A few open-coded sequences equivalent to SPDK_CONTAINEROF() were
scattered around; replace them with the macro from spdk/util.h.
Change-Id: I95c6e6838902f411420573399ced7c58c2e4ef84
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/418126
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Add a check to prevent spdk_mem_map_set_translation() or
spdk_mem_map_clear_translation() calls that start within the valid
address range but specify a size that would access parts of the mem map
outside of the valid region.
spdk_mem_map_translate() is safe without any extra checks since it only
accesses the first entry regardless of size, and the MASK_256TB check
catches out-of-range accesses to that entry.
Change-Id: Ie1437e57b5158363bb98a6b42a26fb41a089bbad
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/418106
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
The arrays for both the map_256tb and map_1gb structures were twice
as large as necessary; fix the sizes and add unit tests for the boundary
conditions to verify that the fix works.
Change-Id: I66bce463f234f54e69cf2a697db9f806d398ca1e
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/418105
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
It's not required. We only use it in virtio-pci.
Change-Id: I61e95d680d00fa3d56ebccbc9a372db7e1db296d
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417002
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Previously when IO was sent to bdev, even without callback yet,
status for task was set to SPDK_SCSI_STATUS_GOOD.
This patch changes it so that it is only set when callback actually
comesback from bdev via spdk_bdev_io_get_scsi_status().
Change-Id: I4a36ec345608257123e1036d31540f5f1090a23b
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-on: https://review.gerrithub.io/417708
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
There is not any plan to support bidirectional command. To understand implementation
of residual count a little easier, remove residual count variables about unsupported
bidirectional command.
Change-Id: I1154e2e2708cc47c5bd4630ce9a91980f433fef8
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/417216
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This bug was caused by the following two changes:
iscsi: Remove duplication of the variable for write completion to bdev
https://review.gerrithub.io/#/c/393582/
iscsi: restore data_transferred accumulation for read
https://review.gerrithub.io/#/c/393713/
For write, bytes_completed is always equal to data_transferred. However,
for read, bytes_completed is based on expected data transfer length and
data_transferred is actual read size. Hence bytes_completed cannot be
used to calculate residual counts.
One of the reason why this bug cannot be found was that there was not any
test case when task->scsi.data_transferred is 0 and task->scsi.length is
not 0. Hence UT code is also added.
Same bug will occur for read sense data. Hence UT code for read sense data
is added in the next patch.
Change-Id: Ib1a283b769e5af0c2d05acb69f90948c5d658087
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/417960
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
During spdk_bdev_init, examine_config is called.
This call can claim bdev synchronously, based on
configuration. On spdk_bdev_start if none module
claimed bdev, examine_disk is called and can
perform I/O before claiming bdev.
Signed-off-by: Piotr Pelplinski <piotr.pelplinski@intel.com>
Change-Id: I1448dd368cf3a24a5daccab387d7af7c3d231127
Reviewed-on: https://review.gerrithub.io/413913
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Maciej Szwed <maciej.szwed@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Right now both virtio-pci and virtio-user allocate
the queue via spdk_dma_malloc, but that's about
to change soon.
Change-Id: I3acdad45cd9a0639f9070bc448fdf8f9d2c706c0
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417000
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This was omitted in the original implementation and the
`del_queue` callback wasn't called anywhere. For virtio-user
we open some eventfds on queue creation and until now they
were never being closed.
Change-Id: Iee1ced1e17a59d5cb13449538c115678a1c1a328
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416999
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The spdk_dma_zmalloc guarantee about physical memory contiguity
is about to be removed soon. A single tracker is page size
aligned and is exactly one page big, so it is physically
contiguous, but we can't assume an array of those is physically
contiguous as well.
Change-Id: I3aa4d14dd677601c30aa2d8f15197886d6c46e58
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416840
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
There was a plan a long time ago to try to output all
unit test results in JSON format for post-processing.
It is used by only one unit test file currently, but the
results aren't used and there are no plans to use it in
the future or extend it to other unit tests.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I72c409090bbd7b8fa7ec307bbc97f97ccf2e397c
Reviewed-on: https://review.gerrithub.io/418112
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Ia413a913857dad00ae091b8cea02a647b8996a5c
Reviewed-on: https://review.gerrithub.io/418108
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
There's no reason to abort IO that have been queued
due to QoS limits, when QoS is switched from enabled
to disabled. Submit them to the bdev instead.
Fixes issue #357.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: If5eafc53418ac686120e1d6a1da884b42cef845e
Reviewed-on: https://review.gerrithub.io/418128
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Seth Howell <seth.howell5141@gmail.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The device might not have this message implemented.
Sending it could result in a connection being terminated.
Change-Id: I53c08f1108ebc7de630569f3983c317cc6510fa4
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417636
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
For vhost-user it's a protocol feature that can simply
be not supported. The subsequent patch introduces an extra
check that may cause config read/write to fail.
Change-Id: I5b0e11845fb6021472c608477f1797dada8ab961
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417458
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
It was introduced a while after the original GET/SET_CONFIG
implementation in QEMU, but within the same QEMU release (2.12?).
It is required by the vhost-user spec. Rte_vhost doesn't check
it, so everything worked so far, but other implementations
might (and should) reject our GET_CONFIG requests right now.
As a part of this feature, we should also check the same
flag before sending GET/SET_CONFIG messages to respect those
devices that really don't implement F_CONFIG. This is done
in a separate patch.
Change-Id: Ib7e9b11a0074f4aee70609af0cad2ef59a8bf427
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417459
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Required by the vhost-user spec.
Change-Id: Id7143a0f6cc34463ad5f22d8db96ac5c51e04081
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417457
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
rte_vhost doesn't respect those, but any other
implementation should.
Change-Id: Id0a0fa031b7c6e9d572cdffeeb3a1e40d824826d
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417456
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
The NVMe PCIe transport only requires physically contiguous allocations
for struct nvme_tracker and the I/O SQ and CQ entries, which are already
handled separately. Change the comments to indicate that struct
nvme_payload's contiguous type only requires the memory to be virtually
contiguous, since nvme_pcie_prp_list_append() already steps through the
buffer and translates each (4K) page independently.
Change-Id: I45ac8dfb2c033a0fcbf2effbe33af4efc1eb23cb
Reported-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/417045
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The IOAT "ring" is actually just a circular linked list of descriptors;
the descriptors do not need to be in a single physically contiguous
region. This can be accomodated by calling spdk_vtophys() on each
descriptor rather than assuming they are all in a single contiguous
region.
Also store the physical address of each descriptor in its associated
software descriptor context to avoid the need to call spdk_vtophys()
during runtime.
Change-Id: Ic8636bbc61deb496a0c6d0ea56b75d298f5f426c
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/417782
Reviewed-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Print an error message instead. The driver can still do
a manual rescan if it wants to see the new SCSI target.
Change-Id: Ieb76ada8625bf00ad068a791b860e4b08ad5cb83
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417268
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I9e7ff07de8b06afb45b1059d455c8fe854035640
Signed-off-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-on: https://review.gerrithub.io/417783
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Also make it correctly account for alignment and automatically
use the internal iov element if necessary.
Change-Id: I0b33ef9444f0693c2d6b0cdaf221c4a5b0ad2cc3
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/416870
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This is potentially useful for more types of commands.
Change-Id: Ifbde7ae35294f581b8360891579836fd6f9573a6
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/416869
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This makes it possible to build SPDK on a machine which doesn't have
CUnit installed.
Change-Id: Icb4c01092d1432fcff4bdbbfa01489d1ddfcdd8b
Signed-off-by: Mike Playle <mplayle@solarflare.com>
Reviewed-on: https://review.gerrithub.io/417663
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
In our current code, when we call the following statement:
spdk_bdev_destruct_done(&lun->bdev, 0);
Actually lun is already freed in bdev_iscsi_lun_cleanup function
(called by iscsi_free_lun). So this patch can be used to
prevent this issue by changing the order.
Change-Id: I5ec02319b8205fafc4d8074511f5a334b9bbb3ad
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/417630
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Poll it every 50 microseconds.
Reason: Ceph's librbd service is not very fast.
So choose a timer period to poll it, instead of
use the syscall to poll it again and again, it
is unnecessary. And this is a default value,
which can be tuned by customers later.
Change-Id: I91c31442da5ddadf3bef43083bd789d01bb1c952
Signed-off-by: Ziye Yang <optimistyzy@gmail.com>
Reviewed-on: https://review.gerrithub.io/417442
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This code was created when spdk_pci_device wasn't
available yet. Now we can use spdk_pci_device* instead
of void* for extra code clarity.
Change-Id: I81d440720b22a484ae3d2739e0510a021bebbafe
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416995
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since delete_bdev should be used only for debug purpose,
this patch adds delete call specific for passthru bdev.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I8ea20b3003dd6539d84123c3b5363bd8bdfd6f7f
Reviewed-on: https://review.gerrithub.io/416535
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since delete_bdev should be used only for debug purpose,
this patch adds delete call specific for pmem bdev.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: Ic526e9ae462f595c4668c2b2612ad074208a7c4e
Reviewed-on: https://review.gerrithub.io/416520
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since delete_bdev should be used only for debug purpose,
this patch adds delete call specific for rbd bdev.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: Ia87459e0cc49a0c408de582bd1f5680c570d42d9
Reviewed-on: https://review.gerrithub.io/416529
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Since delete_bdev should be used only for debug purpose,
this patch adds delete call specific for null bdev.
Changes in spdkcli done accordingly.
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I576322257e5cf70ec03d6b20f8ba43bfce222907
Reviewed-on: https://review.gerrithub.io/416505
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
IO channel for LUN is used for hot removal as it is emphasized in the
previous patches.
In iSCSI, a SCSI device has multiple LUNs. So freeing only the IO channel
of a LUN must be possible.
Change-Id: I5b355200b4e173512a5aa4b7351534faf8839eef
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/417197
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reorder operations of LUN hot removal so that following are satisfied.
Wait for completion of all outstanding tasks first. (After turning lun->removed
on, there will be no new outstanding task.)
Then wait for IO channel being freed. (For VHOST SCSI, IO channel is freed
in the callback handler of hot removal. For iSCSI, IO channel is freed when
the final connection exits. IO channel of LUN is freed only by the allocator.)
Then free LUN finally.
For VHOST SCSI, the callback handler of hot removal will
call spdk_scsi_lun_destruct() in spdk_scsi_dev_destruct(). But lun->removed is
already turned on and spdk_scsi_lun_hot_remove() will be NOP. Hence LUN is
freed safely by the first caller of spdk_scsi_lun_hot_remove().
Change-Id: I276dfed1d4a7767e382003bd9bb389aaff920115
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-on: https://review.gerrithub.io/417196
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
We don't support the kernel vhost. Vhost-SCSI is not even
fully implemented in Linux, so there's no point trying.
Change-Id: Ie564d46c497718081dd2a5c28829fdcf88e1c0a0
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417455
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This fixes intermittent failures with QEMU's virtio-scsi-pci
device.
Apparently, from time to time QEMU enters a broken state in
which it doesn't turn off the NO_NOTIFY flag. This flag should
be set only for the period of time virtqueues are being processed,
but QEMU makes it set all the time. This makes us not signal any
I/O submissions - SPDK thinks the device is currently processing
notifications so it will pick up our I/O automatically, but in
realitly it won't and we end up with a deadlock.
I believe kernel virtio driver doesn't hit this issue because of
event index feature. If negotiated, the NO_NOTIFY flag won't be
used at all. Initial tests show that the issue is indeed gone
with this patch.
Event index for SPDK virtio is not particularly useful when using
a poll-mode device, but it doesn't do any harm. It can be further
optimized in the future, but macro-benchmarks don't show any
performance difference compared with the legacy notification
handling so there's no hurry.
Change-Id: I46e8ab0da170780fcdc0dd239208670ee5ed6357
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/415591
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Pawel Wodkowski <pawelx.wodkowski@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
With upcoming event index patch we will need
to keep track of how many descriptor chains we
have really put into the avail vring. This patch
is a step towards that.
Our virtio layer abstracts away descriptor chains
as "requests". We can start requests, add descriptors
to them, and finally flush them. So far we used to put
any descriptors directly into the virtqueue desc ring,
and made them visible to the device only through
virtqueue_req_flush().
All of our virtio bdev modules currently flush the
virtqueue after adding each single request, but the
docs for the virtio API say it's possible to start
multiple subsequent requests and flush them all at
once. This was conceptually broken so far and only
the last request would be exposed to the device.
It's now fixed and subsequent requests are put
into the avail vring as soon as they're complete
(either the next request is started, or the
virtqueue is flushed).
Change-Id: I76b7db88ab094c38430edd8ff0e65681775dcb31
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/415590
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
There's no need to split a bufffer if it's physically
contiguous. We can now merge buffers that would be
previously split by the nvme_pcie driver and also
separate SGEs provided by the user that happen to be
physically contiguous.
Change-Id: I9c9de31d52a9dc9e384806555cb94609aff0ccf3
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417061
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
This allows NVMe PCIe devices to be used with
physically discontiguous I/O payload buffers.
So far this is just a dumb splitting which
doesn't check for physical contiguity. This is
improved in a subsequent patch.
Change-Id: I0ecc443149225eaa0e4156ddda78613bcf034406
Suggested-by: Daniel Verkamp <daniel.verkamp@intel.com>
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/417060
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This is an spdk_memzone_reserve variant with additional
alignment parameter. Now that memzones must be used for
physically contiguous memory, it will become extremely useful.
Change-Id: Ie48d682217e0e2f5c859a1603bb8a81fd2a7d7df
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416978
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Future DPDK versions may drop physical memory contiguity
guarantee for common memzones. DPDK 18.05 introduces
an RTE_MEMZONE_IOVA_CONTIG (0x00100000) flag, which is
documented as follows:
> RTE_MEMZONE_IOVA_CONTIG - Ensure reserved memzone is IOVA-contiguous.
> This option should be used when allocating
> memory intended for hardware rings etc.
To preserve backward compatibility, SPDK introduces an opposite
flag, SPDK_MEMZONE_NO_IOVA_CONTIG.
Change-Id: I9ea79b096fdb094051f13c9a802740b0e4ccc98e
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416977
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
This const makes the array passed in immutable, but that
isn't helpful or interesting since it just contains
invalid pointer addresses. It may also make sense in the
future to NULL out the addresses in the array in a debug
build. So drop the const.
Change-Id: I921551c7cb1dbf6c765fb301c31906b8b93b7f16
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/417362
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
This was obviously omitted in the original implementation.
Change-Id: Ifba049cdb8d1c1f24e30adb542946a2fa83c7464
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-on: https://review.gerrithub.io/416655
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Within the bdev layer, we want to know specifically what I/O
types the module supports. However, the bdev module may elect
to emulate some commands and report additional support via
the public API.
The bdev layer already emulates WRITE_ZEROES, so correctly
report that fact.
Change-Id: I79bfb1aee1b3e6048f951bb1b2c7d4f7c9ef184d
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.gerrithub.io/416464
Tested-by: SPDK Automated Test System <sys_sgsw@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Daniel Verkamp <daniel.verkamp@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
