When the number of a virtio-blk device's virtqueues is larger than
BITS_PER_LONG, the out-of-bounds access to bitmap[ ] will occur.
Fixes: e21737ab15 ("virtio-blk: multiqueue batch notify")
Cc: qemu-stable@nongnu.org
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Li Hangjing <lihangjing@baidu.com>
Reviewed-by: Xie Yongji <xieyongji@baidu.com>
Reviewed-by: Chai Wen <chaiwen@baidu.com>
Message-id: 20191216023050.48620-1-lihangjing@baidu.com
Message-Id: <20191216023050.48620-1-lihangjing@baidu.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
When 'system_reset' is called, the main loop clear the memory
region cache before the BH has a chance to execute. Later when
the deferred function is called, some assumptions that were
made when scheduling them are no longer true when they actually
execute.
This is what happens using a virtio-blk device (fresh RHEL7.8 install):
$ (sleep 12.3; echo system_reset; sleep 12.3; echo system_reset; sleep 1; echo q) \
| qemu-system-x86_64 -m 4G -smp 8 -boot menu=on \
-device virtio-blk-pci,id=image1,drive=drive_image1 \
-drive file=/var/lib/libvirt/images/rhel78.qcow2,if=none,id=drive_image1,format=qcow2,cache=none \
-device virtio-net-pci,netdev=net0,id=nic0,mac=52:54:00:c4:e7:84 \
-netdev tap,id=net0,script=/bin/true,downscript=/bin/true,vhost=on \
-monitor stdio -serial null -nographic
(qemu) system_reset
(qemu) system_reset
(qemu) qemu-system-x86_64: hw/virtio/virtio.c:225: vring_get_region_caches: Assertion `caches != NULL' failed.
Aborted
(gdb) bt
Thread 1 (Thread 0x7f109c17b680 (LWP 10939)):
#0 0x00005604083296d1 in vring_get_region_caches (vq=0x56040a24bdd0) at hw/virtio/virtio.c:227
#1 0x000056040832972b in vring_avail_flags (vq=0x56040a24bdd0) at hw/virtio/virtio.c:235
#2 0x000056040832d13d in virtio_should_notify (vdev=0x56040a240630, vq=0x56040a24bdd0) at hw/virtio/virtio.c:1648
#3 0x000056040832d1f8 in virtio_notify_irqfd (vdev=0x56040a240630, vq=0x56040a24bdd0) at hw/virtio/virtio.c:1662
#4 0x00005604082d213d in notify_guest_bh (opaque=0x56040a243ec0) at hw/block/dataplane/virtio-blk.c:75
#5 0x000056040883dc35 in aio_bh_call (bh=0x56040a243f10) at util/async.c:90
#6 0x000056040883dccd in aio_bh_poll (ctx=0x560409161980) at util/async.c:118
#7 0x0000560408842af7 in aio_dispatch (ctx=0x560409161980) at util/aio-posix.c:460
#8 0x000056040883e068 in aio_ctx_dispatch (source=0x560409161980, callback=0x0, user_data=0x0) at util/async.c:261
#9 0x00007f10a8fca06d in g_main_context_dispatch () at /lib64/libglib-2.0.so.0
#10 0x0000560408841445 in glib_pollfds_poll () at util/main-loop.c:215
#11 0x00005604088414bf in os_host_main_loop_wait (timeout=0) at util/main-loop.c:238
#12 0x00005604088415c4 in main_loop_wait (nonblocking=0) at util/main-loop.c:514
#13 0x0000560408416b1e in main_loop () at vl.c:1923
#14 0x000056040841e0e8 in main (argc=20, argv=0x7ffc2c3f9c58, envp=0x7ffc2c3f9d00) at vl.c:4578
Fix this by cancelling the BH when the virtio dataplane is stopped.
[This is version of the patch was modified as discussed with Philippe on
the mailing list thread.
--Stefan]
Reported-by: Yihuang Yu <yihyu@redhat.com>
Suggested-by: Stefan Hajnoczi <stefanha@redhat.com>
Fixes: https://bugs.launchpad.net/qemu/+bug/1839428
Signed-off-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20190816171503.24761-1-philmd@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
In my "build everything" tree, changing qemu/main-loop.h triggers a
recompile of some 5600 out of 6600 objects (not counting tests and
objects that don't depend on qemu/osdep.h). It includes block/aio.h,
which in turn includes qemu/event_notifier.h, qemu/notify.h,
qemu/processor.h, qemu/qsp.h, qemu/queue.h, qemu/thread-posix.h,
qemu/thread.h, qemu/timer.h, and a few more.
Include qemu/main-loop.h only where it's needed. Touching it now
recompiles only some 1700 objects. For block/aio.h and
qemu/event_notifier.h, these numbers drop from 5600 to 2800. For the
others, they shrink only slightly.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Message-Id: <20190812052359.30071-21-armbru@redhat.com>
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
In my "build everything" tree, changing hw/hw.h triggers a recompile
of some 2600 out of 6600 objects (not counting tests and objects that
don't depend on qemu/osdep.h).
The previous commits have left only the declaration of hw_error() in
hw/hw.h. This permits dropping most of its inclusions. Touching it
now recompiles less than 200 objects.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Message-Id: <20190812052359.30071-19-armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
This patch introduces a poll callback for event channel fd-s and uses
this to invoke the channel callback function.
To properly support polling, it is necessary for the event channel callback
function to return a boolean saying whether it has done any useful work or
not. Thus xen_block_dataplane_event() is modified to directly invoke
xen_block_handle_requests() and the latter only returns true if it actually
processes any requests. This also means that the call to qemu_bh_schedule()
is moved into xen_block_complete_aio(), which is more intuitive since the
only reason for doing a deferred poll of the shared ring should be because
there were previously insufficient resources to fully complete a previous
poll.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20190408151617.13025-4-paul.durrant@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
This patch adds an AioContext parameter to xen_device_bind_event_channel()
and then uses aio_set_fd_handler() to set the callback rather than
qemu_set_fd_handler().
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20190408151617.13025-3-paul.durrant@citrix.com>
[Call aio_set_fd_handler() with is_external=true]
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
A recent Xen commit [1] clarified the semantics of sector based quantities
used in the blkif protocol such that it is now safe to create a xen-block
device with a logical_block_size != 512, as long as the device only
connects to a frontend advertizing 'feature-large-block-size'.
This patch modifies xen-block accordingly. It also uses a stack variable
for the BlockBackend in xen_block_realize() to avoid repeated dereferencing
of the BlockConf pointer, and changes the parameters of
xen_block_dataplane_create() so that the BlockBackend pointer and sector
size are passed expicitly rather than implicitly via the BlockConf.
These modifications have been tested against a recent Windows PV XENVBD
driver [2] using a xen-disk device with a 4kB logical block size.
[1] http://xenbits.xen.org/gitweb/?p=xen.git;a=commit;h=67e1c050e36b2c9900cca83618e56189effbad98
[2] https://winpvdrvbuild.xenproject.org:8080/job/XENVBD-master/126
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20190409164038.25484-1-paul.durrant@citrix.com>
[Edited error message]
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Add an Error parameter to blk_set_aio_context() and use
bdrv_child_try_set_aio_context() internally to check whether all
involved nodes can actually support the AioContext switch.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The Xen blkif protocol requires that sector based quantities should be
interpreted strictly as multiples of 512 bytes. Specifically:
"first_sect and last_sect in blkif_request_segment, as well as
sector_number in blkif_request, are always expressed in 512-byte units."
Commit fcab2b464e "xen: add header and build dataplane/xen-block.c"
incorrectly modified behaviour to use the block device logical_block_size
property as the scale, instead of correctly shifting values by the
hardcoded BDRV_SECTOR_BITS (and hence scaling them to 512 byte units).
This patch undoes that change and restores compliance with the spec.
Furthermore, this patch also restores the original xen_disk behaviour
of advertizing a hardcoded 'sector-size' value of 512 in xenstore and
scaling 'sectors' accordingly. The realize() method is also modified to
fail if logical_block_size is set to anything other than 512.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20190401121719.27208-1-paul.durrant@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
We spell out sub/dir/ in sub/dir/trace-events' comments pointing to
source files. That's because when trace-events got split up, the
comments were moved verbatim.
Delete the sub/dir/ part from these comments. Gets rid of several
misspellings.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-id: 20190314180929.27722-3-armbru@redhat.com
Message-Id: <20190314180929.27722-3-armbru@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The if() statement is clearly bogus (dead code which should have been
cleaned up when grant mapping was removed).
Spotted by Coverity: CID 1398635
While in the neighbourhood, add a missing 'fall through' annotation.
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Message-Id: <20190215162533.19475-2-paul.durrant@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Some frontend drivers will handle dynamic resizing of PV disks, so set up
the BlockDevOps resize_cb() method during xen_block_realize() to allow
this to be done.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
The xen-block dataplane currently allocates memory to hold the data for
each request as that request is used, and frees it afterwards. Because
it requires page-aligned blocks, this interacts poorly with non-page-
aligned allocations and balloons the heap.
Instead, allocate the maximum possible buffer size required for the
protocol, which is BLKIF_MAX_SEGMENTS_PER_REQUEST (currently 11) pages
when the request structure is created, and keep that buffer until it is
destroyed. Since the requests are re-used via a free list, this should
actually improve memory usage.
Signed-off-by: Tim Smith <tim.smith@citrix.com>
Re-based and commit comment adjusted.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
If the I/O ring is full, the guest cannot send any more requests
until some responses are sent. Only sending all available responses
just before checking for new work does not leave much time for the
guest to supply new work, so this will cause stalls if the ring gets
full. Also, not completing reads as soon as possible adds latency
to the guest.
To alleviate that, complete IO requests as soon as they come back.
xen_block_send_response() already returns a value indicating whether
a notify should be sent, which is all the batching we need.
Signed-off-by: Tim Smith <tim.smith@citrix.com>
Re-based and commit comment adjusted.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
When I/O consists of many small requests, performance is improved by
batching them together in a single io_submit() call. When there are
relatively few requests, the extra overhead is not worth it. This
introduces a check to start batching I/O requests via blk_io_plug()/
blk_io_unplug() in an amount proportional to the number which were
already in flight at the time we started reading the ring.
Signed-off-by: Tim Smith <tim.smith@citrix.com>
Re-based and commit comment adjusted.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony PERARD <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
This is a purely cosmetic patch that purges remaining use of 'blk' and
'ioreq' in local function names, and then makes sure all functions are
prefixed with 'xen_block_'.
No functional change.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
This is a purely cosmetic patch that purges the name 'ioreq' from struct,
variable and field names. (This name has been problematic for a long time
as 'ioreq' is the name used for generic I/O requests coming from Xen).
The patch replaces 'struct ioreq' with a new 'XenBlockRequest' type and
'ioreq' field/variable names with 'request', and then does necessary
fix-up to adhere to coding style.
Function names are not modified by this patch. They will be dealt with in
a subsequent patch.
No functional change.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
This is a purely cosmetic patch that substitutes the old 'struct XenBlkDev'
name with 'XenBlockDataPlane' and 'blkdev' field/variable names with
'dataplane', and then does necessary fix-up to adhere to coding style.
No functional change.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
This patch adds the transformations necessary to get dataplane/xen-block.c
to build against the new XenBus/XenDevice framework. MAINTAINERS is also
updated due to the introduction of dataplane/xen-block.h.
NOTE: Existing data structure names are retained for the moment. These will
be modified by subsequent patches. A typedef for XenBlockDataPlane
has been added to the header (based on the old struct XenBlkDev name
for the moment) so that the old names don't need to leak out of the
dataplane code.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Reviewed-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Not all of the code duplicated from xen_disk.c is required as the basis for
the new dataplane implementation so this patch removes extraneous code,
along with the legacy #includes and calls to the legacy xen_pv_printf()
function. Error messages are changed to be reported using error_report().
NOTE: The code is still not yet built. Further transformations will be
required to make it correctly interface to the new XenBus/XenDevice
framework. They will be delivered in a subsequent patch.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
The new xen-block XenDevice implementation requires the same core
dataplane as the legacy xen_disk implementation it will eventually replace.
This patch therefore copies the legacy xen_disk.c source module into a new
dataplane/xen-block.c source module as the basis for the new dataplane and
adjusts the MAINTAINERS file accordingly.
NOTE: The duplicated code is not yet built. It is simply put into place by
this patch (just fixing style violations) such that the
modifications that will need to be made to the code are not
conflated with code movement, thus making review harder.
Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Acked-by: Anthony Perard <anthony.perard@citrix.com>
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
The preferred way to select the KVM accelerator is to use "-accel kvm"
these days, so let's be consistent in our documentation and help texts.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Message-Id: <1528866321-23886-3-git-send-email-thuth@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Remove those unneeded includes to speed up the compilation
process a little bit. (Continue 7eceff5b5a cleanup)
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20180528232719.4721-13-f4bug@amsat.org>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If the main loop thread invokes .ioeventfd_stop() just as the vq handler
function begins in the IOThread then the handler may lose the race for
the AioContext lock. By the time the vq handler is able to acquire the
AioContext lock the ioeventfd has already been removed and the handler
isn't supposed to run anymore!
Use the new aio_wait_bh_oneshot() function to perform ioeventfd removal
from within the IOThread. This way no races with the vq handler are
possible.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20180307144205.20619-3-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Commit 5b2ffbe4d9 ("virtio-blk: dataplane:
notify guest as a batch") deferred guest notification to a BH in order
batch notifications, with purpose of avoiding flooding the guest with
interruptions.
This optimization came with a cost. The average latency perceived in the
guest is increased by a few microseconds, but also when multiple IO
operations finish at the same time, the guest won't be notified until
all completions from each operation has been run. On the contrary,
virtio-scsi issues the notification at the end of each completion.
On the other hand, nowadays we have the EVENT_IDX feature that allows a
better coordination between QEMU and the Guest OS to avoid sending
unnecessary interruptions.
With this change, virtio-blk/dataplane only batches notifications if the
EVENT_IDX feature is not present.
Some numbers obtained with fio (ioengine=sync, iodepth=1, direct=1):
- Test specs:
* fio-3.4 (ioengine=sync, iodepth=1, direct=1)
* qemu master
* virtio-blk with a dedicated iothread (default poll-max-ns)
* backend: null_blk nr_devices=1 irqmode=2 completion_nsec=280000
* 8 vCPUs pinned to isolated physical cores
* Emulator and iothread also pinned to separate isolated cores
* variance between runs < 1%
- Not patched
* numjobs=1: lat_avg=327.32 irqs=29998
* numjobs=4: lat_avg=337.89 irqs=29073
* numjobs=8: lat_avg=342.98 irqs=28643
- Patched:
* numjobs=1: lat_avg=323.92 irqs=30262
* numjobs=4: lat_avg=332.65 irqs=29520
* numjobs=8: lat_avg=335.54 irqs=29323
Signed-off-by: Sergio Lopez <slp@redhat.com>
Message-id: 20180307114459.26636-1-slp@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The virtio_bus_set_host_notifier function no longer calls
event_notifier_cleanup when a event notifier is removed.
The commit updates the code to match the new behavior and calls
virtio_bus_cleanup_host_notifier after the notifier was de-assign
and no longer in use.
This change is a preparation to allow executing the
virtio_bus_set_host_notifier function in a memory region
transaction.
Signed-off-by: Gal Hammer <ghammer@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Tested-by: Greg Kurz <groug@kaod.org>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
When the function no success value to transmit, it usually make the
function return void. It has turned out not to be a success, because
it means that the extra local_err variable and error_propagate() will
be needed. It leads to cumbersome code, therefore, transmit success/
failure in the return value is worth.
So fix the return type of blkconf_apply_backend_options(),
blkconf_geometry() and virtio_blk_data_plane_create() to avoid it.
Cc: John Snow <jsnow@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Max Reitz <mreitz@redhat.com>
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Mao Zhongyi <maozy.fnst@cn.fujitsu.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-id: ac0edc1fc70c4457e5cec94405eb7d1f89f9c2c1.1511317952.git.maozy.fnst@cn.fujitsu.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
With the move of some docs/ to docs/devel/ on ac06724a71,
no references were updated.
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
In virtio_queue_host_notifier_aio_poll, not all "!virtio_queue_empty()"
cases are making true progress.
Currently the offending one is virtio-scsi event queue, whose handler
does nothing if no event is pending. As a result aio_poll() will spin on
the "non-empty" VQ and take 100% host CPU.
Fix this by reporting actual progress from virtio queue aio handlers.
Reported-by: Ed Swierk <eswierk@skyportsystems.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Tested-by: Ed Swierk <eswierk@skyportsystems.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
There are a number of unused trace events that
scripts/cleanup-trace-events.pl finds. The "hw/vfio/pci-quirks.c"
filename was typoed and "qapi/qapi-visit-core.c" was missing the qapi/
directory prefix.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Message-id: 20170126171613.1399-3-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The trace-events for a given source file should generally
always live in the same directory as the source file.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170125161417.31949-3-berrange@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Dataplane has been omitting forever the step of setting ISR when
an interrupt is raised. This caused little breakage, because the
specification actually says that ISR may not be updated in MSI mode.
Some versions of the Windows drivers however didn't clear MSI mode
correctly, and proceeded using polling mode (using ISR, not the used
ring index!) for crashdump and hibernation. If it were just crashdump
and hibernation it would not be a big deal, but recent releases of
Windows do not really shut down, but rather log out and hibernate to
make the next startup faster. Hence, this manifested as a more serious
hang during shutdown with e.g. Windows 8.1 and virtio-win 1.8.0 RPMs.
Newer versions fixed this, while older versions do not use MSI at all.
The failure has always been there for virtio dataplane, but it became
visible after commits 9ffe337 ("virtio-blk: always use dataplane path
if ioeventfd is active", 2016-10-30) and ad07cd6 ("virtio-scsi: always
use dataplane path if ioeventfd is active", 2016-10-30) made virtio-blk
and virtio-scsi always use the dataplane code under KVM. The good news
therefore is that it was not a bug in the patches---they were doing
exactly what they were meant for, i.e. shake out remaining dataplane bugs.
The fix is not hard, so it's worth arranging for the broken drivers.
The virtio_should_notify+event_notifier_set pair that is common to
virtio-blk and virtio-scsi dataplane is replaced with a new public
function virtio_notify_irqfd that also sets ISR. The irqfd emulation
code now need not set ISR anymore, so virtio_irq is removed.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Farhan Ali <alifm@linux.vnet.ibm.com>
Tested-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Override start_ioeventfd and stop_ioeventfd to start/stop the
whole dataplane logic. This has some positive side effects:
- no need anymore for virtio_add_queue_aio (i.e. a revert of
commit 0ff841f6d1)
- no need anymore to switch from generic ioeventfd handlers to
dataplane
It detects some errors better:
$ qemu-system-x86_64 -object iothread,id=io \
-drive id=null,file=null-aio://,if=none,format=raw \
-device virtio-blk-pci,ioeventfd=off,iothread=io,drive=null
qemu-system-x86_64: -device virtio-blk-pci,ioeventfd=off,iothread=io,drive=null:
ioeventfd is required for iothread
while previously it would have started just fine.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
This simplifies the code and removes the ioeventfd_started
and ioeventfd_set_started callback. The only difference is
in how virtio-ccw handles an error---it doesn't disable
ioeventfd forever anymore. It was the only backend to do
so, and if desired this behavior should be implemented in
virtio-bus.c.
Instead of ioeventfd_started, the ioeventfd_assign callback now
determines whether the virtio bus supports host notifiers.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
No need duplicate the judgment, there is one in function entry.
Cc: Stefan Hajnoczi <stefanha@redhat.com>
Cc: Kevin Wolf <kwolf@redhat.com>
Cc: Max Reitz <mreitz@redhat.com>
Signed-off-by: Cao jin <caoj.fnst@cn.fujitsu.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 1468814749-14510-1-git-send-email-caoj.fnst@cn.fujitsu.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Monitor ioeventfds for all virtqueues in the device's AioContext. This
is not true multiqueue because requests from all virtqueues are
processed in a single IOThread. In the future it will be possible to
use multiple IOThreads when the QEMU block layer supports multiqueue.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 1466511196-12612-7-git-send-email-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Let the virtio_blk_data_plane_notify() caller decide which virtqueue to
notify. This will allow the function to be used with multiqueue.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 1466511196-12612-4-git-send-email-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
The batch notification BH needs to know which virtqueues to notify when
multiqueue is enabled. Use a bitmap to track the virtqueues with
pending notifications.
At this point there is only one virtqueue so hard-code virtqueue index
0. A later patch will switch to real virtqueue indices.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 1466511196-12612-3-git-send-email-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
All users have been converted to the new ioevent callbacks.
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Have vhost and dataplane use the new api for transports that
have been converted.
Signed-off-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Block layer is prepared to unspecialize dataplane, an evidence is this
almost complete list of unblocked operations. It has all types except
two (actually three if DATAPLANE itself counts but blockdev.c makes sure
attaching twice is not possible): MIRROR_TARGET and BACKUP_TARGET.
blockdev-mirror refuses to start if target is attached, so the first is
not a problem.
By removing BACKUP_TARGET, blockdev-backup will become permissive to
write to a virtio-blk dataplane disk, but that is not worse than
non-dataplane given the latter is already possible. In either case,
blockdev.c always checks the target and source are on the same
AioContext, or bring them together if possible.
Signed-off-by: Fam Zheng <famz@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Message-id: 1463969978-24970-4-git-send-email-famz@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Eliminating the reentrancy is actually a nice thing that we can do
with the API that Michael proposed, so let's make it first class.
This also hides the complex assign/set_handler conventions from
callers of virtio_queue_aio_set_host_notifier_handler, which in
fact was always called with assign=true.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
In addition to handling IO in vcpu thread and in io thread, dataplane
introduces yet another mode: handling it by AioContext.
This reuses the same handler as previous modes, which triggers races as
these were not designed to be reentrant. Use a separate handler just
for aio, and disable regular handlers when dataplane is active.
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
We must not call virtio_blk_data_plane_notify if dataplane is
disabled: we would hit a segmentation fault in notify_guest_bh as
s->guest_notifier has not been setup and is NULL.
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Commit 57cb38b included qapi/error.h into qemu/osdep.h to get the
Error typedef. Since then, we've moved to include qemu/osdep.h
everywhere. Its file comment explains: "To avoid getting into
possible circular include dependencies, this file should not include
any other QEMU headers, with the exceptions of config-host.h,
compiler.h, os-posix.h and os-win32.h, all of which are doing a
similar job to this file and are under similar constraints."
qapi/error.h doesn't do a similar job, and it doesn't adhere to
similar constraints: it includes qapi-types.h. That's in excess of
100KiB of crap most .c files don't actually need.
Add the typedef to qemu/typedefs.h, and include that instead of
qapi/error.h. Include qapi/error.h in .c files that need it and don't
get it now. Include qapi-types.h in qom/object.h for uint16List.
Update scripts/clean-includes accordingly. Update it further to match
reality: replace config.h by config-target.h, add sysemu/os-posix.h,
sysemu/os-win32.h. Update the list of includes in the qemu/osdep.h
comment quoted above similarly.
This reduces the number of objects depending on qapi/error.h from "all
of them" to less than a third. Unfortunately, the number depending on
qapi-types.h shrinks only a little. More work is needed for that one.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
[Fix compilation without the spice devel packages. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
In disabled mode, virtio-blk dataplane seems to be enabled, but flow
actually goes through the normal virtio path. This patch simplifies a bit
the handling of disabled mode. In disabled mode, virtio_blk_handle_output
might be called even if s->dataplane is not NULL.
This is a bit tricky, because the current check for s->dataplane will
always trigger, causing a continuous stream of calls to
virtio_blk_data_plane_start. Unfortunately, these calls will not
do anything. To fix this, set the "started" flag even in disabled
mode, and skip virtio_blk_data_plane_start if the started flag is true.
The resulting changes also prepare the code for the next patch, were
virtio-blk dataplane will reuse the same virtio_blk_handle_output function
as "regular" virtio-blk.
Because struct VirtIOBlockDataPlane is opaque in virtio-blk.c, we have
to move s->dataplane->started inside struct VirtIOBlock.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Make the API more similar to the regular virtqueue API. This will
help when modifying the code to not use vring.c anymore.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Cornelia Huck <cornelia.huck@de.ibm.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
The return code of virtqueue_pop/vring_pop is unused except to check for
errors or 0. We can thus easily move allocation inside the functions
and just return a pointer to the VirtQueueElement.
The advantage is that we will be able to allocate only the space that
is needed for the actual size of the s/g list instead of the full
VIRTQUEUE_MAX_SIZE items. Currently VirtQueueElement takes about 48K
of memory, and this kind of allocation puts a lot of stress on malloc.
By cutting the size by two or three orders of magnitude, malloc can
use much more efficient algorithms.
The patch is pretty large, but changes to each device are testable
more or less independently. Splitting it would mostly add churn.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Cornelia Huck <cornelia.huck@de.ibm.com>
