qThis change was made after feedback from upstream, aiming to align with
the style guide for consistent log formatting. No functional changes
were made to the driver, only the formatting of the log messages.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49799
(cherry picked from commit 4494ea5406f79a6cb2d3631a723eb286ca96a9b9)
The driver version has been updated to 8.14.0.2.0 to reflect the
latest release. This is a version-only update with no functional
code changes.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49753
(cherry picked from commit f4d51d3e1a90dabbed26aacf1b58e20e23a19342)
All driver source files have been updated to reflect the year 2025.
This change is cosmetic and does not affect functionality.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49752
(cherry picked from commit 69be9a247b845b35146fed4575dacb565c9c18ba)
The driver now processes LOG Data events in the interrupt top
half instead of the bottom half. This allows LOG events to be
handled immediately upon receipt, reducing latency and enabling
early notification to the Library or applications.
This change ensures LOG event handling occurs closer to the
firmware's generation point, improving alignment with event-driven
diagnostic and monitoring mechanisms.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49751
(cherry picked from commit df8a74fca036a0ba8026f3ae3da830dffb922faa)
This patch ensures the driver sets the Diag Save bit for all
diagnostic fault conditions before issuing a Diagnostic Fault Reset.
The firmware now incorporates logic to manage snapdump saving
and requires the driver to always set the Diag Save bit to enable
this feature. This change supports improved diagnostic data
collection and fault handling.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49750
(cherry picked from commit b854677d970a11e87929fa95d0e4067f911faa62)
The driver previously blocked I/Os only for OS-initiated task
management commands. This patch extends the behavior to also
block I/Os during application-initiated task management
operations (excluding Task Abort).
Before submitting such commands to the firmware, I/O
submissions are paused for the respective device. Once the
command completes, I/O operations are resumed.
This ensures safe and consistent task management handling.
[[ Note: Warner landed this with the pending suggestion
since this change is good enough for 14.3, but chs' suggestion
for better atomics needs to be implemented soon ]]
Discussed with: imp, chs
Differential Revision: https://reviews.freebsd.org/D49749
(cherry picked from commit 49445b54638f7b2d45942edd61c70fe0baa28f37)
The driver now retrieves the Time Stamp value from Driver Page 1
during load and after controller reset. If the value is valid, it
is used to enable periodic host timestamp synchronization.
This adds a tunable NVData parameter to control the behavior of
host time sync, enhancing flexibility and platform-specific control.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49748
(cherry picked from commit e80317babdb9044962dcbf16c69633579cd90b00)
The driver now checks for insufficient power faults during the load
phase and immediately fails initialization instead of retrying.
Additionally, if an insufficient power fault is detected by the watchdog
after the controller is up, the controller is marked as unrecoverable
instead of triggering a reset.
This improves fault handling and avoids unnecessary recovery attempts
in low-power conditions.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49747
(cherry picked from commit 116c8b18a2b5278df0d1982a683193c3dba6f30c)
This patch improves the retry logic during the IOC enable process.
If a controller faults or if the reset history bit is detected during
the ready status check, the driver will retry initialization up to three
times or until 510 seconds have passed.
A soft reset will also be issued if the controller is detected while
waiting for the ready status.
This enhances reliability during controller initialization.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49746
(cherry picked from commit e977ed50bb2b9fce5e285756673286c5fcfb1f31)
This patch updates the driver to set the OSExposure field in DriverCapability
during IOCInit to MPI3_IOCINIT_DRIVERCAP_OSEXPOSURE_NO_SPECIAL(0x1), aligning
with the latest MPI specification version 36.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49745
(cherry picked from commit d14c4ea0caebb9d0ddb613e348cb8c1a19e8d8fc)
Updated the MPI header files to version 36 to align with the latest
MPI specification. This includes updated structures, field definitions,
and constants required for compatibility with updated firmware.
Reviewed by: ssaxena, imp
Differential Revision: https://reviews.freebsd.org/D49743
(cherry picked from commit 92f340d137ba5d6db7610ba1dae35842e2c9c8ea)
The max I/O size that an mpi3mr HBA supports is reported in the IOCFacts
structure (with 0 representing the legacy max I/O size of 1 MB).
By default, set the max I/O size of devices attached to mpi3mr controllers
to the smaller of the HBA's max I/O size and the kernel's maxphys.
Allow this default to be overriden by a global tunable
"hw.mpi3mr.max_sgl_entries" or by a per-controller tunable
"dev.mpi3mr.N.max_sgl_entries".
Sponsored by: Netflix
Reviewed by: imp
Differential Revision: https://reviews.freebsd.org/D49090
(cherry picked from commit f2605f67a13eb1f1427ee58a097791ca25359a5e)
Firmware advertises the transfer lenght for writesame commands to driver during init.
So for any writesame IOs with ndob and unmap bit set and transfer lengh is greater
than the max write same length specified by the firmware, then direct those commands
to firmware instead of hardware otherwise hardware will break.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44452
(cherry picked from commit 3f3a15543a6721100dda0e4219eb48ecbe35731a)
The driver is modified to add FreeBSD OS type in the upper nibble of the
fault/reset reason code for appropriate qualification of the reason code.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44427
(cherry picked from commit 3012fa8faef262c615672fc90319c777b4e3dffc)
Accept any fault as successful for diagnostic fault reset, not just the 0xF000
code. print fault information and return.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44426
(cherry picked from commit 4034d7061a112b78d60cdb581c2d71f7cfa9f74e)
Instead of updating the ConsumerIndex of the Admin and Operational ReplyQueues
after processing all replies in the queue, it will now be periodically updated
after processing every 100 replies.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44425
(cherry picked from commit eb7a4b35de89a0e22f1fb373df33d79135b68dd8)
Post controller reset, If any device removal events arrive, and if
there are any outstanding IOs then the driver will unnecessarily wait
in the loop for 30 seconds before removing the device from the OS.
reset target outstanding IO counter and controller outstanding IO counter
and remove the redundant wait loop.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44424
(cherry picked from commit 571f1d0638fa6b3a08f84118a33e53f941f0820c)
An outstanding IO counter per target check has been added before deleting
the target from the OS which will poll the reply queue if there are any
outstanding IOs are found.
A new flag, named "MPI3MR_DEV_REMOVE_HS_COMPLETED," is added. If a remove event
for a target occurs and before the deletion of the target resource if the add event
for another target arrives reusing the same target ID then this flag will prevent
the removal of the target reference. This flag ensures synchronization between the interrupt
top and bottom half during target removal and addition events.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44423
(cherry picked from commit 701d776c98e015160c2d0e3f4fdaf980d1d5dc83)
The driver removes the drive from the OS if firmware sends 'device added'
event with hidden bit or inaccessible status.
Reviewed by: imp
Approved by: imp
Differential revision: https://reviews.freebsd.org/D44421
(cherry picked from commit 0d7056458db5b5dd7fdc5ccd8abab73e3ee76a20)
Use sc->mpi3mr_debug & MPI3MR_IOT over the | version to test if a bit is
set.
CID: 1529718
Sponsored by: Netflix
(cherry picked from commit 34f0a01b6bd4ff0e61a7066729515148400e0982)
Use a while loop with cancel / drain to make sure that all tasks have
completed before proceeding to reset.
Suggested by: jhb
Sponsored by: Netflix
(cherry picked from commit 272a406042608da9bc3e67e41c6b7fc31d4166b8)
No sense having a variable for this. So use BUS_SPACE_MAXADDR and remove
dma_hiaddr from softc.
Suggested by: jhb
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D42808
(cherry picked from commit 1ec7c672bc2854a1efb1d50f189d4881163c16c6)
Replace the test for DataLength == 0 with an assert. It can't happen,
but an assert doesn't hurt. Emacs removed some trailing white space too.
Sponsored by: Netflix
Differential Revision: https://reviews.freebsd.org/D42807
(cherry picked from commit 2361a0056fc16b5e11df2aa948aaa633ac01d685)
Use the simpler template code for the parent busdma tag for all I/O to
this card.
Reviewed by: mav, jhb, imp
Differential Revision: https://reviews.freebsd.org/D42607
(cherry picked from commit 489eee0d41dce317678adb0dae8d509a5f1d6d93)
These calls "should" all be synchrounous. There's no bouncing that's
needed for them (at least in the typical case that we have a sane card
that has more bits of dma addresses decoded than we have memory), so
there's no errors possible. Ensure these calls are really synchronous
with BUS_DMA_NOWAIT flags (which should never fail now that the
bus_dmamem_alloc() has succeeded).
Reviewed by: mav, jhb, imp
Differential Revision: https://reviews.freebsd.org/D42606
(cherry picked from commit 39a3e6a812ad9c089bd2c4935193f1b3c4c5c35a)
This usage is obsolete. Replace with maximum bus space size. maxphys
will sort itself out at higher levels.
Reviewed by: mav, jhb, imp
Differential Revision: https://reviews.freebsd.org/D42605
(cherry picked from commit 4e6d128bd823e5b5d0b058f918c7036978a0e8bf)
Publish the firmware version on the card like we do for mps/mpr.
Sponsored by: Netflix
Reviewed by: mav
Differential Revision: https://reviews.freebsd.org/D42588
(cherry picked from commit 28a274342ea0b0666b56704477d2d1c17564942e)
The number of signficant bits that are decoded are returned in the flags
field of the IOCFacts structure from the device. Rather than assume the
worst with a pessimal 32-bit maximum, look at this value and pass it
along to all the dma map creation requests.
A lof of those creations are repetitive and could just inherit from the
base tag if we moved to the templated interface. This is called out as
desireable future work not done at this time.
In addition, due to a chicken and an egg problem, we have to allocate
some of the maps with a 32-bit loaddr. These are the ones we need to
read iocfacts. And they are fine to be so restricted: they are little
used after startup, and when they are used, bouncing is fine.
Sponsored by: Netflix
Reviewed by: mav
Differential Revision: https://reviews.freebsd.org/D42559
(cherry picked from commit 91d961356d03465635c4784fab48acdd1304e1e0)
Move enqueueing of commands to bus_dmamap_load_ccb callback
Fix fundamental difference between FreeBSD and Linux. On Linux, your dma
load callback always happends before it returns, so drivers are written
to load the map, then submit to hardware. On FreeBSD, the callback may
be deferred and return EINPROGRESS. This means the callback is
responsible for queueing the request to the hardware is done after the
SGL list is created. Make a number of interrelated cahnages:
At the end of mpi3mr_prepare_sgls, add a call to mpi3mr_enqueue_request.
Split the hardware submission out from the end of mpi3mr_action_scsiio
and move it into a new routine mpi3mr_enqueue_request.
Move all error completion from the end of mpi3mr_action_scsiio to where
the error is detected. We cannot pass errors back from the
mpi3mr_enqueue_request to do this on a 'failed' mpi3mr in a centralized
place (since it has to be fire and forget).
Add comments about zero length SGLs never making it into
mpi3mr_prepare_sgls. Keep the code there for the moment, but we only set
cm->data to non-NULL when scsiio_req->DataLength is not zero. So the
datalength can't be zero and we can't send the zero SGLs.
Add commentts about other "impossible" tests in mpi3mr_prepare_sgls that
really should be simple asserts of some flavor.
Eliminate cm->error_code, since we can't pass data back from the
mpi3mr_prepare_sgl callback anymore.
In mpi3mr_map_request, call mpi3mr_enqueue_request for the no data case.
This seems to work even though we've not done the special zero length
handling that was in mpi3mr_prepare_sgls, giving further evidence to it
not actually being needed. This is needed for SCSI CDBs that have no
data to pass to the drive like TEST UNIT READY.
With this change, and the prior ones, we're now able to run with mpi3mr
on 128GB systems and very heavy disk load (so many buffers land > 4GB:
the driver instructs busdma to never use memory abouve 4GB, which may be
too conservative, but an issue for another time).
Sponsored by: Netflix
Reviewed by: sumit.saxena_broadcom.com, mav, jhb
Differential Revision: https://reviews.freebsd.org/D42543
(cherry picked from commit 3208a189c1e2c4ef35daa432fe45629a043d7047)
More uniformly use mpi3mr_set_ccbstatus in mpi3mr_action_scsiio. The
routine mostly used it, but also has setting of status by hand. In those
cases where we want to error out the request, use this routine.
As part of this, move setting CAM_SIM_QUEUED later in the function to
when we're sure it's been queued. Remove the places we clear it before
this.
Sponsored by: Netflix
Reviewed by: mav, jhb
Differential Revision: https://reviews.freebsd.org/D42542
(cherry picked from commit cf8c23230aabd30aa9251975dbe705da559a2d02)
Since we assume there's a timeout to cancel when this is true, only set
it true when we set the timeout. Otherwise we may try to cancel a timeout
when there's been an error in submission.
Sponsored by: Netflix
Reviewed by: mav
Differential Revision: https://reviews.freebsd.org/D42541
(cherry picked from commit 1cfd01111eb6a28ca5043e928aa8e9099064177a)
Fold two lines to make this more readable.
Sponsored by: Netflix
Reviewed by: mav, jhb
Differential Revision: https://reviews.freebsd.org/D42540
(cherry picked from commit e2b27df9eb324fa6e72e29d3185dcd6b390efa9d)
Reduce the scope of reset_mutext to protect the msleep in the watch dog
thread as well as the MPI3MR_FLAGS_SHUTDOWN bit. Use it to protect the
wakeup in mpi3mr_detach so this thread can exit sooner when we're trying
to do an orderly shutdown. Optimize the flow to check the sleep and
other conditions before going to sleep.
It's an open question if this should protect sc->unrecoverable, and if
we should wakeup the watchdog thread when we set it. We might also want
to move too booleans for the three flags that we have now in
mpi3mr_flags. There are a number of U8s that should really be bools and
we might want to also group them together to pack softc better.
Sponsored by: Netflix
Reviewed by: mav
Differential Revision: https://reviews.freebsd.org/D42539
(cherry picked from commit 7c4913093a759adf2e4c7d65535aee04aadee4df)
All of these fields are either unused, or just initialized. Remove
them. This saves about 1MB of memory for the cards that I have which can
do 8k transactions at once.
Sponsored by: Netflix
Reviewed by: mav, jhb
Differential Revision: https://reviews.freebsd.org/D42538
(cherry picked from commit a2b046620c54db977196128b3c53da2704b9fd20)
Holding fwevt_lock when we call taskqueue_drain can lead to deadlock
because it's draining a queue needs fwevt_lock to do work, so that other
thread will try to take out the lock and block, making the thread never
finish and taskqueue_drain never complete. There's a witness
warning/error for this which was exposed when the lock was converted to
a MTX_DEF lock from a MTX_SPIN prior to committing to the FreeBSD tree.
The lock appears to be to protect against additional items being added
to the event list while we're doing a reset. Since the taskqueue is
blocked, items can get added to the list, but won't be processed during
the reset, but there is still a (likely small) race between the
taskqueue_drain and the taskqueue_block calls where an interrupt could
fire on another CPU, resulting in a task being enqueued and started
before the block can take effect. The only way to fix that race is to
turn off interrupt processing during a reset. So we replace a deadlock
with a smaller race.
Sponsored by: Netflix
Reviewed by: sumit.saxena_broadcom.com, mav, jhb
Differential Revision: https://reviews.freebsd.org/D42537
(cherry picked from commit b411372b7d17ae7e5d6c944732d41b979bde2ac4)
A failed copyin will cause the driver to use the contents of
uninitialized buffers instead, which is unlikely to be the behaviour
that we want. Check for errors.
This is in preparation for annotating copyin() and related functions
with __result_use_check.
Reviewed by: imp
MFC after: 1 week
Differential Revision: https://reviews.freebsd.org/D43098
(cherry picked from commit 6bfb7306ef92aaccffae42ed00ce6d7201b76966)
Move creation of watchdog process from just before we configure the
interrupt config hook to into the config hook itself. This prevents it
from racing the config intr hook and doing an extra reset of the
card. This extra reset is usually harmless, but sometimes it can prevent
discovery of devices if done at just the wrong time. This can lead to no
disks being registered in a box full of disks, for example. Starting it
later eliminates this race, making discovery reliable.
Reviewed by: imp
(cherry picked from commit 7e02c7074c4c6df77b860e0dbcd032a2ea04b98b)
Previously every file that included mpi3mr_app.h but did not use
mpi3mr_mgmt_info reported error: 'mpi3mr_mgmt_info' defined but not
used.
Fixes: 2d1d418e1e ("mpi3mr: 3rd Generation Tri-Mode NVMe/SAS/SATA...")
Reported by: amd64-gcc12 Cirrus-CI job
Sponsored by: The FreeBSD Foundation
This is Broadcom's mpi3mr driver for FreeBSD version 8.6.0.2.0.
The mpi3mr driver supports Broadcom SAS4116-based cards in the 9600
series: 9670W-16i, 9670-24i, 9660-16i, 9620-16i, 9600-24i, 9600-16i,
9600W-16e, 9600-16e, 9600-8i8e.
Initially only available as a module and on amd64/arm64, since that's
how it has been tested to date. Future commits will add it to the kernel
build and may expand the architectures it is supported on.
Co-authored-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Feedback-by: ken (prior versions)
Reviewed-by: imp
RelNotes: yes
Differential-Revision: https://reviews.freebsd.org/D36771
Differential-Revision: https://reviews.freebsd.org/D36772
