get_subdevice(sc->mps_dev);
data->SubsystemVendorId = pci_get_subvendor(sc->mps_dev);
/*
* Get the driver version.
*/
strcpy((char *)&data->DriverVersion[0], MPS_DRIVER_VERSION);
/*
* Need to get BIOS Config Page 3 for the BIOS Version.
*/
data->BiosVersion = 0;
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);
if (mps_config_get_bios_pg3(sc, &mpi_reply, &config_page))
printf("%s: Error while retrieving BIOS Version\n", __func__);
else
data->BiosVersion = config_page.BiosVersion;
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);
}
static void
mps_user_read_pci_info(struct mps_softc *sc, mps_pci_info_t *data)
{
int i;
/*
* Use the PCI interface functions to get the Bus, Device, and Function
* information.
*/
data->BusNumber = pci_get_bus(sc->mps_dev);
data->DeviceNumber = pci_get_slot(sc->mps_dev);
data->FunctionNumber = pci_get_function(sc->mps_dev);
/*
* Now get the interrupt vector and the pci header. The vector can
* only be 0 right now. The header is the first 256 bytes of config
* space.
*/
data->InterruptVector = 0;
for (i = 0; i < sizeof (data->PciHeader); i++) {
data->PciHeader[i] = pci_read_config(sc->mps_dev, i, 1);
}
}
static uint8_t
mps_get_fw_diag_buffer_number(struct mps_softc *sc, uint32_t unique_id)
{
uint8_t index;
for (index = 0; index < MPI2_DIAG_BUF_TYPE_COUNT; index++) {
if (sc->fw_diag_buffer_list[index].unique_id == unique_id) {
return (index);
}
}
return (MPS_FW_DIAGNOSTIC_UID_NOT_FOUND);
}
static int
mps_post_fw_diag_buffer(struct mps_softc *sc,
mps_fw_diagnostic_buffer_t *pBuffer, uint32_t *return_code)
{
MPI2_DIAG_BUFFER_POST_REQUEST *req;
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
MPI2_DIAG_BUFFER_POST_REPLY *reply = NULL;
struct mps_command *cm = NULL;
int i, status;
/*
* If buffer is not enabled, just leave.
*/
*return_code = MPS_FW_DIAG_ERROR_POST_FAILED;
if (!pBuffer->enabled) {
return (MPS_DIAG_FAILURE);
}
/*
* Clear some flags initially.
*/
pBuffer->force_release = FALSE;
pBuffer->valid_data = FALSE;
pBuffer->owned_by_firmware = FALSE;
/*
* Get a command.
*/
cm = mps_alloc_command(sc);
if (cm == NULL) {
mps_printf(sc, "%s: no mps requests\n", __func__);
return (MPS_DIAG_FAILURE);
}
/*
* Build the request for releasing the FW Diag Buffer and send it.
*/
req = (MPI2_DIAG_BUFFER_POST_REQUEST *)cm->cm_req;
req->Function = MPI2_FUNCTION_DIAG_BUFFER_POST;
req->BufferType = pBuffer->buffer_type;
req->ExtendedType = pBuffer->extended_type;
req->BufferLength = pBuffer->size;
for (i = 0; i < (sizeof(req->ProductSpecific) / 4); i++)
req->ProductSpecific[i] = pBuffer->product_specific[i];
mps_from_u64(sc->fw_diag_busaddr, &req->BufferAddress);
cm->cm_data = NULL;
cm->cm_length = 0;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
cm->cm_complete_data = NULL;
/*
* Send command synchronously.
*/
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
status = mps_wait_command(sc, &cm, 30, CAN_SLEEP);
if (status || (cm == NULL)) {
mps_printf(sc, "%s: invalid request: error %d\n", __func__,
status);
status = MPS_DIAG_FAILURE;
goto done;
}
/*
* Process POST reply.
*/
reply = (MPI2_DIAG_BUFFER_POST_REPLY *)cm->cm_reply;
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (reply == NULL) {
mps_printf(sc, "%s: reply is NULL, probably due to "
"reinitialization\n", __func__);
status = MPS_DIAG_FAILURE;
goto done;
}
if ((le16toh(reply->IOCStatus) & MPI2_IOCSTATUS_MASK) !=
MPI2_IOCSTATUS_SUCCESS) {
status = MPS_DIAG_FAILURE;
mps_dprint(sc, MPS_FAULT, "%s: post of FW Diag Buffer failed "
"with IOCStatus = 0x%x, IOCLogInfo = 0x%x and "
"TransferLength = 0x%x\n", __func__,
le16toh(reply->IOCStatus), le32toh(reply->IOCLogInfo),
le32toh(reply->TransferLength));
goto done;
}
/*
* Post was successful.
*/
pBuffer->valid_data = TRUE;
pBuffer->owned_by_firmware = TRUE;
*return_code = MPS_FW_DIAG_ERROR_SUCCESS;
status = MPS_DIAG_SUCCESS;
done:
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
if (cm != NULL)
mps_free_command(sc, cm);
return (status);
}
static int
mps_release_fw_diag_buffer(struct mps_softc *sc,
mps_fw_diagnostic_buffer_t *pBuffer, uint32_t *return_code,
uint32_t diag_type)
{
MPI2_DIAG_RELEASE_REQUEST *req;
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
MPI2_DIAG_RELEASE_REPLY *reply = NULL;
struct mps_command *cm = NULL;
int status;
/*
* If buffer is not enabled, just leave.
*/
*return_code = MPS_FW_DIAG_ERROR_RELEASE_FAILED;
if (!pBuffer->enabled) {
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
mps_dprint(sc, MPS_USER, "%s: This buffer type is not "
"supported by the IOC", __func__);
return (MPS_DIAG_FAILURE);
}
/*
* Clear some flags initially.
*/
pBuffer->force_release = FALSE;
pBuffer->valid_data = FALSE;
pBuffer->owned_by_firmware = FALSE;
/*
* Get a command.
*/
cm = mps_alloc_command(sc);
if (cm == NULL) {
mps_printf(sc, "%s: no mps requests\n", __func__);
return (MPS_DIAG_FAILURE);
}
/*
* Build the request for releasing the FW Diag Buffer and send it.
*/
req = (MPI2_DIAG_RELEASE_REQUEST *)cm->cm_req;
req->Function = MPI2_FUNCTION_DIAG_RELEASE;
req->BufferType = pBuffer->buffer_type;
cm->cm_data = NULL;
cm->cm_length = 0;
cm->cm_desc.Default.RequestFlags = MPI2_REQ_DESCRIPT_FLAGS_DEFAULT_TYPE;
cm->cm_complete_data = NULL;
/*
* Send command synchronously.
*/
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
status = mps_wait_command(sc, &cm, 30, CAN_SLEEP);
if (status || (cm == NULL)) {
mps_printf(sc, "%s: invalid request: error %d\n", __func__,
status);
status = MPS_DIAG_FAILURE;
goto done;
}
/*
* Process RELEASE reply.
*/
reply = (MPI2_DIAG_RELEASE_REPLY *)cm->cm_reply;
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (reply == NULL) {
mps_printf(sc, "%s: reply is NULL, probably due to "
"reinitialization\n", __func__);
status = MPS_DIAG_FAILURE;
goto done;
}
if (((le16toh(reply->IOCStatus) & MPI2_IOCSTATUS_MASK) !=
MPI2_IOCSTATUS_SUCCESS) || pBuffer->owned_by_firmware) {
status = MPS_DIAG_FAILURE;
mps_dprint(sc, MPS_FAULT, "%s: release of FW Diag Buffer "
"failed with IOCStatus = 0x%x and IOCLogInfo = 0x%x\n",
__func__, le16toh(reply->IOCStatus),
le32toh(reply->IOCLogInfo));
goto done;
}
/*
* Release was successful.
*/
*return_code = MPS_FW_DIAG_ERROR_SUCCESS;
status = MPS_DIAG_SUCCESS;
/*
* If this was for an UNREGISTER diag type command, clear the unique ID.
*/
if (diag_type == MPS_FW_DIAG_TYPE_UNREGISTER) {
pBuffer->unique_id = MPS_FW_DIAG_INVALID_UID;
}
done:
Changes to make mps(4) and mpr(4) handle reinit with reallocation. When the mps(4) and mpr(4) drivers need to reinitialize the firmware, they sometimes need to reallocate all of the memory allocated by the driver. The reallocation happens whenever the IOC Facts change. That should only happen after a firmware upgrade. If the reinitialization happens as a result of a timed out command sent to the card, the command that timed out and triggered the reinit may have been freed if iocfacts_allocate() reallocated all memory. If the caller attempts to access the command after that, the kernel will panic because the caller will be dereferencing freed memory. The solution is to set a flag in the softc when we reallocate, and avoid dereferencing the command strucure if we've reallocated. The changes are largely the same in both drivers, since mpr(4) is a derivative of mps(4). o In iocfacts_allocate(), if the IOC Facts have changed and we need to reallocate, set the REALLOCATED flag in the softc. o Change wait_command() to take a struct mps_command ** instead of a struct mps_command *. This allows us to NULL out the caller's command pointer if we have to reinit the controller and the data structures get reallocated. (The REALLOCATED flag will be set in the softc if that has happened.) o In every place that calls wait_command(), make sure we handle the case where the command is NULL after the call. o The mpr(4) driver has mpr_request_polled() which can also reinitialize the card. Also check for reallocation there. Reviewed by: scottl, slm MFC after: 1 week Sponsored by: Spectra Logic
2017-08-10 14:59:17 +00:00
if (cm != NULL)
mps_free_command(sc, cm);
return (status);
}
static int
mps_diag_register(struct mps_softc *sc, mps_fw_diag_register_t *diag_register,
uint32_t *return_code)
{
mps_fw_diagnostic_buffer_t *pBuffer;
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
struct mps_busdma_context *ctx;
uint8_t extended_type, buffer_type, i;
uint32_t buffer_size;
uint32_t unique_id;
int status;
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
int error;
extended_type = diag_register->ExtendedType;
buffer_type = diag_register->BufferType;
buffer_size = diag_register->RequestedBufferSize;
unique_id = diag_register->UniqueId;
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
ctx = NULL;
error = 0;
/*
* Check for valid buffer type
*/
if (buffer_type >= MPI2_DIAG_BUF_TYPE_COUNT) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
return (MPS_DIAG_FAILURE);
}
/*
* Get the current buffer and look up the unique ID. The unique ID
* should not be found. If it is, the ID is already in use.
*/
i = mps_get_fw_diag_buffer_number(sc, unique_id);
pBuffer = &sc->fw_diag_buffer_list[buffer_type];
if (i != MPS_FW_DIAGNOSTIC_UID_NOT_FOUND) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
/*
* The buffer's unique ID should not be registered yet, and the given
* unique ID cannot be 0.
*/
if ((pBuffer->unique_id != MPS_FW_DIAG_INVALID_UID) ||
(unique_id == MPS_FW_DIAG_INVALID_UID)) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
/*
* If this buffer is already posted as immediate, just change owner.
*/
if (pBuffer->immediate && pBuffer->owned_by_firmware &&
(pBuffer->unique_id == MPS_FW_DIAG_INVALID_UID)) {
pBuffer->immediate = FALSE;
pBuffer->unique_id = unique_id;
return (MPS_DIAG_SUCCESS);
}
/*
* Post a new buffer after checking if it's enabled. The DMA buffer
* that is allocated will be contiguous (nsegments = 1).
*/
if (!pBuffer->enabled) {
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
return (MPS_DIAG_FAILURE);
}
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (bus_dma_tag_create( sc->mps_parent_dmat, /* parent */
1, 0, /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT,/* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
buffer_size, /* maxsize */
1, /* nsegments */
buffer_size, /* maxsegsize */
0, /* flags */
NULL, NULL, /* lockfunc, lockarg */
&sc->fw_diag_dmat)) {
mps_dprint(sc, MPS_ERROR,
"Cannot allocate FW diag buffer DMA tag\n");
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
status = MPS_DIAG_FAILURE;
goto bailout;
}
if (bus_dmamem_alloc(sc->fw_diag_dmat, (void **)&sc->fw_diag_buffer,
BUS_DMA_NOWAIT, &sc->fw_diag_map)) {
mps_dprint(sc, MPS_ERROR,
"Cannot allocate FW diag buffer memory\n");
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
status = MPS_DIAG_FAILURE;
goto bailout;
}
bzero(sc->fw_diag_buffer, buffer_size);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
ctx = malloc(sizeof(*ctx), M_MPSUSER, M_WAITOK | M_ZERO);
if (ctx == NULL) {
device_printf(sc->mps_dev, "%s: context malloc failed\n",
__func__);
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
status = MPS_DIAG_FAILURE;
goto bailout;
}
ctx->addr = &sc->fw_diag_busaddr;
ctx->buffer_dmat = sc->fw_diag_dmat;
ctx->buffer_dmamap = sc->fw_diag_map;
ctx->softc = sc;
error = bus_dmamap_load(sc->fw_diag_dmat, sc->fw_diag_map,
sc->fw_diag_buffer, buffer_size, mps_memaddr_wait_cb,
ctx, 0);
if (error == EINPROGRESS) {
/* XXX KDM */
device_printf(sc->mps_dev, "%s: Deferred bus_dmamap_load\n",
__func__);
/*
* Wait for the load to complete. If we're interrupted,
* bail out.
*/
mps_lock(sc);
if (ctx->completed == 0) {
error = msleep(ctx, &sc->mps_mtx, PCATCH, "mpswait", 0);
if (error != 0) {
/*
* We got an error from msleep(9). This is
* most likely due to a signal. Tell
* mpr_memaddr_wait_cb() that we've abandoned
* the context, so it needs to clean up when
* it is called.
*/
ctx->abandoned = 1;
/* The callback will free this memory */
ctx = NULL;
mps_unlock(sc);
device_printf(sc->mps_dev, "Cannot "
"bus_dmamap_load FW diag buffer, error = "
"%d returned from msleep\n", error);
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
status = MPS_DIAG_FAILURE;
goto bailout;
}
}
mps_unlock(sc);
}
if ((error != 0) || (ctx->error != 0)) {
device_printf(sc->mps_dev, "Cannot bus_dmamap_load FW diag "
"buffer, %serror = %d\n", error ? "" : "callback ",
error ? error : ctx->error);
*return_code = MPS_FW_DIAG_ERROR_NO_BUFFER;
status = MPS_DIAG_FAILURE;
goto bailout;
}
bus_dmamap_sync(sc->fw_diag_dmat, sc->fw_diag_map, BUS_DMASYNC_PREREAD);
pBuffer->size = buffer_size;
/*
* Copy the given info to the diag buffer and post the buffer.
*/
pBuffer->buffer_type = buffer_type;
pBuffer->immediate = FALSE;
if (buffer_type == MPI2_DIAG_BUF_TYPE_TRACE) {
for (i = 0; i < (sizeof (pBuffer->product_specific) / 4);
i++) {
pBuffer->product_specific[i] =
diag_register->ProductSpecific[i];
}
}
pBuffer->extended_type = extended_type;
pBuffer->unique_id = unique_id;
status = mps_post_fw_diag_buffer(sc, pBuffer, return_code);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
bailout:
/*
* In case there was a failure, free the DMA buffer.
*/
if (status == MPS_DIAG_FAILURE) {
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (sc->fw_diag_busaddr != 0) {
bus_dmamap_unload(sc->fw_diag_dmat, sc->fw_diag_map);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_busaddr = 0;
}
if (sc->fw_diag_buffer != NULL) {
bus_dmamem_free(sc->fw_diag_dmat, sc->fw_diag_buffer,
sc->fw_diag_map);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_buffer = NULL;
}
if (sc->fw_diag_dmat != NULL) {
bus_dma_tag_destroy(sc->fw_diag_dmat);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_dmat = NULL;
}
}
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (ctx != NULL)
free(ctx, M_MPSUSER);
return (status);
}
static int
mps_diag_unregister(struct mps_softc *sc,
mps_fw_diag_unregister_t *diag_unregister, uint32_t *return_code)
{
mps_fw_diagnostic_buffer_t *pBuffer;
uint8_t i;
uint32_t unique_id;
int status;
unique_id = diag_unregister->UniqueId;
/*
* Get the current buffer and look up the unique ID. The unique ID
* should be there.
*/
i = mps_get_fw_diag_buffer_number(sc, unique_id);
if (i == MPS_FW_DIAGNOSTIC_UID_NOT_FOUND) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
pBuffer = &sc->fw_diag_buffer_list[i];
/*
* Try to release the buffer from FW before freeing it. If release
* fails, don't free the DMA buffer in case FW tries to access it
* later. If buffer is not owned by firmware, can't release it.
*/
if (!pBuffer->owned_by_firmware) {
status = MPS_DIAG_SUCCESS;
} else {
status = mps_release_fw_diag_buffer(sc, pBuffer, return_code,
MPS_FW_DIAG_TYPE_UNREGISTER);
}
/*
* At this point, return the current status no matter what happens with
* the DMA buffer.
*/
pBuffer->unique_id = MPS_FW_DIAG_INVALID_UID;
if (status == MPS_DIAG_SUCCESS) {
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
if (sc->fw_diag_busaddr != 0) {
bus_dmamap_unload(sc->fw_diag_dmat, sc->fw_diag_map);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_busaddr = 0;
}
if (sc->fw_diag_buffer != NULL) {
bus_dmamem_free(sc->fw_diag_dmat, sc->fw_diag_buffer,
sc->fw_diag_map);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_buffer = NULL;
}
if (sc->fw_diag_dmat != NULL) {
bus_dma_tag_destroy(sc->fw_diag_dmat);
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
sc->fw_diag_dmat = NULL;
}
}
return (status);
}
static int
mps_diag_query(struct mps_softc *sc, mps_fw_diag_query_t *diag_query,
uint32_t *return_code)
{
mps_fw_diagnostic_buffer_t *pBuffer;
uint8_t i;
uint32_t unique_id;
unique_id = diag_query->UniqueId;
/*
* If ID is valid, query on ID.
* If ID is invalid, query on buffer type.
*/
if (unique_id == MPS_FW_DIAG_INVALID_UID) {
i = diag_query->BufferType;
if (i >= MPI2_DIAG_BUF_TYPE_COUNT) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
} else {
i = mps_get_fw_diag_buffer_number(sc, unique_id);
if (i == MPS_FW_DIAGNOSTIC_UID_NOT_FOUND) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
}
/*
* Fill query structure with the diag buffer info.
*/
pBuffer = &sc->fw_diag_buffer_list[i];
diag_query->BufferType = pBuffer->buffer_type;
diag_query->ExtendedType = pBuffer->extended_type;
if (diag_query->BufferType == MPI2_DIAG_BUF_TYPE_TRACE) {
for (i = 0; i < (sizeof(diag_query->ProductSpecific) / 4);
i++) {
diag_query->ProductSpecific[i] =
pBuffer->product_specific[i];
}
}
diag_query->TotalBufferSize = pBuffer->size;
diag_query->DriverAddedBufferSize = 0;
diag_query->UniqueId = pBuffer->unique_id;
diag_query->ApplicationFlags = 0;
diag_query->DiagnosticFlags = 0;
/*
* Set/Clear application flags
*/
if (pBuffer->immediate) {
diag_query->ApplicationFlags &= ~MPS_FW_DIAG_FLAG_APP_OWNED;
} else {
diag_query->ApplicationFlags |= MPS_FW_DIAG_FLAG_APP_OWNED;
}
if (pBuffer->valid_data || pBuffer->owned_by_firmware) {
diag_query->ApplicationFlags |= MPS_FW_DIAG_FLAG_BUFFER_VALID;
} else {
diag_query->ApplicationFlags &= ~MPS_FW_DIAG_FLAG_BUFFER_VALID;
}
if (pBuffer->owned_by_firmware) {
diag_query->ApplicationFlags |=
MPS_FW_DIAG_FLAG_FW_BUFFER_ACCESS;
} else {
diag_query->ApplicationFlags &=
~MPS_FW_DIAG_FLAG_FW_BUFFER_ACCESS;
}
return (MPS_DIAG_SUCCESS);
}
static int
mps_diag_read_buffer(struct mps_softc *sc,
mps_diag_read_buffer_t *diag_read_buffer, uint8_t *ioctl_buf,
uint32_t *return_code)
{
mps_fw_diagnostic_buffer_t *pBuffer;
uint8_t i, *pData;
uint32_t unique_id;
int status;
unique_id = diag_read_buffer->UniqueId;
/*
* Get the current buffer and look up the unique ID. The unique ID
* should be there.
*/
i = mps_get_fw_diag_buffer_number(sc, unique_id);
if (i == MPS_FW_DIAGNOSTIC_UID_NOT_FOUND) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
pBuffer = &sc->fw_diag_buffer_list[i];
/*
* Make sure requested read is within limits
*/
if (diag_read_buffer->StartingOffset + diag_read_buffer->BytesToRead >
pBuffer->size) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
return (MPS_DIAG_FAILURE);
}
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers. In mp{r,s}_diag_register(), which is used to register diagnostic buffers with the mp{r,s}(4) firmware, we allocate DMAable memory. There were several issues here: o No checking of the bus_dmamap_load() return value. If the load failed or got deferred, mp{r,s}_diag_register() continued on as if nothing had happened. We now check the return value and bail out if it fails. o No waiting for a deferred load callback. bus_dmamap_load() calls a supplied callback when the mapping is done. This is generally done immediately, but it can be deferred. mp{r,s}_diag_register() did not check to see whether the callback was already done before proceeding on. We now sleep until the callback is done if it is deferred. o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the memory is allocated and loaded. This is necessary on some platforms to synchronize host memory that is going to be updated by a device. Both drivers would also panic if the firmware was reinitialized while a diagnostic buffer operation was in progress. This fixes that problem as well. (The driver will reinitialize the firmware in various circumstances, but the problem I ran into was that the firmware would generate an IOC Fault due to a PCIe error.) mp{r,s}var.h: Add a new structure, struct mpr_busdma_context, that is used for deferred busdma load callbacks. Add a prototype for mp{r,s}_memaddr_wait_cb(). mp{r,s}.c: Add a new busdma callback function, mp{r,s}_memaddr_wait_cb(). This provides synchronization for callers that want to wait on a deferred bus_dmamap_load() callback. mp{r,s}_user.c: In bus_dmamap_register(), add a call to bus_dmamap_sync() with the BUS_DMASYNC_PREREAD flag set after an allocation is loaded. Also, check the return value of bus_dmamap_load(). If it fails, bail out. If it is EINPROGRESS, wait for the callback to happen. We use an interruptible sleep (msleep with PCATCH) and let the callback clean things up if we get interrupted. In mpr_diag_read_buffer() and mps_diag_read_buffer(), call bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying the data out to make sure the data is in stable storage. In mp{r,s}_post_fw_diag_buffer() and mp{r,s}_release_fw_diag_buffer(), check the reply to see whether it is NULL. It can be NULL (and the command non-NULL) if the controller gets reinitialized while we're waiting for the command to complete but the driver structures aren't reallocated. The driver structures generally won't be reallocated unless there is a firmware upgrade that changes one of the IOCFacts. When freeing diagnostic buffers in mp{r,s}_diag_register() and mp{r,s}_diag_unregister(), zero/NULL out the buffer after freeing it. This will prevent a duplicate free in some situations. Sponsored by: Spectra Logic Reviewed by: mav, scottl MFC after: 1 week Differential Revision: D13453
2018-02-06 15:58:22 +00:00
/* Sync the DMA map before we copy to userland. */
bus_dmamap_sync(sc->fw_diag_dmat, sc->fw_diag_map,
BUS_DMASYNC_POSTREAD);
/*
* Copy the requested data from DMA to the diag_read_buffer. The DMA
* buffer that was allocated is one contiguous buffer.
*/
pData = (uint8_t *)(sc->fw_diag_buffer +
diag_read_buffer->StartingOffset);
if (copyout(pData, ioctl_buf, diag_read_buffer->BytesToRead) != 0)
return (MPS_DIAG_FAILURE);
diag_read_buffer->Status = 0;
/*
* Set or clear the Force Release flag.
*/
if (pBuffer->force_release) {
diag_read_buffer->Flags |= MPS_FW_DIAG_FLAG_FORCE_RELEASE;
} else {
diag_read_buffer->Flags &= ~MPS_FW_DIAG_FLAG_FORCE_RELEASE;
}
/*
* If buffer is to be reregistered, make sure it's not already owned by
* firmware first.
*/
status = MPS_DIAG_SUCCESS;
if (!pBuffer->owned_by_firmware) {
if (diag_read_buffer->Flags & MPS_FW_DIAG_FLAG_REREGISTER) {
status = mps_post_fw_diag_buffer(sc, pBuffer,
return_code);
}
}
return (status);
}
static int
mps_diag_release(struct mps_softc *sc, mps_fw_diag_release_t *diag_release,
uint32_t *return_code)
{
mps_fw_diagnostic_buffer_t *pBuffer;
uint8_t i;
uint32_t unique_id;
int status;
unique_id = diag_release->UniqueId;
/*
* Get the current buffer and look up the unique ID. The unique ID
* should be there.
*/
i = mps_get_fw_diag_buffer_number(sc, unique_id);
if (i == MPS_FW_DIAGNOSTIC_UID_NOT_FOUND) {
*return_code = MPS_FW_DIAG_ERROR_INVALID_UID;
return (MPS_DIAG_FAILURE);
}
pBuffer = &sc->fw_diag_buffer_list[i];
/*
* If buffer is not owned by firmware, it's already been released.
*/
if (!pBuffer->owned_by_firmware) {
*return_code = MPS_FW_DIAG_ERROR_ALREADY_RELEASED;
return (MPS_DIAG_FAILURE);
}
/*
* Release the buffer.
*/
status = mps_release_fw_diag_buffer(sc, pBuffer, return_code,
MPS_FW_DIAG_TYPE_RELEASE);
return (status);
}
static int
mps_do_diag_action(struct mps_softc *sc, uint32_t action, uint8_t *diag_action,
uint32_t length, uint32_t *return_code)
{
mps_fw_diag_register_t diag_register;
mps_fw_diag_unregister_t diag_unregister;
mps_fw_diag_query_t diag_query;
mps_diag_read_buffer_t diag_read_buffer;
mps_fw_diag_release_t diag_release;
int status = MPS_DIAG_SUCCESS;
uint32_t original_return_code;
original_return_code = *return_code;
*return_code = MPS_FW_DIAG_ERROR_SUCCESS;
switch (action) {
case MPS_FW_DIAG_TYPE_REGISTER:
if (!length) {
*return_code =
MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
if (copyin(diag_action, &diag_register,
sizeof(diag_register)) != 0)
return (MPS_DIAG_FAILURE);
status = mps_diag_register(sc, &diag_register,
return_code);
break;
case MPS_FW_DIAG_TYPE_UNREGISTER:
if (length < sizeof(diag_unregister)) {
*return_code =
MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
if (copyin(diag_action, &diag_unregister,
sizeof(diag_unregister)) != 0)
return (MPS_DIAG_FAILURE);
status = mps_diag_unregister(sc, &diag_unregister,
return_code);
break;
case MPS_FW_DIAG_TYPE_QUERY:
if (length < sizeof (diag_query)) {
*return_code =
MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
if (copyin(diag_action, &diag_query, sizeof(diag_query))
!= 0)
return (MPS_DIAG_FAILURE);
status = mps_diag_query(sc, &diag_query, return_code);
if (status == MPS_DIAG_SUCCESS)
if (copyout(&diag_query, diag_action,
sizeof (diag_query)) != 0)
return (MPS_DIAG_FAILURE);
break;
case MPS_FW_DIAG_TYPE_READ_BUFFER:
if (copyin(diag_action, &diag_read_buffer,
sizeof(diag_read_buffer)) != 0)
return (MPS_DIAG_FAILURE);
if (length < diag_read_buffer.BytesToRead) {
*return_code =
MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
status = mps_diag_read_buffer(sc, &diag_read_buffer,
PTRIN(diag_read_buffer.PtrDataBuffer),
return_code);
if (status == MPS_DIAG_SUCCESS) {
if (copyout(&diag_read_buffer, diag_action,
sizeof(diag_read_buffer) -
sizeof(diag_read_buffer.PtrDataBuffer)) !=
0)
return (MPS_DIAG_FAILURE);
}
break;
case MPS_FW_DIAG_TYPE_RELEASE:
if (length < sizeof(diag_release)) {
*return_code =
MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
if (copyin(diag_action, &diag_release,
sizeof(diag_release)) != 0)
return (MPS_DIAG_FAILURE);
status = mps_diag_release(sc, &diag_release,
return_code);
break;
default:
*return_code = MPS_FW_DIAG_ERROR_INVALID_PARAMETER;
status = MPS_DIAG_FAILURE;
break;
}
if ((status == MPS_DIAG_FAILURE) &&
(original_return_code == MPS_FW_DIAG_NEW) &&
(*return_code != MPS_FW_DIAG_ERROR_SUCCESS))
status = MPS_DIAG_SUCCESS;
return (status);
}
static int
mps_user_diag_action(struct mps_softc *sc, mps_diag_action_t *data)
{
int status;
/*
* Only allow one diag action at one time.
*/
if (sc->mps_flags & MPS_FLAGS_BUSY) {
mps_dprint(sc, MPS_USER, "%s: Only one FW diag command "
"allowed at a single time.", __func__);
return (EBUSY);
}
sc->mps_flags |= MPS_FLAGS_BUSY;
/*
* Send diag action request
*/
if (data->Action == MPS_FW_DIAG_TYPE_REGISTER ||
data->Action == MPS_FW_DIAG_TYPE_UNREGISTER ||
data->Action == MPS_FW_DIAG_TYPE_QUERY ||
data->Action == MPS_FW_DIAG_TYPE_READ_BUFFER ||
data->Action == MPS_FW_DIAG_TYPE_RELEASE) {
status = mps_do_diag_action(sc, data->Action,
PTRIN(data->PtrDiagAction), data->Length,
&data->ReturnCode);
} else
status = EINVAL;
sc->mps_flags &= ~MPS_FLAGS_BUSY;
return (status);
}
/*
* Copy the event recording mask and the event queue size out. For
* clarification, the event recording mask (events_to_record) is not the same
* thing as the event mask (event_mask). events_to_record has a bit set for
* every event type that is to be recorded by the driver, and event_mask has a
* bit cleared for every event that is allowed into the driver from the IOC.
* They really have nothing to do with each other.
*/
static void
mps_user_event_query(struct mps_softc *sc, mps_event_query_t *data)
{
uint8_t i;
mps_lock(sc);
data->Entries = MPS_EVENT_QUEUE_SIZE;
for (i = 0; i < 4; i++) {
data->Types[i] = sc->events_to_record[i];
}
mps_unlock(sc);
}
/*
* Set the driver's event mask according to what's been given. See
* mps_user_event_query for explanation of the event recording mask and the IOC
* event mask. It's the app's responsibility to enable event logging by setting
* the bits in events_to_record. Initially, no events will be logged.
*/
static void
mps_user_event_enable(struct mps_softc *sc, mps_event_enable_t *data)
{
uint8_t i;
mps_lock(sc);
for (i = 0; i < 4; i++) {
sc->events_to_record[i] = data->Types[i];
}
mps_unlock(sc);
}
/*
* Copy out the events that have been recorded, up to the max events allowed.
*/
static int
mps_user_event_report(struct mps_softc *sc, mps_event_report_t *data)
{
int status = 0;
uint32_t size;
mps_lock(sc);
size = data->Size;
if ((size >= sizeof(sc->recorded_events)) && (status == 0)) {
mps_unlock(sc);
if (copyout((void *)sc->recorded_events,
PTRIN(data->PtrEvents), size) != 0)
status = EFAULT;
mps_lock(sc);
} else {
/*
* data->Size value is not large enough to copy event data.
*/
status = EFAULT;
}
/*
* Change size value to match the number of bytes that were copied.
*/
if (status == 0)
data->Size = sizeof(sc->recorded_events);
mps_unlock(sc);
return (status);
}
/*
* Record events into the driver from the IOC if they are not masked.
*/
void
mpssas_record_event(struct mps_softc *sc,
MPI2_EVENT_NOTIFICATION_REPLY *event_reply)
{
uint32_t event;
int i, j;
uint16_t event_data_len;
boolean_t sendAEN = FALSE;
event = event_reply->Event;
/*
* Generate a system event to let anyone who cares know that a
* LOG_ENTRY_ADDED event has occurred. This is sent no matter what the
* event mask is set to.
*/
if (event == MPI2_EVENT_LOG_ENTRY_ADDED) {
sendAEN = TRUE;
}
/*
* Record the event only if its corresponding bit is set in
* events_to_record. event_index is the index into recorded_events and
* event_number is the overall number of an event being recorded since
* start-of-day. event_index will roll over; event_number will never
* roll over.
*/
i = (uint8_t)(event / 32);
j = (uint8_t)(event % 32);
if ((i < 4) && ((1 << j) & sc->events_to_record[i])) {
i = sc->event_index;
sc->recorded_events[i].Type = event;
sc->recorded_events[i].Number = ++sc->event_number;
bzero(sc->recorded_events[i].Data, MPS_MAX_EVENT_DATA_LENGTH *
4);
event_data_len = event_reply->EventDataLength;
if (event_data_len > 0) {
/*
* Limit data to size in m_event entry
*/
if (event_data_len > MPS_MAX_EVENT_DATA_LENGTH) {
event_data_len = MPS_MAX_EVENT_DATA_LENGTH;
}
for (j = 0; j < event_data_len; j++) {
sc->recorded_events[i].Data[j] =
event_reply->EventData[j];
}
/*
* check for index wrap-around
*/
if (++i == MPS_EVENT_QUEUE_SIZE) {
i = 0;
}
sc->event_index = (uint8_t)i;
/*
* Set flag to send the event.
*/
sendAEN = TRUE;
}
}
/*
* Generate a system event if flag is set to let anyone who cares know
* that an event has occurred.
*/
if (sendAEN) {
//SLM-how to send a system event (see kqueue, kevent)
// (void) ddi_log_sysevent(mpt->m_dip, DDI_VENDOR_LSI, "MPT_SAS",
// "SAS", NULL, NULL, DDI_NOSLEEP);
}
}
static int
mps_user_reg_access(struct mps_softc *sc, mps_reg_access_t *data)
{
int status = 0;
switch (data->Command) {
/*
* IO access is not supported.
*/
case REG_IO_READ:
case REG_IO_WRITE:
mps_dprint(sc, MPS_USER, "IO access is not supported. "
"Use memory access.");
status = EINVAL;
break;
case REG_MEM_READ:
data->RegData = mps_regread(sc, data->RegOffset);
break;
case REG_MEM_WRITE:
mps_regwrite(sc, data->RegOffset, data->RegData);
break;
default:
status = EINVAL;
break;
}
return (status);
}
static int
mps_user_btdh(struct mps_softc *sc, mps_btdh_mapping_t *data)
{
uint8_t bt2dh = FALSE;
uint8_t dh2bt = FALSE;
uint16_t dev_handle, bus, target;
bus = data->Bus;
target = data->TargetID;
dev_handle = data->DevHandle;
/*
* When DevHandle is 0xFFFF and Bus/Target are not 0xFFFF, use Bus/
* Target to get DevHandle. When Bus/Target are 0xFFFF and DevHandle is
* not 0xFFFF, use DevHandle to get Bus/Target. Anything else is
* invalid.
*/
if ((bus == 0xFFFF) && (target == 0xFFFF) && (dev_handle != 0xFFFF))
dh2bt = TRUE;
if ((dev_handle == 0xFFFF) && (bus != 0xFFFF) && (target != 0xFFFF))
bt2dh = TRUE;
if (!dh2bt && !bt2dh)
return (EINVAL);
/*
* Only handle bus of 0. Make sure target is within range.
*/
if (bt2dh) {
if (bus != 0)
return (EINVAL);
if (target > sc->max_devices) {
mps_dprint(sc, MPS_FAULT, "Target ID is out of range "
"for Bus/Target to DevHandle mapping.");
return (EINVAL);
}
dev_handle = sc->mapping_table[target].dev_handle;
if (dev_handle)
data->DevHandle = dev_handle;
} else {
bus = 0;
target = mps_mapping_get_tid_from_handle(sc, dev_handle);
data->Bus = bus;
data->TargetID = target;
}
return (0);
}
static int
mps_ioctl(struct cdev *dev, u_long cmd, void *arg, int flag,
struct thread *td)
{
struct mps_softc *sc;
struct mps_cfg_page_req *page_req;
struct mps_ext_cfg_page_req *ext_page_req;
void *mps_page;
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
int error, msleep_ret;
mps_page = NULL;
sc = dev->si_drv1;
page_req = (void *)arg;
ext_page_req = (void *)arg;
switch (cmd) {
case MPSIO_READ_CFG_HEADER:
mps_lock(sc);
error = mps_user_read_cfg_header(sc, page_req);
mps_unlock(sc);
break;
case MPSIO_READ_CFG_PAGE:
mps_page = malloc(page_req->len, M_MPSUSER, M_WAITOK | M_ZERO);
error = copyin(page_req->buf, mps_page,
sizeof(MPI2_CONFIG_PAGE_HEADER));
if (error)
break;
mps_lock(sc);
error = mps_user_read_cfg_page(sc, page_req, mps_page);
mps_unlock(sc);
if (error)
break;
error = copyout(mps_page, page_req->buf, page_req->len);
break;
case MPSIO_READ_EXT_CFG_HEADER:
mps_lock(sc);
error = mps_user_read_extcfg_header(sc, ext_page_req);
mps_unlock(sc);
break;
case MPSIO_READ_EXT_CFG_PAGE:
mps_page = malloc(ext_page_req->len, M_MPSUSER, M_WAITOK|M_ZERO);
error = copyin(ext_page_req->buf, mps_page,
sizeof(MPI2_CONFIG_EXTENDED_PAGE_HEADER));
if (error)
break;
mps_lock(sc);
error = mps_user_read_extcfg_page(sc, ext_page_req, mps_page);
mps_unlock(sc);
if (error)
break;
error = copyout(mps_page, ext_page_req->buf, ext_page_req->len);
break;
case MPSIO_WRITE_CFG_PAGE:
mps_page = malloc(page_req->len, M_MPSUSER, M_WAITOK|M_ZERO);
error = copyin(page_req->buf, mps_page, page_req->len);
if (error)
break;
mps_lock(sc);
error = mps_user_write_cfg_page(sc, page_req, mps_page);
mps_unlock(sc);
break;
case MPSIO_MPS_COMMAND:
error = mps_user_command(sc, (struct mps_usr_command *)arg);
break;
case MPTIOCTL_PASS_THRU:
/*
* The user has requested to pass through a command to be
* executed by the MPT firmware. Call our routine which does
* this. Only allow one passthru IOCTL at one time.
*/
error = mps_user_pass_thru(sc, (mps_pass_thru_t *)arg);
break;
case MPTIOCTL_GET_ADAPTER_DATA:
/*
* The user has requested to read adapter data. Call our
* routine which does this.
*/
error = 0;
mps_user_get_adapter_data(sc, (mps_adapter_data_t *)arg);
break;
case MPTIOCTL_GET_PCI_INFO:
/*
* The user has requested to read pci info. Call
* our routine which does this.
*/
mps_lock(sc);
error = 0;
mps_user_read_pci_info(sc, (mps_pci_info_t *)arg);
mps_unlock(sc);
break;
case MPTIOCTL_RESET_ADAPTER:
mps_lock(sc);
sc->port_enable_complete = 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
uint32_t reinit_start = time_uptime;
error = mps_reinit(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
/* Sleep for 300 second. */
msleep_ret = msleep(&sc->port_enable_complete, &sc->mps_mtx, PRIBIO,
"mps_porten", 300 * hz);
mps_unlock(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
if (msleep_ret)
printf("Port Enable did not complete after Diag "
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
"Reset msleep error %d.\n", msleep_ret);
else
mps_dprint(sc, MPS_USER,
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
"Hard Reset with Port Enable completed in %d seconds.\n",
(uint32_t) (time_uptime - reinit_start));
break;
case MPTIOCTL_DIAG_ACTION:
/*
* The user has done a diag buffer action. Call our routine
* which does this. Only allow one diag action at one time.
*/
mps_lock(sc);
error = mps_user_diag_action(sc, (mps_diag_action_t *)arg);
mps_unlock(sc);
break;
case MPTIOCTL_EVENT_QUERY:
/*
* The user has done an event query. Call our routine which does
* this.
*/
error = 0;
mps_user_event_query(sc, (mps_event_query_t *)arg);
break;
case MPTIOCTL_EVENT_ENABLE:
/*
* The user has done an event enable. Call our routine which
* does this.
*/
error = 0;
mps_user_event_enable(sc, (mps_event_enable_t *)arg);
break;
case MPTIOCTL_EVENT_REPORT:
/*
* The user has done an event report. Call our routine which
* does this.
*/
error = mps_user_event_report(sc, (mps_event_report_t *)arg);
break;
case MPTIOCTL_REG_ACCESS:
/*
* The user has requested register access. Call our routine
* which does this.
*/
mps_lock(sc);
error = mps_user_reg_access(sc, (mps_reg_access_t *)arg);
mps_unlock(sc);
break;
case MPTIOCTL_BTDH_MAPPING:
/*
* The user has requested to translate a bus/target to a
* DevHandle or a DevHandle to a bus/target. Call our routine
* which does this.
*/
error = mps_user_btdh(sc, (mps_btdh_mapping_t *)arg);
break;
default:
error = ENOIOCTL;
break;
}
if (mps_page != NULL)
free(mps_page, M_MPSUSER);
return (error);
}
#ifdef COMPAT_FREEBSD32
struct mps_cfg_page_req32 {
MPI2_CONFIG_PAGE_HEADER header;
uint32_t page_address;
uint32_t buf;
int len;
uint16_t ioc_status;
};
struct mps_ext_cfg_page_req32 {
MPI2_CONFIG_EXTENDED_PAGE_HEADER header;
uint32_t page_address;
uint32_t buf;
int len;
uint16_t ioc_status;
};
struct mps_raid_action32 {
uint8_t action;
uint8_t volume_bus;
uint8_t volume_id;
uint8_t phys_disk_num;
uint32_t action_data_word;
uint32_t buf;
int len;
uint32_t volume_status;
uint32_t action_data[4];
uint16_t action_status;
uint16_t ioc_status;
uint8_t write;
};
struct mps_usr_command32 {
uint32_t req;
uint32_t req_len;
uint32_t rpl;
uint32_t rpl_len;
uint32_t buf;
int len;
uint32_t flags;
};
#define MPSIO_READ_CFG_HEADER32 _IOWR('M', 200, struct mps_cfg_page_req32)
#define MPSIO_READ_CFG_PAGE32 _IOWR('M', 201, struct mps_cfg_page_req32)
#define MPSIO_READ_EXT_CFG_HEADER32 _IOWR('M', 202, struct mps_ext_cfg_page_req32)
#define MPSIO_READ_EXT_CFG_PAGE32 _IOWR('M', 203, struct mps_ext_cfg_page_req32)
#define MPSIO_WRITE_CFG_PAGE32 _IOWR('M', 204, struct mps_cfg_page_req32)
#define MPSIO_RAID_ACTION32 _IOWR('M', 205, struct mps_raid_action32)
#define MPSIO_MPS_COMMAND32 _IOWR('M', 210, struct mps_usr_command32)
static int
mps_ioctl32(struct cdev *dev, u_long cmd32, void *_arg, int flag,
struct thread *td)
{
struct mps_cfg_page_req32 *page32 = _arg;
struct mps_ext_cfg_page_req32 *ext32 = _arg;
struct mps_raid_action32 *raid32 = _arg;
struct mps_usr_command32 *user32 = _arg;
union {
struct mps_cfg_page_req page;
struct mps_ext_cfg_page_req ext;
struct mps_raid_action raid;
struct mps_usr_command user;
} arg;
u_long cmd;
int error;
switch (cmd32) {
case MPSIO_READ_CFG_HEADER32:
case MPSIO_READ_CFG_PAGE32:
case MPSIO_WRITE_CFG_PAGE32:
if (cmd32 == MPSIO_READ_CFG_HEADER32)
cmd = MPSIO_READ_CFG_HEADER;
else if (cmd32 == MPSIO_READ_CFG_PAGE32)
cmd = MPSIO_READ_CFG_PAGE;
else
cmd = MPSIO_WRITE_CFG_PAGE;
CP(*page32, arg.page, header);
CP(*page32, arg.page, page_address);
PTRIN_CP(*page32, arg.page, buf);
CP(*page32, arg.page, len);
CP(*page32, arg.page, ioc_status);
break;
case MPSIO_READ_EXT_CFG_HEADER32:
case MPSIO_READ_EXT_CFG_PAGE32:
if (cmd32 == MPSIO_READ_EXT_CFG_HEADER32)
cmd = MPSIO_READ_EXT_CFG_HEADER;
else
cmd = MPSIO_READ_EXT_CFG_PAGE;
CP(*ext32, arg.ext, header);
CP(*ext32, arg.ext, page_address);
PTRIN_CP(*ext32, arg.ext, buf);
CP(*ext32, arg.ext, len);
CP(*ext32, arg.ext, ioc_status);
break;
case MPSIO_RAID_ACTION32:
cmd = MPSIO_RAID_ACTION;
CP(*raid32, arg.raid, action);
CP(*raid32, arg.raid, volume_bus);
CP(*raid32, arg.raid, volume_id);
CP(*raid32, arg.raid, phys_disk_num);
CP(*raid32, arg.raid, action_data_word);
PTRIN_CP(*raid32, arg.raid, buf);
CP(*raid32, arg.raid, len);
CP(*raid32, arg.raid, volume_status);
bcopy(raid32->action_data, arg.raid.action_data,
sizeof arg.raid.action_data);
CP(*raid32, arg.raid, ioc_status);
CP(*raid32, arg.raid, write);
break;
case MPSIO_MPS_COMMAND32:
cmd = MPSIO_MPS_COMMAND;
PTRIN_CP(*user32, arg.user, req);
CP(*user32, arg.user, req_len);
PTRIN_CP(*user32, arg.user, rpl);
CP(*user32, arg.user, rpl_len);
PTRIN_CP(*user32, arg.user, buf);
CP(*user32, arg.user, len);
CP(*user32, arg.user, flags);
break;
default:
return (ENOIOCTL);
}
error = mps_ioctl(dev, cmd, &arg, flag, td);
if (error == 0 && (cmd32 & IOC_OUT) != 0) {
switch (cmd32) {
case MPSIO_READ_CFG_HEADER32:
case MPSIO_READ_CFG_PAGE32:
case MPSIO_WRITE_CFG_PAGE32:
CP(arg.page, *page32, header);
CP(arg.page, *page32, page_address);
PTROUT_CP(arg.page, *page32, buf);
CP(arg.page, *page32, len);
CP(arg.page, *page32, ioc_status);
break;
case MPSIO_READ_EXT_CFG_HEADER32:
case MPSIO_READ_EXT_CFG_PAGE32:
CP(arg.ext, *ext32, header);
CP(arg.ext, *ext32, page_address);
PTROUT_CP(arg.ext, *ext32, buf);
CP(arg.ext, *ext32, len);
CP(arg.ext, *ext32, ioc_status);
break;
case MPSIO_RAID_ACTION32:
CP(arg.raid, *raid32, action);
CP(arg.raid, *raid32, volume_bus);
CP(arg.raid, *raid32, volume_id);
CP(arg.raid, *raid32, phys_disk_num);
CP(arg.raid, *raid32, action_data_word);
PTROUT_CP(arg.raid, *raid32, buf);
CP(arg.raid, *raid32, len);
CP(arg.raid, *raid32, volume_status);
bcopy(arg.raid.action_data, raid32->action_data,
sizeof arg.raid.action_data);
CP(arg.raid, *raid32, ioc_status);
CP(arg.raid, *raid32, write);
break;
case MPSIO_MPS_COMMAND32:
PTROUT_CP(arg.user, *user32, req);
CP(arg.user, *user32, req_len);
PTROUT_CP(arg.user, *user32, rpl);
CP(arg.user, *user32, rpl_len);
PTROUT_CP(arg.user, *user32, buf);
CP(arg.user, *user32, len);
CP(arg.user, *user32, flags);
break;
}
}
return (error);
}
#endif /* COMPAT_FREEBSD32 */
static int
mps_ioctl_devsw(struct cdev *dev, u_long com, caddr_t arg, int flag,
struct thread *td)
{
#ifdef COMPAT_FREEBSD32
if (SV_CURPROC_FLAG(SV_ILP32))
return (mps_ioctl32(dev, com, arg, flag, td));
#endif
return (mps_ioctl(dev, com, arg, flag, td));
}