d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
16969d1448
freebsd-dev/sys/cam/ata/ata_pmp.c

864 lines
23 KiB
C
Raw Normal View History

MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
/*-
sys/cam: further adoption of SPDX licensing ID tags. Mainly focus on files that use BSD 2-Clause license, however the tool I was using misidentified many licenses so this was mostly a manual - error prone - task. The Software Package Data Exchange (SPDX) group provides a specification to make it easier for automated tools to detect and summarize well known opensource licenses. We are gradually adopting the specification, noting that the tags are considered only advisory and do not, in any way, superceed or replace the license texts.
2017-11-27 15:12:43 +00:00
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
* Copyright (c) 2009 Alexander Motin <mav@FreeBSD.org>
* 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,
* without modification, immediately at the beginning of the file.
* 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 ``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 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.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#ifdef _KERNEL
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/bio.h>
#include <sys/sysctl.h>
#include <sys/taskqueue.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/conf.h>
#include <sys/devicestat.h>
#include <sys/eventhandler.h>
#include <sys/malloc.h>
#include <sys/cons.h>
#include <geom/geom_disk.h>
#endif /* _KERNEL */
#ifndef _KERNEL
#include <stdio.h>
#include <string.h>
#endif /* _KERNEL */
#include <cam/cam.h>
#include <cam/cam_ccb.h>
#include <cam/cam_periph.h>
#include <cam/cam_xpt_periph.h>
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
#include <cam/cam_xpt_internal.h>
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
#include <cam/cam_sim.h>
#include <cam/ata/ata_all.h>
#ifdef _KERNEL
typedef enum {
PMP_STATE_NORMAL,
PMP_STATE_PORTS,
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
PMP_STATE_PM_QUIRKS_1,
PMP_STATE_PM_QUIRKS_2,
PMP_STATE_PM_QUIRKS_3,
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
PMP_STATE_PRECONFIG,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
PMP_STATE_RESET,
PMP_STATE_CONNECT,
PMP_STATE_CHECK,
PMP_STATE_CLEAR,
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
PMP_STATE_CONFIG,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
PMP_STATE_SCAN
} pmp_state;
typedef enum {
PMP_FLAG_SCTX_INIT = 0x200
} pmp_flags;
typedef enum {
PMP_CCB_PROBE = 0x01,
} pmp_ccb_state;
/* Offsets into our private area for storing information */
#define ccb_state ppriv_field0
#define ccb_bp ppriv_ptr1
struct pmp_softc {
SLIST_ENTRY(pmp_softc) links;
pmp_state state;
pmp_flags flags;
uint32_t pm_pid;
uint32_t pm_prv;
int pm_ports;
int pm_step;
int pm_try;
int found;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
int reset;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
int frozen;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
int restart;
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
int events;
#define PMP_EV_RESET 1
#define PMP_EV_RESCAN 2
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
u_int caps;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
struct task sysctl_task;
struct sysctl_ctx_list sysctl_ctx;
struct sysctl_oid *sysctl_tree;
};
static periph_init_t pmpinit;
static void pmpasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg);
static void pmpsysctlinit(void *context, int pending);
static periph_ctor_t pmpregister;
static periph_dtor_t pmpcleanup;
static periph_start_t pmpstart;
static periph_oninv_t pmponinvalidate;
static void pmpdone(struct cam_periph *periph,
union ccb *done_ccb);
#ifndef PMP_DEFAULT_TIMEOUT
#define PMP_DEFAULT_TIMEOUT 30 /* Timeout in seconds */
#endif
#ifndef PMP_DEFAULT_RETRY
#define PMP_DEFAULT_RETRY 1
#endif
Add tunable/sysctl kern.cam.pmp.hide_special, controlling whether special PMP ports such as PMP configuration or SEMB should be exposed or hidden. These ports were always hidden before as useless and sometimes promatic. But with updated ses driver supporting SEMB it is no longer so straight. Keep ports hidden by default to avoid probe request ttimeouts if SEP is not connected to PMP's SEMB via I2C, that is very often situation.
2012-05-25 08:30:09 +00:00
#ifndef PMP_DEFAULT_HIDE_SPECIAL
#define PMP_DEFAULT_HIDE_SPECIAL 1
#endif
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
static int pmp_retry_count = PMP_DEFAULT_RETRY;
static int pmp_default_timeout = PMP_DEFAULT_TIMEOUT;
Add tunable/sysctl kern.cam.pmp.hide_special, controlling whether special PMP ports such as PMP configuration or SEMB should be exposed or hidden. These ports were always hidden before as useless and sometimes promatic. But with updated ses driver supporting SEMB it is no longer so straight. Keep ports hidden by default to avoid probe request ttimeouts if SEP is not connected to PMP's SEMB via I2C, that is very often situation.
2012-05-25 08:30:09 +00:00
static int pmp_hide_special = PMP_DEFAULT_HIDE_SPECIAL;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (17 of many) r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are still not MPSAFE (or already are but aren’t properly marked). Use it in preparation for a general review of all nodes. This is non-functional change that adds annotations to SYSCTL_NODE and SYSCTL_PROC nodes using one of the soon-to-be-required flags. Mark all obvious cases as MPSAFE. All entries that haven't been marked as MPSAFE before are by default marked as NEEDGIANT Approved by: kib (mentor, blanket) Commented by: kib, gallatin, melifaro Differential Revision: https://reviews.freebsd.org/D23718
2020-02-26 14:26:36 +00:00
static SYSCTL_NODE(_kern_cam, OID_AUTO, pmp, CTLFLAG_RD | CTLFLAG_MPSAFE, 0,
"CAM Direct Access Disk driver");
Pull in r267961 and r267973 again. Fix for issues reported will follow.
2014-06-28 03:56:17 +00:00
SYSCTL_INT(_kern_cam_pmp, OID_AUTO, retry_count, CTLFLAG_RWTUN,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
&pmp_retry_count, 0, "Normal I/O retry count");
Pull in r267961 and r267973 again. Fix for issues reported will follow.
2014-06-28 03:56:17 +00:00
SYSCTL_INT(_kern_cam_pmp, OID_AUTO, default_timeout, CTLFLAG_RWTUN,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
&pmp_default_timeout, 0, "Normal I/O timeout (in seconds)");
Pull in r267961 and r267973 again. Fix for issues reported will follow.
2014-06-28 03:56:17 +00:00
SYSCTL_INT(_kern_cam_pmp, OID_AUTO, hide_special, CTLFLAG_RWTUN,
Add tunable/sysctl kern.cam.pmp.hide_special, controlling whether special PMP ports such as PMP configuration or SEMB should be exposed or hidden. These ports were always hidden before as useless and sometimes promatic. But with updated ses driver supporting SEMB it is no longer so straight. Keep ports hidden by default to avoid probe request ttimeouts if SEP is not connected to PMP's SEMB via I2C, that is very often situation.
2012-05-25 08:30:09 +00:00
&pmp_hide_special, 0, "Hide extra ports");
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
static struct periph_driver pmpdriver =
{
pmpinit, "pmp",
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
TAILQ_HEAD_INITIALIZER(pmpdriver.units), /* generation */ 0,
CAM_PERIPH_DRV_EARLY
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
};
PERIPHDRIVER_DECLARE(pmp, pmpdriver);
static void
pmpinit(void)
{
cam_status status;
/*
* Install a global async callback. This callback will
* receive async callbacks like "new device found".
*/
status = xpt_register_async(AC_FOUND_DEVICE, pmpasync, NULL, NULL);
if (status != CAM_REQ_CMP) {
printf("pmp: Failed to attach master async callback "
"due to status 0x%x!\n", status);
}
}
static void
pmpfreeze(struct cam_periph *periph, int mask)
{
struct pmp_softc *softc = (struct pmp_softc *)periph->softc;
struct cam_path *dpath;
int i;
mask &= ~softc->frozen;
for (i = 0; i < 15; i++) {
if ((mask & (1 << i)) == 0)
continue;
if (xpt_create_path(&dpath, periph,
xpt_path_path_id(periph->path),
i, 0) == CAM_REQ_CMP) {
softc->frozen |= (1 << i);
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
xpt_acquire_device(dpath->device);
MFprojects/camlock r248894: Use full freeze while PMP does hard reset. This is only cosmetical change.
2013-04-13 14:03:44 +00:00
cam_freeze_devq(dpath);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
xpt_free_path(dpath);
}
}
}
static void
pmprelease(struct cam_periph *periph, int mask)
{
struct pmp_softc *softc = (struct pmp_softc *)periph->softc;
struct cam_path *dpath;
int i;
mask &= softc->frozen;
for (i = 0; i < 15; i++) {
if ((mask & (1 << i)) == 0)
continue;
if (xpt_create_path(&dpath, periph,
xpt_path_path_id(periph->path),
i, 0) == CAM_REQ_CMP) {
softc->frozen &= ~(1 << i);
MFprojects/camlock r248894: Use full freeze while PMP does hard reset. This is only cosmetical change.
2013-04-13 14:03:44 +00:00
cam_release_devq(dpath, 0, 0, 0, FALSE);
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
xpt_release_device(dpath->device);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
xpt_free_path(dpath);
}
}
}
static void
pmponinvalidate(struct cam_periph *periph)
{
struct cam_path *dpath;
int i;
/*
* De-register any async callbacks.
*/
xpt_register_async(0, pmpasync, periph, periph->path);
for (i = 0; i < 15; i++) {
if (xpt_create_path(&dpath, periph,
xpt_path_path_id(periph->path),
i, 0) == CAM_REQ_CMP) {
xpt_async(AC_LOST_DEVICE, dpath, NULL);
xpt_free_path(dpath);
}
}
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
pmprelease(periph, -1);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
static void
pmpcleanup(struct cam_periph *periph)
{
struct pmp_softc *softc;
softc = (struct pmp_softc *)periph->softc;
cam_periph_unlock(periph);
/*
* If we can't free the sysctl tree, oh well...
*/
if ((softc->flags & PMP_FLAG_SCTX_INIT) != 0
&& sysctl_ctx_free(&softc->sysctl_ctx) != 0) {
xpt_print(periph->path, "can't remove sysctl context\n");
}
free(softc, M_DEVBUF);
cam_periph_lock(periph);
}
static void
pmpasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg)
{
struct cam_periph *periph;
struct pmp_softc *softc;
periph = (struct cam_periph *)callback_arg;
switch (code) {
case AC_FOUND_DEVICE:
{
struct ccb_getdev *cgd;
cam_status status;
cam: clean up empty lines in .c and .h files
2020-09-01 22:13:48 +00:00
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
cgd = (struct ccb_getdev *)arg;
if (cgd == NULL)
break;
if (cgd->protocol != PROTO_SATAPM)
break;
/*
* Allocate a peripheral instance for
* this device and start the probe
* process.
*/
status = cam_periph_alloc(pmpregister, pmponinvalidate,
pmpcleanup, pmpstart,
"pmp", CAM_PERIPH_BIO,
Merge CAM locking changes from the projects/camlock branch to radically reduce lock congestion and improve SMP scalability of the SCSI/ATA stack, preparing the ground for the coming next GEOM direct dispatch support. Replace big per-SIM locks with bunch of smaller ones: - per-LUN locks to protect device and peripheral drivers state; - per-target locks to protect list of LUNs on target; - per-bus locks to protect reference counting; - per-send queue locks to protect queue of CCBs to be sent; - per-done queue locks to protect queue of completed CCBs; - remaining per-SIM locks now protect only HBA driver internals. While holding LUN lock it is allowed (while not recommended for performance reasons) to take SIM lock. The opposite acquisition order is forbidden. All the other locks are leaf locks, that can be taken anywhere, but should not be cascaded. Many functions, such as: xpt_action(), xpt_done(), xpt_async(), xpt_create_path(), etc. are no longer require (but allow) SIM lock to be held. To keep compatibility and solve cases where SIM lock can't be dropped, all xpt_async() calls in addition to xpt_done() calls are queued to completion threads for async processing in clean environment without SIM lock held. Instead of single CAM SWI thread, used for commands completion processing before, use multiple (depending on number of CPUs) threads. Load balanced between them using "hash" of the device B:T:L address. HBA drivers that can drop SIM lock during completion processing and have sufficient number of completion threads to efficiently scale to multiple CPUs can use new function xpt_done_direct() to avoid extra context switch. Make ahci(4) driver to use this mechanism depending on hardware setup. Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-21 12:00:26 +00:00
path, pmpasync,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
AC_FOUND_DEVICE, cgd);
if (status != CAM_REQ_CMP
&& status != CAM_REQ_INPROG)
printf("pmpasync: Unable to attach to new device "
"due to status 0x%x\n", status);
break;
}
case AC_SCSI_AEN:
case AC_SENT_BDR:
case AC_BUS_RESET:
softc = (struct pmp_softc *)periph->softc;
cam_periph_async(periph, code, path, arg);
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
if (code == AC_SCSI_AEN)
softc->events |= PMP_EV_RESCAN;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
else
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
softc->events |= PMP_EV_RESET;
if (code == AC_SCSI_AEN && softc->state != PMP_STATE_NORMAL)
break;
xpt_hold_boot();
pmpfreeze(periph, softc->found);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
if (code == AC_SENT_BDR || code == AC_BUS_RESET)
softc->found = 0; /* We have to reset everything. */
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
if (softc->state == PMP_STATE_NORMAL) {
Return a C errno for cam_periph_acquire(). There's no compelling reason to return a cam_status type for this function and doing so only creates confusion with normal C coding practices. It's technically an API change, but the periph API isn't widely used. No efffective change to operation. Reviewed by: imp, mav, ken Sponsored by: Netflix Differential Revision: D14063
2018-02-06 06:42:25 +00:00
if (cam_periph_acquire(periph) == 0) {
Merge CAM locking changes from the projects/camlock branch to radically reduce lock congestion and improve SMP scalability of the SCSI/ATA stack, preparing the ground for the coming next GEOM direct dispatch support. Replace big per-SIM locks with bunch of smaller ones: - per-LUN locks to protect device and peripheral drivers state; - per-target locks to protect list of LUNs on target; - per-bus locks to protect reference counting; - per-send queue locks to protect queue of CCBs to be sent; - per-done queue locks to protect queue of completed CCBs; - remaining per-SIM locks now protect only HBA driver internals. While holding LUN lock it is allowed (while not recommended for performance reasons) to take SIM lock. The opposite acquisition order is forbidden. All the other locks are leaf locks, that can be taken anywhere, but should not be cascaded. Many functions, such as: xpt_action(), xpt_done(), xpt_async(), xpt_create_path(), etc. are no longer require (but allow) SIM lock to be held. To keep compatibility and solve cases where SIM lock can't be dropped, all xpt_async() calls in addition to xpt_done() calls are queued to completion threads for async processing in clean environment without SIM lock held. Instead of single CAM SWI thread, used for commands completion processing before, use multiple (depending on number of CPUs) threads. Load balanced between them using "hash" of the device B:T:L address. HBA drivers that can drop SIM lock during completion processing and have sufficient number of completion threads to efficiently scale to multiple CPUs can use new function xpt_done_direct() to avoid extra context switch. Make ahci(4) driver to use this mechanism depending on hardware setup. Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-21 12:00:26 +00:00
if (softc->pm_pid == 0x37261095 ||
softc->pm_pid == 0x38261095)
softc->state = PMP_STATE_PM_QUIRKS_1;
else
softc->state = PMP_STATE_PRECONFIG;
xpt_schedule(periph, CAM_PRIORITY_DEV);
} else {
pmprelease(periph, softc->found);
xpt_release_boot();
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
} else
softc->restart = 1;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
break;
default:
cam_periph_async(periph, code, path, arg);
break;
}
}
static void
pmpsysctlinit(void *context, int pending)
{
struct cam_periph *periph;
struct pmp_softc *softc;
Reduce size of several on-stack string buffers. Submitted by: Dmitry Luhtionov <dmitryluhtionov@gmail.com> MFC after: 2 weeks
2017-12-13 21:17:00 +00:00
char tmpstr[32], tmpstr2[16];
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
periph = (struct cam_periph *)context;
Return a C errno for cam_periph_acquire(). There's no compelling reason to return a cam_status type for this function and doing so only creates confusion with normal C coding practices. It's technically an API change, but the periph API isn't widely used. No efffective change to operation. Reviewed by: imp, mav, ken Sponsored by: Netflix Differential Revision: D14063
2018-02-06 06:42:25 +00:00
if (cam_periph_acquire(periph) != 0)
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
return;
softc = (struct pmp_softc *)periph->softc;
snprintf(tmpstr, sizeof(tmpstr), "CAM PMP unit %d", periph->unit_number);
snprintf(tmpstr2, sizeof(tmpstr2), "%d", periph->unit_number);
sysctl_ctx_init(&softc->sysctl_ctx);
softc->flags |= PMP_FLAG_SCTX_INIT;
Add label annotations to CAM sysctls. Under kern.cam we have certain sysctls that are per-device, such as the ones under kern.cam.ada.[0-9]+.*. Add a "device_index" label annotation to such sysctls, so that the Prometheus metrics exporter will give all of those metrics the same name. The device number will be added to the metric name as the "device_index" label. Reviewed by: cem Differential Revision: https://reviews.freebsd.org/D8775
2016-12-14 12:53:33 +00:00
softc->sysctl_tree = SYSCTL_ADD_NODE_WITH_LABEL(&softc->sysctl_ctx,
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
SYSCTL_STATIC_CHILDREN(_kern_cam_pmp), OID_AUTO, tmpstr2,
Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (17 of many) r357614 added CTLFLAG_NEEDGIANT to make it easier to find nodes that are still not MPSAFE (or already are but aren’t properly marked). Use it in preparation for a general review of all nodes. This is non-functional change that adds annotations to SYSCTL_NODE and SYSCTL_PROC nodes using one of the soon-to-be-required flags. Mark all obvious cases as MPSAFE. All entries that haven't been marked as MPSAFE before are by default marked as NEEDGIANT Approved by: kib (mentor, blanket) Commented by: kib, gallatin, melifaro Differential Revision: https://reviews.freebsd.org/D23718
2020-02-26 14:26:36 +00:00
CTLFLAG_RD | CTLFLAG_MPSAFE, 0, tmpstr, "device_index");
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
if (softc->sysctl_tree == NULL) {
printf("pmpsysctlinit: unable to allocate sysctl tree\n");
cam_periph_release(periph);
return;
}
cam_periph_release(periph);
}
static cam_status
pmpregister(struct cam_periph *periph, void *arg)
{
struct pmp_softc *softc;
struct ccb_getdev *cgd;
cgd = (struct ccb_getdev *)arg;
if (cgd == NULL) {
printf("pmpregister: no getdev CCB, can't register device\n");
return(CAM_REQ_CMP_ERR);
}
softc = (struct pmp_softc *)malloc(sizeof(*softc), M_DEVBUF,
M_NOWAIT|M_ZERO);
if (softc == NULL) {
printf("pmpregister: Unable to probe new device. "
"Unable to allocate softc\n");
return(CAM_REQ_CMP_ERR);
}
periph->softc = softc;
softc->pm_pid = ((uint32_t *)&cgd->ident_data)[0];
softc->pm_prv = ((uint32_t *)&cgd->ident_data)[1];
TASK_INIT(&softc->sysctl_task, 0, pmpsysctlinit, periph);
xpt_announce_periph(periph, NULL);
/*
* Add async callbacks for bus reset and
* bus device reset calls. I don't bother
* checking if this fails as, in most cases,
* the system will function just fine without
* them and the only alternative would be to
* not attach the device on failure.
*/
xpt_register_async(AC_SENT_BDR | AC_BUS_RESET | AC_LOST_DEVICE |
AC_SCSI_AEN, pmpasync, periph, periph->path);
/*
* Take an exclusive refcount on the periph while pmpstart is called
* to finish the probe. The reference will be dropped in pmpdone at
* the end of probe.
*/
(void)cam_periph_acquire(periph);
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
xpt_hold_boot();
softc->state = PMP_STATE_PORTS;
softc->events = PMP_EV_RESCAN;
xpt_schedule(periph, CAM_PRIORITY_DEV);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
return(CAM_REQ_CMP);
}
static void
pmpstart(struct cam_periph *periph, union ccb *start_ccb)
{
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
struct ccb_trans_settings cts;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
struct ccb_ataio *ataio;
struct pmp_softc *softc;
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
struct cam_path *dpath;
int revision = 0;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
softc = (struct pmp_softc *)periph->softc;
ataio = &start_ccb->ataio;
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("pmpstart\n"));
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
if (softc->restart) {
softc->restart = 0;
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
if (softc->pm_pid == 0x37261095 || softc->pm_pid == 0x38261095)
softc->state = min(softc->state, PMP_STATE_PM_QUIRKS_1);
else
softc->state = min(softc->state, PMP_STATE_PRECONFIG);
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
}
/* Fetch user wanted device speed. */
if (softc->state == PMP_STATE_RESET ||
softc->state == PMP_STATE_CONNECT) {
if (xpt_create_path(&dpath, periph,
xpt_path_path_id(periph->path),
softc->pm_step, 0) == CAM_REQ_CMP) {
bzero(&cts, sizeof(cts));
xpt_setup_ccb(&cts.ccb_h, dpath, CAM_PRIORITY_NONE);
cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
cts.type = CTS_TYPE_USER_SETTINGS;
xpt_action((union ccb *)&cts);
if (cts.xport_specific.sata.valid & CTS_SATA_VALID_REVISION)
revision = cts.xport_specific.sata.revision;
xpt_free_path(dpath);
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
}
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
switch (softc->state) {
case PMP_STATE_PORTS:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_read_cmd(ataio, 2, 15);
break;
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
case PMP_STATE_PM_QUIRKS_1:
case PMP_STATE_PM_QUIRKS_3:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_read_cmd(ataio, 129, 15);
break;
case PMP_STATE_PM_QUIRKS_2:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_write_cmd(ataio, 129, 15, softc->caps & ~0x1);
break;
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
case PMP_STATE_PRECONFIG:
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
/* Get/update host SATA capabilities. */
bzero(&cts, sizeof(cts));
xpt_setup_ccb(&cts.ccb_h, periph->path, CAM_PRIORITY_NONE);
cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
cts.type = CTS_TYPE_CURRENT_SETTINGS;
xpt_action((union ccb *)&cts);
if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
softc->caps = cts.xport_specific.sata.caps;
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
else
softc->caps = 0;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
ata_pm_write_cmd(ataio, 0x60, 15, 0x0);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
break;
case PMP_STATE_RESET:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_write_cmd(ataio, 2, softc->pm_step,
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
(revision << 4) |
((softc->found & (1 << softc->pm_step)) ? 0 : 1));
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
break;
case PMP_STATE_CONNECT:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
ata_pm_write_cmd(ataio, 2, softc->pm_step,
(revision << 4));
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
break;
case PMP_STATE_CHECK:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_read_cmd(ataio, 0, softc->pm_step);
break;
case PMP_STATE_CLEAR:
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->reset = 0;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
ata_pm_write_cmd(ataio, 1, softc->pm_step, 0xFFFFFFFF);
break;
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
case PMP_STATE_CONFIG:
cam_fill_ataio(ataio,
pmp_retry_count,
pmpdone,
/*flags*/CAM_DIR_NONE,
0,
/*data_ptr*/NULL,
/*dxfer_len*/0,
pmp_default_timeout * 1000);
Improve SATA Asynchronous Notification feature support in CAM: - make SATA SIMs announce capabilities to handle SDB with Notification bit; - make PMP driver honor this SIMs capability; - make SATA XPT to negotiate and enable this feature for ATAPI devices. This feature allows supporting SATA ATAPI devices to inform system about some events happened, that may require attention. In my case this allows LG GH22LS50 SATA DVR-RW drive to report tray open/close events. Events reported to CAM in form of AC_SCSI_AEN async. Further they could be used as a hints for checking device status and reporting media change to upper layers, for example, via spoiling mechanism of GEOM.
2011-04-13 16:20:54 +00:00
ata_pm_write_cmd(ataio, 0x60, 15, 0x07 |
((softc->caps & CTS_SATA_CAPS_H_AN) ? 0x08 : 0));
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
break;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
default:
break;
}
xpt_action(start_ccb);
}
static void
pmpdone(struct cam_periph *periph, union ccb *done_ccb)
{
MFp4: - Extend XPT-SIM transfer settings control API. Now it allows to report to SATA SIM number of tags supported by each device, implement ATA mode and SATA revision negotiation for both SATA and PATA SIMs. - Make ahci(4) and siis(4) to use submitted maximum tag number, when scheduling requests. It allows to support NCQ on devices with lower tags count then controller supports. - Make PMP driver to report attached devices connection speeds. - Implement ATA mode negotiation between user settings, device and controller capabilities.
2009-11-24 12:47:58 +00:00
struct ccb_trans_settings cts;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
struct pmp_softc *softc;
struct ccb_ataio *ataio;
Remove some used variables. Found with: Clang Static Analyzer
2011-04-18 13:59:56 +00:00
struct cam_path *dpath;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
u_int32_t priority, res;
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
int i;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
softc = (struct pmp_softc *)periph->softc;
ataio = &done_ccb->ataio;
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("pmpdone\n"));
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
priority = done_ccb->ccb_h.pinfo.priority;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
if ((done_ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) {
Remove unused 4th argument to match the standard error routines. Sponsored by: Netflix Differential Revision: https://reviews.freebsd.org/D13386
2017-12-06 00:29:50 +00:00
if (cam_periph_error(done_ccb, 0, 0) == ERESTART) {
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
return;
} else if ((done_ccb->ccb_h.status & CAM_DEV_QFRZN) != 0) {
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
cam_release_devq(done_ccb->ccb_h.path,
/*relsim_flags*/0,
/*reduction*/0,
/*timeout*/0,
/*getcount_only*/0);
}
goto done;
}
if (softc->restart) {
softc->restart = 0;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
xpt_release_ccb(done_ccb);
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
if (softc->pm_pid == 0x37261095 || softc->pm_pid == 0x38261095)
softc->state = min(softc->state, PMP_STATE_PM_QUIRKS_1);
else
softc->state = min(softc->state, PMP_STATE_PRECONFIG);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
return;
}
switch (softc->state) {
case PMP_STATE_PORTS:
Remove some used variables. Found with: Clang Static Analyzer
2011-04-18 13:59:56 +00:00
softc->pm_ports = (ataio->res.lba_high << 24) +
(ataio->res.lba_mid << 16) +
(ataio->res.lba_low << 8) +
ataio->res.sector_count;
Add tunable/sysctl kern.cam.pmp.hide_special, controlling whether special PMP ports such as PMP configuration or SEMB should be exposed or hidden. These ports were always hidden before as useless and sometimes promatic. But with updated ses driver supporting SEMB it is no longer so straight. Keep ports hidden by default to avoid probe request ttimeouts if SEP is not connected to PMP's SEMB via I2C, that is very often situation.
2012-05-25 08:30:09 +00:00
if (pmp_hide_special) {
/*
* This PMP declares 6 ports, while only 5 of them
* are real. Port 5 is a SEMB port, probing which
* causes timeouts if external SEP is not connected
* to PMP over I2C.
*/
Hide SEMB port of the SiI3826 Port Multiplier by default to avoid extra errors while it tries to talk via I2C to usually missing external SEP. There is tunable to enable it back when needed.
2013-02-22 19:53:12 +00:00
if ((softc->pm_pid == 0x37261095 ||
softc->pm_pid == 0x38261095) &&
softc->pm_ports == 6)
Add tunable/sysctl kern.cam.pmp.hide_special, controlling whether special PMP ports such as PMP configuration or SEMB should be exposed or hidden. These ports were always hidden before as useless and sometimes promatic. But with updated ses driver supporting SEMB it is no longer so straight. Keep ports hidden by default to avoid probe request ttimeouts if SEP is not connected to PMP's SEMB via I2C, that is very often situation.
2012-05-25 08:30:09 +00:00
softc->pm_ports = 5;
/*
* This PMP declares 7 ports, while only 5 of them
* are real. Port 5 is a fake "Config Disk" with
* 640 sectors size. Port 6 is a SEMB port.
*/
if (softc->pm_pid == 0x47261095 && softc->pm_ports == 7)
softc->pm_ports = 5;
/*
* These PMPs have extra configuration port.
*/
if (softc->pm_pid == 0x57231095 ||
softc->pm_pid == 0x57331095 ||
softc->pm_pid == 0x57341095 ||
softc->pm_pid == 0x57441095)
softc->pm_ports--;
}
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
printf("%s%d: %d fan-out ports\n",
periph->periph_name, periph->unit_number,
softc->pm_ports);
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
if (softc->pm_pid == 0x37261095 || softc->pm_pid == 0x38261095)
softc->state = PMP_STATE_PM_QUIRKS_1;
else
softc->state = PMP_STATE_PRECONFIG;
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
case PMP_STATE_PM_QUIRKS_1:
softc->caps = (ataio->res.lba_high << 24) +
(ataio->res.lba_mid << 16) +
(ataio->res.lba_low << 8) +
ataio->res.sector_count;
if (softc->caps & 0x1)
softc->state = PMP_STATE_PM_QUIRKS_2;
else
softc->state = PMP_STATE_PRECONFIG;
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
case PMP_STATE_PM_QUIRKS_2:
if (bootverbose)
softc->state = PMP_STATE_PM_QUIRKS_3;
else
softc->state = PMP_STATE_PRECONFIG;
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
case PMP_STATE_PM_QUIRKS_3:
res = (ataio->res.lba_high << 24) +
(ataio->res.lba_mid << 16) +
(ataio->res.lba_low << 8) +
ataio->res.sector_count;
printf("%s%d: Disabling SiI3x26 R_OK in GSCR_POLL: %x->%x\n",
periph->periph_name, periph->unit_number, softc->caps, res);
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
softc->state = PMP_STATE_PRECONFIG;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
Disable sending Early R_OK on SiI3726/SiI3826 port multipliers. With "cached read" HDD testing and multiple ports busy on a SATA host controller, 3726/3826 PMP will very rarely drop a deferred R_OK that was intended for the host. Symptom will be all 5 drives under test will timeout, get reset, and recover. Submitted by: Rich Futyma <rich.futyma@sanmina.com> MFC after: 2 weeks
2013-05-11 13:21:31 +00:00
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
case PMP_STATE_PRECONFIG:
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->pm_step = 0;
softc->state = PMP_STATE_RESET;
softc->reset |= ~softc->found;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
xpt_release_ccb(done_ccb);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
return;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
case PMP_STATE_RESET:
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->pm_step++;
if (softc->pm_step >= softc->pm_ports) {
softc->pm_step = 0;
cam_freeze_devq(periph->path);
cam_release_devq(periph->path,
RELSIM_RELEASE_AFTER_TIMEOUT,
/*reduction*/0,
/*timeout*/5,
/*getcount_only*/0);
softc->state = PMP_STATE_CONNECT;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
xpt_release_ccb(done_ccb);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
return;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
case PMP_STATE_CONNECT:
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->pm_step++;
if (softc->pm_step >= softc->pm_ports) {
softc->pm_step = 0;
softc->pm_try = 0;
cam_freeze_devq(periph->path);
cam_release_devq(periph->path,
RELSIM_RELEASE_AFTER_TIMEOUT,
/*reduction*/0,
/*timeout*/10,
/*getcount_only*/0);
softc->state = PMP_STATE_CHECK;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
xpt_release_ccb(done_ccb);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
return;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
case PMP_STATE_CHECK:
Remove some used variables. Found with: Clang Static Analyzer
2011-04-18 13:59:56 +00:00
res = (ataio->res.lba_high << 24) +
(ataio->res.lba_mid << 16) +
(ataio->res.lba_low << 8) +
ataio->res.sector_count;
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
if (((res & 0xf0f) == 0x103 && (res & 0x0f0) != 0) ||
(res & 0x600) != 0) {
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
if (bootverbose) {
printf("%s%d: port %d status: %08x\n",
periph->periph_name, periph->unit_number,
softc->pm_step, res);
}
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
/* Report device speed if it is online. */
if ((res & 0xf0f) == 0x103 &&
xpt_create_path(&dpath, periph,
MFp4: - Extend XPT-SIM transfer settings control API. Now it allows to report to SATA SIM number of tags supported by each device, implement ATA mode and SATA revision negotiation for both SATA and PATA SIMs. - Make ahci(4) and siis(4) to use submitted maximum tag number, when scheduling requests. It allows to support NCQ on devices with lower tags count then controller supports. - Make PMP driver to report attached devices connection speeds. - Implement ATA mode negotiation between user settings, device and controller capabilities.
2009-11-24 12:47:58 +00:00
xpt_path_path_id(periph->path),
softc->pm_step, 0) == CAM_REQ_CMP) {
bzero(&cts, sizeof(cts));
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
xpt_setup_ccb(&cts.ccb_h, dpath, CAM_PRIORITY_NONE);
MFp4: - Extend XPT-SIM transfer settings control API. Now it allows to report to SATA SIM number of tags supported by each device, implement ATA mode and SATA revision negotiation for both SATA and PATA SIMs. - Make ahci(4) and siis(4) to use submitted maximum tag number, when scheduling requests. It allows to support NCQ on devices with lower tags count then controller supports. - Make PMP driver to report attached devices connection speeds. - Implement ATA mode negotiation between user settings, device and controller capabilities.
2009-11-24 12:47:58 +00:00
cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
cts.type = CTS_TYPE_CURRENT_SETTINGS;
cts.xport_specific.sata.revision = (res & 0x0f0) >> 4;
cts.xport_specific.sata.valid = CTS_SATA_VALID_REVISION;
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
cts.xport_specific.sata.caps = softc->caps &
Improve SATA Asynchronous Notification feature support in CAM: - make SATA SIMs announce capabilities to handle SDB with Notification bit; - make PMP driver honor this SIMs capability; - make SATA XPT to negotiate and enable this feature for ATAPI devices. This feature allows supporting SATA ATAPI devices to inform system about some events happened, that may require attention. In my case this allows LG GH22LS50 SATA DVR-RW drive to report tray open/close events. Events reported to CAM in form of AC_SCSI_AEN async. Further they could be used as a hints for checking device status and reporting media change to upper layers, for example, via spoiling mechanism of GEOM.
2011-04-13 16:20:54 +00:00
(CTS_SATA_CAPS_H_PMREQ |
CTS_SATA_CAPS_H_DMAAA |
CTS_SATA_CAPS_H_AN);
Make SATA XPT negotiate and enable some additional SATA features, such as: - device initiated power management (some devices support only this way); - Automatic Partial to Slumber Transition (more power saving); - DMA auto-activation (expected to slightly improve performance). More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
cts.xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
MFp4: - Extend XPT-SIM transfer settings control API. Now it allows to report to SATA SIM number of tags supported by each device, implement ATA mode and SATA revision negotiation for both SATA and PATA SIMs. - Make ahci(4) and siis(4) to use submitted maximum tag number, when scheduling requests. It allows to support NCQ on devices with lower tags count then controller supports. - Make PMP driver to report attached devices connection speeds. - Implement ATA mode negotiation between user settings, device and controller capabilities.
2009-11-24 12:47:58 +00:00
xpt_action((union ccb *)&cts);
xpt_free_path(dpath);
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->found |= (1 << softc->pm_step);
softc->pm_step++;
} else {
if (softc->pm_try < 10) {
cam_freeze_devq(periph->path);
cam_release_devq(periph->path,
RELSIM_RELEASE_AFTER_TIMEOUT,
/*reduction*/0,
/*timeout*/10,
/*getcount_only*/0);
softc->pm_try++;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
} else {
MFp4; - Cleanup kernel messages, mostly PMP. - Took references on devices, while PMP reinitializes them, to not let them go and distort freeze reference counting.
2009-12-07 16:23:25 +00:00
if (bootverbose) {
printf("%s%d: port %d status: %08x\n",
periph->periph_name, periph->unit_number,
softc->pm_step, res);
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->found &= ~(1 << softc->pm_step);
if (xpt_create_path(&dpath, periph,
done_ccb->ccb_h.path_id,
softc->pm_step, 0) == CAM_REQ_CMP) {
xpt_async(AC_LOST_DEVICE, dpath, NULL);
xpt_free_path(dpath);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->pm_step++;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
}
if (softc->pm_step >= softc->pm_ports) {
if (softc->reset & softc->found) {
cam_freeze_devq(periph->path);
cam_release_devq(periph->path,
RELSIM_RELEASE_AFTER_TIMEOUT,
/*reduction*/0,
/*timeout*/1000,
/*getcount_only*/0);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->state = PMP_STATE_CLEAR;
softc->pm_step = 0;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
xpt_release_ccb(done_ccb);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
return;
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
case PMP_STATE_CLEAR:
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
softc->pm_step++;
Disable PortMultiplier Async Notifications for time of ports reset. They are useless at that time, but confuse Marvell AHCI. Add quirk for SiI57XX Port Multipliers, to hide extra port.
2009-11-16 15:18:02 +00:00
if (softc->pm_step >= softc->pm_ports) {
softc->state = PMP_STATE_CONFIG;
softc->pm_step = 0;
}
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
case PMP_STATE_CONFIG:
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
for (i = 0; i < softc->pm_ports; i++) {
union ccb *ccb;
if ((softc->found & (1 << i)) == 0)
continue;
if (xpt_create_path(&dpath, periph,
xpt_path_path_id(periph->path),
i, 0) != CAM_REQ_CMP) {
Fix double free on error. Found with: Coverity Prevent(tm) CID: 4573
2010-06-05 08:36:37 +00:00
printf("pmpdone: xpt_create_path failed\n");
continue;
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
}
/* If we did hard reset to this device, inform XPT. */
if ((softc->reset & softc->found & (1 << i)) != 0)
xpt_async(AC_SENT_BDR, dpath, NULL);
/* If rescan requested, scan this device. */
if (softc->events & PMP_EV_RESCAN) {
ccb = xpt_alloc_ccb_nowait();
if (ccb == NULL) {
xpt_free_path(dpath);
goto done;
}
xpt_setup_ccb(&ccb->ccb_h, dpath, CAM_PRIORITY_XPT);
xpt_rescan(ccb);
} else
xpt_free_path(dpath);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
}
break;
default:
break;
}
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
done:
xpt_release_ccb(done_ccb);
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
softc->state = PMP_STATE_NORMAL;
MFp4: Large set of CAM inprovements. - Unify bus reset/probe sequence. Whenever bus attached at boot or later, CAM will automatically reset and scan it. It allows to remove duplicate code from many drivers. - Any bus, attached before CAM completed it's boot-time initialization, will equally join to the process, delaying boot if needed. - New kern.cam.boot_delay loader tunable should help controllers that are still unable to register their buses in time (such as slow USB/ PCCard/ CardBus devices), by adding one more event to wait on boot. - To allow synchronization between different CAM levels, concept of requests priorities was extended. Priorities now split between several "run levels". Device can be freezed at specified level, allowing higher priority requests to pass. For example, no payload requests allowed, until PMP driver enable port. ATA XPT negotiate transfer parameters, periph driver configure caching and so on. - Frozen requests are no more counted by request allocation scheduler. It fixes deadlocks, when frozen low priority payload requests occupying slots, required by higher levels to manage theit execution. - Two last changes were holding proper ATA reinitialization and error recovery implementation. Now it is done: SATA controllers and Port Multipliers now implement automatic hot-plug and should correctly recover from timeouts and bus resets. - Improve SCSI error recovery for devices on buses without automatic sense reporting, such as ATAPI or USB. For example, it allows CAM to wait, while CD drive loads disk, instead of immediately return error status. - Decapitalize diagnostic messages and make them more readable and sensible. - Teach PMP driver to limit maximum speed on fan-out ports. - Make boot wait for PMP scan completes, and make rescan more reliable. - Fix pass driver, to return CCB to user level in case of error. - Increase number of retries in cd driver, as device may return several UAs.
2010-01-28 08:41:30 +00:00
softc->events = 0;
xpt_release_boot();
MFp4: Move Port Multiplier support code out of ATA XPT into pmp periph driver. This is convinient, as PMP itself is a bus target and has own state.
2009-10-23 12:36:42 +00:00
pmprelease(periph, -1);
cam_periph_release_locked(periph);
}
#endif /* _KERNEL */
Reference in New Issue Copy Permalink