Commit Graph

926 Commits

Author SHA1 Message Date
mav
b91310c282 Decrease poll interval from 1000 to 100us. This significantly reduces
kernel dump time, especially with minidump.
2010-09-22 05:17:18 +00:00
mdf
e9f1044ef2 Use destroy_dev_sched(9) instead of destroy_dev(9) in passcleanup() as
it is indirectly a d_close method.

Prompted by:	kib
Reviewed by:	mav
MFC after:	2 weeks
2010-09-20 19:42:14 +00:00
gibbs
6833acab2d Correct bioq_disksort so that bioq_insert_tail() offers barrier semantic.
Add the BIO_ORDERED flag for struct bio and update bio clients to use it.

The barrier semantics of bioq_insert_tail() were broken in two ways:

 o In bioq_disksort(), an added bio could be inserted at the head of
   the queue, even when a barrier was present, if the sort key for
   the new entry was less than that of the last queued barrier bio.

 o The last_offset used to generate the sort key for newly queued bios
   did not stay at the position of the barrier until either the
   barrier was de-queued, or a new barrier (which updates last_offset)
   was queued.  When a barrier is in effect, we know that the disk
   will pass through the barrier position just before the
   "blocked bios" are released, so using the barrier's offset for
   last_offset is the optimal choice.

sys/geom/sched/subr_disk.c:
sys/kern/subr_disk.c:
	o Update last_offset in bioq_insert_tail().

	o Only update last_offset in bioq_remove() if the removed bio is
	  at the head of the queue (typically due to a call via
	  bioq_takefirst()) and no barrier is active.

	o In bioq_disksort(), if we have a barrier (insert_point is non-NULL),
	  set prev to the barrier and cur to it's next element.  Now that
	  last_offset is kept at the barrier position, this change isn't
	  strictly necessary, but since we have to take a decision branch
	  anyway, it does avoid one, no-op, loop iteration in the while
	  loop that immediately follows.

	o In bioq_disksort(), bypass the normal sort for bios with the
	  BIO_ORDERED attribute and instead insert them into the queue
	  with bioq_insert_tail().  bioq_insert_tail() not only gives
	  the desired command order during insertion, but also provides
	  barrier semantics so that commands disksorted in the future
	  cannot pass the just enqueued transaction.

sys/sys/bio.h:
	Add BIO_ORDERED as bit 4 of the bio_flags field in struct bio.

sys/cam/ata/ata_da.c:
sys/cam/scsi/scsi_da.c
	Use an ordered command for SCSI/ATA-NCQ commands issued in
	response to bios with the BIO_ORDERED flag set.

sys/cam/scsi/scsi_da.c
	Use an ordered tag when issuing a synchronize cache command.

	Wrap some lines to 80 columns.

sys/cddl/contrib/opensolaris/uts/common/fs/zfs/vdev_geom.c
sys/geom/geom_io.c
	Mark bios with the BIO_FLUSH command as BIO_ORDERED.

Sponsored by:	Spectra Logic Corporation
MFC after:	1 month
2010-09-02 19:40:28 +00:00
mjacob
6f82627f96 Revert r211434. Offline discussions have convinced me that this should
be left alone for now.
2010-08-20 17:20:05 +00:00
mjacob
6ed7860b67 Now is as good a time as any to find out if we induce breakage
by issueing aborts for any pending commands when we're decommssioning
a disk.

MFC after:	3 months
2010-08-17 17:11:15 +00:00
bcr
28966cf21c Fix two (very common) occurrences of s/sytem/system in comments.
Approved by:    mav
MFC after:      3 days
2010-08-02 18:06:49 +00:00
mav
68b26f6649 Export PCI IDs of ATA/SATA controllers through CAM and ata(4) layers to
GEOM. This information needed for proper soft-RAID's on-disk metadata
reading and writing.
2010-07-25 15:43:52 +00:00
mav
cc2181f051 ATA device reset starts probe sequence from the beginning. If reset caused
by timeout/error of one of probe commands, process may continue infinitely.
Make CAM ATA more robust to faulty devices and false positive detections,
abort probe after two restarts on timeouts or ten on other errors.
2010-07-06 19:05:41 +00:00
mjacob
566f36f961 Don't lock buses around a call to xptperiphlistmatch- the buses will be
locked at appropriate places.

MFC after:	1 week
X-MFC:		208752
2010-06-29 17:10:55 +00:00
ken
56a01a64f1 For the target port groups structures, don't allocate the initial element.
This makes things easier for target implementations to calculate how many
elements they need to allocate.

Discussed with:	mjacob, gibbs
MFC after:	1 week
2010-06-14 22:02:18 +00:00
mjacob
8b02ad3564 Rearrange how things are done to avoid dereferencing stale pointers in
the case of immediate unconfigure after configure. Hold the periph an
extra count while we have the task to create sysctl context outstanding
so that the periph doesn't go away unexpectedly.

Sponsored by:	Panasas
Reviewed by:	scsi@
MFC after:	1 month
2010-06-08 22:46:44 +00:00
mjacob
c62422a6e6 One byte off for scsi_target_group cdb.
MFC after:	2 weeks
2010-06-08 17:44:49 +00:00
mjacob
dcb98911e2 Implement the usage of Report Luns as part of SCSI probing for SCP3 or
better devices. This can be disabled on a per-device basis using quirks as
well.

This also handles the case where there is actually no connected LUN 0
(which can definitely be the case for storage arrays).

Reviewed by:	scsi@
MFC after:	1 month
2010-06-08 16:17:25 +00:00
mjacob
3597cbf7d3 Make additional definitions up to and including SPC-4. Add in definitions
for REPORT and SET TARGET PORT GROUP commands (foundations for future work).

Regularize opcodes to be upper case hex.

Pick *one* of tab or space after #define (tab) and stick with that.

MFC after:	2 weeks
2010-06-08 05:14:18 +00:00
mjacob
710ba0d853 redfacedly noting a missing part of a patch.
Pointy Hat To:	Me
MFC after:	2 Week
X-MFC:		208896
2010-06-07 22:43:37 +00:00
mjacob
de50224e92 Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists
for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
mav
ced689a069 Add allocation error hadling.
Found with:   Coverity Prevent(tm)
CID:          3897
2010-06-05 09:31:13 +00:00
mav
508015d598 Fix double free on error.
Found with:   Coverity Prevent(tm)
CID:          4573
2010-06-05 08:36:37 +00:00
avg
0dfde2cbbd scsi_cd: pass correct pointer to free()
Found with:	Coverity Prevent(tm)
CID:		2986
MFC after:	4 days
2010-06-04 13:55:55 +00:00
mjacob
cfb9348b69 Protect periph drivers list and rearrange things to minimize the chance of
stepping oneself during probing.

Don't blindly decrement a periph probe count.

Reviewed by:	scsi@
Obtained from:	Alexander Motin, Atillio Rao, Others
MFC after:	1 month
2010-06-02 18:06:32 +00:00
mjacob
39b696507f Add a new primitive, XPT_SCAN_TGT, to cover the range between scanning a
whole bus (XPT_SCAN_BUS) and a single lun on that bus (XPT_SCAN_LUN).

It's less resource comsumptive than scanning a whole bus when the
caller knows only one target has changes.

Reviewed by:	scsi@
Sponsored by:	Panasas
MFC after:	1 month
2010-05-26 22:49:42 +00:00
marius
427e94121a Change ad_firmware_geom_adjust() to operate on a struct disk * only and
hook it up to ada(4) also. While at it, rename *ad_firmware_geom_adjust()
to *ata_disk_firmware_geom_adjust() etc now that these are no longer
limited to ad(4).

Reviewed by:	mav
MFC after:	3 days
2010-05-20 12:46:19 +00:00
marius
985b980574 Remove unnecessary headers which were inherited from cam_xpt.c without
reason.

Reviewed by:	mav
MFC after:	3 days
2010-05-20 12:44:06 +00:00
mjacob
1898e1308d Pick up the right change, not it's close cousin. The one
previously submitted was wrong.

Point hat:      mjacob
X-MFC:          207933
MFC after:	1 week
2010-05-11 22:51:13 +00:00
mjacob
1676f7d82f Deal sensibly with more than 26 sg devices. It isn't a complete
solution.

Sponsored by:   Panasas
MFC after:	1 week
2010-05-11 22:22:01 +00:00
mav
90843ddc5c Make SATA XPT negotiate and enable some additional SATA features, such as:
- device initiated power management (some devices support only this way);
 - Automatic Partial to Slumber Transition (more power saving);
 - DMA auto-activation (expected to slightly improve performance).
More features could be added later, when hardware supports.
2010-05-02 12:07:47 +00:00
mav
113965a783 Add xpt_schedule_dev_sendq() call, lost at r203108. It is not needed in
usual operation, but required in some conditions to make queue running
after being shrinked.

MFC after:	3 days
2010-05-02 04:16:39 +00:00
mav
99446b1a22 Revert r198705.
As scottl@ noticed, max_target/max_lun was intended to be only a hint for
existing bus scanner. Some FC/SAS SIMs report fake values there, that are
smaller then maximum supported IDs. In that case this check makes impossible
manual scan outside hinted range.

For ATA/SATA SIMs respective check was instead implemented at SIM level.
Newer SCSI SIMs expected to have these checks at driver or firmware level.
Some older SCSI SIMs have no this check and the issues will get back there.
2010-04-30 08:57:03 +00:00
mav
828890193b Report PMP absence using target 15, same as for precence (not a wildcard),
to not confuse target ID checks at SIMs.
2010-04-30 07:56:21 +00:00
mav
e766c487b0 Update device identify data and serial number when device change detected.
Reprobe immediately following this should have fresh data.
2010-04-27 15:59:38 +00:00
mav
a5b7da1286 MFp4:
Move PI_TAG_ABLE check from ada driver to ATA XPT.
2010-04-26 12:03:55 +00:00
avg
701362298a scsi_cd: CD_FLAG_VALID_MEDIA is sufficient to set d_sectorsize and
d_mediasize

CD_FLAG_VALID_TOC is not required for setting those media properties.

PR:		kern/145385
Submitted by:	Juergen Lock <nox@jelal.kn-bremen.de>
		a slightly different version
Tested by:	Pavel Sukhoy <sukhoy@ripn.net>,
		Markus Wild <m.wild@cybernet.ch>,
		Juergen Lock <nox@jelal.kn-bremen.de>,
		uqs
MFC after:	1 week
2010-04-15 08:29:14 +00:00
mjacob
4de64106dd Nullify our periph pointer in dacleanup to try and catch
race conditions in callbacks/downcalls.
2010-04-02 20:08:44 +00:00
mjacob
452876c6d9 Add a couple missing basic mode page codes.
MFC after:	1 week
2010-04-01 01:49:43 +00:00
mjacob
8b3f44d23f For unhandled actions in xpt_action_default, remember to call
xpt_done for queued requests. This solves the problem of
indefinite hangs for unspecified transports when XPT_SCAN_BUS
is called.

A few minor cosmetics elsewhere.

MFC after:	1 week
2010-03-31 17:47:57 +00:00
mjacob
bc081f5e7d We actually can generate a host number.
MFC after:	1 month
2010-03-17 18:53:58 +00:00
mav
859e0a0f68 Make PUIS detection more strict. Previous implementation caused false
positives on VMWare's virtual CD-ROMs.
2010-02-26 10:42:46 +00:00
mav
119ce28bd7 Fix newlines broken at r204220. 2010-02-26 10:33:48 +00:00
mav
c847086ff1 Make xpt_rescan() more correct, as it was before r197208: do not use
XPT_SCAN_LUN for wildcard LUN, fall back to XPT_SCAN_BUS.
2010-02-23 18:42:07 +00:00
mav
be2b9d35a5 Virtualize transport part of periph announcement. 2010-02-22 19:17:17 +00:00
mav
6941483833 Improve output for controllers that doesn't report SATA speed. 2010-02-22 10:45:40 +00:00
mav
6be9a30d07 Make CD driver a bit more robust and predictable to unreported errors. 2010-02-15 18:14:51 +00:00
mav
744e8a4159 On probe error, if restart requested, skip any retries and recovery.
Just restart probe from the beginning immediately.
2010-02-04 18:56:38 +00:00
mav
3ec3bd1c03 Disable kern.cam.power_down ATM. It doesn't work fine on some controllers. 2010-02-03 20:00:56 +00:00
mav
45cc6dda97 MFp4:
Add Power Up In Stand-by feature support. Device with PUIS enabled
require explicit command to do initial spin-up. Mark that command
with CAM_HIGH_POWER flag, to allow CAM manage staggered spin-up.
2010-02-03 10:06:03 +00:00
mav
fab0a47626 MFp4:
Make CAM to stop all attached devices on system shutdown.
It allows devices to park heads, reducing stress on power loss.
Add `kern.cam.power_down` tunable and sysctl to controll it.
2010-02-03 08:42:08 +00:00
mav
cce36e4d2f Change the way in which fake async events generated. Do not use
taskqueue for lock decoupling, as it causes unwanted races.
2010-02-02 18:07:16 +00:00
mav
d8af202f1a - Use separate buffer for identify data fetching. We can't use main buffer
here if device already running, as data need to be formatted before use.
- Remove some saved_ccb variables. They are unused now.
2010-02-02 18:03:21 +00:00
mav
cb4c470cd9 Some missed parts for r203376. 2010-02-02 17:56:54 +00:00
mav
b5b6d6b0e0 - Give ATA/SATA SIMs info about ATAPI packet size, supported by device.
- Make ATA XPT to reject longer SCSI CDBs then supported by device, or
any SCSI CDBs, if device doesn't support ATAPI.
2010-02-02 11:09:28 +00:00