| Age | Commit message (Collapse) | Author |
|---|
|
Pull virtio updates from Michael Tsirkin:
"vhost,virtio,vdpa: features, fixes, cleanups.
vdpa/mlx5:
- VHOST_BACKEND_F_ENABLE_AFTER_DRIVER_OK
- new maintainer
vdpa:
- support for vq descriptor mappings
- decouple reset of iotlb mapping from device reset
and fixes, cleanups all over the place"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (34 commits)
vdpa_sim: implement .reset_map support
vdpa/mlx5: implement .reset_map driver op
vhost-vdpa: clean iotlb map during reset for older userspace
vdpa: introduce .compat_reset operation callback
vhost-vdpa: introduce IOTLB_PERSIST backend feature bit
vhost-vdpa: reset vendor specific mapping to initial state in .release
vdpa: introduce .reset_map operation callback
virtio_pci: add check for common cfg size
virtio-blk: fix implicit overflow on virtio_max_dma_size
virtio_pci: add build offset check for the new common cfg items
virtio: add definition of VIRTIO_F_NOTIF_CONFIG_DATA feature bit
vduse: make vduse_class constant
vhost-scsi: Spelling s/preceeding/preceding/g
virtio: kdoc for struct virtio_pci_modern_device
vdpa: Update sysfs ABI documentation
MAINTAINERS: Add myself as mlx5_vdpa driver
virtio-balloon: correct the comment of virtballoon_migratepage()
mlx5_vdpa: offer VHOST_BACKEND_F_ENABLE_AFTER_DRIVER_OK
vdpa/mlx5: Update cvq iotlb mapping on ASID change
vdpa/mlx5: Make iotlb helper functions more generic
...
|
|
Fix a misspelling of "preceding".
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Message-Id: <b57b882675809f1f9dacbf42cf6b920b2bea9cba.1695903476.git.geert+renesas@glider.be>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This allows userspace to request the fabric drivers do direct submissions
if they support it. With the new device file, submit_type, users can
write 0 - 2 to control how commands are submitted to the backend:
0 - TARGET_FABRIC_DEFAULT_SUBMIT - LIO will use the fabric's default
submission type. This is the default for compat.
1 - TARGET_DIRECT_SUBMIT - LIO will submit the cmd to the backend from the
calling context if the fabric the cmd was received on supports it,
else it will use the fabric's default type.
2 - TARGET_QUEUE_SUBMIT - LIO will queue the cmd to the LIO submission
workqueue which will pass it to the backend.
When using an NVMe drive and vhost-scsi with direct submission we see
around a 20% improvement in 4K I/Os:
fio jobs 1 2 4 8 10
--------------------------------------------------
defer 94K 190K 394K 770K 890K
direct 128K 252K 488K 950K -
And when using the queueing mode, we now no longer see issues like where
the iSCSI tx thread is blocked in the block layer waiting on a tag so it
can't respond to a nop or perform I/Os for other LUs.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230928020907.5730-6-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In some cases, like with multiple LUN targets or where the target has to
respond to transport level requests from the receiving context it can be
better to defer cmd submission to a helper thread. If the backend driver
blocks on something like request/tag allocation it can block the entire
target submission path and other LUs and transport IO on that session.
In other cases like single LUN targets with storage that can support all
the commands that the target can queue, then it's best to submit the cmd
to the backend from the target's cmd receiving context.
Subsequent commits will allow the user to config what they prefer, but
drivers like loop can't directly submit because they can be called from a
context that can't sleep. And, drivers like vhost-scsi can support direct
submission, but need to keep their default behavior of deferring execution
to avoid possible regressions where the backend can block.
Make the drivers tell LIO core if they support direct submissions and their
current default, so we can prevent users from misconfiguring the system and
initialize devices correctly.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20230928020907.5730-2-michael.christie@oracle.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Rename vhost_scsi_iov_to_sgl to vhost_scsi_map_iov_to_sgl so it matches
matches the naming style used for vhost_scsi_copy_iov_to_sgl.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230709202859.138387-3-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
The linux block layer requires bios/requests to have lengths with a 512
byte alignment. Some drivers/layers like dm-crypt and the directi IO code
will test for it and just fail. Other drivers like SCSI just assume the
requirement is met and will end up in infinte retry loops. The problem
for drivers like SCSI is that it uses functions like blk_rq_cur_sectors
and blk_rq_sectors which divide the request's length by 512. If there's
lefovers then it just gets dropped. But other code in the block/scsi
layer may use blk_rq_bytes/blk_rq_cur_bytes and end up thinking there is
still data left and try to retry the cmd. We can then end up getting
stuck in retry loops where part of the block/scsi thinks there is data
left, but other parts think we want to do IOs of zero length.
Linux will always check for alignment, but windows will not. When
vhost-scsi then translates the iovec it gets from a windows guest to a
scatterlist, we can end up with sg items where the sg->length is not
divisible by 512 due to the misaligned offset:
sg[0].offset = 255;
sg[0].length = 3841;
sg...
sg[N].offset = 0;
sg[N].length = 255;
When the lio backends then convert the SG to bios or other iovecs, we
end up sending them with the same misaligned values and can hit the
issues above.
This just has us drop down to allocating a temp page and copying the data
when we detect a misaligned buffer and the IO is large enough that it
will get split into multiple bad IOs.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230709202859.138387-2-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This has vhost-scsi support the worker ioctls by calling the
vhost_worker_ioctl helper.
With a single worker, the single thread becomes a bottlneck when trying
to use 3 or more virtqueues like:
fio --filename=/dev/sdb --direct=1 --rw=randrw --bs=4k \
--ioengine=libaio --iodepth=128 --numjobs=3
With the patches and doing a worker per vq, we can scale to at least
16 vCPUs/vqs (that's my system limit) with the same command fio command
above with numjobs=16:
fio --filename=/dev/sdb --direct=1 --rw=randrw --bs=4k \
--ioengine=libaio --iodepth=64 --numjobs=16
which gives around 2002K IOPs.
Note that for testing I dropped depth to 64 above because the vhost/virt
layer supports only 1024 total commands per device. And the only tuning I
did was set LIO's emulate_pr to 0 to avoid LIO's PR lock in the main IO
path which becomes an issue at around 12 jobs/virtqueues.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230626232307.97930-17-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
With one worker we will always send the scsi cmd responses then send the
TMF rsp, because LIO will always complete the scsi cmds first then call
into us to send the TMF response.
With multiple workers, the IO vq workers could be running while the
TMF/ctl vq worker is running so this has us do a flush before completing
the TMF to make sure cmds are completed when it's work is later queued
and run.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230626232307.97930-12-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
Convert from vhost_work_queue to vhost_vq_work_queue so we can
remove vhost_work_queue.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230626232307.97930-11-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
This patch separates the scsi cmd completion code paths so we can complete
cmds based on their vq instead of having all cmds complete on the same
worker/CPU. This will be useful with the next patches that allow us to
create mulitple worker threads and bind them to different vqs, and we can
have completions running on different threads/CPUs.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230626232307.97930-10-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
Pull virtio updates from Michael Tsirkin:
"virtio,vhost,vdpa: features, fixes, and cleanups:
- reduction in interrupt rate in virtio
- perf improvement for VDUSE
- scalability for vhost-scsi
- non power of 2 ring support for packed rings
- better management for mlx5 vdpa
- suspend for snet
- VIRTIO_F_NOTIFICATION_DATA
- shared backend with vdpa-sim-blk
- user VA support in vdpa-sim
- better struct packing for virtio
and fixes, cleanups all over the place"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (52 commits)
vhost_vdpa: fix unmap process in no-batch mode
MAINTAINERS: make me a reviewer of VIRTIO CORE AND NET DRIVERS
tools/virtio: fix build caused by virtio_ring changes
virtio_ring: add a struct device forward declaration
vdpa_sim_blk: support shared backend
vdpa_sim: move buffer allocation in the devices
vdpa/snet: use likely/unlikely macros in hot functions
vdpa/snet: implement kick_vq_with_data callback
virtio-vdpa: add VIRTIO_F_NOTIFICATION_DATA feature support
virtio: add VIRTIO_F_NOTIFICATION_DATA feature support
vdpa/snet: support the suspend vDPA callback
vdpa/snet: support getting and setting VQ state
MAINTAINERS: add vringh.h to Virtio Core and Net Drivers
vringh: address kdoc warnings
vdpa: address kdoc warnings
virtio_ring: don't update event idx on get_buf
vdpa_sim: add support for user VA
vdpa_sim: replace the spinlock with a mutex to protect the state
vdpa_sim: use kthread worker
vdpa_sim: make devices agnostic for work management
...
|
|
Pull SCSI updates from James Bottomley:
"Updates to the usual drivers (megaraid_sas, scsi_debug, lpfc, target,
mpi3mr, hisi_sas, arcmsr).
The major core change is the constification of the host templates
(which touches everything) along with other minor fixups and clean
ups"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (207 commits)
scsi: ufs: mcq: Use pointer arithmetic in ufshcd_send_command()
scsi: ufs: mcq: Annotate ufshcd_inc_sq_tail() appropriately
scsi: cxlflash: s/semahpore/semaphore/
scsi: lpfc: Silence an incorrect device output
scsi: mpi3mr: Use IRQ save variants of spinlock to protect chain frame allocation
scsi: scsi_debug: Fix missing error code in scsi_debug_init()
scsi: hisi_sas: Work around build failure in suspend function
scsi: lpfc: Fix ioremap issues in lpfc_sli4_pci_mem_setup()
scsi: mpt3sas: Fix an issue when driver is being removed
scsi: mpt3sas: Remove HBA BIOS version in the kernel log
scsi: target: core: Fix invalid memory access
scsi: scsi_debug: Drop sdebug_queue
scsi: scsi_debug: Only allow sdebug_max_queue be modified when no shosts
scsi: scsi_debug: Use scsi_host_busy() in delay_store() and ndelay_store()
scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in stop_all_queued()
scsi: scsi_debug: Use blk_mq_tagset_busy_iter() in sdebug_blk_mq_poll()
scsi: scsi_debug: Dynamically allocate sdebug_queued_cmd
scsi: scsi_debug: Use scsi_block_requests() to block queues
scsi: scsi_debug: Protect block_unblock_all_queues() with mutex
scsi: scsi_debug: Change shost list lock to a mutex
...
|
|
Pull ITER_UBUF updates from Jens Axboe:
"This turns singe vector imports into ITER_UBUF, rather than
ITER_IOVEC.
The former is more trivial to iterate and advance, and hence a bit
more efficient. From some very unscientific testing, ~60% of all iovec
imports are single vector"
* tag 'iter-ubuf.2-2023-04-21' of git://git.kernel.dk/linux:
iov_iter: Mark copy_compat_iovec_from_user() noinline
iov_iter: import single vector iovecs as ITER_UBUF
iov_iter: convert import_single_range() to ITER_UBUF
iov_iter: overlay struct iovec and ubuf/len
iov_iter: set nr_segs = 1 for ITER_UBUF
iov_iter: remove iov_iter_iovec()
iov_iter: add iter_iov_addr() and iter_iov_len() helpers
ALSA: pcm: check for user backed iterator, not specific iterator type
IB/qib: check for user backed iterator, not specific iterator type
IB/hfi1: check for user backed iterator, not specific iterator type
iov_iter: add iter_iovec() helper
block: ensure bio_alloc_map_data() deals with ITER_UBUF correctly
|
|
We on longer need to hold the vhost_scsi_mutex the entire time we
set/clear the endpoint. The tv_tpg_mutex handles tpg accesses not related
to the tpg list, the port link/unlink functions use the tv_tpg_mutex while
accessing the tpg->vhost_scsi pointer, vhost_scsi_do_plug will no longer
queue events after the virtqueue's backend has been cleared and flushed,
and we don't drop our refcount to the tpg until after we have stopped
cmds and wait for outstanding cmds to complete.
This moves the vhost_scsi_mutex use to it's documented use of being used
to access the tpg list. We then don't need to hold it while a flush is
being performed causing other device's vhost_scsi_set_endpoint
and vhost_scsi_make_tpg/vhost_scsi_drop_tpg calls to have to wait on a
flakey device.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-8-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
We are using the vhost_scsi_mutex to make sure vhost_scsi_port_link and
vhost_scsi_port_unlink see if vhost_scsi_clear_endpoint has cleared
tpg->vhost_scsi and it can't be freed while they are using.
However, we currently set the tpg->vhost_scsi pointer while holding
tv_tpg_mutex. So, we can just hold that while calling
vhost_scsi_hotplug/hotunplug. We then don't need to hold the
vhost_scsi_mutex while vhost_scsi_clear_endpoint is holding it and doing
a flush which could cause the LUN map/unmap to have to wait on another
device's flush.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-7-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
We currenly hold the vhost_scsi_mutex while clearing the endpoint and
while performing vhost_scsi_do_plug, so tpg->vhost_scsi can't be freed
from uder us, and to make sure anything queued is handled by the
full call in vhost_scsi_clear_endpoint.
This patch removes the need for the vhost_scsi_mutex for the latter
case. In the next patches, we won't hold the vhost_scsi_mutex while
flushing so this patch adds a check for the clearing of the virtqueue
from vhost_scsi_clear_endpoint. We then know that once
vhost_scsi_clear_endpoint has cleared the backend that no new events
will be queued, and the flush after the vhost_vq_set_backend(vq, NULL)
call will see everything that's been queued to that point. So the flush
will then handle all events without the need for the vhost_scsi_mutex.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-6-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
We don't need the device mutex in vhost_scsi_do_plug because:
1. we have the vhost_scsi_mutex so the tpg->vhost_scsi pointer will not
change on us and the vhost_scsi can't be freed from under us if it was
set.
2. vhost_scsi_clear_endpoint will stop the virtqueues and flush them while
holding the vhost_scsi_mutex so we know once vhost_scsi_clear_endpoint
has completed that vhost_scsi_do_plug can't send new events and any
queued ones have completed.
So this patch drops the device mutex use in vhost_scsi_do_plug.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-5-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
We currently hold the vhost_scsi_mutex the entire time we are running
vhost_scsi_clear_endpoint. One of the reasons for this is that it prevents
userspace from being able to free the se_tpg from under us after we have
called target_undepend_item. However, it forces management operations for
for other devices to have to wait on a flakey device's:
vhost_scsi_clear_endpoint -> vhost_scsi_flush()
call which can which can take a long time.
This moves the target_undepend_item call and the tpg unsetup code to after
we have stopped new IO from starting up and after we have waited on
running IO. We can then release our refcount on the tpg and session
knowing our device is no longer accessing them. We can then drop the
vhost_scsi_mutex use during thee flush call in later patches in this set,
when we have removed other reasons for holding it.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-4-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
We normally clear the endpoint then unmap LUNs so the devices are fully
shutdown when the LUN is unmapped, but it's legal to unmap before
clearing. If the user does that while TMFs are running then we can end
up crashing.
vhost_scsi_port_unlink assumes that the LUN's tmf struct will always be on
the tmf_queue list. However, if a TMF is running then it will have been
removed while it's executing. If we do a LUN unmap at this time, then
we assume the entry is on the list and just start accessing it and free
it.
This fixes the bug by just allocating the vhost_scsi_tmf struct when it's
needed like is done with the se_tmr struct that's needed when we submit
the TMF. In this path perf is not an issue and we can use GFP_KERNEL
since it won't swing directly back on us, so we don't need to preallocate
the struct.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-3-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
If vhost_scsi_setup_vq_cmds fails we leave the tpg->vhost_scsi pointer
set. If the device is freed and then the user unmaps the LUN, the call to
vhost_scsi_port_unlink -> vhost_scsi_hotunplug will see the that
tpg->vhost_scsi is still set and try to use it.
This has us clear the vhost_scsi pointer in the failure path. It also
has us take tv_tpg_mutex in this failure path, because tv_tpg_vhost_count
is accessed under this mutex in vhost_scsi_drop_nexus and in the future
we will want to serialize access to tpg->vhost_scsi with that mutex
instead of the vhost_scsi_mutex.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20230321020624.13323-2-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
This returns a pointer to the current iovec entry in the iterator. Only
useful with ITER_IOVEC right now, but it prepares us to treat ITER_UBUF
and ITER_IOVEC identically for the first segment.
Rename struct iov_iter->iov to iov_iter->__iov to find any potentially
troublesome spots, and also to prevent anyone from adding new code that
accesses iter->iov directly.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Remove callouts that are identical to the default implementations in TCM
Core.
Signed-off-by: Dmitry Bogdanov <d.bogdanov@yadro.com>
Link: https://lore.kernel.org/r/20230313181110.20566-9-d.bogdanov@yadro.com
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Follow the advice of the Documentation/filesystems/sysfs.rst
and show() should only use sysfs_emit() or sysfs_emit_at()
when formatting the value to be returned to user space.
Signed-off-by: Bo Liu <liubo03@inspur.com>
Message-Id: <20230129091145.2837-1-liubo03@inspur.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
Al Viro said:
"""
Since "vhost/scsi: fix reuse of &vq->iov[out] in response"
we have this:
cmd->tvc_resp_iov = vq->iov[vc.out];
cmd->tvc_in_iovs = vc.in;
combined with
iov_iter_init(&iov_iter, ITER_DEST, &cmd->tvc_resp_iov,
cmd->tvc_in_iovs, sizeof(v_rsp));
in vhost_scsi_complete_cmd_work(). We used to have ->tvc_resp_iov
_pointing_ to vq->iov[vc.out]; back then iov_iter_init() asked to
set an iovec-backed iov_iter over the tail of vq->iov[], with
length being the amount of iovecs in the tail.
Now we have a copy of one element of that array. Fortunately, the members
following it in the containing structure are two non-NULL kernel pointers,
so copy_to_iter() will not copy anything beyond the first iovec - kernel
pointer is not (on the majority of architectures) going to be accepted by
access_ok() in copyout() and it won't be skipped since the "length" (in
reality - another non-NULL kernel pointer) won't be zero.
So it's not going to give a guest-to-qemu escalation, but it's definitely
a bug. Frankly, my preference would be to verify that the very first iovec
is long enough to hold rsp_size. Due to the above, any users that try to
give us vq->iov[vc.out].iov_len < sizeof(struct virtio_scsi_cmd_resp)
would currently get a failure in vhost_scsi_complete_cmd_work()
anyway.
"""
However, the spec doesn't say anything about the legacy descriptor
layout for the respone. So this patch tries to not assume the response
to reside in a single separate descriptor which is what commit
79c14141a487 ("vhost/scsi: Convert completion path to use") tries to
achieve towards to ANY_LAYOUT.
This is done by allocating and using dedicate resp iov in the
command. To be safety, start with UIO_MAXIOV to be consistent with the
limitation that we advertise to the vhost_get_vq_desc().
Testing with the hacked virtio-scsi driver that use 1 descriptor for 1
byte in the response.
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Cc: Benjamin Coddington <bcodding@redhat.com>
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Fixes: a77ec83a5789 ("vhost/scsi: fix reuse of &vq->iov[out] in response")
Signed-off-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20230119073647.76467-1-jasowang@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
READ/WRITE proved to be actively confusing - the meanings are
"data destination, as used with read(2)" and "data source, as
used with write(2)", but people keep interpreting those as
"we read data from it" and "we write data to it", i.e. exactly
the wrong way.
Call them ITER_DEST and ITER_SOURCE - at least that is harder
to misinterpret...
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
Pull virtio updates from Michael Tsirkin:
- A huge patchset supporting vq resize using the new vq reset
capability
- Features, fixes, and cleanups all over the place
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (88 commits)
vdpa/mlx5: Fix possible uninitialized return value
vdpa_sim_blk: add support for discard and write-zeroes
vdpa_sim_blk: add support for VIRTIO_BLK_T_FLUSH
vdpa_sim_blk: make vdpasim_blk_check_range usable by other requests
vdpa_sim_blk: check if sector is 0 for commands other than read or write
vdpa_sim: Implement suspend vdpa op
vhost-vdpa: uAPI to suspend the device
vhost-vdpa: introduce SUSPEND backend feature bit
vdpa: Add suspend operation
virtio-blk: Avoid use-after-free on suspend/resume
virtio_vdpa: support the arg sizes of find_vqs()
vhost-vdpa: Call ida_simple_remove() when failed
vDPA: fix 'cast to restricted le16' warnings in vdpa.c
vDPA: !FEATURES_OK should not block querying device config space
vDPA/ifcvf: support userspace to query features and MQ of a management device
vDPA/ifcvf: get_config_size should return a value no greater than dev implementation
vhost scsi: Allow user to control num virtqueues
vhost-scsi: Fix max number of virtqueues
vdpa/mlx5: Support different address spaces for control and data
vdpa/mlx5: Implement susupend virtqueue callback
...
|
|
We are currently hard coded to always create 128 IO virtqueues, so this
adds a modparam to control it. For large systems where we are ok with
using memory for virtqueues it allows us to add up to 1024. This limit
was just selected becuase that's qemu's limit.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20220708030525.5065-3-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
|
|
Qemu takes it's num_queues limit then adds the fixed queues (control and
event) to the total it will request from the kernel. So when a user
requests 128 (or qemu does it's num_queues calculation based on vCPUS
and other system limits), we hit errors due to userspace trying to setup
130 queues when vhost-scsi has a hard coded limit of 128.
This has vhost-scsi adjust it's max so we can do a total of 130 virtqueues
(128 IO and 2 fixed). For the case where the user has 128 vCPUs the guest
OS can then nicely map each IO virtqueue to a vCPU and not have the odd case
where 2 vCPUs share a virtqueue.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Message-Id: <20220708030525.5065-2-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
|
|
Most of the users immediately follow successful iov_iter_get_pages()
with advancing by the amount it had returned.
Provide inline wrappers doing that, convert trivial open-coded
uses of those.
BTW, iov_iter_get_pages() never returns more than it had been asked
to; such checks in cifs ought to be removed someday...
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
This patch renames vhost_work_dev_flush to just vhost_dev_flush to
relfect that it flushes everything on the device and that drivers
don't know/care that polls are based on vhost_works. Drivers just
flush the entire device and polls, and works for vhost-scsi
management TMFs and IO net virtqueues, etc all are flushed.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Message-Id: <20220517180850.198915-9-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
The flush after vhost_dev_cleanup is not needed because:
1. It doesn't do anything. vhost_dev_cleanup will stop the worker thread
so the flush call will just return since the worker has not device.
2. It's not needed for the re-queue case. vhost_scsi_evt_handle_kick grabs
the mutex and if the backend is NULL will return without queueing a work.
vhost_scsi_clear_endpoint will set the backend to NULL under the vq->mutex
then drops the mutex and does a flush. So we know when
vhost_scsi_clear_endpoint has dropped the mutex after clearing the backend
no evt related work will be able to requeue. The flush would then make sure
any queued evts are run and return.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Message-Id: <20220517180850.198915-7-michael.christie@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
use SPDX-License-Identifier instead of a verbose license text
Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Link: https://lore.kernel.org/r/20210821123320.734-1-caihuoqing@baidu.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
vhost_scsi_flush will flush everything, so we can clear the backends then
flush, then destroy. We don't need to flush before each vq destruction
because after the flush we will have made sure there can be no new cmds
started and there are no running cmds.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20210525174733.6212-4-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
The vhost work flush function was flushing the entire work queue, so
there is no need for the double vhost_work_dev_flush calls in
vhost_scsi_flush.
And we do not need to call vhost_poll_flush for each poller because
that call also ends up flushing the same work queue thread the
vhost_work_dev_flush call flushed.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20210525174733.6212-3-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
vhost_work_flush doesn't do anything with the work arg. This patch drops
it and then renames vhost_work_flush to vhost_work_dev_flush to reflect
that the function flushes all the works in the dev and not just a
specific queue or work item.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Chaitanya Kulkarni <chaitanya.kulkarni@wdc.com>
Link: https://lore.kernel.org/r/20210525174733.6212-2-michael.christie@oracle.com
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
Convert vhost-scsi to use the LIO wq cmd submission helper.
Link: https://lore.kernel.org/r/20210227170006.5077-18-michael.christie@oracle.com
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
tcm_loop could be used like a normal block device, so we can't use
GFP_KERNEL and should use GFP_NOIO. This adds a gfp_t arg to
target_cmd_init_cdb() and converts the users. For every driver but loop
GFP_KERNEL is kept.
This will also be useful in subsequent patches where loop needs to do
target_submit_prep() from interrupt context to get a ref to the se_device,
and so it will need to use GFP_ATOMIC.
Link: https://lore.kernel.org/r/20210227170006.5077-16-michael.christie@oracle.com
Tested-by: Laurence Oberman <loberman@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
target_submit_cmd_map_sgls() is being removed, so convert vhost-scsi to the
new submission API. This has it use target_init_cmd(),
target_submit_prep(), target_submit() because we need to have LIO core map
sgls which is now done in target_submit_prep(), and in the next patches we
will do the target_submit step from the LIO workqueue.
Note: vhost-scsi never calls target_stop_session() so
target_submit_cmd_map_sgls() never failed (in the new API target_init_cmd()
handles target_stop_session() being called when cmds are being
submitted). If it were to have used target_stop_session() and got an error,
we would have hit a refcount bug like xen and usb, because it does:
if (rc < 0) {
transport_send_check_condition_and_sense(se_cmd,
TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE, 0);
transport_generic_free_cmd(se_cmd, 0);
}
transport_send_check_condition_and_sense() calls queue_status which does
transport_generic_free_cmd(), and then we do an extra
transport_generic_free_cmd() call above which would have dropped the
refcount to -1 and the refcount code would spit out errors.
Link: https://lore.kernel.org/r/20210227170006.5077-12-michael.christie@oracle.com
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Allocation hint should have belonged to sbitmap. Also, when sbitmap's depth
is high and there is no need to use mulitple wakeup queues, user can
benefit from percpu allocation hint too.
Move allocation hint into sbitmap, then SCSI device queue can benefit from
allocation hint when converting to plain sbitmap.
Convert vhost/scsi.c to use sbitmap allocation with percpu alloc hint. This
is more efficient than the previous approach.
Link: https://lore.kernel.org/r/20210122023317.687987-5-ming.lei@redhat.com
Cc: Omar Sandoval <osandov@fb.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Cc: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Cc: Mike Christie <michael.christie@oracle.com>
Cc: virtualization@lists.linux-foundation.org
Tested-by: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Currently the allocation round_robin info is maintained by sbitmap_queue.
However, bit allocation really belongs to sbitmap. Move it there.
Link: https://lore.kernel.org/r/20210122023317.687987-3-ming.lei@redhat.com
Cc: Omar Sandoval <osandov@fb.com>
Cc: Kashyap Desai <kashyap.desai@broadcom.com>
Cc: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: virtualization@lists.linux-foundation.org
Tested-by: Sumanesh Samanta <sumanesh.samanta@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The size of 'struct vhost_scsi' is order-10 (~2.3MB). It may take long time
delay by kzalloc() to compact memory pages by retrying multiple times when
there is a lack of high-order pages. As a result, there is latency to
create a VM (with vhost-scsi) or to hotadd vhost-scsi-based storage.
The prior commit 595cb754983d ("vhost/scsi: use vmalloc for order-10
allocation") prefers to fallback only when really needed, while this patch
allocates with kvzalloc() with __GFP_NORETRY implicitly set to avoid
retrying memory pages compact for multiple times.
The __GFP_NORETRY is implicitly set if the size to allocate is more than
PAGE_SZIE and when __GFP_RETRY_MAYFAIL is not explicitly set.
Cc: Aruna Ramakrishna <aruna.ramakrishna@oracle.com>
Cc: Joe Jin <joe.jin@oracle.com>
Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>
Link: https://lore.kernel.org/r/20210123080853.4214-1-dongli.zhang@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
|
|
Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.
Fixes: 25b98b64e284 ("vhost scsi: alloc cmds per vq instead of session")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1607071411-33484-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
vhost scsi owns the scsi se_cmd but lio frees the se_cmd->se_tmr
before calling release_cmd, so while with normal cmd completion we
can access the se_cmd from the vhost work, we can't do the same with
se_cmd->se_tmr. This has us copy the tmf response in
vhost_scsi_queue_tm_rsp to our internal vhost-scsi tmf struct for
when it gets sent to the guest from our worker thread.
Fixes: efd838fec17b ("vhost scsi: Add support for LUN resets.")
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Link: https://lore.kernel.org/r/1605887459-3864-1-git-send-email-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
In newer versions of virtio-scsi we just reset the timer when an a
command times out, so TMFs are never sent for the cmd time out case.
However, in older kernels and for the TMF inject cases, we can still get
resets and we end up just failing immediately so the guest might see the
device get offlined and IO errors.
For the older kernel cases, we want the same end result as the
modern virtio-scsi driver where we let the lower levels fire their error
handling and handle the problem. And at the upper levels we want to
wait. This patch ties the LUN reset handling into the LIO TMF code which
will just wait for outstanding commands to complete like we are doing in
the modern virtio-scsi case.
Note: I did not handle the ABORT case to keep this simple. For ABORTs
LIO just waits on the cmd like how it does for the RESET case. If
an ABORT fails, the guest OS ends up escalating to LUN RESET, so in
the end we get the same behavior where we wait on the outstanding
cmds.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/1604986403-4931-6-git-send-email-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Move code to parse lun from req's lun_buf to helper, so tmf code
can use it in the next patch.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/1604986403-4931-5-git-send-email-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
We might not do the final se_cmd put from vhost_scsi_complete_cmd_work.
When the last put happens a little later then we could race where
vhost_scsi_complete_cmd_work does vhost_signal, the guest runs and sends
more IO, and vhost_scsi_handle_vq runs but does not find any free cmds.
This patch has us delay completing the cmd until the last lio core ref
is dropped. We then know that once we signal to the guest that the cmd
is completed that if it queues a new command it will find a free cmd.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Reviewed-by: Maurizio Lombardi <mlombard@redhat.com>
Link: https://lore.kernel.org/r/1604986403-4931-4-git-send-email-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
We currently are limited to 256 cmds per session. This leads to problems
where if the user has increased virtqueue_size to more than 2 or
cmd_per_lun to more than 256 vhost_scsi_get_tag can fail and the guest
will get IO errors.
This patch moves the cmd allocation to per vq so we can easily match
whatever the user has specified for num_queues and
virtqueue_size/cmd_per_lun. It also makes it easier to control how much
memory we preallocate. For cases, where perf is not as important and
we can use the current defaults (1 vq and 128 cmds per vq) memory use
from preallocate cmds is cut in half. For cases, where we are willing
to use more memory for higher perf, cmd mem use will now increase as
the num queues and queue depth increases.
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/1604986403-4931-3-git-send-email-michael.christie@oracle.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Maurizio Lombardi <mlombard@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
vhost/scsi doesn't handle type conversion correctly
for request type when using virtio 1.0 and up for BE,
or cross-endian platforms.
Fix it up using vhost_32_to_cpu.
Cc: stable@vger.kernel.org
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Pull virtio updates from Michael Tsirkin:
- virtio-mem: paravirtualized memory hotplug
- support doorbell mapping for vdpa
- config interrupt support in ifc
- fixes all over the place
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (40 commits)
vhost/test: fix up after API change
virtio_mem: convert device block size into 64bit
virtio-mem: drop unnecessary initialization
ifcvf: implement config interrupt in IFCVF
vhost: replace -1 with VHOST_FILE_UNBIND in ioctls
vhost_vdpa: Support config interrupt in vdpa
ifcvf: ignore continuous setting same status value
virtio-mem: Don't rely on implicit compiler padding for requests
virtio-mem: Try to unplug the complete online memory block first
virtio-mem: Use -ETXTBSY as error code if the device is busy
virtio-mem: Unplug subblocks right-to-left
virtio-mem: Drop manual check for already present memory
virtio-mem: Add parent resource for all added "System RAM"
virtio-mem: Better retry handling
virtio-mem: Offline and remove completely unplugged memory blocks
mm/memory_hotplug: Introduce offline_and_remove_memory()
virtio-mem: Allow to offline partially unplugged memory blocks
mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE
virtio-mem: Paravirtualized memory hotunplug part 2
virtio-mem: Paravirtualized memory hotunplug part 1
...
|
|
Pull SCSI updates from James Bottomley:
:This series consists of the usual driver updates (qla2xxx, ufs, zfcp,
target, scsi_debug, lpfc, qedi, qedf, hisi_sas, mpt3sas) plus a host
of other minor updates.
There are no major core changes in this series apart from a
refactoring in scsi_lib.c"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (207 commits)
scsi: ufs: ti-j721e-ufs: Fix unwinding of pm_runtime changes
scsi: cxgb3i: Fix some leaks in init_act_open()
scsi: ibmvscsi: Make some functions static
scsi: iscsi: Fix deadlock on recovery path during GFP_IO reclaim
scsi: ufs: Fix WriteBooster flush during runtime suspend
scsi: ufs: Fix index of attributes query for WriteBooster feature
scsi: ufs: Allow WriteBooster on UFS 2.2 devices
scsi: ufs: Remove unnecessary memset for dev_info
scsi: ufs-qcom: Fix scheduling while atomic issue
scsi: mpt3sas: Fix reply queue count in non RDPQ mode
scsi: lpfc: Fix lpfc_nodelist leak when processing unsolicited event
scsi: target: tcmu: Fix a use after free in tcmu_check_expired_queue_cmd()
scsi: vhost: Notify TCM about the maximum sg entries supported per command
scsi: qla2xxx: Remove return value from qla_nvme_ls()
scsi: qla2xxx: Remove an unused function
scsi: iscsi: Register sysfs for iscsi workqueue
scsi: scsi_debug: Parser tables and code interaction
scsi: core: Refactor scsi_mq_setup_tags function
scsi: core: Fix incorrect usage of shost_for_each_device
scsi: qla2xxx: Fix endianness annotations in source files
...
|
