Commit Graph

56 Commits

Author SHA1 Message Date
Alexander Motin
154c6ffd71 Build NVMe CAM transport unrelated to NVMe SIM.
Before this I suppose it was impossible load CAM-based NVMe as module.
Plus this appeared to be needed to build r345815 without NVMe driver.

MFC after:	2 weeks
2019-04-02 20:27:56 +00:00
Alexander Motin
e40d8dbbcb Make cam_error_print() decode NVMe commands.
MFC after:	2 weeks
2019-04-02 19:37:52 +00:00
Warner Losh
52467047aa Regularize the Netflix copyright
Use recent best practices for Copyright form at the top of
the license:
1. Remove all the All Rights Reserved clauses on our stuff. Where we
   piggybacked others, use a separate line to make things clear.
2. Use "Netflix, Inc." everywhere.
3. Use a single line for the copyright for grep friendliness.
4. Use date ranges in all places for our stuff.

Approved by: Netflix Legal (who gave me the form), adrian@ (pmc files)
2019-02-04 21:28:25 +00:00
Chuck Tuffli
87b3975e36 nda(4) fix check for Dataset Management support
In the nda(4) driver, only set DISKFLAG_CANDELETE (a.k.a. can support
BIO_DELETE) if the drive supports Dataset Management. There are reports
that without this check, VMWare Workstation does not work reliably.

Fix is to check the ONCS field in the NVMe Controller Data structure for
support. This check previously existed but did not survive the
big-endian changes.

Reported by: yuripv@yuripv.net
Reviewed by: imp, mav, jimharris
Approved by: imp (mentor)
MFC after: 2 weeks
Differential Revision: https://reviews.freebsd.org/D18493
2018-12-13 13:25:37 +00:00
Warner Losh
ea657f2c76 Add statistics for TRIM comands
Add a counter for the LBAs, Ranges and hardware commands so that we
can provide additional color to the statistics we provide to vendors.

Sponsored by: Netflix, Inc
2018-10-26 16:23:51 +00:00
Chuck Tuffli
9544e6dcf1 Make NVMe compatible with the original API
The original NVMe API used bit-fields to represent fields in data
structures defined by the specification (e.g. the op-code in the command
data structure). The implementation targeted x86_64 processors and
defined the bit fields for little endian dwords (i.e. 32 bits).

This approach does not work as-is for big endian architectures and was
changed to use a combination of bit shifts and masks to support PowerPC.
Unfortunately, this changed the NVMe API and forces #ifdef's based on
the OS revision level in user space code.

This change reverts to something that looks like the original API, but
it uses bytes instead of bit-fields inside the packed command structure.
As a bonus, this works as-is for both big and little endian CPU
architectures.

Bump __FreeBSD_version to 1200081 due to API change

Reviewed by: imp, kbowling, smh, mav
Approved by: imp (mentor)
Differential Revision: https://reviews.freebsd.org/D16404
2018-08-22 04:29:24 +00:00
Conrad Meyer
f053ca1f08 Walk back r337554 while discussion continues
The idea was to get the uncontroversial mechanical change out of the way,
then get the meatier functional changes reviewed subsequently.  I had not
realized that the immediately adjacent issue was addressed in a different
direction in r334506 (see Warner's guidance in D15592).

Discussion continues, trying to determine if there is a secondary issue
still[1] and how best to fix it.  With 12-related activities coming up,
while that is ongoing, just take this back for now.

[1]: Shutdown-time eventhandler events fire normally during panic's reboot
path.  Driver callbacks that attempt to issue and wait on interrupt-
completed IO may never complete, hanging the system.  This is particularly
obnoxious in the shutdown/panic path, as the debugger cannot be entered
anymore and the hang prevents reboot restoring availability.

(There's nothing CAM-specific about this problem -- any shutdown
event-triggered driver could do something like this during panic.  But most
NICs, etc.  don't try to send spin-down commands at shutdown. ;-))

Discussed with:	imp, markj
2018-08-10 19:19:07 +00:00
Conrad Meyer
2077be2b73 cam(4): Add an xpt-neutral flag indicating a valid panic CCB
No functional change.

Note that this change is careful to set the CCB header xflags after
foo_fill_bar() routines, which generally zero existing flags.  An earlier
version of this patch mistakenly set the flag before the fill routines.

Submitted by:	Scott Ferris <sferris AT isilon.com>, jhibbits@
Reviewed by:	bdrewery@, markj@, and non-committer FreeBSD contributor Anton Rang
Sponsored by:	Dell EMC Isilon
2018-08-09 21:53:32 +00:00
Andriy Gapon
b0af06052c remove unneeded inclusion of sys/interrupt.h from several files
It's likely that the header was needed in the past for swi(9).
But now that code does not use swi(9) or any other interfaces defined
in sys/interrupt.h.

MFC after:	1 week
2018-07-04 09:07:18 +00:00
Eric van Gyzen
2ebb808f8c cam nvme: fix array overrun
Fix a classic array overrun where the index could be one past the end.

Reported by:	Coverity
CID:		1356596
MFC after:	3 days
Sponsored by:	Dell EMC
2018-05-28 03:14:36 +00:00
Alexander Motin
f439e3a4ff Refactor NVMe CAM integration.
- Remove layering violation, when NVMe SIM code accessed CAM internal
device structures to set pointers on controller and namespace data.
Instead make NVMe XPT probe fetch the data directly from hardware.
 - Cleanup NVMe SIM code, fixing support for multiple namespaces per
controller (reporting them as LUNs) and adding controller detach support
and run-time namespace change notifications.
 - Add initial support for namespace change async events.  So far only
in CAM mode, but it allows run-time namespace arrival and departure.
 - Add missing nvme_notify_fail_consumers() call on controller detach.
Together with previous changes this allows NVMe device detach/unplug.

Non-CAM mode still requires a lot of love to stay on par, but at least
CAM mode code should not stay in the way so much, becoming much more
self-sufficient.

Reviewed by:	imp
MFC after:	1 month
Sponsored by:	iXsystems, Inc.
2018-05-25 03:34:33 +00:00
Warner Losh
b1988d44b3 We can't release the refcount outside of the periph lock.
We're dropping the periph lock then dropping the refcount. However,
that violates the locking protocol and is racy. This seems to be
the cause of weird occasional panics with a bogus assert.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D15517
2018-05-24 16:31:18 +00:00
Warner Losh
d9a7a61b2b Hold the reference count until the CCB is released
When a disk disappears and the periph is invalidated, any I/Os that
are pending with the controller can cause a crash when they
complete. Move to holding the softc reference count taken in dastart()
until the I/O is complete rather than only until xpt_action()
returns. (This approach was suggested by Ken Merry.) This extends
the method used in da to ada, nda, and mda.

Sponsored by: Netflix
Submitted by: Chuck Silvers
2018-05-15 22:22:10 +00:00
Warner Losh
041f49aece Remove the 'All Rights Reserved' clause from some of the stuff I've
done for Netflix, since I'm in the neighborhood.
2018-05-09 20:32:23 +00:00
Warner Losh
400326b667 Kill assert I shouldn't have committed 2018-03-20 13:14:10 +00:00
Warner Losh
afdbfe1e1b Starting LBA is a 64bit number, so use htole64 instead of htole32. The
latter casts the LBA to a 32-bit number before assigning it to the 64
bit structure entity. This works fine on the first 2TB of TRIMs, but
terrible beyond that due to trucation.

Also, add an assert to make sure we don't end too many DSM TRIM
entries in one request.

Sponsored by: Netflix
2018-03-20 03:37:14 +00:00
Warner Losh
6f591d13fd Make kern.cam.nda.num_trim tunable to limit the number of BIO_DELETE
requests that we'll collapse into one DSM_TRIM. By default it is a
256, which is the max that will fit into a 4k page.

Sponsored by: Netflix
2018-03-20 03:37:09 +00:00
Warner Losh
fdfc0a83a3 Remove some redundant MPSAFE flags.
This was pointed out in a code review I'm having trouble finding right
now, but go ahead and eliminate these.

Sponsored by: Netfix
2018-03-20 03:37:04 +00:00
Warner Losh
d38677d23c Create a sysctl kern.cam.{,a,n}da.X.invalidate
kern.cam.{,a,n}da.X.invalidate=1 forces *daX to detach by calling
cam_periph_invalidate on the underlying periph. This is for testing
purposes only. Include only with options CAM_TEST_FAILURE and rename
the former [AN]DA_TEST_FAILURE, and fix nda to compile with it set.
We're using it at work to harden geom and the buffer cache to be
resilient in the face of drive failure. Today, it far too often
results in a panic. While much work was done on SIM initiated removal
for the USB thumnb drive removal work, little has been done for periph
initiated removal. This simulates what *daerror() does for some errors
nicely: we get the same panics with it that we do with failing drives.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D14581
2018-03-14 17:53:37 +00:00
Warner Losh
807e94b2c3 Implement trim collapsing in nda
When multiple trims are in the queue, collapse them as much as
possible. At present, this usually results in only a few trims being
collapsed together, but more work on that will make it possible to do
hundreds (up to some configurable max).

Sponsored by: Netflix
2018-03-14 16:44:50 +00:00
Conrad Meyer
2e1fccf2cf nvme_da: Fix minor memory leak in error case
Reported by:	cppcheck
Sponsored by:	Dell EMC Isilon
2018-03-10 01:28:55 +00:00
Wojciech Macek
0d787e9b35 NVMe: Add big-endian support
Remove bitfields from defined structures as they are not portable.
Instead use shift and mask macros in the driver and nvmecontrol application.

NVMe is now working on powerpc64 host.

Submitted by:          Michal Stanek <mst@semihalf.com>
Obtained from:         Semihalf
Reviewed by:           imp, wma
Sponsored by:          IBM, QCM Technologies
Differential revision: https://reviews.freebsd.org/D13916
2018-02-22 13:32:31 +00:00
Warner Losh
0028abe633 Backout r329818, r329816 and r329815.
These aren't the commits I thought I was testing prior to
commit. Revert until I can sort out what happened and fix it.
2018-02-22 11:18:33 +00:00
Warner Losh
4d87e27125 Combine BIO_DELETE requests for nda devices
Now that we're queueing BIO_DELETE requests in the CAM I/O scheduler,
it make sense to try to combine as many as possible into a single
request to send down to hardware. Hopefully, lots of larger requests
like this are better than lots of individual transactions.

Note for future: need to limit based on total size of the trim
request. Should also collapse adjacent ranges where possible to
increase the size of the max payload.

Sponsored by: Netflix
2018-02-22 05:44:00 +00:00
Warner Losh
c5fe3ae9b8 Introduce capacity flags for periphs
Introduce flags word to describe the capacities of the peripheral.
First bit will describe if the periph driver allows multiple
outstanding TRIMS to be active in a device.

Modify the I/O scheduler so that the nda driver can queue trims
for a while after the first one arrives. We'll queue until we see
a I/O scheduler tick, then we'll schedule as many TRIMs as allowed
by other factors (currently this is slocts in the NVMe controller).
This mariginally helps the read latency issues we see with reads,
but sets the stage for the nda driver to do TRIM collapsing like the
da and ada drivers do today.

Sponsored by: Netflix
2018-02-22 05:43:55 +00:00
Warner Losh
6ffe523f18 Minor formatting nits. 2018-02-21 23:49:35 +00:00
Warner Losh
9d602e4e3d Fix cut and pasted comments to reflect differences in code from the
original source.

Sponsored by: Netflix
2018-02-07 18:33:46 +00:00
Scott Long
99e7a4ad9e Return a C errno for cam_periph_acquire().
There's no compelling reason to return a cam_status type for this
function and doing so only creates confusion with normal C
coding practices. It's technically an API change, but the periph API
isn't widely used. No efffective change to operation.

Reviewed by:	imp, mav, ken
Sponsored by:	Netflix
Differential Revision:	D14063
2018-02-06 06:42:25 +00:00
Warner Losh
60b7691d56 Fix a sleepable malloc in ndastart. We shouldn't be sleeping
here. Return ENOMEM when we can't malloc a buffer for the DSM
TRIM. This should fix the WITNESS warnings similar to the following:

uma_zalloc_arg: zone "16" with the following non-sleepable locks held:
exclusive sleep mutex CAM device lock (CAM device lock) r = 0 (0xfffff800080c34d0) locked @ /usr/src/sys/cam/nvme/nvme_da.c:351

Reviewed by: scottl@
Sponsored by: Netflix
2018-01-26 23:14:46 +00:00
Pedro F. Giffuni
f24882eca5 SPDX: finish tagging sys/cam. 2018-01-16 23:19:57 +00:00
Scott Long
876f6a6af2 Release the held refcount on the probe periph when probing is
done, now that r327741 lets this happen.

Obtained from:	Netflix
2018-01-09 21:24:05 +00:00
Scott Long
bff0b56cdf Don't hold the periph locks during dump.
Obtained from:	Netflix
2018-01-09 00:17:15 +00:00
Alexander Motin
4d4709520a Reduce size of several on-stack string buffers.
Submitted by:	Dmitry Luhtionov <dmitryluhtionov@gmail.com>
MFC after:	2 weeks
2017-12-13 21:17:00 +00:00
Warner Losh
762a7f4f5f Define xpt_path_inq.
This provides a nice wrarpper around the XPT_PATH_INQ ccb creation and
calling.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D13387
2017-12-06 23:05:22 +00:00
Warner Losh
2b31251a64 Now that cam_periph_runccb() can be called from situations where the
kernel scheduler is stopped, replace the by hand calling of
xpt_polled_action() with it.

Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D13388
2017-12-06 23:05:15 +00:00
Warner Losh
553484ae07 Remove unused 4th argument to match the standard error routines.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D13386
2017-12-06 00:29:50 +00:00
Warner Losh
5e8a39f6ab Properly decode NVMe state of the drive and print out the information
in the attach to more closely match what SCSI and ATA attached
storage provides.

Sponsored by: Netflix
2017-11-14 05:05:26 +00:00
Warner Losh
4e3b274457 Provide link speed data in XPT_GET_TRAN_SETTINGS. Provide full version
information for that and XPT_PATH_INQ. Provide macros to encode/decode
major/minor versions.  Read the link speed and lane count to compute
the base_transfer_speed for XPT_PATH_INQ.

Sponsored by: Netflix
2017-11-14 05:05:16 +00:00
Warner Losh
712ad71996 nvd alias has caused some problems, revert it for the moment.
Sponsored by: Netflix
2017-10-27 14:57:38 +00:00
Warner Losh
fa271a5d09 Closer examination shows that nvme and CAM both normally zero-fill
allocations (for req and ccb, which ultimately contain the
nvme_cmd). As such, we can micro-optimize these routines. Add a
comment to this effect, and bzero the ccb used to make the requests
for the nda dump rotuine so it more closely matches a ccb allocated
with xpt_get_ccb().

Sponsored by: Netflix
2017-10-15 23:53:55 +00:00
Warner Losh
c4231018d0 Be nicer on the dump stack by allocating only a ccb_nvmeio rather than
a full ccb. This saves a few hundre bytes, which might be important
during a crash dump...

Sponsored by: Netflix
Suggested by: scottl@
2017-10-15 16:18:03 +00:00
Warner Losh
717bff5d85 Update comment to reflect actual default timeout.
Sponsored by: Netflix
2017-10-15 16:17:55 +00:00
Warner Losh
851063e16a Allow multiple TRIMs to be done for nda
Don't call cam_iosched_trim_done or cam_iosched_submit_trim for nda
since its hardware can handle almost an arbitrary number of TRIMs and
we don't have to be careful to only ever do one.

Sponsored by: Netflix
2017-09-15 20:16:06 +00:00
Warner Losh
9f8ed7e40b Fix NVMe's use of XPT_GDEV_TYPE
This patch changes the way XPT_GDEV_TYPE works for NVMe. The current
ccb_getdev structure includes pointers to the NVMe Identify Controller
and Namespace structures, but these are kernel virtual addresses which
are not accessible from user space.

As an alternative, the patch changes the pointers into padding in
ccb_getdev and adds two new types to ccb_dev_advinfo to retrieve the
Identify Controller (CDAI_TYPE_NVME_CNTRL) and Namespace
(CDAI_TYPE_NVME_NS) data structures.

Reviewed By: rpokala, imp
Differential Revision: https://reviews.freebsd.org/D10466
Submitted by: Chuck Tuffli
2017-08-29 17:03:30 +00:00
Warner Losh
519772814d Add CAM/NVMe support for CAM_DATA_SG
This adds support in pass(4) for data to be described with a
scatter-gather list (sglist) to augment the existing (single) virtual
address.

Differential Revision: https://reviews.freebsd.org/D11361
Submitted by: Chuck Tuffli
Reviewed by: imp@, scottl@, kenm@
2017-08-29 15:29:57 +00:00
Warner Losh
9754579b01 Add comment about where we need to place this routine, and why.
Sponsored by: Netflix
2017-08-28 19:27:33 +00:00
Warner Losh
d45e16744f Add nvd alias to nda ndoes.
All ndaX and ndaXpY nodes will appear as nvdX and nvdXpY as well
(through symlinks in devfs via the normal disk aliasing mechanism in
GEOM).

Differential Revision: https://reviews.freebsd.org/D11873
2017-08-07 21:12:43 +00:00
Warner Losh
1d6e811063 Namespace is 32-bits, don't cast it to 16 here 2017-06-27 16:48:05 +00:00
Alexander Motin
17160457b4 Report random flash storage as non-rotating to GEOM_DISK.
While doing it, introduce respective constants in geom_disk.h.

MFC after:	1 week
2017-01-12 08:53:10 +00:00
Ed Schouten
4c484fd216 Add label annotations to CAM sysctls.
Under kern.cam we have certain sysctls that are per-device, such as the
ones under kern.cam.ada.[0-9]+.*. Add a "device_index" label annotation
to such sysctls, so that the Prometheus metrics exporter will give all
of those metrics the same name. The device number will be added to the
metric name as the "device_index" label.

Reviewed by:	cem
Differential Revision:	https://reviews.freebsd.org/D8775
2016-12-14 12:53:33 +00:00