A Write Zeroes commands should not be counted in either the 'Data Units
Written' or in 'Host Write Commands' SMART/Health Information Log page.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
The 'len' member of the nvme_compare_ctx struct is redundant since the
same information is available in the 'iov' member.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Dataset Management is not subject to MDTS, but exceeded a certain size
per range causes internal looping. Report this limit (DMRSL) in the NVM
command set specific identify controller data structure.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Add a trace event for the offline zone condition when checking zone
read.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
[k.jensen: split commit]
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
assert may be compiled to a noop and we could end up returning an
uninitialized status.
Fix this by always returning Internal Device Error as a fallback.
Note that, as pointed out by Philippe, per commit 262a69f428 ("osdep.h:
Prohibit disabling assert() in supported builds") this shouldn't be
possible. But clean it up so we don't worry about it again.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
[k.jensen: split commit]
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Add a trace event for the Identify command.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Remove an unnecessary le_to_cpu conversion in Identify.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
ZASL (Zone Append Size Limit) is defined exactly like MDTS (Maximum Data
Transfer Size), that is, it is a value in units of the minimum memory
page size (CAP.MPSMIN) and is reported as a power of two.
The 'mdts' nvme device parameter is specified as in the spec, but the
'zoned.append_size_limit' parameter is specified in bytes. This is
suboptimal for a number of reasons:
1. It is just plain confusing wrt. the definition of mdts.
2. There is a lot of complexity involved in validating the value; it
must be a power of two, it should be larger than 4k, if it is zero
we set it internally to mdts, but still report it as zero.
3. While "hw/block/nvme: improve invalid zasl value reporting"
slightly improved the handling of the parameter, the validation is
still wrong; it does not depend on CC.MPS, it depends on
CAP.MPSMIN. And we are not even checking that it is actually less
than or equal to MDTS, which is kinda the *one* condition it must
satisfy.
Fix this by defining zasl exactly like mdts and checking the one thing
that it must satisfy (that it is less than or equal to mdts). Also,
change the default value from 128KiB to 0 (aka, whatever mdts is).
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Add support for using the broadcast nsid to issue a flush on all
namespaces through a single command.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Commit 6eb7a07129 ("hw/block/nvme: change controller pci id") changed
the controller to use a Red Hat assigned PCI Device and Vendor ID, but
did not change the IEEE OUI away from the Intel IEEE OUI.
Fix that and use the locally assigned QEMU IEEE OUI instead if the
`use-intel-id` parameter is not explicitly set. Also reverse the Intel
IEEE OUI bytes.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
The Zone Append Size Limit (ZASL) must be at least 4096 bytes, so
improve the user experience by adding an early parameter check in
nvme_check_constraints.
When ZASL is still too small due to the host configuring the device for
an even larger page size, convert the trace point in nvme_start_ctrl to
an NVME_GUEST_ERR such that this is logged by QEMU instead of only
traced.
Reported-by: Corne <info@dantalion.nl>
Cc: Dmitry Fomichev <Dmitry.Fomichev@wdc.com>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Implicitly and Explicitly Open zones can be closed by Close Zone
management function. This got broken by a recent commit ("hw/block/nvme:
refactor zone resource management") and now such commands fail with
Invalid Zone State Transition status.
Modify nvm_zrm_close() function to make Close Zone work correctly.
Signed-off-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Add support for TP 4065a ("Simple Copy Command"), v2020.05.04
("Ratified").
The implementation uses a bounce buffer to first read in the source
logical blocks, then issue a write of that bounce buffer. The default
maximum number of source logical blocks is 128, translating to 512 KiB
for 4k logical blocks which aligns with the default value of MDTS.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
In preparation for Simple Copy, pull write pointer advancement into a
separate function that is independent off an NvmeRequest.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Zone transition handling and resource management is open coded (and
semi-duplicated in the case of open, close and finish).
In preparation for Simple Copy command support (which also needs to open
zones for writing), consolidate into a set of 'nvme_zrm' functions and
in the process fix a bug with the controller not closing an open zone to
allow another zone to be explicitly opened.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Remove the unused NvmeCtrl parameter in nvme_check_zone_write.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
nvme-ns device is registered to a nvme controller device during the
initialization in nvme_register_namespace() in case that 'bus' property
is given which means it's mapped to a single controller.
This patch introduced a new property 'subsys' just like the controller
device instance did to map a namespace to a NVMe subsystem.
If 'subsys' property is given to the nvme-ns device, it will belong to
the specified subsystem and will be attached to all controllers in that
subsystem by enabling shared namespace capability in NMIC(Namespace
Multi-path I/O and Namespace Capabilities) in Identify Namespace.
Usage:
-device nvme-subsys,id=subsys0
-device nvme,serial=foo,id=nvme0,subsys=subsys0
-device nvme,serial=bar,id=nvme1,subsys=subsys0
-device nvme,serial=baz,id=nvme2,subsys=subsys0
-device nvme-ns,id=ns1,drive=<drv>,nsid=1,subsys=subsys0 # Shared
-device nvme-ns,id=ns2,drive=<drv>,nsid=2,bus=nvme2 # Non-shared
In the above example, 'ns1' will be shared to 'nvme0' and 'nvme1' in
the same subsystem. On the other hand, 'ns2' will be attached to the
'nvme2' only as a private namespace in that subsystem.
All the namespace with 'subsys' parameter will attach all controllers in
the subsystem to the namespace by default.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
We have nvme-subsys and nvme devices mapped together. To support
multi-controller scheme to this setup, controller identifier(id) has to
be managed. Earlier, cntlid(controller id) used to be always 0 because
we didn't have any subsystem scheme that controller id matters.
This patch introduced 'cntlid' attribute to the nvme controller
instance(NvmeCtrl) and make it allocated by the nvme-subsys device
mapped to the controller. If nvme-subsys is not given to the
controller, then it will always be 0 as it was.
Added 'ctrls' array in the nvme-subsys instance to manage attached
controllers to the subsystem with a limit(32). This patch didn't take
list for the controllers to make it seamless with nvme-ns device.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
nvme controller(nvme) can be mapped to a NVMe subsystem(nvme-subsys).
This patch maps a controller to a subsystem by adding a parameter
'subsys' to the nvme device.
To map a controller to a subsystem, we need to put nvme-subsys first and
then maps the subsystem to the controller:
-device nvme-subsys,id=subsys0
-device nvme,serial=foo,id=nvme0,subsys=subsys0
If 'subsys' property is not given to the nvme controller, then subsystem
NQN will be created with serial (e.g., 'foo' in above example),
Otherwise, it will be based on subsys id (e.g., 'subsys0' in above
example).
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
To support multi-path in QEMU NVMe device model, We need to have NVMe
subsystem hierarchy to map controllers and namespaces to a NVMe
subsystem.
This patch introduced a simple nvme-subsys device model. The subsystem
will be prepared with subsystem NQN with <subsys_id> provided in
nvme-subsys device:
ex) -device nvme-subsys,id=subsys0: nqn.2019-08.org.qemu:subsys0
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
[k.jensen: added 'nqn' device parameter per request]
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Current QEMU HEAD nvme.c does not compile with the default GCC 5.4
on a Ubuntu 16.04 host:
hw/block/nvme.c:3242:9: error: ‘result’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
trace_pci_nvme_getfeat_vwcache(result ? "enabled" : "disabled");
^
hw/block/nvme.c:3150:14: note: ‘result’ was declared here
uint32_t result;
^
Explicitly initialize the result to fix it.
Fixes: aa5e55e3b0 ("hw/block/nvme: open code for volatile write cache")
Fixes: Coverity CID 1446371
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Moving namespace registration to the nvme-ns realization function had
the unintended side-effect of breaking legacy namespace registration.
Fix this.
Fixes: 15d024d4aa ("hw/block/nvme: split setup and register for namespace")
Reported-by: Alexander Graf <agraf@csgraf.de>
Cc: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Alexander Graf <agraf@csgraf.de>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Refactor the zone write check logic such that the most "meaningful"
error is returned first. That is, first, if the zone is not writable,
return an appropriate status code for that. Then, make sure we are
actually writing at the write pointer and finally check that we do not
cross the zone write boundary. This aligns with the "priority" of status
codes for zone read checks.
Also add a couple of additional descriptive trace events and remove an
always true assert.
Cc: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Tested-by: Niklas Cassel <niklas.cassel@wdc.com>
Tested-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
When a zone append is processed the controller checks that validity of
the write before assigning the LBA to the append command. This causes
the boundary check to be wrong.
Fix this by checking the write *after* assigning the LBA. Remove the
append special case from the nvme_check_zone_write and open code it in
nvme_do_write, assigning the slba when basic sanity checks have been
performed. Then check the validity of the resulting write like any other
write command.
In the process, also fix a missing endianness conversion for the zone
append ALBA.
Reported-by: Niklas Cassel <Niklas.Cassel@wdc.com>
Cc: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Tested-by: Niklas Cassel <niklas.cassel@wdc.com>
Tested-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
The actual parameter name is 'cross_read' rather than 'cross_zone_read'.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Change status checks to align with the existing style and remove the
explicit check against NVME_SUCCESS.
Cc: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Currently, no features are saveable, so the current check is not wrong,
but add a check against the feature capabilities to make sure this will
not regress if saveable features are added later.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Only enable DULBE if the namespace supports it.
Signed-off-by: Gollu Appalanaidu <anaidu.gollu@samsung.com>
Reviewed-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
The controller now implements v1.4 and we can lift the restrictions on
CMB Data Pointer and Command Independent Locations Support (CDPCILS) and
CMB Data Pointer Mixed Locations Support (CDPMLS) since the device
really does not care about mixed host/cmb pointers in those cases.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
With the new CMB logic in place, bump the implemented specification
version to v1.4 by default.
This requires adding the setting the CNTRLTYPE field and modifying the
VWC field since 0x00 is no longer a valid value for bits 2:1.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Implement v1.4 logic for configuring the Controller Memory Buffer. By
default, the v1.4 scheme will be used (CMB must be explicitly enabled by
the host), so drivers that only support v1.3 will not be able to use the
CMB anymore.
To retain the v1.3 behavior, set the boolean 'legacy-cmb' nvme device
parameter.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Padmakar Kalghatgi <p.kalghatgi@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Add support for the PMRMSCL and PMRMSCU MMIO registers. This allows
adding RDS/WDS support for PMR as well.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Naveen Nagar <naveen.n1@samsung.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
The PMR should not be enabled at boot up. Disable the PMR MemoryRegion
initially and implement MMIO for PMRCTL, allowing the host to enable the
PMR explicitly.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
The controller registers are initially zero. Remove the redundant
zeroing.
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Use the correct field names.
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
With BAR 4 now free to use, allow PMR and CMB to be enabled
simultaneously.
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
In the interest of supporting both CMB and PMR to be enabled on the same
device, move the MSI-X table and pending bit array out of BAR 4 and into
BAR 0.
This is a simplified version of the patch contributed by Andrzej
Jakowski (see [1]). Leaving the CMB at offset 0 removes the need for
changes to CMB address mapping code.
[1]: https://lore.kernel.org/qemu-devel/20200729220107.37758-3-andrzej.jakowski@linux.intel.com/
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Tested-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
This patch sets CMBS bit in controller capabilities register when user
configures NVMe driver with CMB support, so capabilites are correctly
reported to guest OS.
Signed-off-by: Andrzej Jakowski <andrzej.jakowski@linux.intel.com>
Reviewed-by: Maxim Levitsky <mlevitsky@gmail.com>
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
64 bit registers like ASQ and ACQ should be writable by both a hi/lo 32
bit write combination as well as a plain 64 bit write. The spec does not
define ordering on the hi/lo split, but the code currently assumes that
the low order bits are written first. Additionally, the code does not
consider that another address might already have been written into the
register, causing the OR'ing to result in a bad address.
Fix this by explicitly overwriting only the low or high order bits for
32 bit writes.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Add the size of the mmio read/write to the trace event.
Reviewed-by: Minwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
During smart critical warning injection by setting property from QMP
command, also try to trigger asynchronous event.
Suggested by Keith, if a event has already been raised, there is no
need to enqueue the duplicate event any more.
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
[k.jensen: fix typo in commit message]
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
There is a very low probability that hitting physical NVMe disk
hardware critical warning case, it's hard to write & test a monitor
agent service.
For debugging purposes, add a new 'smart_critical_warning' property
to emulate this situation.
The orignal version of this change is implemented by adding a fixed
property which could be initialized by QEMU command line. Suggested
by Philippe & Klaus, rework like current version.
Test with this patch:
1, change smart_critical_warning property for a running VM:
#virsh qemu-monitor-command nvme-upstream '{ "execute": "qom-set",
"arguments": { "path": "/machine/peripheral-anon/device[0]",
"property": "smart_critical_warning", "value":16 } }'
2, run smartctl in guest
#smartctl -H -l error /dev/nvme0n1
=== START OF SMART DATA SECTION ===
SMART overall-health self-assessment test result: FAILED!
- volatile memory backup device has failed
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Signed-off-by: zhenwei pi <pizhenwei@bytedance.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
The zone write pointer is unconditionally advanced, even for write
faults. Make sure that the zone is always transitioned to Full if the
write pointer reaches zone capacity.
Cc: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
nvme_ns_setup() finally does not have nothing to do with NvmeCtrl
instance.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Volatile Write Cache(VWC) feature is set in nvme_ns_setup() in the
initial time. This feature is related to block device backed, but this
feature is controlled in controller level via Set/Get Features command.
This patch removed dependency between nvme and nvme-ns to manage the VWC
flag value. Also, it open coded the Get Features for VWC to check all
namespaces attached to the controller, and if false detected, return
directly false.
Signed-off-by: Minwoo Im <minwoo.im.dev@gmail.com>
[k.jensen: report write cache preset if present on ANY namespace]
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
TP 4053 says (in section 2.3.1.1) -
... if a Zone Append command specifies a ZSLBA that is not the lowest
logical block address in that zone, then the controller shall abort
that command with a status code of Invalid Field In Command.
In the code, Zone Invalid Write is returned instead, fix this.
Signed-off-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
nvme_io_cmd already checks if the namespace supports the Zone Append
command, so the removed check is dead code.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Tested-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
The zoned command set specification states that "All logical blocks in a
zone *shall* be marked as deallocated when [the zone is reset]". Since
the device guarantees 0x00 to be read from deallocated blocks we have to
issue a pwrite_zeroes since we cannot be sure that a discard will do
anything. But typically, this will be achieved with an efficient
unmap/discard operation.
Signed-off-by: Klaus Jensen <k.jensen@samsung.com>
Tested-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
Reviewed-by: Dmitry Fomichev <dmitry.fomichev@wdc.com>
