| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Introduce check function for setting read_only flags. Will return < 0 on
error, with appropriate Error value set. Does not alter any flags.
Signed-off-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: e2bba34ac3bc76a0c42adc390413f358ae0566e8.1491597120.git.jcody@redhat.com
|
|
A few block drivers will set the BDS read_only flag from their
.bdrv_open() function. This means the bs->read_only flag could
be set after we enable copy_on_read, as the BDRV_O_COPY_ON_READ
flag check occurs prior to the call to bdrv->bdrv_open().
This adds an error return to bdrv_set_read_only(), and an error will be
return if we try to set the BDS to read_only while copy_on_read is
enabled.
This patch also changes the behavior of vvfat. Before, vvfat could
override the drive 'readonly' flag with its own, internal 'rw' flag.
For instance, this -drive parameter would result in a writable image:
"-drive format=vvfat,dir=/tmp/vvfat,rw,if=virtio,readonly=on"
This is not correct. Now, attempting to use the above -drive parameter
will result in an error (i.e., 'rw' is incompatible with 'readonly=on').
Signed-off-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 0c5b4c1cc2c651471b131f21376dfd5ea24d2196.1491597120.git.jcody@redhat.com
|
|
We have a helper wrapper for checking for the BDS read_only flag,
add a helper wrapper to set the read_only flag as well.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Jeff Cody <jcody@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Message-id: 9b18972d05f5fa2ac16c014f0af98d680553048d.1491597120.git.jcody@redhat.com
|
|
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-Id: <20170421122710.15373-9-famz@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
|
|
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-Id: <20170421122710.15373-6-famz@redhat.com>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
|
|
During block job completion, nothing is preventing
block_job_defer_to_main_loop_bh from being called in a nested
aio_poll(), which is a trouble, such as in this code path:
qmp_block_commit
commit_active_start
bdrv_reopen
bdrv_reopen_multiple
bdrv_reopen_prepare
bdrv_flush
aio_poll
aio_bh_poll
aio_bh_call
block_job_defer_to_main_loop_bh
stream_complete
bdrv_reopen
block_job_defer_to_main_loop_bh is the last step of the stream job,
which should have been "paused" by the bdrv_drained_begin/end in
bdrv_reopen_multiple, but it is not done because it's in the form of a
main loop BH.
Similar to why block jobs should be paused between drained_begin and
drained_end, BHs they schedule must be excluded as well. To achieve
this, this patch forces draining the BH in BDRV_POLL_WHILE.
As a side effect this fixes a hang in block_job_detach_aio_context
during system_reset when a block job is ready:
#0 0x0000555555aa79f3 in bdrv_drain_recurse
#1 0x0000555555aa825d in bdrv_drained_begin
#2 0x0000555555aa8449 in bdrv_drain
#3 0x0000555555a9c356 in blk_drain
#4 0x0000555555aa3cfd in mirror_drain
#5 0x0000555555a66e11 in block_job_detach_aio_context
#6 0x0000555555a62f4d in bdrv_detach_aio_context
#7 0x0000555555a63116 in bdrv_set_aio_context
#8 0x0000555555a9d326 in blk_set_aio_context
#9 0x00005555557e38da in virtio_blk_data_plane_stop
#10 0x00005555559f9d5f in virtio_bus_stop_ioeventfd
#11 0x00005555559fa49b in virtio_bus_stop_ioeventfd
#12 0x00005555559f6a18 in virtio_pci_stop_ioeventfd
#13 0x00005555559f6a18 in virtio_pci_reset
#14 0x00005555559139a9 in qdev_reset_one
#15 0x0000555555916738 in qbus_walk_children
#16 0x0000555555913318 in qdev_walk_children
#17 0x0000555555916738 in qbus_walk_children
#18 0x00005555559168ca in qemu_devices_reset
#19 0x000055555581fcbb in pc_machine_reset
#20 0x00005555558a4d96 in qemu_system_reset
#21 0x000055555577157a in main_loop_should_exit
#22 0x000055555577157a in main_loop
#23 0x000055555577157a in main
The rationale is that the loop in block_job_detach_aio_context cannot
make any progress in pausing/completing the job, because bs->in_flight
is 0, so bdrv_drain doesn't process the block_job_defer_to_main_loop
BH. With this patch, it does.
Reported-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
Message-Id: <20170418143044.12187-3-famz@redhat.com>
Reviewed-by: Jeff Cody <jcody@redhat.com>
Tested-by: Jeff Cody <jcody@redhat.com>
Signed-off-by: Fam Zheng <famz@redhat.com>
|
|
Signed-off-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
|
|
They start the coroutine on the specified context.
Signed-off-by: Fam Zheng <famz@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
|
|
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
|
|
Usually guest devices don't like other writers to the same image, so
they use blk_set_perm() to prevent this from happening. In the migration
phase before the VM is actually running, though, they don't have a
problem with writes to the image. On the other hand, storage migration
needs to be able to write to the image in this phase, so the restrictive
blk_set_perm() call of qdev devices breaks it.
This patch flags all BlockBackends with a qdev device as
blk->disable_perm during incoming migration, which means that the
requested permissions are stored in the BlockBackend, but not actually
applied to its root node yet.
Once migration has finished and the VM should be resumed, the
permissions are applied. If they cannot be applied (e.g. because the NBD
server used for block migration hasn't been shut down), resuming the VM
fails.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Tested-by: Kashyap Chamarthy <kchamart@redhat.com>
|
|
bdrv_child_set_perm alone is not very usable because the caller must
call bdrv_child_check_perm first. This is already encapsulated
conveniently in bdrv_child_try_set_perm, so remove the other prototypes
from the header and fix the one wrong caller, block/mirror.c.
Signed-off-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
When adding an Error parameter, bdrv_replace_in_backing_chain() would
become nothing more than a wrapper around change_parent_backing_link().
So make the latter public, renamed as bdrv_replace_node(), and remove
bdrv_replace_in_backing_chain().
Most of the callers just remove a node from the graph that they just
inserted, so they can use &error_abort, but completion of a mirror job
with 'replaces' set can actually fail.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
|
|
change_parent_backing_link() will need to update multiple BdrvChild
objects at once. Checking permissions reference by reference doesn't
work because permissions need to be consistent only with all parents
moved to the new child.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
|
|
Aborting on error in bdrv_append() isn't correct. This patch fixes it
and lets the callers handle failures.
Test case 085 needs a reference output update. This is caused by the
reversed order of bdrv_set_backing_hd() and change_parent_backing_link()
in bdrv_append(): When the backing file of the new node is set, the
parent nodes are still pointing to the old top, so the backing blocker
is now initialised with the node name rather than the BlockBackend name.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Not all callers of bdrv_set_backing_hd() know for sure that attaching
the backing file will be allowed by the permission system. Return the
error from the function rather than aborting.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Management tools need to be able to know about every node in the graph
and need a way to address them. Changing the graph structure was okay
because libvirt doesn't really manage the node level yet, but future
libvirt versions need to deal with both new and old version of qemu.
This new option to blockdev-commit allows the client to set a node-name
for the automatically inserted filter driver, and at the same time
serves as a witness for a future libvirt that this version of qemu does
automatically insert a filter driver.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Management tools need to be able to know about every node in the graph
and need a way to address them. Changing the graph structure was okay
because libvirt doesn't really manage the node level yet, but future
libvirt versions need to deal with both new and old version of qemu.
This new option to blockdev-mirror allows the client to set a node-name
for the automatically inserted filter driver, and at the same time
serves as a witness for a future libvirt that this version of qemu does
automatically insert a filter driver.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
In some cases, we want to remove op blockers on intermediate nodes
before the whole block job transaction has completed (because they block
restoring the final graph state during completion). Provide a function
for this.
The whole block job lifecycle is a bit messed up and it's hard to
actually do all things in the right order, but I'll leave simplifying
this for another day.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Backing files are somewhat special compared to other kinds of children
because they are attached and detached using bdrv_set_backing_hd()
rather than the normal set of functions, which does a few more things
like setting backing blockers, toggling the BDRV_O_NO_BACKING flag,
setting parent_bs->backing_file, etc.
These special features are a reason why change_parent_backing_link()
can't handle backing files yet. With abstracting the additional features
into .attach/.detach callbacks, we get a step closer to a function that
can actually deal with this.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
Block jobs don't actually do I/O through the the reference they create
with block_job_add_bdrv(), but they might want to use the permisssion
system to express what the block job does to intermediate nodes. This
adds permissions to block_job_add_bdrv() to provide the means to request
permissions.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
When the parents' child links are updated in bdrv_append() or
bdrv_replace_in_backing_chain(), this should affect all child links of
BlockBackends or other nodes, but not on child links held for other
purposes (like for setting permissions). This patch allows to control
the behaviour per BdrvChildRole.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
For meaningful error messages in the permission system, we need to get
some human-readable description of the parent of a BdrvChild.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
This functions creates a BlockBackend internally, so the block jobs need
to tell it what they want to do with the BB.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
Some devices allow a media change between read-only and read-write
media. They need to adapt the permissions in their .change_media_cb()
implementation, which can fail. So add an Error parameter to the
function.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
blk_new_open() is a convenience function that processes flags rather
than QDict options as a simple way to just open an image file.
In order to keep it convenient in the future, it must automatically
request the necessary permissions. This can easily be inferred from the
flags for read and write, but we need another flag that tells us whether
to get the resize permission.
We can't just always request it because that means that no block jobs
can run on the resulting BlockBackend (which is something that e.g.
qemu-img commit wants to do), but we also can't request it never because
most of the .bdrv_create() implementations call blk_truncate().
The solution is to introduce another flag that is passed by all users
that want to resize the image.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
vvfat is the last remaining driver that can have children, but doesn't
implement .bdrv_child_perm() yet. The default handlers aren't suitable
here, so let's implement a very simple driver-specific one that protects
the internal child from being used by other users as good as our
permissions permit.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
Almost all format drivers have the same characteristics as far as
permissions are concerned: They have one or more children for storing
their own data and, more importantly, metadata (can be written to and
grow even without external write requests, must be protected against
other writers and present consistent data) and optionally a backing file
(this is just data, so like for a filter, it only depends on what the
parent nodes need).
This provides a default implementation that can be shared by most of
our format drivers.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Most filters need permissions related to read and write for their
children, but only if the node has a parent that wants to use the same
operation on the filter. The same is true for resize.
This adds a default implementation that simply forwards all necessary
permissions to all children of the node and leaves the other permissions
unchanged.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
In many cases, the required permissions of one node on its children
depend on what its parents require from it. For example, the raw format
or most filter drivers only need to request consistent reads if that's
something that one of their parents wants.
In order to achieve this, this patch introduces two new BlockDriver
callbacks. The first one lets drivers first check (recursively) whether
the requested permissions can be set; the second one actually sets the
new permission bitmask.
Also add helper functions that drivers can use in their implementation
of the callbacks to update their permissions on a specific child.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
When attaching a node as a child to a new parent, the required and
shared permissions for this parent are checked against all other parents
of the node now, and an error is returned if there is a conflict.
This allows error returns to a function that previously always
succeeded, and the same is true for quite a few callers and their
callers. Converting all of them within the same patch would be too much,
so for now everyone tells that they don't need any permissions and allow
everyone else to do anything. This way we can use &error_abort initially
and convert caller by caller to pass actual permission requirements and
implement error handling.
All these places are marked with FIXME comments and it will be the job
of the next patches to clean them up again.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
It will have to return an error soon, so prepare the callers for it.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
This patch defines the permission categories that will be used by the
new op blocker system.
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
Acked-by: Fam Zheng <famz@redhat.com>
|
|
This function allows to create more or less normal BlockDriverStates
even for BlockDrivers that aren't globally registered (e.g. helper
filters for block jobs).
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Max Reitz <mreitz@redhat.com>
|
|
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-19-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This patch prepares for the removal of unnecessary lockcnt inc/dec pairs.
Extract the dispatching loop for file descriptor handlers into a new
function aio_dispatch_handlers, and then inline aio_dispatch into
aio_poll.
aio_dispatch can now become void.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Reviewed-by: Daniel P. Berrange <berrange@redhat.com>
Message-id: 20170213135235.12274-17-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
aio_co_wake provides the infrastructure to start a coroutine on a "home"
AioContext. It will be used by CoMutex and CoQueue, so that coroutines
don't jump from one context to another when they go to sleep on a
mutex or waitqueue. However, it can also be used as a more efficient
alternative to one-shot bottom halves, and saves the effort of tracking
which AioContext a coroutine is running on.
aio_co_schedule is the part of aio_co_wake that starts a coroutine
on a remove AioContext, but it is also useful to implement e.g.
bdrv_set_aio_context callbacks.
The implementation of aio_co_schedule is based on a lock-free
multiple-producer, single-consumer queue. The multiple producers use
cmpxchg to add to a LIFO stack. The consumer (a per-AioContext bottom
half) grabs all items added so far, inverts the list to make it FIFO,
and goes through it one item at a time until it's empty. The data
structure was inspired by OSv, which uses it in the very code we'll
"port" to QEMU for the thread-safe CoMutex.
Most of the new code is really tests.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170213135235.12274-3-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Commit 6f6071745bd0 ("raw-posix: Fetch max sectors for host block device")
introduced a routine to call the kernel BLKSECTGET ioctl, which stores the
result back to user space. However, the size of the data returned depends
on the routine handling the ioctl. The (compat_)blkdev_ioctl returns a
short, while sg_ioctl returns an int. Thus, on big-endian systems, we can
find ourselves accidentally shifting the result to a much larger value.
(On s390x, a short is 16 bits while an int is 32 bits.)
Also, the two ioctl handlers return values in different scales (block
returns sectors, while sg returns bytes), so some tweaking of the outputs
is required such that hdev_get_max_transfer_length returns a value in a
consistent set of units.
Signed-off-by: Eric Farman <farman@linux.vnet.ibm.com>
Message-Id: <20170120162527.66075-3-farman@linux.vnet.ibm.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
|
|
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170112180800.21085-10-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This will make it possible to walk the list of bottom halves without
holding the AioContext lock---and in turn to call bottom half
handlers without holding the lock.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170112180800.21085-4-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
This will be used for AioHandlers too. There is going to be little
or no contention, so it is better to reuse the same lock.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Fam Zheng <famz@redhat.com>
Message-id: 20170112180800.21085-2-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
bdrv_io_plug and bdrv_io_unplug are only called (via their
BlockBackend equivalents) after starting asynchronous I/O.
bdrv_drain is not going to be called while they are running,
because---even if a coroutine runs for some reason---it will
only drain in the next iteration of the event loop through
bdrv_co_yield_to_drain.
So this mechanism is unnecessary, get rid of it.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161129113334.605-1-pbonzini@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
These files deal with the file protocol, not the raw format (the
file protocol is often used with other formats, and the raw
format is not forced to use the file protocol). Rename things
to make it a bit easier to follow.
Suggested-by: Daniel P. Berrange <berrange@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
Reviewed-by: John Snow <jsnow@redhat.com>
Reviewed-by: Laszlo Ersek <lersek@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
|
|
This patch is based on the algorithm for the kvm.ko halt_poll_ns
parameter in Linux. The initial polling time is zero.
If the event loop is woken up within the maximum polling time it means
polling could be effective, so grow polling time.
If the event loop is woken up beyond the maximum polling time it means
polling is not effective, so shrink polling time.
If the event loop makes progress within the current polling time then
the sweet spot has been reached.
This algorithm adjusts the polling time so it can adapt to variations in
workloads. The goal is to reach the sweet spot while also recognizing
when polling would hurt more than help.
Two new trace events, poll_grow and poll_shrink, are added for observing
polling time adjustment.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161201192652.9509-13-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
The begin and end callbacks can be used to prepare for the polling loop
and clean up when polling stops. Note that they may only be called once
for multiple aio_poll() calls if polling continues to succeed. Once
polling fails the end callback is invoked before aio_poll() resumes file
descriptor monitoring.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161201192652.9509-11-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
The AioContext event loop uses ppoll(2) or epoll_wait(2) to monitor file
descriptors or until a timer expires. In cases like virtqueues, Linux
AIO, and ThreadPool it is technically possible to wait for events via
polling (i.e. continuously checking for events without blocking).
Polling can be faster than blocking syscalls because file descriptors,
the process scheduler, and system calls are bypassed.
The main disadvantage to polling is that it increases CPU utilization.
In classic polling configuration a full host CPU thread might run at
100% to respond to events as quickly as possible. This patch implements
a timeout so we fall back to blocking syscalls if polling detects no
activity. After the timeout no CPU cycles are wasted on polling until
the next event loop iteration.
The run_poll_handlers_begin() and run_poll_handlers_end() trace events
are added to aid performance analysis and troubleshooting. If you need
to know whether polling mode is being used, trace these events to find
out.
Note that the AioContext is now re-acquired before disabling notify_me
in the non-polling case. This makes the code cleaner since notify_me
was enabled outside the non-polling AioContext release region. This
change is correct since it's safe to keep notify_me enabled longer
(disabling is an optimization) but potentially causes unnecessary
event_notifer_set() calls. I think the chance of performance regression
is small here.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161201192652.9509-4-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
The new AioPollFn io_poll() argument to aio_set_fd_handler() and
aio_set_event_handler() is used in the next patch.
Keep this code change separate due to the number of files it touches.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161201192652.9509-3-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Polling mode will not call ppoll(2)/epoll_wait(2). Therefore we know
there are no fds ready and should avoid looping over fd handlers in
aio_dispatch().
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-id: 20161201192652.9509-2-stefanha@redhat.com
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
|
|
Commit 87f68d318222563822b5c6b28192215fc4b4e441 (block: drop aio
functions that operate on the main AioContext) drops qemu_aio_wait
function references mostly while leaves these behind, clean up them.
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Message-Id: <1480566640-27264-3-git-send-email-baiyaowei@cmss.chinamobile.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
|
|
Refactor backup_start as backup_job_create, which only creates the job,
but does not automatically start it. The old interface, 'backup_start',
is not kept in favor of limiting the number of nearly-identical interfaces
that would have to be edited to keep up with QAPI changes in the future.
Callers that wish to synchronously start the backup_block_job can
instead just call block_job_start immediately after calling
backup_job_create.
Transactions are updated to use the new interface, calling block_job_start
only during the .commit phase, which helps prevent race conditions where
jobs may finish before we even finish building the transaction. This may
happen, for instance, during empty block backup jobs.
Reported-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Signed-off-by: John Snow <jsnow@redhat.com>
Message-id: 1478587839-9834-6-git-send-email-jsnow@redhat.com
Signed-off-by: Jeff Cody <jcody@redhat.com>
|
