Requests that complete in an IOThread use irqfd to notify the guest
while requests that complete in the main loop thread use the traditional
qdev irq code path. The reason for this conditional is that the irq code
path requires the BQL:
if (s->ioeventfd_started && !s->ioeventfd_disabled) {
virtio_notify_irqfd(vdev, req->vq);
} else {
virtio_notify(vdev, req->vq);
}
There is a corner case where the conditional invokes the irq code path
instead of the irqfd code path:
static void virtio_blk_stop_ioeventfd(VirtIODevice *vdev)
{
...
/*
* Set ->ioeventfd_started to false before draining so that host notifiers
* are not detached/attached anymore.
*/
s->ioeventfd_started = false;
/* Wait for virtio_blk_dma_restart_bh() and in flight I/O to complete */
blk_drain(s->conf.conf.blk);
During blk_drain() the conditional produces the wrong result because
ioeventfd_started is false.
Use qemu_in_iothread() instead of checking the ioeventfd state.
Cc: qemu-stable@nongnu.org
Buglink: https://issues.redhat.com/browse/RHEL-15394
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240122172625.415386-1-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Commit d3f6f294ae ("virtio-blk: always set
ioeventfd during startup") has made virtio_blk_start_ioeventfd() always
kick the virtqueue (set the ioeventfd), regardless of whether the BB is
drained. That is no longer necessary, because attaching the host
notifier will now set the ioeventfd, too; this happens either
immediately right here in virtio_blk_start_ioeventfd(), or later when
the drain ends, in virtio_blk_ioeventfd_attach().
With event_notifier_set() removed, the code becomes the same as the one
in virtio_blk_ioeventfd_attach(), so we can reuse that function.
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20240202153158.788922-4-hreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
During drain, we do not care about virtqueue notifications, which is why
we remove the handlers on it. When removing those handlers, whether vq
notifications are enabled or not depends on whether we were in polling
mode or not; if not, they are enabled (by default); if so, they have
been disabled by the io_poll_start callback.
Because we do not care about those notifications after removing the
handlers, this is fine. However, we have to explicitly ensure they are
enabled when re-attaching the handlers, so we will resume receiving
notifications. We do this in virtio_queue_aio_attach_host_notifier*().
If such a function is called while we are in a polling section,
attaching the notifiers will then invoke the io_poll_start callback,
re-disabling notifications.
Because we will always miss virtqueue updates in the drained section, we
also need to poll the virtqueue once after attaching the notifiers.
Buglink: https://issues.redhat.com/browse/RHEL-3934
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20240202153158.788922-3-hreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
As of commit 38738f7dbb ("virtio-scsi:
don't waste CPU polling the event virtqueue"), we only attach an io_read
notifier for the virtio-scsi event virtqueue instead, and no polling
notifiers. During operation, the event virtqueue is typically
non-empty, but none of the buffers are intended to be used immediately.
Instead, they only get used when certain events occur. Therefore, it
makes no sense to continuously poll it when non-empty, because it is
supposed to be and stay non-empty.
We do this by using virtio_queue_aio_attach_host_notifier_no_poll()
instead of virtio_queue_aio_attach_host_notifier() for the event
virtqueue.
Commit 766aa2de0f ("virtio-scsi: implement
BlockDevOps->drained_begin()") however has virtio_scsi_drained_end() use
virtio_queue_aio_attach_host_notifier() for all virtqueues, including
the event virtqueue. This can lead to it being polled again, undoing
the benefit of commit 38738f7dbb.
Fix it by using virtio_queue_aio_attach_host_notifier_no_poll() for the
event virtqueue.
Reported-by: Fiona Ebner <f.ebner@proxmox.com>
Fixes: 766aa2de0f
("virtio-scsi: implement BlockDevOps->drained_begin()")
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Tested-by: Fiona Ebner <f.ebner@proxmox.com>
Reviewed-by: Fiona Ebner <f.ebner@proxmox.com>
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20240202153158.788922-2-hreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
usb-storage is for the most part just a wrapper around an internally
created scsi-disk device. It uses DEFINE_BLOCK_PROPERTIES() to offer all
of the usual block device properties to the user, but then only forwards
a few select properties to the internal device while the rest is
silently ignored.
This changes scsi_bus_legacy_add_drive() to accept a whole BlockConf
instead of some individual values inside of it so that usb-storage can
now pass the whole configuration to the internal scsi-disk. This enables
the remaining block device properties, e.g. logical/physical_block_size
or discard_granularity.
Buglink: https://issues.redhat.com/browse/RHEL-22375
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Message-ID: <20240131130607.24117-1-kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
QEMU's coding style generally forbids C99 mixed declarations.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240206140410.65650-1-stefanha@redhat.com>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
scsi_device_for_each_req_async() currently does not provide any way to
be awaited. One of its callers is scsi_device_purge_requests(), which
therefore currently does not guarantee that all requests are fully
settled when it returns.
We want all requests to be settled, because scsi_device_purge_requests()
is called through the unrealize path, including the one invoked by
virtio_scsi_hotunplug() through qdev_simple_device_unplug_cb(), which
most likely assumes that all SCSI requests are done then.
In fact, scsi_device_purge_requests() already contains a blk_drain(),
but this will not fully await scsi_device_for_each_req_async(), only the
I/O requests it potentially cancels (not the non-I/O requests).
However, we can have scsi_device_for_each_req_async() increment the BB
in-flight counter, and have scsi_device_for_each_req_async_bh()
decrement it when it is done. This way, the blk_drain() will fully
await all SCSI requests to be purged.
This also removes the need for scsi_device_for_each_req_async_bh() to
double-check the current context and potentially re-schedule itself,
should it now differ from the BB's context: Changing a BB's AioContext
with a root node is done through bdrv_try_change_aio_context(), which
creates a drained section. With this patch, we keep the BB in-flight
counter elevated throughout, so we know the BB's context cannot change.
Signed-off-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20240202144755.671354-3-hreitz@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Hanna Czenczek <hreitz@redhat.com> noted that the array index in
virtio_blk_dma_restart_cb() is not bounds-checked:
g_autofree VirtIOBlockReq **vq_rq = g_new0(VirtIOBlockReq *, num_queues);
...
while (rq) {
VirtIOBlockReq *next = rq->next;
uint16_t idx = virtio_get_queue_index(rq->vq);
rq->next = vq_rq[idx];
^^^^^^^^^^
The code is correct because both rq->vq and vq_rq[] depend on
num_queues, but this is indirect and not 100% obvious. Add an assertion.
Suggested-by: Hanna Czenczek <hreitz@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240206190610.107963-4-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
It is not possible to instantiate a virtio-blk device with 0 virtqueues.
The following check is located in ->realize():
if (!conf->num_queues) {
error_setg(errp, "num-queues property must be larger than 0");
return;
}
Later on we access s->vq_aio_context[0] under the assumption that there
is as least one virtqueue. Hanna Czenczek <hreitz@redhat.com> noted that
it would help to show that the array index is already valid.
Add an assertion to document that s->vq_aio_context[0] is always
safe...and catch future code changes that break this assumption.
Suggested-by: Hanna Czenczek <hreitz@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240206190610.107963-3-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
Hanna Czenczek <hreitz@redhat.com> noticed that the safety of
`vq_aio_context[vq->value] = ctx;` with user-defined vq->value inputs is
not obvious.
The code is structured in validate() + apply() steps so input validation
is there, but it happens way earlier and there is nothing that
guarantees apply() can only be called with validated inputs.
This patch moves the validate() call inside the apply() function so
validation is guaranteed. I also added the bounds checking assertion
that Hanna suggested.
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Manos Pitsidianakis <manos.pitsidianakis@linaro.org>
Reviewed-by: Hanna Czenczek <hreitz@redhat.com>
Message-ID: <20240206190610.107963-2-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
QEMU initializes preallocated backend memory as the objects are parsed from
the command line. This is not optimal in some cases (e.g. memory spanning
multiple NUMA nodes) because the memory objects are initialized in series.
Allow the initialization to occur in parallel (asynchronously). In order to
ensure optimal thread placement, asynchronous initialization requires prealloc
context threads to be in use.
Signed-off-by: Mark Kanda <mark.kanda@oracle.com>
Message-ID: <20240131165327.3154970-2-mark.kanda@oracle.com>
Tested-by: Mario Casquero <mcasquer@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
We used to check that the memory region size is multiples of the overall
requested address alignment for the device memory address.
We removed that check, because there are cases (i.e., hv-balloon) where
devices unconditionally request an address alignment that has a very large
alignment (i.e., 32 GiB), but the actual memory device size might not be
multiples of that alignment.
However, this change:
(a) allows for some practically impossible DIMM sizes, like "1GB+1 byte".
(b) allows for DIMMs that partially cover hugetlb pages, previously
reported in [1].
Both scenarios don't make any sense: we might even waste memory.
So let's reintroduce that check, but only check that the
memory region size is multiples of the memory region alignment (i.e.,
page size, huge page size), but not any additional memory device
requirements communicated using md->get_min_alignment().
The following examples now fail again as expected:
(a) 1M with 2M THP
qemu-system-x86_64 -m 4g,maxmem=16g,slots=1 -S -nodefaults -nographic \
-object memory-backend-ram,id=mem1,size=1M \
-device pc-dimm,id=dimm1,memdev=mem1
-> backend memory size must be multiple of 0x200000
(b) 1G+1byte
qemu-system-x86_64 -m 4g,maxmem=16g,slots=1 -S -nodefaults -nographic \
-object memory-backend-ram,id=mem1,size=1073741825B \
-device pc-dimm,id=dimm1,memdev=mem1
-> backend memory size must be multiple of 0x200000
(c) Unliagned hugetlb size (2M)
qemu-system-x86_64 -m 4g,maxmem=16g,slots=1 -S -nodefaults -nographic \
-object memory-backend-file,id=mem1,mem-path=/dev/hugepages/tmp,size=511M \
-device pc-dimm,id=dimm1,memdev=mem1
backend memory size must be multiple of 0x200000
(d) Unliagned hugetlb size (1G)
qemu-system-x86_64 -m 4g,maxmem=16g,slots=1 -S -nodefaults -nographic \
-object memory-backend-file,id=mem1,mem-path=/dev/hugepages1G/tmp,size=2047M \
-device pc-dimm,id=dimm1,memdev=mem1
-> backend memory size must be multiple of 0x40000000
Note that this fix depends on a hv-balloon change to communicate its
additional alignment requirements using get_min_alignment() instead of
through the memory region.
[1] https://lkml.kernel.org/r/f77d641d500324525ac036fe1827b3070de75fc1.1701088320.git.mprivozn@redhat.com
Message-ID: <20240117135554.787344-3-david@redhat.com>
Reported-by: Zhenyu Zhang <zhenyzha@redhat.com>
Reported-by: Michal Privoznik <mprivozn@redhat.com>
Fixes: eb1b7c4bd4 ("memory-device: Drop size alignment check")
Tested-by: Zhenyu Zhang <zhenyzha@redhat.com>
Tested-by: Mario Casquero <mcasquer@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
The character "+" is now forbidden in QOM device names (see commit
b447378e12 - "Limit type names to alphanumerical and some few special
characters"). For the "power5+" and "power7+" CPU names, there is
currently a hack in type_name_is_valid() to still allow them for
compatibility reasons. However, there is a much nicer solution for this:
Simply use aliases! This way we can still support the old names without
the need for the ugly hack in type_name_is_valid().
Message-ID: <20240117141054.73841-2-thuth@redhat.com>
Reviewed-by: Cédric Le Goater <clg@kaod.org>
Reviewed-by: Harsh Prateek Bora <harshpb@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
When the maximum count of SCRIPTS instructions is reached, the code
stops execution and returns, but fails to decrement the reentrancy
counter. This effectively renders the SCSI controller unusable
because on next entry the reentrancy counter is still above the limit.
This bug was seen on HP-UX 10.20 which seems to trigger SCRIPTS
loops.
Fixes: b987718bbb ("hw/scsi/lsi53c895a: Fix reentrancy issues in the LSI controller (CVE-2023-0330)")
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Message-ID: <20240128202214.2644768-1-svens@stackframe.org>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Tested-by: Helge Deller <deller@gmx.de>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Let's implement the get_min_alignment() callback for memory devices, and
copy for the device memory region the alignment of the host memory
region. This mimics what virtio-mem does, and allows for re-introducing
proper alignment checks for the memory region size (where we don't care
about additional device requirements) in memory device core.
Message-ID: <20240117135554.787344-2-david@redhat.com>
Reviewed-by: Maciej S. Szmigiero <maciej.szmigiero@oracle.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
This function is no longer used, as all its callers have been converted
to use pci_init_nic_devices() instead.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Obtain the MAC address from the NIC configuration if there is one, or
generate one explicitly so that it can be placed in the PROM.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Create the device only if there is a corresponding NIC config for it.
Remove the explicit check on nd_table[0].used from hw/hppa/machine.c
which (since commit d8a3220005) tries to do the same thing.
The lasi_82596 support has been disabled since it was first introduced,
since enable_lasi_lan() has always been zero. This allows the user to
enable it by explicitly requesting a NIC model 'lasi_82596' or just
using the alias 'lasi'. Otherwise, it defaults to a PCI NIC as before.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
When converting to the shiny build-system-du-jour, a typo prevented the
last_i82596 driver from being built. Correct the config option name to
re-enable the build. And include "sysemu/sysemu.h" so it actually builds.
Fixes: b1419fa665 ("meson: convert hw/net")
Resolves: https://gitlab.com/qemu-project/qemu/-/issues/2144
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
The MIPS SIM platform instantiates its NIC only if a corresponding
configuration exists for it. Use qemu_create_nic_device() function for
that.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
If a corresponding NIC configuration was found, it will have a MAC address
already assigned, so use that. Else, generate and assign a default one.
Using qemu_find_nic_info() is simpler than the alternative of using
qemu_configure_nic_device() and then having to fetch the "mac" property
as a string and convert it.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Rather than just using qemu_configure_nic_device(), populate the MAC
address in the system-registers device by peeking at the NICInfo before
it's assigned to the device.
Generate the MAC address early, if there is no matching -nic option.
Otherwise the MAC address wouldn't be generated until net_client_init1()
runs.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Also update the test to specify which device to attach the test socket
to, and remove the comment lamenting the fact that we can't do so.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Some callers instantiate the device unconditionally, others will do so only
if there is a NICInfo to go with it. This appears to be fairly random, but
preseve the existing behaviour for now.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Some callers instantiate the device unconditionally, others will do so only
if there is a NICInfo to go with it. This appears to be fairly random, but
preserve the existing behaviour of each caller for now.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
The first sunhme NIC gets placed a function 1 on slot 1 of PCI bus A,
and the rest are dynamically assigned on PCI bus B.
Previously, any PCI NIC would get the special treatment purely by
virtue of being first in the list.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Previously, the first PCI NIC would be assigned to slot 2 even if the
user override the model and made it something other than an rtl8139
which is the default. Everything else would be dynamically assigned.
Now, the first rtl8139 gets slot 2 and everything else is dynamic.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Yoshinori Sato <ysato@users.sourceforge.jp>
Avoid directly referencing nd_table[] by first instantiating any
spapr-vlan devices using a qemu_get_nic_info() loop, then calling
pci_init_nic_devices() to do the rest.
No functional change intended.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Previously, the first PCI NIC would be placed in PCI slot 3 and the rest
would be dynamically assigned. Even if the user overrode the default NIC
type and made it something other than PCNet.
Now, the first PCNet NIC (that is, anything not explicitly specified
to be anything different) will go to slot 3 even if it isn't the first
NIC specified on the command line. And anything else will be dynamically
assigned.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
The Malta board setup code would previously place the first NIC into PCI
slot 11 if was a PCNet card, and the rest (including the first if it was
anything other than a PCNet card) would be dynamically assigned.
Now it will place any PCNet NIC into slot 11, and then anything else will
be dynamically assigned.
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
Reviewed-by: Thomas Huth <thuth@redhat.com>
