Add some tests for various cases of named-field use, both ones that
should work and ones that should be diagnosed as errors.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230523120447.728365-7-peter.maydell@linaro.org>
Implement support for named fields, i.e. where one field is defined
in terms of another, rather than directly in terms of bits extracted
from the instruction.
The new method referenced_fields() on all the Field classes returns a
list of fields that this field references. This just passes through,
except for the new NamedField class.
We can then use referenced_fields() to:
* construct a list of 'dangling references' for a format or
pattern, which is the fields that the format/pattern uses but
doesn't define itself
* do a topological sort, so that we output "field = value"
assignments in an order that means that we assign a field before
we reference it in a subsequent assignment
* check when we output the code for a pattern whether we need to
fill in the format fields before or after the pattern fields, and
do other error checking
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230523120447.728365-6-peter.maydell@linaro.org>
To support named fields, we will need to be able to do a topological
sort (so that we ensure that we output the assignment to field A
before the assignment to field B if field B refers to field A by
name). The good news is that there is a tsort in the python standard
library; the bad news is that it was only added in Python 3.9.
To bridge the gap between our current minimum supported Python
version and 3.9, provide a local implementation that has the
same API as the stdlib version for the parts we care about.
In future when QEMU's minimum Python version requirement reaches
3.9 we can delete this code and replace it with an 'import' line.
The core of this implementation is based on
https://code.activestate.com/recipes/578272-topological-sort/
which is MIT-licensed.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Acked-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230523120447.728365-5-peter.maydell@linaro.org>
To support referring to other named fields in field definitions, we
need to pass the str_extract() method a function which tells it how
to emit the code for a previously initialized named field. (In
Pattern::output_code() the other field will be "u.f_foo.field", and
in Format::output_extract() it is "a->field".)
Refactor the two callsites that currently do "output code to
initialize each field", and have them pass a lambda that defines how
to format the lvalue in each case. This is then used both in
emitting the LHS of the assignment and also passed down to
str_extract() as a new argument (unused at the moment, but will be
used in the following patch).
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230523120447.728365-4-peter.maydell@linaro.org>
Document the named field syntax that we want to implement for the
decodetree script. This allows a field to be defined in terms of
some other field that the instruction pattern has already set, for
example:
%sz_imm 10:3 sz:3 !function=expand_sz_imm
to allow a function to be passed both an immediate field from the
instruction and also a sz value which might have been specified by
the instruction pattern directly (sz=1, etc) rather than being a
simple field within the instruction.
Note that the restriction on not having the format referring to the
pattern and the pattern referring to the format simultaneously is a
restriction of the decoder generator rather than inherently being a
silly thing to do.
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Reviewed-by: Richard Henderson <richard.henderson@linaro.org>
Message-Id: <20230523120447.728365-3-peter.maydell@linaro.org>
Nor report any PermissionError on remove.
The primary purpose is testing with -o /dev/null.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Test err_pattern_group_empty.decode failed with exception:
Traceback (most recent call last):
File "./scripts/decodetree.py", line 1424, in <module> main()
File "./scripts/decodetree.py", line 1342, in main toppat.build_tree()
File "./scripts/decodetree.py", line 627, in build_tree
self.tree = self.__build_tree(self.pats, self.fixedbits,
File "./scripts/decodetree.py", line 607, in __build_tree
fb = i.fixedbits & innermask
TypeError: unsupported operand type(s) for &: 'NoneType' and 'int'
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The last use was removed by e77c89fb08.
Fixes: e77c89fb08 ("cputlb: Remove static tlb sizing")
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use LPQ/STPQ when 16-byte atomicity is required.
Note that these instructions require 16-byte alignment.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Use LQ/STQ with ISA v2.07, and 16-byte atomicity is required.
Note that these instructions do not require 16-byte alignment.
Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With FEAT_LSE2, LDP/STP suffices. Without FEAT_LSE2, use LDXP+STXP
16-byte atomicity is required and LDP/STP otherwise.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Adjust the softmmu tlb to use TMP[0-2], not any of the normally available
registers. Since we handle overlap betwen inputs and helper arguments,
we can allow any allocatable reg.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
We will need to allocate a second general-purpose temporary.
Rename the existing temps to add a distinguishing number.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
With CPUINFO_ATOMIC_VMOVDQA, we can perform proper atomic
load/store without cmpxchg16b.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Older versions of clang have missing runtime functions for arithmetic
with -fsanitize=undefined (see 464e3671f9), so we cannot use
__int128_t for implementing Int128. But __int128_t is present,
data movement works, and it can be used for atomic128.
Probe for both CONFIG_INT128_TYPE and CONFIG_INT128, adjust
qemu/int128.h to define Int128Alias if CONFIG_INT128_TYPE,
and adjust the meson probe for atomics to use has_int128_type.
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
PAGE_WRITE is current writability, as modified by TB protection;
PAGE_WRITE_ORG is the original page writability.
Fixes: cdfac37be0 ("accel/tcg: Honor atomicity of loads")
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
The first move was incorrectly using TCG_TYPE_I32 while the second
move was correctly using TCG_TYPE_REG. This prevents a 64-bit host
from moving all 128-bits of the return value.
Fixes: ebebea53ef ("tcg: Support TCG_TYPE_I128 in tcg_out_{ld,st}_helper_{args,ret}")
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
- Fix blockdev-create with iothreads
- Remove aio_disable_external() API
-----BEGIN PGP SIGNATURE-----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=yuER
-----END PGP SIGNATURE-----
Merge tag 'for-upstream' of https://repo.or.cz/qemu/kevin into staging
Block layer patches
- Fix blockdev-create with iothreads
- Remove aio_disable_external() API
# -----BEGIN PGP SIGNATURE-----
#
# iQJFBAABCAAvFiEE3D3rFZqa+V09dFb+fwmycsiPL9YFAmR2JIARHGt3b2xmQHJl
# ZGhhdC5jb20ACgkQfwmycsiPL9brtA/9HVdAdtJxW78J60TE2lTqE9XlqMOEHBZl
# 8GN72trjP2geY/9mVsv/XoFie4ecqFsYjwAWWUuXZwLgAo53jh7oFN7gBH5iGyyD
# +EukYEfjqoykX5BkoK0gbMZZUe5Y4Dr2CNXYw4bNg8kDzj2RLifGA1XhdL3HoiVt
# PHZrhwBR7ddww6gVOnyJrfGL8fMkW/ZNeKRhrTZuSP+63oDOeGTsTumD+YKJzfPs
# p5WlwkuPjcqbO+w32FeVOHVhNI4swkN5svz3fkr8NuflfA7kH6nBQ5wymObbaTLc
# Erx03lrtP1+6nw43V11UnYt6iDMg4EBUQwtzNaKFnk3rMIdjoQYxIM5FTBWL2rYD
# Dg6PhkncXQ1WNWhUaFqpTFLB52XAYsSa4/y2QAGP6nWbqAUAUknQ3exaMvWiq7Z0
# nZeyyhIWvpJIHGCArWRdqqh+zsBdsmUVuPGyZnZgL/cXoJboYiHMyMJSUWE0XxML
# NGrncwxdsBXkVGGwTdHpBT64dcu3ENRgwtraqRLQm+tp5MKNTJB/+Ug2/p1vonHT
# UOoHz//UPskn8sHIyevoHXeu2Ns0uIHzrAXr+7Ay+9UYyIH6a07F4b2BGqkfyi/i
# 8wQsDmJ/idx5C4q1+jS+GuIbpnjIx6nxXwXMqpscUXZmM4Am8OMkiKxQAa1wExGF
# paId+HHwyks=
# =yuER
# -----END PGP SIGNATURE-----
# gpg: Signature made Tue 30 May 2023 09:29:52 AM PDT
# gpg: using RSA key DC3DEB159A9AF95D3D7456FE7F09B272C88F2FD6
# gpg: issuer "kwolf@redhat.com"
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>" [full]
* tag 'for-upstream' of https://repo.or.cz/qemu/kevin: (32 commits)
aio: remove aio_disable_external() API
virtio: do not set is_external=true on host notifiers
virtio-scsi: implement BlockDevOps->drained_begin()
virtio-blk: implement BlockDevOps->drained_begin()
virtio: make it possible to detach host notifier from any thread
block/fuse: do not set is_external=true on FUSE fd
block/export: don't require AioContext lock around blk_exp_ref/unref()
block/export: rewrite vduse-blk drain code
hw/xen: do not set is_external=true on evtchn fds
xen-block: implement BlockDevOps->drained_begin()
block: drain from main loop thread in bdrv_co_yield_to_drain()
block: add blk_in_drain() API
hw/xen: do not use aio_set_fd_handler(is_external=true) in xen_xenstore
block/export: stop using is_external in vhost-user-blk server
block/export: wait for vhost-user-blk requests when draining
util/vhost-user-server: rename refcount to in_flight counter
virtio-scsi: stop using aio_disable_external() during unplug
virtio-scsi: avoid race between unplug and transport event
hw/qdev: introduce qdev_is_realized() helper
block-backend: split blk_do_set_aio_context()
...
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
All callers now pass is_external=false to aio_set_fd_handler() and
aio_set_event_notifier(). The aio_disable_external() API that
temporarily disables fd handlers that were registered is_external=true
is therefore dead code.
Remove aio_disable_external(), aio_enable_external(), and the
is_external arguments to aio_set_fd_handler() and
aio_set_event_notifier().
The entire test-fdmon-epoll test is removed because its sole purpose was
testing aio_disable_external().
Parts of this patch were generated using the following coccinelle
(https://coccinelle.lip6.fr/) semantic patch:
@@
expression ctx, fd, is_external, io_read, io_write, io_poll, io_poll_ready, opaque;
@@
- aio_set_fd_handler(ctx, fd, is_external, io_read, io_write, io_poll, io_poll_ready, opaque)
+ aio_set_fd_handler(ctx, fd, io_read, io_write, io_poll, io_poll_ready, opaque)
@@
expression ctx, notifier, is_external, io_read, io_poll, io_poll_ready;
@@
- aio_set_event_notifier(ctx, notifier, is_external, io_read, io_poll, io_poll_ready)
+ aio_set_event_notifier(ctx, notifier, io_read, io_poll, io_poll_ready)
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-21-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Host notifiers can now use is_external=false since virtio-blk and
virtio-scsi no longer rely on is_external=true for drained sections.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-20-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The virtio-scsi Host Bus Adapter provides access to devices on a SCSI
bus. Those SCSI devices typically have a BlockBackend. When the
BlockBackend enters a drained section, the SCSI device must temporarily
stop submitting new I/O requests.
Implement this behavior by temporarily stopping virtio-scsi virtqueue
processing when one of the SCSI devices enters a drained section. The
new scsi_device_drained_begin() API allows scsi-disk to message the
virtio-scsi HBA.
scsi_device_drained_begin() uses a drain counter so that multiple SCSI
devices can have overlapping drained sections. The HBA only sees one
pair of .drained_begin/end() calls.
After this commit, virtio-scsi no longer depends on hw/virtio's
ioeventfd aio_set_event_notifier(is_external=true). This commit is a
step towards removing the aio_disable_external() API.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-19-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Detach ioeventfds during drained sections to stop I/O submission from
the guest. virtio-blk is no longer reliant on aio_disable_external()
after this patch. This will allow us to remove the
aio_disable_external() API once all other code that relies on it is
converted.
Take extra care to avoid attaching/detaching ioeventfds if the data
plane is started/stopped during a drained section. This should be rare,
but maybe the mirror block job can trigger it.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-18-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
virtio_queue_aio_detach_host_notifier() does two things:
1. It removes the fd handler from the event loop.
2. It processes the virtqueue one last time.
The first step can be peformed by any thread and without taking the
AioContext lock.
The second step may need the AioContext lock (depending on the device
implementation) and runs in the thread where request processing takes
place. virtio-blk and virtio-scsi therefore call
virtio_queue_aio_detach_host_notifier() from a BH that is scheduled in
AioContext.
The next patch will introduce a .drained_begin() function that needs to
call virtio_queue_aio_detach_host_notifier(). .drained_begin() functions
cannot call aio_poll() to wait synchronously for the BH. It is possible
for a .drained_poll() callback to asynchronously wait for the BH, but
that is more complex than necessary here.
Move the virtqueue processing out to the callers of
virtio_queue_aio_detach_host_notifier() so that the function can be
called from any thread. This is in preparation for the next patch.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-Id: <20230516190238.8401-17-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This is part of ongoing work to remove the aio_disable_external() API.
Use BlockDevOps .drained_begin/end/poll() instead of
aio_set_fd_handler(is_external=true).
As a side-effect the FUSE export now follows AioContext changes like the
other export types.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-16-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The FUSE export calls blk_exp_ref/unref() without the AioContext lock.
Instead of fixing the FUSE export, adjust blk_exp_ref/unref() so they
work without the AioContext lock. This way it's less error-prone.
Suggested-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-15-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
vduse_blk_detach_ctx() waits for in-flight requests using
AIO_WAIT_WHILE(). This is not allowed according to a comment in
bdrv_set_aio_context_commit():
/*
* Take the old AioContex when detaching it from bs.
* At this point, new_context lock is already acquired, and we are now
* also taking old_context. This is safe as long as bdrv_detach_aio_context
* does not call AIO_POLL_WHILE().
*/
Use this opportunity to rewrite the drain code in vduse-blk:
- Use the BlockExport refcount so that vduse_blk_exp_delete() is only
called when there are no more requests in flight.
- Implement .drained_poll() so in-flight request coroutines are stopped
by the time .bdrv_detach_aio_context() is called.
- Remove AIO_WAIT_WHILE() from vduse_blk_detach_ctx() to solve the
.bdrv_detach_aio_context() constraint violation. It's no longer
needed due to the previous changes.
- Always handle the VDUSE file descriptor, even in drained sections. The
VDUSE file descriptor doesn't submit I/O, so it's safe to handle it in
drained sections. This ensures that the VDUSE kernel code gets a fast
response.
- Suspend virtqueue fd handlers in .drained_begin() and resume them in
.drained_end(). This eliminates the need for the
aio_set_fd_handler(is_external=true) flag, which is being removed from
QEMU.
This is a long list but splitting it into individual commits would
probably lead to git bisect failures - the changes are all related.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-14-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
is_external=true suspends fd handlers between aio_disable_external() and
aio_enable_external(). The block layer's drain operation uses this
mechanism to prevent new I/O from sneaking in between
bdrv_drained_begin() and bdrv_drained_end().
The previous commit converted the xen-block device to use BlockDevOps
.drained_begin/end() callbacks. It no longer relies on is_external=true
so it is safe to pass is_external=false.
This is part of ongoing work to remove the aio_disable_external() API.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-13-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Detach event channels during drained sections to stop I/O submission
from the ring. xen-block is no longer reliant on aio_disable_external()
after this patch. This will allow us to remove the
aio_disable_external() API once all other code that relies on it is
converted.
Extend xen_device_set_event_channel_context() to allow ctx=NULL. The
event channel still exists but the event loop does not monitor the file
descriptor. Event channel processing can resume by calling
xen_device_set_event_channel_context() with a non-NULL ctx.
Factor out xen_device_set_event_channel_context() calls in
hw/block/dataplane/xen-block.c into attach/detach helper functions.
Incidentally, these don't require the AioContext lock because
aio_set_fd_handler() is thread-safe.
It's safer to register BlockDevOps after the dataplane instance has been
created. The BlockDevOps .drained_begin/end() callbacks depend on the
dataplane instance, so move the blk_set_dev_ops() call after
xen_block_dataplane_create().
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-12-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
For simplicity, always run BlockDevOps .drained_begin/end/poll()
callbacks in the main loop thread. This makes it easier to implement the
callbacks and avoids extra locks.
Move the function pointer declarations from the I/O Code section to the
Global State section for BlockDevOps, BdrvChildClass, and BlockDriver.
Narrow IO_OR_GS_CODE() to GLOBAL_STATE_CODE() where appropriate.
The test-bdrv-drain test case calls bdrv_drain() from an IOThread. This
is now only allowed from coroutine context, so update the test case to
run in a coroutine.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-11-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The BlockBackend quiesce_counter is greater than zero during drained
sections. Add an API to check whether the BlockBackend is in a drained
section.
The next patch will use this API.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-10-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
There is no need to suspend activity between aio_disable_external() and
aio_enable_external(), which is mainly used for the block layer's drain
operation.
This is part of ongoing work to remove the aio_disable_external() API.
Reviewed-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-9-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
vhost-user activity must be suspended during bdrv_drained_begin/end().
This prevents new requests from interfering with whatever is happening
in the drained section.
Previously this was done using aio_set_fd_handler()'s is_external
argument. In a multi-queue block layer world the aio_disable_external()
API cannot be used since multiple AioContext may be processing I/O, not
just one.
Switch to BlockDevOps->drained_begin/end() callbacks.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-8-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Each vhost-user-blk request runs in a coroutine. When the BlockBackend
enters a drained section we need to enter a quiescent state. Currently
any in-flight requests race with bdrv_drained_begin() because it is
unaware of vhost-user-blk requests.
When blk_co_preadv/pwritev()/etc returns it wakes the
bdrv_drained_begin() thread but vhost-user-blk request processing has
not yet finished. The request coroutine continues executing while the
main loop thread thinks it is in a drained section.
One example where this is unsafe is for blk_set_aio_context() where
bdrv_drained_begin() is called before .aio_context_detached() and
.aio_context_attach(). If request coroutines are still running after
bdrv_drained_begin(), then the AioContext could change underneath them
and they race with new requests processed in the new AioContext. This
could lead to virtqueue corruption, for example.
(This example is theoretical, I came across this while reading the
code and have not tried to reproduce it.)
It's easy to make bdrv_drained_begin() wait for in-flight requests: add
a .drained_poll() callback that checks the VuServer's in-flight counter.
VuServer just needs an API that returns true when there are requests in
flight. The in-flight counter needs to be atomic.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-7-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
The VuServer object has a refcount field and ref/unref APIs. The name is
confusing because it's actually an in-flight request counter instead of
a refcount.
Normally a refcount destroys the object upon reaching zero. The VuServer
counter is used to wake up the vhost-user coroutine when there are no
more requests.
Avoid confusing by renaming refcount and ref/unref to in_flight and
inc/dec.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-6-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
This patch is part of an effort to remove the aio_disable_external()
API because it does not fit in a multi-queue block layer world where
many AioContexts may be submitting requests to the same disk.
The SCSI emulation code is already in good shape to stop using
aio_disable_external(). It was only used by commit 9c5aad84da
("virtio-scsi: fixed virtio_scsi_ctx_check failed when detaching scsi
disk") to ensure that virtio_scsi_hotunplug() works while the guest
driver is submitting I/O.
Ensure virtio_scsi_hotunplug() is safe as follows:
1. qdev_simple_device_unplug_cb() -> qdev_unrealize() ->
device_set_realized() calls qatomic_set(&dev->realized, false) so
that future scsi_device_get() calls return NULL because they exclude
SCSIDevices with realized=false.
That means virtio-scsi will reject new I/O requests to this
SCSIDevice with VIRTIO_SCSI_S_BAD_TARGET even while
virtio_scsi_hotunplug() is still executing. We are protected against
new requests!
2. scsi_qdev_unrealize() already contains a call to
scsi_device_purge_requests() so that in-flight requests are cancelled
synchronously. This ensures that no in-flight requests remain once
qdev_simple_device_unplug_cb() returns.
Thanks to these two conditions we don't need aio_disable_external()
anymore.
Cc: Zhengui Li <lizhengui@huawei.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-5-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Only report a transport reset event to the guest after the SCSIDevice
has been unrealized by qdev_simple_device_unplug_cb().
qdev_simple_device_unplug_cb() sets the SCSIDevice's qdev.realized field
to false so that scsi_device_find/get() no longer see it.
scsi_target_emulate_report_luns() also needs to be updated to filter out
SCSIDevices that are unrealized.
Change virtio_scsi_push_event() to take event information as an argument
instead of the SCSIDevice. This allows virtio_scsi_hotunplug() to emit a
VIRTIO_SCSI_T_TRANSPORT_RESET event after the SCSIDevice has already
been unrealized.
These changes ensure that the guest driver does not see the SCSIDevice
that's being unplugged if it responds very quickly to the transport
reset event.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Daniil Tatianin <d-tatianin@yandex-team.ru>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-4-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Add a helper function to check whether the device is realized without
requiring the Big QEMU Lock. The next patch adds a second caller. The
goal is to avoid spreading DeviceState field accesses throughout the
code.
Suggested-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Reviewed-by: Philippe Mathieu-Daudé <philmd@linaro.org>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-3-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
blk_set_aio_context() is not fully transactional because
blk_do_set_aio_context() updates blk->ctx outside the transaction. Most
of the time this goes unnoticed but a BlockDevOps.drained_end() callback
that invokes blk_get_aio_context() fails assert(ctx == blk->ctx). This
happens because blk->ctx is only assigned after
BlockDevOps.drained_end() is called and we're in an intermediate state
where BlockDrvierState nodes already have the new context and the
BlockBackend still has the old context.
Making blk_set_aio_context() fully transactional solves this assertion
failure because the BlockBackend's context is updated as part of the
transaction (before BlockDevOps.drained_end() is called).
Split blk_do_set_aio_context() in order to solve this assertion failure.
This helper function actually serves two different purposes:
1. It drives blk_set_aio_context().
2. It responds to BdrvChildClass->change_aio_ctx().
Get rid of the helper function. Do #1 inside blk_set_aio_context() and
do #2 inside blk_root_set_aio_ctx_commit(). This simplifies the code.
The only drawback of the fully transactional approach is that
blk_set_aio_context() must contend with blk_root_set_aio_ctx_commit()
being invoked as part of the AioContext change propagation. This can be
solved by temporarily setting blk->allow_aio_context_change to true.
Future patches call blk_get_aio_context() from
BlockDevOps->drained_end(), so this patch will become necessary.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230516190238.8401-2-stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
If blockdev-create references an existing node in an iothread (e.g. as
it's 'file' child), then suddenly all of the image creation code must
run in that AioContext, too. Test that this actually works.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-13-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
It has no internal callers, so its only use is being called from
individual test cases. If the name starts with an underscore, it is
considered private and linters warn against calling it. 256 only gets
away with it currently because it's on the exception list for linters.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-Id: <20230525124713.401149-12-kwolf@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
