qemu-e2k/hw/char/virtio-console.c
Markus Armbruster b69c3c21a5 qdev: Unrealize must not fail
Devices may have component devices and buses.

Device realization may fail.  Realization is recursive: a device's
realize() method realizes its components, and device_set_realized()
realizes its buses (which should in turn realize the devices on that
bus, except bus_set_realized() doesn't implement that, yet).

When realization of a component or bus fails, we need to roll back:
unrealize everything we realized so far.  If any of these unrealizes
failed, the device would be left in an inconsistent state.  Must not
happen.

device_set_realized() lets it happen: it ignores errors in the roll
back code starting at label child_realize_fail.

Since realization is recursive, unrealization must be recursive, too.
But how could a partly failed unrealize be rolled back?  We'd have to
re-realize, which can fail.  This design is fundamentally broken.

device_set_realized() does not roll back at all.  Instead, it keeps
unrealizing, ignoring further errors.

It can screw up even for a device with no buses: if the lone
dc->unrealize() fails, it still unregisters vmstate, and calls
listeners' unrealize() callback.

bus_set_realized() does not roll back either.  Instead, it stops
unrealizing.

Fortunately, no unrealize method can fail, as we'll see below.

To fix the design error, drop parameter @errp from all the unrealize
methods.

Any unrealize method that uses @errp now needs an update.  This leads
us to unrealize() methods that can fail.  Merely passing it to another
unrealize method cannot cause failure, though.  Here are the ones that
do other things with @errp:

* virtio_serial_device_unrealize()

  Fails when qbus_set_hotplug_handler() fails, but still does all the
  other work.  On failure, the device would stay realized with its
  resources completely gone.  Oops.  Can't happen, because
  qbus_set_hotplug_handler() can't actually fail here.  Pass
  &error_abort to qbus_set_hotplug_handler() instead.

* hw/ppc/spapr_drc.c's unrealize()

  Fails when object_property_del() fails, but all the other work is
  already done.  On failure, the device would stay realized with its
  vmstate registration gone.  Oops.  Can't happen, because
  object_property_del() can't actually fail here.  Pass &error_abort
  to object_property_del() instead.

* spapr_phb_unrealize()

  Fails and bails out when remove_drcs() fails, but other work is
  already done.  On failure, the device would stay realized with some
  of its resources gone.  Oops.  remove_drcs() fails only when
  chassis_from_bus()'s object_property_get_uint() fails, and it can't
  here.  Pass &error_abort to remove_drcs() instead.

Therefore, no unrealize method can fail before this patch.

device_set_realized()'s recursive unrealization via bus uses
object_property_set_bool().  Can't drop @errp there, so pass
&error_abort.

We similarly unrealize with object_property_set_bool() elsewhere,
always ignoring errors.  Pass &error_abort instead.

Several unrealize methods no longer handle errors from other unrealize
methods: virtio_9p_device_unrealize(),
virtio_input_device_unrealize(), scsi_qdev_unrealize(), ...
Much of the deleted error handling looks wrong anyway.

One unrealize methods no longer ignore such errors:
usb_ehci_pci_exit().

Several realize methods no longer ignore errors when rolling back:
v9fs_device_realize_common(), pci_qdev_unrealize(),
spapr_phb_realize(), usb_qdev_realize(), vfio_ccw_realize(),
virtio_device_realize().

Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-17-armbru@redhat.com>
2020-05-15 07:08:14 +02:00

307 lines
9.4 KiB
C

/*
* Virtio Console and Generic Serial Port Devices
*
* Copyright Red Hat, Inc. 2009, 2010
*
* Authors:
* Amit Shah <amit.shah@redhat.com>
*
* This work is licensed under the terms of the GNU GPL, version 2. See
* the COPYING file in the top-level directory.
*/
#include "qemu/osdep.h"
#include "chardev/char-fe.h"
#include "qemu/error-report.h"
#include "qemu/module.h"
#include "trace.h"
#include "hw/qdev-properties.h"
#include "hw/virtio/virtio-serial.h"
#include "qapi/error.h"
#include "qapi/qapi-events-char.h"
#define TYPE_VIRTIO_CONSOLE_SERIAL_PORT "virtserialport"
#define VIRTIO_CONSOLE(obj) \
OBJECT_CHECK(VirtConsole, (obj), TYPE_VIRTIO_CONSOLE_SERIAL_PORT)
typedef struct VirtConsole {
VirtIOSerialPort parent_obj;
CharBackend chr;
guint watch;
} VirtConsole;
/*
* Callback function that's called from chardevs when backend becomes
* writable.
*/
static gboolean chr_write_unblocked(GIOChannel *chan, GIOCondition cond,
void *opaque)
{
VirtConsole *vcon = opaque;
vcon->watch = 0;
virtio_serial_throttle_port(VIRTIO_SERIAL_PORT(vcon), false);
return FALSE;
}
/* Callback function that's called when the guest sends us data */
static ssize_t flush_buf(VirtIOSerialPort *port,
const uint8_t *buf, ssize_t len)
{
VirtConsole *vcon = VIRTIO_CONSOLE(port);
ssize_t ret;
if (!qemu_chr_fe_backend_connected(&vcon->chr)) {
/* If there's no backend, we can just say we consumed all data. */
return len;
}
ret = qemu_chr_fe_write(&vcon->chr, buf, len);
trace_virtio_console_flush_buf(port->id, len, ret);
if (ret < len) {
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_GET_CLASS(port);
/*
* Ideally we'd get a better error code than just -1, but
* that's what the chardev interface gives us right now. If
* we had a finer-grained message, like -EPIPE, we could close
* this connection.
*/
if (ret < 0)
ret = 0;
/* XXX we should be queuing data to send later for the
* console devices too rather than silently dropping
* console data on EAGAIN. The Linux virtio-console
* hvc driver though does sends with spinlocks held,
* so if we enable throttling that'll stall the entire
* guest kernel, not merely the process writing to the
* console.
*
* While we could queue data for later write without
* enabling throttling, this would result in the guest
* being able to trigger arbitrary memory usage in QEMU
* buffering data for later writes.
*
* So fixing this problem likely requires fixing the
* Linux virtio-console hvc driver to not hold spinlocks
* while writing, and instead merely block the process
* that's writing. QEMU would then need some way to detect
* if the guest had the fixed driver too, before we can
* use throttling on host side.
*/
if (!k->is_console) {
virtio_serial_throttle_port(port, true);
if (!vcon->watch) {
vcon->watch = qemu_chr_fe_add_watch(&vcon->chr,
G_IO_OUT|G_IO_HUP,
chr_write_unblocked, vcon);
}
}
}
return ret;
}
/* Callback function that's called when the guest opens/closes the port */
static void set_guest_connected(VirtIOSerialPort *port, int guest_connected)
{
VirtConsole *vcon = VIRTIO_CONSOLE(port);
DeviceState *dev = DEVICE(port);
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_GET_CLASS(port);
if (!k->is_console) {
qemu_chr_fe_set_open(&vcon->chr, guest_connected);
}
if (dev->id) {
qapi_event_send_vserport_change(dev->id, guest_connected);
}
}
static void guest_writable(VirtIOSerialPort *port)
{
VirtConsole *vcon = VIRTIO_CONSOLE(port);
qemu_chr_fe_accept_input(&vcon->chr);
}
/* Readiness of the guest to accept data on a port */
static int chr_can_read(void *opaque)
{
VirtConsole *vcon = opaque;
return virtio_serial_guest_ready(VIRTIO_SERIAL_PORT(vcon));
}
/* Send data from a char device over to the guest */
static void chr_read(void *opaque, const uint8_t *buf, int size)
{
VirtConsole *vcon = opaque;
VirtIOSerialPort *port = VIRTIO_SERIAL_PORT(vcon);
trace_virtio_console_chr_read(port->id, size);
virtio_serial_write(port, buf, size);
}
static void chr_event(void *opaque, QEMUChrEvent event)
{
VirtConsole *vcon = opaque;
VirtIOSerialPort *port = VIRTIO_SERIAL_PORT(vcon);
trace_virtio_console_chr_event(port->id, event);
switch (event) {
case CHR_EVENT_OPENED:
virtio_serial_open(port);
break;
case CHR_EVENT_CLOSED:
if (vcon->watch) {
g_source_remove(vcon->watch);
vcon->watch = 0;
}
virtio_serial_close(port);
break;
case CHR_EVENT_BREAK:
case CHR_EVENT_MUX_IN:
case CHR_EVENT_MUX_OUT:
/* Ignore */
break;
}
}
static int chr_be_change(void *opaque)
{
VirtConsole *vcon = opaque;
VirtIOSerialPort *port = VIRTIO_SERIAL_PORT(vcon);
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_GET_CLASS(port);
if (k->is_console) {
qemu_chr_fe_set_handlers(&vcon->chr, chr_can_read, chr_read,
NULL, chr_be_change, vcon, NULL, true);
} else {
qemu_chr_fe_set_handlers(&vcon->chr, chr_can_read, chr_read,
chr_event, chr_be_change, vcon, NULL, false);
}
if (vcon->watch) {
g_source_remove(vcon->watch);
vcon->watch = qemu_chr_fe_add_watch(&vcon->chr,
G_IO_OUT | G_IO_HUP,
chr_write_unblocked, vcon);
}
return 0;
}
static void virtconsole_enable_backend(VirtIOSerialPort *port, bool enable)
{
VirtConsole *vcon = VIRTIO_CONSOLE(port);
if (!qemu_chr_fe_backend_connected(&vcon->chr)) {
return;
}
if (enable) {
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_GET_CLASS(port);
qemu_chr_fe_set_handlers(&vcon->chr, chr_can_read, chr_read,
k->is_console ? NULL : chr_event,
chr_be_change, vcon, NULL, false);
} else {
qemu_chr_fe_set_handlers(&vcon->chr, NULL, NULL, NULL,
NULL, NULL, NULL, false);
}
}
static void virtconsole_realize(DeviceState *dev, Error **errp)
{
VirtIOSerialPort *port = VIRTIO_SERIAL_PORT(dev);
VirtConsole *vcon = VIRTIO_CONSOLE(dev);
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_GET_CLASS(dev);
if (port->id == 0 && !k->is_console) {
error_setg(errp, "Port number 0 on virtio-serial devices reserved "
"for virtconsole devices for backward compatibility.");
return;
}
if (qemu_chr_fe_backend_connected(&vcon->chr)) {
/*
* For consoles we don't block guest data transfer just
* because nothing is connected - we'll just let it go
* whetherever the chardev wants - /dev/null probably.
*
* For serial ports we need 100% reliable data transfer
* so we use the opened/closed signals from chardev to
* trigger open/close of the device
*/
if (k->is_console) {
qemu_chr_fe_set_handlers(&vcon->chr, chr_can_read, chr_read,
NULL, chr_be_change,
vcon, NULL, true);
virtio_serial_open(port);
} else {
qemu_chr_fe_set_handlers(&vcon->chr, chr_can_read, chr_read,
chr_event, chr_be_change,
vcon, NULL, false);
}
}
}
static void virtconsole_unrealize(DeviceState *dev)
{
VirtConsole *vcon = VIRTIO_CONSOLE(dev);
if (vcon->watch) {
g_source_remove(vcon->watch);
}
}
static void virtconsole_class_init(ObjectClass *klass, void *data)
{
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_CLASS(klass);
k->is_console = true;
}
static const TypeInfo virtconsole_info = {
.name = "virtconsole",
.parent = TYPE_VIRTIO_CONSOLE_SERIAL_PORT,
.class_init = virtconsole_class_init,
};
static Property virtserialport_properties[] = {
DEFINE_PROP_CHR("chardev", VirtConsole, chr),
DEFINE_PROP_END_OF_LIST(),
};
static void virtserialport_class_init(ObjectClass *klass, void *data)
{
DeviceClass *dc = DEVICE_CLASS(klass);
VirtIOSerialPortClass *k = VIRTIO_SERIAL_PORT_CLASS(klass);
k->realize = virtconsole_realize;
k->unrealize = virtconsole_unrealize;
k->have_data = flush_buf;
k->set_guest_connected = set_guest_connected;
k->enable_backend = virtconsole_enable_backend;
k->guest_writable = guest_writable;
device_class_set_props(dc, virtserialport_properties);
}
static const TypeInfo virtserialport_info = {
.name = TYPE_VIRTIO_CONSOLE_SERIAL_PORT,
.parent = TYPE_VIRTIO_SERIAL_PORT,
.instance_size = sizeof(VirtConsole),
.class_init = virtserialport_class_init,
};
static void virtconsole_register_types(void)
{
type_register_static(&virtserialport_info);
type_register_static(&virtconsole_info);
}
type_init(virtconsole_register_types)