2003-08-24 09:22:26 +00:00
|
|
|
|
/*-
|
2008-04-10 13:05:05 +00:00
|
|
|
|
* Copyright (c) 1998 - 2008 S<EFBFBD>ren Schmidt <sos@FreeBSD.org>
|
2003-08-24 09:22:26 +00:00
|
|
|
|
* All rights reserved.
|
|
|
|
|
*
|
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
|
* are met:
|
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
|
* notice, this list of conditions and the following disclaimer,
|
|
|
|
|
* 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.
|
|
|
|
|
*/
|
|
|
|
|
|
2003-08-24 17:55:58 +00:00
|
|
|
|
#include <sys/cdefs.h>
|
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
2003-08-24 09:22:26 +00:00
|
|
|
|
#include "opt_ata.h"
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
|
#include <sys/systm.h>
|
|
|
|
|
#include <sys/ata.h>
|
|
|
|
|
#include <sys/kernel.h>
|
2004-01-28 20:38:51 +00:00
|
|
|
|
#include <sys/bio.h>
|
2003-08-24 09:22:26 +00:00
|
|
|
|
#include <sys/bus.h>
|
|
|
|
|
#include <sys/conf.h>
|
2004-01-11 22:08:34 +00:00
|
|
|
|
#include <sys/sema.h>
|
2003-08-24 09:22:26 +00:00
|
|
|
|
#include <sys/taskqueue.h>
|
2004-01-14 21:26:35 +00:00
|
|
|
|
#include <vm/uma.h>
|
2003-08-24 09:22:26 +00:00
|
|
|
|
#include <machine/bus.h>
|
|
|
|
|
#include <sys/rman.h>
|
|
|
|
|
#include <dev/ata/ata-all.h>
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
#include <ata_if.h>
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
|
|
|
|
/* prototypes */
|
2003-08-25 09:01:49 +00:00
|
|
|
|
static void ata_completed(void *, int);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
static void ata_sort_queue(struct ata_channel *ch, struct ata_request *request);
|
2003-09-18 09:22:14 +00:00
|
|
|
|
static char *ata_skey2str(u_int8_t);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
ata_queue_request(struct ata_request *request)
|
|
|
|
|
{
|
2006-03-31 08:09:05 +00:00
|
|
|
|
struct ata_channel *ch;
|
2004-09-26 11:48:43 +00:00
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
/* treat request as virgin (this might be an ATA_R_REQUEUE) */
|
2003-08-24 09:22:26 +00:00
|
|
|
|
request->result = request->status = request->error = 0;
|
2004-09-26 11:48:43 +00:00
|
|
|
|
|
2008-04-11 11:30:27 +00:00
|
|
|
|
/* check that the device is still valid */
|
2006-03-31 08:09:05 +00:00
|
|
|
|
if (!(request->parent = device_get_parent(request->dev))) {
|
|
|
|
|
request->result = ENXIO;
|
|
|
|
|
if (request->callback)
|
|
|
|
|
(request->callback)(request);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
ch = device_get_softc(request->parent);
|
|
|
|
|
callout_init_mtx(&request->callout, &ch->state_mtx, CALLOUT_RETURNUNLOCKED);
|
2004-01-19 15:20:00 +00:00
|
|
|
|
if (!request->callback && !(request->flags & ATA_R_REQUEUE))
|
2004-01-14 21:26:35 +00:00
|
|
|
|
sema_init(&request->done, 0, "ATA request done");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2006-02-25 17:27:33 +00:00
|
|
|
|
/* in ATA_STALL_QUEUE state we call HW directly */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if ((ch->state & ATA_STALL_QUEUE) && (request->flags & ATA_R_CONTROL)) {
|
2005-04-21 11:13:39 +00:00
|
|
|
|
mtx_lock(&ch->state_mtx);
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ch->running = request;
|
|
|
|
|
if (ch->hw.begin_transaction(request) == ATA_OP_FINISHED) {
|
|
|
|
|
ch->running = NULL;
|
2004-01-19 15:20:00 +00:00
|
|
|
|
if (!request->callback)
|
|
|
|
|
sema_destroy(&request->done);
|
2005-04-21 11:13:39 +00:00
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
2004-01-19 15:20:00 +00:00
|
|
|
|
return;
|
2004-01-11 22:08:34 +00:00
|
|
|
|
}
|
2005-04-21 11:13:39 +00:00
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
}
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* otherwise put request on the locked queue at the specified location */
|
2004-01-11 22:08:34 +00:00
|
|
|
|
else {
|
2004-09-26 11:48:43 +00:00
|
|
|
|
mtx_lock(&ch->queue_mtx);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (request->flags & ATA_R_AT_HEAD)
|
2004-09-26 11:48:43 +00:00
|
|
|
|
TAILQ_INSERT_HEAD(&ch->ata_queue, request, chain);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
else if (request->flags & ATA_R_ORDERED)
|
|
|
|
|
ata_sort_queue(ch, request);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
else
|
2004-09-26 11:48:43 +00:00
|
|
|
|
TAILQ_INSERT_TAIL(&ch->ata_queue, request, chain);
|
|
|
|
|
mtx_unlock(&ch->queue_mtx);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "queued");
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
ata_start(ch->dev);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
2004-01-19 15:20:00 +00:00
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* if this is a requeued request callback/sleep we're done */
|
2004-01-19 15:20:00 +00:00
|
|
|
|
if (request->flags & ATA_R_REQUEUE)
|
|
|
|
|
return;
|
|
|
|
|
|
|
|
|
|
/* if this is not a callback wait until request is completed */
|
|
|
|
|
if (!request->callback) {
|
2006-12-23 12:40:54 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "wait for completion");
|
2006-03-01 07:48:41 +00:00
|
|
|
|
if (!dumping &&
|
|
|
|
|
sema_timedwait(&request->done, request->timeout * hz * 4)) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev,
|
2006-03-01 07:48:41 +00:00
|
|
|
|
"WARNING - %s taskqueue timeout "
|
|
|
|
|
"- completing request directly\n",
|
|
|
|
|
ata_cmd2str(request));
|
|
|
|
|
request->flags |= ATA_R_DANGER1;
|
|
|
|
|
ata_completed(request, 0);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
}
|
2004-01-14 21:26:35 +00:00
|
|
|
|
sema_destroy(&request->done);
|
2004-01-19 15:20:00 +00:00
|
|
|
|
}
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int
|
2005-04-30 16:22:07 +00:00
|
|
|
|
ata_controlcmd(device_t dev, u_int8_t command, u_int16_t feature,
|
2003-08-24 09:22:26 +00:00
|
|
|
|
u_int64_t lba, u_int16_t count)
|
|
|
|
|
{
|
2009-02-17 23:20:04 +00:00
|
|
|
|
struct ata_device *atadev = device_get_softc(dev);
|
2008-04-17 12:29:35 +00:00
|
|
|
|
struct ata_request *request = ata_alloc_request();
|
2003-08-24 09:22:26 +00:00
|
|
|
|
int error = ENOMEM;
|
|
|
|
|
|
|
|
|
|
if (request) {
|
2008-04-17 12:29:35 +00:00
|
|
|
|
request->dev = dev;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
request->u.ata.command = command;
|
|
|
|
|
request->u.ata.lba = lba;
|
|
|
|
|
request->u.ata.count = count;
|
|
|
|
|
request->u.ata.feature = feature;
|
|
|
|
|
request->flags = ATA_R_CONTROL;
|
2009-02-17 23:20:04 +00:00
|
|
|
|
if (atadev->spindown_state) {
|
|
|
|
|
device_printf(dev, "request while spun down, starting.\n");
|
|
|
|
|
atadev->spindown_state = 0;
|
|
|
|
|
request->timeout = 31;
|
|
|
|
|
} else {
|
|
|
|
|
request->timeout = 5;
|
|
|
|
|
}
|
2004-08-16 09:32:35 +00:00
|
|
|
|
request->retries = 0;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
ata_queue_request(request);
|
|
|
|
|
error = request->result;
|
|
|
|
|
ata_free_request(request);
|
|
|
|
|
}
|
|
|
|
|
return error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
int
|
2005-04-30 16:22:07 +00:00
|
|
|
|
ata_atapicmd(device_t dev, u_int8_t *ccb, caddr_t data,
|
2003-08-24 09:22:26 +00:00
|
|
|
|
int count, int flags, int timeout)
|
|
|
|
|
{
|
2008-04-17 12:29:35 +00:00
|
|
|
|
struct ata_request *request = ata_alloc_request();
|
2005-04-30 16:22:07 +00:00
|
|
|
|
struct ata_device *atadev = device_get_softc(dev);
|
2004-02-28 17:47:27 +00:00
|
|
|
|
int error = ENOMEM;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
|
|
|
|
if (request) {
|
2008-04-17 12:29:35 +00:00
|
|
|
|
request->dev = dev;
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if ((atadev->param.config & ATA_PROTO_MASK) == ATA_PROTO_ATAPI_12)
|
2004-02-28 17:47:27 +00:00
|
|
|
|
bcopy(ccb, request->u.atapi.ccb, 12);
|
2004-01-12 09:33:10 +00:00
|
|
|
|
else
|
2004-02-28 17:47:27 +00:00
|
|
|
|
bcopy(ccb, request->u.atapi.ccb, 16);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
request->data = data;
|
|
|
|
|
request->bytecount = count;
|
|
|
|
|
request->transfersize = min(request->bytecount, 65534);
|
|
|
|
|
request->flags = flags | ATA_R_ATAPI;
|
|
|
|
|
request->timeout = timeout;
|
2005-05-03 07:55:07 +00:00
|
|
|
|
request->retries = 0;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
ata_queue_request(request);
|
|
|
|
|
error = request->result;
|
|
|
|
|
ata_free_request(request);
|
|
|
|
|
}
|
|
|
|
|
return error;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
ata_start(device_t dev)
|
2003-08-24 09:22:26 +00:00
|
|
|
|
{
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
struct ata_channel *ch = device_get_softc(dev);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
struct ata_request *request;
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
struct ata_composite *cptr;
|
|
|
|
|
int dependencies = 0;
|
2004-08-05 21:13:41 +00:00
|
|
|
|
|
2004-10-13 15:16:35 +00:00
|
|
|
|
/* if we have a request on the queue try to get it running */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
mtx_lock(&ch->queue_mtx);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
if ((request = TAILQ_FIRST(&ch->ata_queue))) {
|
2004-01-11 22:08:34 +00:00
|
|
|
|
|
2004-10-13 15:16:35 +00:00
|
|
|
|
/* we need the locking function to get the lock for this channel */
|
2005-03-31 15:05:40 +00:00
|
|
|
|
if (ATA_LOCKING(dev, ATA_LF_LOCK) == ch->unit) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
|
|
|
|
|
/* check for composite dependencies */
|
|
|
|
|
if ((cptr = request->composite)) {
|
|
|
|
|
mtx_lock(&cptr->lock);
|
|
|
|
|
if ((request->flags & ATA_R_WRITE) &&
|
|
|
|
|
(cptr->wr_depend & cptr->rd_done) != cptr->wr_depend) {
|
|
|
|
|
dependencies = 1;
|
|
|
|
|
}
|
|
|
|
|
mtx_unlock(&cptr->lock);
|
|
|
|
|
}
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* check we are in the right state and has no dependencies */
|
2004-09-26 11:48:43 +00:00
|
|
|
|
mtx_lock(&ch->state_mtx);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (ch->state == ATA_IDLE && !dependencies) {
|
2004-10-19 20:11:23 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "starting");
|
2004-10-13 15:16:35 +00:00
|
|
|
|
TAILQ_REMOVE(&ch->ata_queue, request, chain);
|
|
|
|
|
ch->running = request;
|
2004-10-19 20:11:23 +00:00
|
|
|
|
ch->state = ATA_ACTIVE;
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
|
|
|
|
|
/* if we are the freezing point release it */
|
|
|
|
|
if (ch->freezepoint == request)
|
|
|
|
|
ch->freezepoint = NULL;
|
|
|
|
|
|
2004-10-13 15:16:35 +00:00
|
|
|
|
if (ch->hw.begin_transaction(request) == ATA_OP_FINISHED) {
|
|
|
|
|
ch->running = NULL;
|
|
|
|
|
ch->state = ATA_IDLE;
|
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
mtx_unlock(&ch->queue_mtx);
|
2005-03-31 15:05:40 +00:00
|
|
|
|
ATA_LOCKING(dev, ATA_LF_UNLOCK);
|
2004-10-13 15:16:35 +00:00
|
|
|
|
ata_finish(request);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2006-01-18 09:14:55 +00:00
|
|
|
|
if (dumping) {
|
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
|
|
|
|
mtx_unlock(&ch->queue_mtx);
|
2009-02-17 21:17:21 +00:00
|
|
|
|
while (ch->running) {
|
|
|
|
|
ata_interrupt(ch);
|
2006-01-18 13:10:17 +00:00
|
|
|
|
DELAY(10);
|
2009-02-17 21:17:21 +00:00
|
|
|
|
}
|
2006-01-18 09:14:55 +00:00
|
|
|
|
return;
|
2006-01-18 13:10:17 +00:00
|
|
|
|
}
|
2004-10-13 15:16:35 +00:00
|
|
|
|
}
|
2004-09-26 11:48:43 +00:00
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
|
|
|
|
}
|
2004-08-05 21:13:41 +00:00
|
|
|
|
}
|
2004-10-13 15:16:35 +00:00
|
|
|
|
mtx_unlock(&ch->queue_mtx);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
|
|
|
|
ata_finish(struct ata_request *request)
|
|
|
|
|
{
|
2006-02-23 20:15:22 +00:00
|
|
|
|
struct ata_channel *ch = device_get_softc(request->parent);
|
2004-08-16 09:32:35 +00:00
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/*
|
|
|
|
|
* if in ATA_STALL_QUEUE state or request has ATA_R_DIRECT flags set
|
|
|
|
|
* we need to call ata_complete() directly here (no taskqueue involvement)
|
|
|
|
|
*/
|
2006-01-18 09:14:55 +00:00
|
|
|
|
if (dumping ||
|
|
|
|
|
(ch->state & ATA_STALL_QUEUE) || (request->flags & ATA_R_DIRECT)) {
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "finish directly");
|
2004-01-11 22:08:34 +00:00
|
|
|
|
ata_completed(request, 0);
|
|
|
|
|
}
|
|
|
|
|
else {
|
2006-12-23 12:40:54 +00:00
|
|
|
|
/* put request on the proper taskqueue for completion */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (request->bio && !(request->flags & (ATA_R_THREAD | ATA_R_TIMEOUT))){
|
2004-10-13 15:16:35 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "finish bio_taskqueue");
|
2004-01-28 20:38:51 +00:00
|
|
|
|
bio_taskqueue(request->bio, (bio_task_t *)ata_completed, request);
|
2004-09-26 11:48:43 +00:00
|
|
|
|
}
|
2004-01-28 20:38:51 +00:00
|
|
|
|
else {
|
|
|
|
|
TASK_INIT(&request->task, 0, ata_completed, request);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "finish taskqueue_swi");
|
|
|
|
|
taskqueue_enqueue(taskqueue_swi, &request->task);
|
2004-01-28 20:38:51 +00:00
|
|
|
|
}
|
2004-01-11 22:08:34 +00:00
|
|
|
|
}
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
2004-01-28 20:38:51 +00:00
|
|
|
|
ata_completed(void *context, int dummy)
|
2003-08-24 09:22:26 +00:00
|
|
|
|
{
|
|
|
|
|
struct ata_request *request = (struct ata_request *)context;
|
2006-02-23 20:15:22 +00:00
|
|
|
|
struct ata_channel *ch = device_get_softc(request->parent);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
struct ata_device *atadev = device_get_softc(request->dev);
|
|
|
|
|
struct ata_composite *composite;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2006-03-01 07:48:41 +00:00
|
|
|
|
if (request->flags & ATA_R_DANGER2) {
|
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"WARNING - %s freeing taskqueue zombie request\n",
|
|
|
|
|
ata_cmd2str(request));
|
|
|
|
|
request->flags &= ~(ATA_R_DANGER1 | ATA_R_DANGER2);
|
|
|
|
|
ata_free_request(request);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
if (request->flags & ATA_R_DANGER1)
|
2006-04-17 10:47:01 +00:00
|
|
|
|
request->flags |= ATA_R_DANGER2;
|
2006-03-01 07:48:41 +00:00
|
|
|
|
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "completed entered");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2004-10-13 15:16:35 +00:00
|
|
|
|
/* if we had a timeout, reinit channel and deal with the falldown */
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if (request->flags & ATA_R_TIMEOUT) {
|
2004-10-20 10:11:05 +00:00
|
|
|
|
/*
|
2006-03-09 10:25:35 +00:00
|
|
|
|
* if the channel is still present and
|
|
|
|
|
* reinit succeeds and
|
|
|
|
|
* the device doesn't get detached and
|
2005-04-21 11:13:39 +00:00
|
|
|
|
* there are retries left we reinject this request
|
2004-10-20 10:11:05 +00:00
|
|
|
|
*/
|
2006-03-09 10:25:35 +00:00
|
|
|
|
if (ch && !ata_reinit(ch->dev) && !request->result &&
|
2005-09-14 12:45:06 +00:00
|
|
|
|
(request->retries-- > 0)) {
|
2005-04-21 11:13:39 +00:00
|
|
|
|
if (!(request->flags & ATA_R_QUIET)) {
|
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"TIMEOUT - %s retrying (%d retr%s left)",
|
|
|
|
|
ata_cmd2str(request), request->retries,
|
|
|
|
|
request->retries == 1 ? "y" : "ies");
|
|
|
|
|
if (!(request->flags & (ATA_R_ATAPI | ATA_R_CONTROL)))
|
2006-03-09 08:34:44 +00:00
|
|
|
|
printf(" LBA=%ju", request->u.ata.lba);
|
2005-04-21 11:13:39 +00:00
|
|
|
|
printf("\n");
|
|
|
|
|
}
|
2004-09-26 11:48:43 +00:00
|
|
|
|
request->flags &= ~(ATA_R_TIMEOUT | ATA_R_DEBUG);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
request->flags |= (ATA_R_AT_HEAD | ATA_R_REQUEUE);
|
2004-10-13 15:16:35 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "completed reinject");
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ata_queue_request(request);
|
|
|
|
|
return;
|
2004-01-11 22:08:34 +00:00
|
|
|
|
}
|
2004-09-26 11:48:43 +00:00
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* ran out of good intentions so finish with error */
|
2005-09-14 12:45:06 +00:00
|
|
|
|
if (!request->result) {
|
|
|
|
|
if (!(request->flags & ATA_R_QUIET)) {
|
|
|
|
|
if (request->dev) {
|
|
|
|
|
device_printf(request->dev, "FAILURE - %s timed out",
|
|
|
|
|
ata_cmd2str(request));
|
|
|
|
|
if (!(request->flags & (ATA_R_ATAPI | ATA_R_CONTROL)))
|
2006-03-09 08:34:44 +00:00
|
|
|
|
printf(" LBA=%ju", request->u.ata.lba);
|
2005-09-14 12:45:06 +00:00
|
|
|
|
printf("\n");
|
|
|
|
|
}
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
}
|
2004-09-26 11:48:43 +00:00
|
|
|
|
request->result = EIO;
|
2005-09-14 12:45:06 +00:00
|
|
|
|
}
|
2004-01-11 22:08:34 +00:00
|
|
|
|
}
|
2006-03-07 14:27:54 +00:00
|
|
|
|
else if (!(request->flags & ATA_R_ATAPI) ){
|
2004-10-13 15:16:35 +00:00
|
|
|
|
/* if this is a soft ECC error warn about it */
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* XXX SOS we could do WARF here */
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if ((request->status & (ATA_S_CORR | ATA_S_ERROR)) == ATA_S_CORR) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"WARNING - %s soft error (ECC corrected)",
|
|
|
|
|
ata_cmd2str(request));
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if (!(request->flags & (ATA_R_ATAPI | ATA_R_CONTROL)))
|
2006-03-09 08:34:44 +00:00
|
|
|
|
printf(" LBA=%ju", request->u.ata.lba);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
printf("\n");
|
|
|
|
|
}
|
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
/* if this is a UDMA CRC error we reinject if there are retries left */
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if (request->flags & ATA_R_DMA && request->error & ATA_E_ICRC) {
|
|
|
|
|
if (request->retries-- > 0) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"WARNING - %s UDMA ICRC error (retrying request)",
|
|
|
|
|
ata_cmd2str(request));
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if (!(request->flags & (ATA_R_ATAPI | ATA_R_CONTROL)))
|
2006-03-09 08:34:44 +00:00
|
|
|
|
printf(" LBA=%ju", request->u.ata.lba);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
printf("\n");
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
request->flags |= (ATA_R_AT_HEAD | ATA_R_REQUEUE);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
ata_queue_request(request);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
}
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
switch (request->flags & ATA_R_ATAPI) {
|
2004-01-11 22:08:34 +00:00
|
|
|
|
|
2003-08-24 09:22:26 +00:00
|
|
|
|
/* ATA errors */
|
|
|
|
|
default:
|
2004-06-01 11:37:24 +00:00
|
|
|
|
if (!request->result && request->status & ATA_S_ERROR) {
|
2003-08-24 09:22:26 +00:00
|
|
|
|
if (!(request->flags & ATA_R_QUIET)) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"FAILURE - %s status=%b error=%b",
|
|
|
|
|
ata_cmd2str(request),
|
|
|
|
|
request->status, "\20\10BUSY\7READY\6DMA_READY"
|
|
|
|
|
"\5DSC\4DRQ\3CORRECTABLE\2INDEX\1ERROR",
|
|
|
|
|
request->error, "\20\10ICRC\7UNCORRECTABLE"
|
|
|
|
|
"\6MEDIA_CHANGED\5NID_NOT_FOUND"
|
|
|
|
|
"\4MEDIA_CHANGE_REQEST"
|
|
|
|
|
"\3ABORTED\2NO_MEDIA\1ILLEGAL_LENGTH");
|
2008-11-27 03:37:46 +00:00
|
|
|
|
if ((request->flags & ATA_R_DMA) && request->dma &&
|
2008-04-17 12:29:35 +00:00
|
|
|
|
(request->dma->status & ATA_BMSTAT_ERROR))
|
|
|
|
|
printf(" dma=0x%02x", request->dma->status);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
if (!(request->flags & (ATA_R_ATAPI | ATA_R_CONTROL)))
|
2006-03-09 08:34:44 +00:00
|
|
|
|
printf(" LBA=%ju", request->u.ata.lba);
|
2003-12-15 20:43:17 +00:00
|
|
|
|
printf("\n");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
request->result = EIO;
|
|
|
|
|
}
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
/* ATAPI errors */
|
|
|
|
|
case ATA_R_ATAPI:
|
2003-09-18 09:22:14 +00:00
|
|
|
|
/* skip if result already set */
|
2003-08-24 09:22:26 +00:00
|
|
|
|
if (request->result)
|
|
|
|
|
break;
|
|
|
|
|
|
2004-02-28 17:47:27 +00:00
|
|
|
|
/* if we have a sensekey -> request sense from device */
|
2006-03-31 08:09:05 +00:00
|
|
|
|
if ((request->error & ATA_E_ATAPI_SENSE_MASK) &&
|
|
|
|
|
(request->u.atapi.ccb[0] != ATAPI_REQUEST_SENSE)) {
|
2004-02-28 17:47:27 +00:00
|
|
|
|
static u_int8_t ccb[16] = { ATAPI_REQUEST_SENSE, 0, 0, 0,
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
sizeof(struct atapi_sense),
|
2004-02-28 17:47:27 +00:00
|
|
|
|
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 };
|
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
request->u.atapi.saved_cmd = request->u.atapi.ccb[0];
|
2004-03-02 14:05:12 +00:00
|
|
|
|
bcopy(ccb, request->u.atapi.ccb, 16);
|
2006-03-31 08:09:05 +00:00
|
|
|
|
request->data = (caddr_t)&request->u.atapi.sense;
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->bytecount = sizeof(struct atapi_sense);
|
2005-06-28 09:06:52 +00:00
|
|
|
|
request->donecount = 0;
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->transfersize = sizeof(struct atapi_sense);
|
|
|
|
|
request->timeout = 5;
|
2007-03-13 20:31:56 +00:00
|
|
|
|
request->flags &= (ATA_R_ATAPI | ATA_R_QUIET | ATA_R_DEBUG);
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
request->flags |= (ATA_R_READ | ATA_R_AT_HEAD | ATA_R_REQUEUE);
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "autoissue request sense");
|
2004-02-28 17:47:27 +00:00
|
|
|
|
ata_queue_request(request);
|
|
|
|
|
return;
|
|
|
|
|
}
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
switch (request->u.atapi.sense.key & ATA_SENSE_KEY_MASK) {
|
|
|
|
|
case ATA_SENSE_RECOVERED_ERROR:
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev, "WARNING - %s recovered error\n",
|
|
|
|
|
ata_cmd2str(request));
|
2004-02-28 17:47:27 +00:00
|
|
|
|
/* FALLTHROUGH */
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
case ATA_SENSE_NO_SENSE:
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->result = 0;
|
|
|
|
|
break;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
case ATA_SENSE_NOT_READY:
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->result = EBUSY;
|
|
|
|
|
break;
|
|
|
|
|
|
2006-03-31 08:09:05 +00:00
|
|
|
|
case ATA_SENSE_UNIT_ATTENTION:
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
atadev->flags |= ATA_D_MEDIA_CHANGED;
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->result = EIO;
|
|
|
|
|
break;
|
|
|
|
|
|
|
|
|
|
default:
|
|
|
|
|
request->result = EIO;
|
|
|
|
|
if (request->flags & ATA_R_QUIET)
|
2003-08-24 09:22:26 +00:00
|
|
|
|
break;
|
|
|
|
|
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
device_printf(request->dev,
|
|
|
|
|
"FAILURE - %s %s asc=0x%02x ascq=0x%02x ",
|
|
|
|
|
ata_cmd2str(request), ata_skey2str(
|
2006-03-31 08:09:05 +00:00
|
|
|
|
(request->u.atapi.sense.key & ATA_SENSE_KEY_MASK)),
|
|
|
|
|
request->u.atapi.sense.asc,
|
|
|
|
|
request->u.atapi.sense.ascq);
|
|
|
|
|
if (request->u.atapi.sense.specific & ATA_SENSE_SPEC_VALID)
|
|
|
|
|
printf("sks=0x%02x 0x%02x 0x%02x\n",
|
|
|
|
|
request->u.atapi.sense.specific & ATA_SENSE_SPEC_MASK,
|
|
|
|
|
request->u.atapi.sense.specific1,
|
|
|
|
|
request->u.atapi.sense.specific2);
|
2006-07-21 19:13:05 +00:00
|
|
|
|
else
|
|
|
|
|
printf("\n");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
2004-02-28 17:47:27 +00:00
|
|
|
|
|
2009-01-12 17:18:58 +00:00
|
|
|
|
if (!request->result &&
|
|
|
|
|
(request->u.atapi.sense.key & ATA_SENSE_KEY_MASK ?
|
2006-03-31 08:09:05 +00:00
|
|
|
|
request->u.atapi.sense.key & ATA_SENSE_KEY_MASK :
|
|
|
|
|
request->error))
|
2004-02-28 17:47:27 +00:00
|
|
|
|
request->result = EIO;
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
2004-01-11 22:08:34 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "completed callback/wakeup");
|
|
|
|
|
|
2005-12-27 16:49:33 +00:00
|
|
|
|
/* if we are part of a composite operation we need to maintain progress */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if ((composite = request->composite)) {
|
|
|
|
|
int index = 0;
|
|
|
|
|
|
|
|
|
|
mtx_lock(&composite->lock);
|
2005-12-27 16:49:33 +00:00
|
|
|
|
|
|
|
|
|
/* update whats done */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (request->flags & ATA_R_READ)
|
|
|
|
|
composite->rd_done |= (1 << request->this);
|
|
|
|
|
if (request->flags & ATA_R_WRITE)
|
|
|
|
|
composite->wr_done |= (1 << request->this);
|
|
|
|
|
|
2005-12-27 16:49:33 +00:00
|
|
|
|
/* find ready to go dependencies */
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (composite->wr_depend &&
|
|
|
|
|
(composite->rd_done & composite->wr_depend)==composite->wr_depend &&
|
|
|
|
|
(composite->wr_needed & (~composite->wr_done))) {
|
2005-12-27 16:49:33 +00:00
|
|
|
|
index = composite->wr_needed & ~composite->wr_done;
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
}
|
2005-12-27 16:49:33 +00:00
|
|
|
|
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
mtx_unlock(&composite->lock);
|
2005-12-27 16:49:33 +00:00
|
|
|
|
|
|
|
|
|
/* if we have any ready candidates kick them off */
|
|
|
|
|
if (index) {
|
|
|
|
|
int bit;
|
|
|
|
|
|
|
|
|
|
for (bit = 0; bit < MAX_COMPOSITES; bit++) {
|
|
|
|
|
if (index & (1 << bit))
|
2006-01-18 13:10:17 +00:00
|
|
|
|
ata_start(device_get_parent(composite->request[bit]->dev));
|
2005-12-27 16:49:33 +00:00
|
|
|
|
}
|
|
|
|
|
}
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
2005-12-27 16:49:33 +00:00
|
|
|
|
/* get results back to the initiator for this request */
|
2003-08-24 09:22:26 +00:00
|
|
|
|
if (request->callback)
|
|
|
|
|
(request->callback)(request);
|
|
|
|
|
else
|
2004-01-11 22:08:34 +00:00
|
|
|
|
sema_post(&request->done);
|
|
|
|
|
|
2006-03-09 10:25:35 +00:00
|
|
|
|
/* only call ata_start if channel is present */
|
|
|
|
|
if (ch)
|
|
|
|
|
ata_start(ch->dev);
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
|
2005-04-21 11:13:39 +00:00
|
|
|
|
void
|
2003-08-24 09:22:26 +00:00
|
|
|
|
ata_timeout(struct ata_request *request)
|
|
|
|
|
{
|
2006-02-23 20:15:22 +00:00
|
|
|
|
struct ata_channel *ch = device_get_softc(request->parent);
|
2004-09-26 11:48:43 +00:00
|
|
|
|
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
//request->flags |= ATA_R_DEBUG;
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ATA_DEBUG_RQ(request, "timeout");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
|
2004-09-26 11:48:43 +00:00
|
|
|
|
/*
|
2005-09-14 12:45:06 +00:00
|
|
|
|
* if we have an ATA_ACTIVE request running, we flag the request
|
|
|
|
|
* ATA_R_TIMEOUT so ata_finish will handle it correctly
|
|
|
|
|
* also NULL out the running request so we wont loose
|
|
|
|
|
* the race with an eventual interrupt arriving late
|
2004-09-26 11:48:43 +00:00
|
|
|
|
*/
|
2005-09-14 12:45:06 +00:00
|
|
|
|
if (ch->state == ATA_ACTIVE) {
|
2004-09-26 11:48:43 +00:00
|
|
|
|
request->flags |= ATA_R_TIMEOUT;
|
2004-10-13 15:16:35 +00:00
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
2005-03-31 15:05:40 +00:00
|
|
|
|
ATA_LOCKING(ch->dev, ATA_LF_UNLOCK);
|
2008-10-27 09:26:24 +00:00
|
|
|
|
if (ch->dma.unload)
|
|
|
|
|
ch->dma.unload(request);
|
2004-09-26 11:48:43 +00:00
|
|
|
|
ata_finish(request);
|
|
|
|
|
}
|
2004-10-13 15:16:35 +00:00
|
|
|
|
else {
|
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
|
|
|
|
}
|
2004-08-27 14:48:32 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
void
|
2005-09-14 12:45:06 +00:00
|
|
|
|
ata_fail_requests(device_t dev)
|
2004-08-27 14:48:32 +00:00
|
|
|
|
{
|
2005-09-14 12:45:06 +00:00
|
|
|
|
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
|
2006-02-23 20:15:22 +00:00
|
|
|
|
struct ata_request *request, *tmp;
|
|
|
|
|
TAILQ_HEAD(, ata_request) fail_requests;
|
|
|
|
|
TAILQ_INIT(&fail_requests);
|
2004-08-27 14:48:32 +00:00
|
|
|
|
|
2006-02-23 20:15:22 +00:00
|
|
|
|
/* grap all channel locks to avoid races */
|
|
|
|
|
mtx_lock(&ch->queue_mtx);
|
2004-10-13 15:16:35 +00:00
|
|
|
|
mtx_lock(&ch->state_mtx);
|
2006-02-23 20:15:22 +00:00
|
|
|
|
|
|
|
|
|
/* do we have any running request to care about ? */
|
2005-09-14 12:45:06 +00:00
|
|
|
|
if ((request = ch->running) && (!dev || request->dev == dev)) {
|
2005-04-30 16:22:07 +00:00
|
|
|
|
callout_stop(&request->callout);
|
2005-09-14 12:45:06 +00:00
|
|
|
|
ch->running = NULL;
|
|
|
|
|
request->result = ENXIO;
|
2006-02-23 20:15:22 +00:00
|
|
|
|
TAILQ_INSERT_TAIL(&fail_requests, request, chain);
|
2004-08-27 14:48:32 +00:00
|
|
|
|
}
|
2004-03-15 12:03:48 +00:00
|
|
|
|
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
/* fail all requests queued on this channel for device dev if !NULL */
|
2006-02-23 20:15:22 +00:00
|
|
|
|
TAILQ_FOREACH_SAFE(request, &ch->ata_queue, chain, tmp) {
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
if (!dev || request->dev == dev) {
|
2004-03-15 12:03:48 +00:00
|
|
|
|
TAILQ_REMOVE(&ch->ata_queue, request, chain);
|
|
|
|
|
request->result = ENXIO;
|
2006-02-23 20:15:22 +00:00
|
|
|
|
TAILQ_INSERT_TAIL(&fail_requests, request, chain);
|
2004-03-15 12:03:48 +00:00
|
|
|
|
}
|
|
|
|
|
}
|
2006-02-23 20:15:22 +00:00
|
|
|
|
|
|
|
|
|
mtx_unlock(&ch->state_mtx);
|
2004-03-15 12:03:48 +00:00
|
|
|
|
mtx_unlock(&ch->queue_mtx);
|
2006-02-23 20:15:22 +00:00
|
|
|
|
|
|
|
|
|
/* finish up all requests collected above */
|
|
|
|
|
TAILQ_FOREACH_SAFE(request, &fail_requests, chain, tmp) {
|
|
|
|
|
TAILQ_REMOVE(&fail_requests, request, chain);
|
|
|
|
|
ata_finish(request);
|
|
|
|
|
}
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static u_int64_t
|
|
|
|
|
ata_get_lba(struct ata_request *request)
|
|
|
|
|
{
|
|
|
|
|
if (request->flags & ATA_R_ATAPI) {
|
|
|
|
|
switch (request->u.atapi.ccb[0]) {
|
|
|
|
|
case ATAPI_READ_BIG:
|
|
|
|
|
case ATAPI_WRITE_BIG:
|
|
|
|
|
case ATAPI_READ_CD:
|
|
|
|
|
return (request->u.atapi.ccb[5]) | (request->u.atapi.ccb[4]<<8) |
|
|
|
|
|
(request->u.atapi.ccb[3]<<16)|(request->u.atapi.ccb[2]<<24);
|
|
|
|
|
case ATAPI_READ:
|
|
|
|
|
case ATAPI_WRITE:
|
|
|
|
|
return (request->u.atapi.ccb[4]) | (request->u.atapi.ccb[3]<<8) |
|
|
|
|
|
(request->u.atapi.ccb[2]<<16);
|
|
|
|
|
default:
|
|
|
|
|
return 0;
|
|
|
|
|
}
|
2004-03-15 12:03:48 +00:00
|
|
|
|
}
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
else
|
|
|
|
|
return request->u.ata.lba;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static void
|
|
|
|
|
ata_sort_queue(struct ata_channel *ch, struct ata_request *request)
|
|
|
|
|
{
|
|
|
|
|
struct ata_request *this, *next;
|
|
|
|
|
|
|
|
|
|
this = TAILQ_FIRST(&ch->ata_queue);
|
|
|
|
|
|
|
|
|
|
/* if the queue is empty just insert */
|
|
|
|
|
if (!this) {
|
|
|
|
|
if (request->composite)
|
|
|
|
|
ch->freezepoint = request;
|
|
|
|
|
TAILQ_INSERT_TAIL(&ch->ata_queue, request, chain);
|
|
|
|
|
return;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* dont sort frozen parts of the queue */
|
|
|
|
|
if (ch->freezepoint)
|
|
|
|
|
this = ch->freezepoint;
|
|
|
|
|
|
|
|
|
|
/* if position is less than head we add after tipping point */
|
|
|
|
|
if (ata_get_lba(request) < ata_get_lba(this)) {
|
|
|
|
|
while ((next = TAILQ_NEXT(this, chain))) {
|
|
|
|
|
|
|
|
|
|
/* have we reached the tipping point */
|
|
|
|
|
if (ata_get_lba(next) < ata_get_lba(this)) {
|
|
|
|
|
|
|
|
|
|
/* sort the insert */
|
|
|
|
|
do {
|
|
|
|
|
if (ata_get_lba(request) < ata_get_lba(next))
|
|
|
|
|
break;
|
|
|
|
|
this = next;
|
|
|
|
|
} while ((next = TAILQ_NEXT(this, chain)));
|
|
|
|
|
break;
|
|
|
|
|
}
|
|
|
|
|
this = next;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
/* we are after head so sort the insert before tipping point */
|
|
|
|
|
else {
|
|
|
|
|
while ((next = TAILQ_NEXT(this, chain))) {
|
|
|
|
|
if (ata_get_lba(next) < ata_get_lba(this) ||
|
|
|
|
|
ata_get_lba(request) < ata_get_lba(next))
|
|
|
|
|
break;
|
|
|
|
|
this = next;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
if (request->composite)
|
|
|
|
|
ch->freezepoint = request;
|
|
|
|
|
TAILQ_INSERT_AFTER(&ch->ata_queue, this, request, chain);
|
2004-03-15 12:03:48 +00:00
|
|
|
|
}
|
|
|
|
|
|
2003-08-24 09:22:26 +00:00
|
|
|
|
char *
|
|
|
|
|
ata_cmd2str(struct ata_request *request)
|
|
|
|
|
{
|
|
|
|
|
static char buffer[20];
|
|
|
|
|
|
|
|
|
|
if (request->flags & ATA_R_ATAPI) {
|
2006-03-31 08:09:05 +00:00
|
|
|
|
switch (request->u.atapi.sense.key ?
|
|
|
|
|
request->u.atapi.saved_cmd : request->u.atapi.ccb[0]) {
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0x00: return ("TEST_UNIT_READY");
|
|
|
|
|
case 0x01: return ("REZERO");
|
|
|
|
|
case 0x03: return ("REQUEST_SENSE");
|
|
|
|
|
case 0x04: return ("FORMAT");
|
|
|
|
|
case 0x08: return ("READ");
|
|
|
|
|
case 0x0a: return ("WRITE");
|
|
|
|
|
case 0x10: return ("WEOF");
|
|
|
|
|
case 0x11: return ("SPACE");
|
2003-09-19 12:46:12 +00:00
|
|
|
|
case 0x12: return ("INQUIRY");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0x15: return ("MODE_SELECT");
|
|
|
|
|
case 0x19: return ("ERASE");
|
|
|
|
|
case 0x1a: return ("MODE_SENSE");
|
|
|
|
|
case 0x1b: return ("START_STOP");
|
|
|
|
|
case 0x1e: return ("PREVENT_ALLOW");
|
|
|
|
|
case 0x23: return ("ATAPI_READ_FORMAT_CAPACITIES");
|
|
|
|
|
case 0x25: return ("READ_CAPACITY");
|
|
|
|
|
case 0x28: return ("READ_BIG");
|
|
|
|
|
case 0x2a: return ("WRITE_BIG");
|
|
|
|
|
case 0x2b: return ("LOCATE");
|
|
|
|
|
case 0x34: return ("READ_POSITION");
|
|
|
|
|
case 0x35: return ("SYNCHRONIZE_CACHE");
|
|
|
|
|
case 0x3b: return ("WRITE_BUFFER");
|
|
|
|
|
case 0x3c: return ("READ_BUFFER");
|
|
|
|
|
case 0x42: return ("READ_SUBCHANNEL");
|
|
|
|
|
case 0x43: return ("READ_TOC");
|
|
|
|
|
case 0x45: return ("PLAY_10");
|
|
|
|
|
case 0x47: return ("PLAY_MSF");
|
|
|
|
|
case 0x48: return ("PLAY_TRACK");
|
|
|
|
|
case 0x4b: return ("PAUSE");
|
|
|
|
|
case 0x51: return ("READ_DISK_INFO");
|
|
|
|
|
case 0x52: return ("READ_TRACK_INFO");
|
|
|
|
|
case 0x53: return ("RESERVE_TRACK");
|
|
|
|
|
case 0x54: return ("SEND_OPC_INFO");
|
|
|
|
|
case 0x55: return ("MODE_SELECT_BIG");
|
|
|
|
|
case 0x58: return ("REPAIR_TRACK");
|
|
|
|
|
case 0x59: return ("READ_MASTER_CUE");
|
|
|
|
|
case 0x5a: return ("MODE_SENSE_BIG");
|
|
|
|
|
case 0x5b: return ("CLOSE_TRACK/SESSION");
|
|
|
|
|
case 0x5c: return ("READ_BUFFER_CAPACITY");
|
|
|
|
|
case 0x5d: return ("SEND_CUE_SHEET");
|
2006-03-05 20:30:54 +00:00
|
|
|
|
case 0x96: return ("SERVICE_ACTION_IN");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0xa1: return ("BLANK_CMD");
|
|
|
|
|
case 0xa3: return ("SEND_KEY");
|
|
|
|
|
case 0xa4: return ("REPORT_KEY");
|
|
|
|
|
case 0xa5: return ("PLAY_12");
|
|
|
|
|
case 0xa6: return ("LOAD_UNLOAD");
|
|
|
|
|
case 0xad: return ("READ_DVD_STRUCTURE");
|
|
|
|
|
case 0xb4: return ("PLAY_CD");
|
|
|
|
|
case 0xbb: return ("SET_SPEED");
|
|
|
|
|
case 0xbd: return ("MECH_STATUS");
|
|
|
|
|
case 0xbe: return ("READ_CD");
|
|
|
|
|
case 0xff: return ("POLL_DSC");
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
else {
|
|
|
|
|
switch (request->u.ata.command) {
|
|
|
|
|
case 0x00: return ("NOP");
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
case 0x08: return ("DEVICE_RESET");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0x20: return ("READ");
|
|
|
|
|
case 0x24: return ("READ48");
|
|
|
|
|
case 0x25: return ("READ_DMA48");
|
|
|
|
|
case 0x26: return ("READ_DMA_QUEUED48");
|
2008-04-10 13:05:05 +00:00
|
|
|
|
case 0x27: return ("READ_NATIVE_MAX_ADDRESS48");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0x29: return ("READ_MUL48");
|
|
|
|
|
case 0x30: return ("WRITE");
|
|
|
|
|
case 0x34: return ("WRITE48");
|
|
|
|
|
case 0x35: return ("WRITE_DMA48");
|
|
|
|
|
case 0x36: return ("WRITE_DMA_QUEUED48");
|
2008-04-10 13:05:05 +00:00
|
|
|
|
case 0x37: return ("SET_MAX_ADDRESS48");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0x39: return ("WRITE_MUL48");
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
case 0x70: return ("SEEK");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0xa0: return ("PACKET_CMD");
|
|
|
|
|
case 0xa1: return ("ATAPI_IDENTIFY");
|
|
|
|
|
case 0xa2: return ("SERVICE");
|
2007-01-27 21:15:59 +00:00
|
|
|
|
case 0xb0: return ("SMART");
|
2005-04-21 11:13:39 +00:00
|
|
|
|
case 0xc0: return ("CFA ERASE");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
case 0xc4: return ("READ_MUL");
|
|
|
|
|
case 0xc5: return ("WRITE_MUL");
|
|
|
|
|
case 0xc6: return ("SET_MULTI");
|
|
|
|
|
case 0xc7: return ("READ_DMA_QUEUED");
|
|
|
|
|
case 0xc8: return ("READ_DMA");
|
|
|
|
|
case 0xca: return ("WRITE_DMA");
|
|
|
|
|
case 0xcc: return ("WRITE_DMA_QUEUED");
|
|
|
|
|
case 0xe6: return ("SLEEP");
|
|
|
|
|
case 0xe7: return ("FLUSHCACHE");
|
|
|
|
|
case 0xea: return ("FLUSHCACHE48");
|
|
|
|
|
case 0xec: return ("ATA_IDENTIFY");
|
2004-01-11 22:08:34 +00:00
|
|
|
|
case 0xef:
|
|
|
|
|
switch (request->u.ata.feature) {
|
|
|
|
|
case 0x03: return ("SETFEATURES SET TRANSFER MODE");
|
|
|
|
|
case 0x02: return ("SETFEATURES ENABLE WCACHE");
|
|
|
|
|
case 0x82: return ("SETFEATURES DISABLE WCACHE");
|
|
|
|
|
case 0xaa: return ("SETFEATURES ENABLE RCACHE");
|
|
|
|
|
case 0x55: return ("SETFEATURES DISABLE RCACHE");
|
|
|
|
|
}
|
This is the much rumoured ATA mkIII update that I've been working on.
o ATA is now fully newbus'd and split into modules.
This means that on a modern system you just load "atapci and ata"
to get the base support, and then one or more of the device
subdrivers "atadisk atapicd atapifd atapist ataraid".
All can be loaded/unloaded anytime, but for obvious reasons you
dont want to unload atadisk when you have mounted filesystems.
o The device identify part of the probe has been rewritten to fix
the problems with odd devices the old had, and to try to remove
so of the long delays some HW could provoke. Also probing is done
without the need for interrupts, making earlier probing possible.
o SATA devices can be hot inserted/removed and devices will be created/
removed in /dev accordingly.
NOTE: only supported on controllers that has this feature:
Promise and Silicon Image for now.
On other controllers the usual atacontrol detach/attach dance is
still needed.
o Support for "atomic" composite ATA requests used for RAID.
o ATA RAID support has been rewritten and and now supports these
metadata formats:
"Adaptec HostRAID"
"Highpoint V2 RocketRAID"
"Highpoint V3 RocketRAID"
"Intel MatrixRAID"
"Integrated Technology Express"
"LSILogic V2 MegaRAID"
"LSILogic V3 MegaRAID"
"Promise FastTrak"
"Silicon Image Medley"
"FreeBSD PseudoRAID"
o Update the ioctl API to match new RAID levels etc.
o Update atacontrol to know about the new RAID levels etc
NOTE: you need to recompile atacontrol with the new sys/ata.h,
make world will take care of that.
NOTE2: that rebuild is done differently from the old system as
the rebuild is now done piggybacked on read requests to the
array, so atacontrol simply starts a background "dd" to rebuild
the array.
o The reinit code has been worked over to be much more robust.
o The timeout code has been overhauled for races.
o Support of new chipsets.
o Lots of fixes for bugs found while doing the modulerization and
reviewing the old code.
Missing or changed features from current ATA:
o atapi-cd no longer has support for ATAPI changers. Todays its
much cheaper and alot faster to copy those CD images to disk
and serve them from there. Besides they dont seem to be made
anymore, maybe for that exact reason.
o ATA RAID can only read metadata from all the above metadata formats,
not write all of them (Promise and Highpoint V2 so far). This means
that arrays can be picked up from the BIOS, but they cannot be
created from FreeBSD. There is more to it than just the missing
write metadata support, those formats are not unique to a given
controller like Promise and Highpoint formats, instead they exist
for several types, and even worse, some controllers can have
different formats and its impossible to tell which one.
The outcome is that we cannot reliably create the metadata of those
formats and be sure the controller BIOS will understand it.
However write support is needed to update/fail/rebuild the arrays
properly so it sits fairly high on the TODO list.
o So far atapicam is not supported with these changes. When/if this
will change is up to the maintainer of atapi-cam so go there for
questions.
HW donated by: Webveveriet AS
HW donated by: Frode Nordahl
HW donated by: Yahoo!
HW donated by: Sentex
Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
|
|
|
|
sprintf(buffer, "SETFEATURES 0x%02x", request->u.ata.feature);
|
2004-01-11 22:08:34 +00:00
|
|
|
|
return buffer;
|
2008-04-10 13:05:05 +00:00
|
|
|
|
case 0xf5: return ("SECURITY_FREE_LOCK");
|
|
|
|
|
case 0xf8: return ("READ_NATIVE_MAX_ADDRESS");
|
|
|
|
|
case 0xf9: return ("SET_MAX_ADDRESS");
|
2003-08-24 09:22:26 +00:00
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
sprintf(buffer, "unknown CMD (0x%02x)", request->u.ata.command);
|
|
|
|
|
return buffer;
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
static char *
|
2003-09-18 09:22:14 +00:00
|
|
|
|
ata_skey2str(u_int8_t skey)
|
2003-08-24 09:22:26 +00:00
|
|
|
|
{
|
|
|
|
|
switch (skey) {
|
|
|
|
|
case 0x00: return ("NO SENSE");
|
|
|
|
|
case 0x01: return ("RECOVERED ERROR");
|
|
|
|
|
case 0x02: return ("NOT READY");
|
|
|
|
|
case 0x03: return ("MEDIUM ERROR");
|
|
|
|
|
case 0x04: return ("HARDWARE ERROR");
|
|
|
|
|
case 0x05: return ("ILLEGAL REQUEST");
|
|
|
|
|
case 0x06: return ("UNIT ATTENTION");
|
|
|
|
|
case 0x07: return ("DATA PROTECT");
|
|
|
|
|
case 0x08: return ("BLANK CHECK");
|
|
|
|
|
case 0x09: return ("VENDOR SPECIFIC");
|
|
|
|
|
case 0x0a: return ("COPY ABORTED");
|
|
|
|
|
case 0x0b: return ("ABORTED COMMAND");
|
|
|
|
|
case 0x0c: return ("EQUAL");
|
|
|
|
|
case 0x0d: return ("VOLUME OVERFLOW");
|
|
|
|
|
case 0x0e: return ("MISCOMPARE");
|
|
|
|
|
case 0x0f: return ("RESERVED");
|
|
|
|
|
default: return("UNKNOWN");
|
|
|
|
|
}
|
|
|
|
|
}
|