With a reconnect socket, qemu_char_open() will start a background
thread. It should keep a reference on the chardev.
Fixes invalid read:
READ of size 8 at 0x6040000ac858 thread T7
#0 0x5555598d37b8 in unix_connect_saddr /home/elmarco/src/qq/util/qemu-sockets.c:954
#1 0x5555598d4751 in socket_connect /home/elmarco/src/qq/util/qemu-sockets.c:1109
#2 0x555559707c34 in qio_channel_socket_connect_sync /home/elmarco/src/qq/io/channel-socket.c:145
#3 0x5555596adebb in tcp_chr_connect_client_task /home/elmarco/src/qq/chardev/char-socket.c:1104
#4 0x555559723d55 in qio_task_thread_worker /home/elmarco/src/qq/io/task.c:123
#5 0x5555598a6731 in qemu_thread_start /home/elmarco/src/qq/util/qemu-thread-posix.c:519
#6 0x7ffff40d4431 in start_thread (/lib64/libpthread.so.0+0x9431)
#7 0x7ffff40029d2 in __clone (/lib64/libc.so.6+0x1019d2)
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20200420112012.567284-1-marcandre.lureau@redhat.com>
When the disconnect event is triggered in the connecting stage,
the tcp_chr_disconnect_locked may be called twice.
The first call:
#0 qemu_chr_socket_restart_timer (chr=0x55555582ee90) at chardev/char-socket.c:120
#1 0x000055555558e38c in tcp_chr_disconnect_locked (chr=<optimized out>) at chardev/char-socket.c:490
#2 0x000055555558e3cd in tcp_chr_disconnect (chr=0x55555582ee90) at chardev/char-socket.c:497
#3 0x000055555558ea32 in tcp_chr_new_client (chr=chr@entry=0x55555582ee90, sioc=sioc@entry=0x55555582f0b0) at chardev/char-socket.c:892
#4 0x000055555558eeb8 in qemu_chr_socket_connected (task=0x55555582f300, opaque=<optimized out>) at chardev/char-socket.c:1090
#5 0x0000555555574352 in qio_task_complete (task=task@entry=0x55555582f300) at io/task.c:196
#6 0x00005555555745f4 in qio_task_thread_result (opaque=0x55555582f300) at io/task.c:111
#7 qio_task_wait_thread (task=0x55555582f300) at io/task.c:190
#8 0x000055555558f17e in tcp_chr_wait_connected (chr=0x55555582ee90, errp=0x555555802a08 <error_abort>) at chardev/char-socket.c:1013
#9 0x0000555555567cbd in char_socket_client_reconnect_test (opaque=0x5555557fe020 <client8unix>) at tests/test-char.c:1152
The second call:
#0 0x00007ffff5ac3277 in raise () from /lib64/libc.so.6
#1 0x00007ffff5ac4968 in abort () from /lib64/libc.so.6
#2 0x00007ffff5abc096 in __assert_fail_base () from /lib64/libc.so.6
#3 0x00007ffff5abc142 in __assert_fail () from /lib64/libc.so.6
#4 0x000055555558d10a in qemu_chr_socket_restart_timer (chr=0x55555582ee90) at chardev/char-socket.c:125
#5 0x000055555558df0c in tcp_chr_disconnect_locked (chr=<optimized out>) at chardev/char-socket.c:490
#6 0x000055555558df4d in tcp_chr_disconnect (chr=0x55555582ee90) at chardev/char-socket.c:497
#7 0x000055555558e5b2 in tcp_chr_new_client (chr=chr@entry=0x55555582ee90, sioc=sioc@entry=0x55555582f0b0) at chardev/char-socket.c:892
#8 0x000055555558e93a in tcp_chr_connect_client_sync (chr=chr@entry=0x55555582ee90, errp=errp@entry=0x7fffffffd178) at chardev/char-socket.c:944
#9 0x000055555558ec78 in tcp_chr_wait_connected (chr=0x55555582ee90, errp=0x555555802a08 <error_abort>) at chardev/char-socket.c:1035
#10 0x000055555556804b in char_socket_client_test (opaque=0x5555557fe020 <client8unix>) at tests/test-char.c:1023
Run test/test-char to reproduce this issue.
test-char: chardev/char-socket.c:125: qemu_chr_socket_restart_timer: Assertion `!s->reconnect_timer' failed.
Signed-off-by: Li Feng <fengli@smartx.com>
Acked-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20200522025554.41063-1-fengli@smartx.com>
Receiving the error in a local variable only to free it is less clear
(and also less efficient) than passing NULL. Clean up.
Cc: Daniel P. Berrange <berrange@redhat.com>
Cc: Jerome Forissier <jerome@forissier.org>
CC: Greg Kurz <groug@kaod.org>
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Greg Kurz <groug@kaod.org>
Message-Id: <20200630090351.1247703-4-armbru@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Errors are already freed by error_report_err, so we only need to call
error_free when that function is not called.
Cc: qemu-stable@nongnu.org
Signed-off-by: lichun <lichun@ruijie.com.cn>
Message-Id: <20200621213017.17978-1-lichun@ruijie.com.cn>
Reviewed-by: Markus Armbruster <armbru@redhat.com>
[Commit message improved, cc: qemu-stable]
Signed-off-by: Markus Armbruster <armbru@redhat.com>
In tcp_chr_sync_read function, there is a possibility of socket
disconnection during blocking read, then tcp_chr_hup function would clean up
the qio channel pointers(i.e ioc, sioc).
Signed-off-by: Sai Pavan Boddu <sai.pavan.boddu@xilinx.com>
Message-Id: <1587289900-29485-1-git-send-email-sai.pavan.boddu@xilinx.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
During testing of the vhost-user-blk reconnect functionality the qemu
SIGSEGV was triggered:
start qemu as:
x86_64-softmmu/qemu-system-x86_64 -m 1024M -M q35 \
-object memory-backend-file,id=ram-node0,size=1024M,mem-path=/dev/shm/qemu,share=on \
-numa node,cpus=0,memdev=ram-node0 \
-chardev socket,id=chardev0,path=./vhost.sock,noserver,reconnect=1 \
-device vhost-user-blk-pci,chardev=chardev0,num-queues=4 --enable-kvm
start vhost-user-blk daemon:
./vhost-user-blk -s ./vhost.sock -b test-img.raw
If vhost-user-blk will be killed during the vhost initialization
process, for instance after getting VHOST_SET_VRING_CALL command, then
QEMU will fail with the following backtrace:
Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault.
0x00005555559272bb in vhost_user_read (dev=0x7fffef2d53e0, msg=0x7fffffffd5b0)
at ./hw/virtio/vhost-user.c:260
260 CharBackend *chr = u->user->chr;
#0 0x00005555559272bb in vhost_user_read (dev=0x7fffef2d53e0, msg=0x7fffffffd5b0)
at ./hw/virtio/vhost-user.c:260
#1 0x000055555592acb8 in vhost_user_get_config (dev=0x7fffef2d53e0, config=0x7fffef2d5394 "", config_len=60)
at ./hw/virtio/vhost-user.c:1645
#2 0x0000555555925525 in vhost_dev_get_config (hdev=0x7fffef2d53e0, config=0x7fffef2d5394 "", config_len=60)
at ./hw/virtio/vhost.c:1490
#3 0x00005555558cc46b in vhost_user_blk_device_realize (dev=0x7fffef2d51a0, errp=0x7fffffffd8f0)
at ./hw/block/vhost-user-blk.c:429
#4 0x0000555555920090 in virtio_device_realize (dev=0x7fffef2d51a0, errp=0x7fffffffd948)
at ./hw/virtio/virtio.c:3615
#5 0x0000555555a9779c in device_set_realized (obj=0x7fffef2d51a0, value=true, errp=0x7fffffffdb88)
at ./hw/core/qdev.c:891
...
The problem is that vhost_user_write doesn't get an error after
disconnect and try to call vhost_user_read(). The tcp_chr_write()
routine should return -1 in case of disconnect. Indicate the EIO error
if this routine is called in the disconnected state.
Signed-off-by: Dima Stepanov <dimastep@yandex-team.ru>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <aeb7806bfc945faadf09f64dcfa30f59de3ac053.1590396396.git.dimastep@yandex-team.ru>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Replace
error_report("...: %s", ..., error_get_pretty(err));
by
error_reportf_err(err, "...: ", ...);
One of the replaced messages lacked a colon. Add it.
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
Message-Id: <20200505101908.6207-6-armbru@redhat.com>
unix_listen/connect_saddr now support abstract address types
two aditional BOOL switches are introduced:
tight: whether to set @addrlen to the minimal string length,
or the maximum sun_path length. default is TRUE
abstract: whether we use abstract address. default is FALSE
cli example:
-monitor unix:/tmp/unix.socket,abstract,tight=off
OR
-chardev socket,path=/tmp/unix.socket,id=unix1,abstract,tight=on
Signed-off-by: xiaoqiang zhao <zxq_yx_007@163.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
The only way object_property_add() can fail is when a property with
the same name already exists. Since our property names are all
hardcoded, failure is a programming error, and the appropriate way to
handle it is passing &error_abort.
Same for its variants, except for object_property_add_child(), which
additionally fails when the child already has a parent. Parentage is
also under program control, so this is a programming error, too.
We have a bit over 500 callers. Almost half of them pass
&error_abort, slightly fewer ignore errors, one test case handles
errors, and the remaining few callers pass them to their own callers.
The previous few commits demonstrated once again that ignoring
programming errors is a bad idea.
Of the few ones that pass on errors, several violate the Error API.
The Error ** argument must be NULL, &error_abort, &error_fatal, or a
pointer to a variable containing NULL. Passing an argument of the
latter kind twice without clearing it in between is wrong: if the
first call sets an error, it no longer points to NULL for the second
call. ich9_pm_add_properties(), sparc32_ledma_realize(),
sparc32_dma_realize(), xilinx_axidma_realize(), xilinx_enet_realize()
are wrong that way.
When the one appropriate choice of argument is &error_abort, letting
users pick the argument is a bad idea.
Drop parameter @errp and assert the preconditions instead.
There's one exception to "duplicate property name is a programming
error": the way object_property_add() implements the magic (and
undocumented) "automatic arrayification". Don't drop @errp there.
Instead, rename object_property_add() to object_property_try_add(),
and add the obvious wrapper object_property_add().
Signed-off-by: Markus Armbruster <armbru@redhat.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20200505152926.18877-15-armbru@redhat.com>
[Two semantic rebase conflicts resolved]
There's a race condition in which the tcp_chr_read() ioc handler can
close a connection that is being written to from another thread.
Running iotest 136 in a loop triggers this problem and crashes QEMU.
(gdb) bt
#0 0x00005558b842902d in object_get_class (obj=0x0) at qom/object.c:860
#1 0x00005558b84f92db in qio_channel_writev_full (ioc=0x0, iov=0x7ffc355decf0, niov=1, fds=0x0, nfds=0, errp=0x0) at io/channel.c:76
#2 0x00005558b84e0e9e in io_channel_send_full (ioc=0x0, buf=0x5558baf5beb0, len=138, fds=0x0, nfds=0) at chardev/char-io.c:123
#3 0x00005558b84e4a69 in tcp_chr_write (chr=0x5558ba460380, buf=0x5558baf5beb0 "...", len=138) at chardev/char-socket.c:135
#4 0x00005558b84dca55 in qemu_chr_write_buffer (s=0x5558ba460380, buf=0x5558baf5beb0 "...", len=138, offset=0x7ffc355dedd0, write_all=false) at chardev/char.c:112
#5 0x00005558b84dcbc2 in qemu_chr_write (s=0x5558ba460380, buf=0x5558baf5beb0 "...", len=138, write_all=false) at chardev/char.c:147
#6 0x00005558b84dfb26 in qemu_chr_fe_write (be=0x5558ba476610, buf=0x5558baf5beb0 "...", len=138) at chardev/char-fe.c:42
#7 0x00005558b8088c86 in monitor_flush_locked (mon=0x5558ba476610) at monitor.c:406
#8 0x00005558b8088e8c in monitor_puts (mon=0x5558ba476610, str=0x5558ba921e49 "") at monitor.c:449
#9 0x00005558b8089178 in qmp_send_response (mon=0x5558ba476610, rsp=0x5558bb161600) at monitor.c:498
#10 0x00005558b808920c in monitor_qapi_event_emit (event=QAPI_EVENT_SHUTDOWN, qdict=0x5558bb161600) at monitor.c:526
#11 0x00005558b8089307 in monitor_qapi_event_queue_no_reenter (event=QAPI_EVENT_SHUTDOWN, qdict=0x5558bb161600) at monitor.c:551
#12 0x00005558b80896c0 in qapi_event_emit (event=QAPI_EVENT_SHUTDOWN, qdict=0x5558bb161600) at monitor.c:626
#13 0x00005558b855f23b in qapi_event_send_shutdown (guest=false, reason=SHUTDOWN_CAUSE_HOST_QMP_QUIT) at qapi/qapi-events-run-state.c:43
#14 0x00005558b81911ef in qemu_system_shutdown (cause=SHUTDOWN_CAUSE_HOST_QMP_QUIT) at vl.c:1837
#15 0x00005558b8191308 in main_loop_should_exit () at vl.c:1885
#16 0x00005558b819140d in main_loop () at vl.c:1924
#17 0x00005558b8198c84 in main (argc=18, argv=0x7ffc355df3f8, envp=0x7ffc355df490) at vl.c:4665
This patch adds a lock to protect tcp_chr_disconnect() and
socket_reconnect_timeout()
Signed-off-by: Alberto Garcia <berto@igalia.com>
Signed-off-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Message-Id: <1565625509-404969-3-git-send-email-andrey.shinkevich@virtuozzo.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Commit 767abe7 ("chardev: forbid 'wait' option with client sockets")
is a bit too strict. Current libvirt always set wait=false, and will
thus fail to add client chardev.
Make the code more permissive, allowing wait=false with client socket
chardevs. Deprecate usage of 'wait' with client sockets.
Fixes: 767abe7f49
Cc: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Message-id: 20190415163337.2795-1-marcandre.lureau@redhat.com
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
Currently any client which can complete the TLS handshake is able to use
a chardev server. The server admin can turn on the 'verify-peer' option
for the x509 creds to require the client to provide a x509
certificate. This means the client will have to acquire a certificate
from the CA before they are permitted to use the chardev server. This is
still a fairly low bar.
This adds a 'tls-authz=OBJECT-ID' option to the socket chardev backend
which takes the ID of a previously added 'QAuthZ' object instance. This
will be used to validate the client's x509 distinguished name. Clients
failing the check will not be permitted to use the chardev server.
For example to setup authorization that only allows connection from a
client whose x509 certificate distinguished name contains 'CN=fred', you
would use:
$QEMU -object tls-creds-x509,id=tls0,dir=/home/berrange/qemutls,\
endpoint=server,verify-peer=yes \
-object authz-simple,id=authz0,identity=CN=laptop.example.com,,\
O=Example Org,,L=London,,ST=London,,C=GB \
-chardev socket,host=127.0.0.1,port=9000,server,\
tls-creds=tls0,tls-authz=authz0 \
...other qemu args...
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
If the socket is connecting or connected, tcp_chr_update_read_handler will
be called but it should not set the NetListener's callbacks again.
Otherwise, tcp_chr_accept is invoked while the socket is in connected
state and you get an assertion failure.
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When the 'reconnect' option is given for a client connection, the
qmp_chardev_open_socket_client method will run an asynchronous
connection attempt. The QIOChannel socket executes this is a single use
background thread, so the connection will succeed immediately (assuming
the server is listening). The chardev, however, won't get the result
from this background thread until the main loop starts running and
processes idle callbacks.
Thus when tcp_chr_wait_connected is run s->ioc will be NULL, but the
state will be TCP_CHARDEV_STATE_CONNECTING, and there may already
be an established connection that will be associated with the chardev
by the pending idle callback. tcp_chr_wait_connected doesn't check the
state, only s->ioc, so attempts to establish another connection
synchronously.
If the server allows multiple connections this is unhelpful but not a
fatal problem as the duplicate connection will get ignored by the
tcp_chr_new_client method when it sees the state is already connected.
If the server only supports a single connection, however, the
tcp_chr_wait_connected method will hang forever because the server will
not accept its synchronous connection attempt until the first connection
is closed.
To deal with this tcp_chr_wait_connected needs to synchronize with the
completion of the background connection task. To do this it needs to
create the QIOTask directly and use the qio_task_wait_thread method.
This will cancel the pending idle callback and directly dispatch the
task completion callback, allowing the connection to be associated
with the chardev. If the background connection failed, it can still
attempt a new synchronous connection.
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-15-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
In the previous commit
commit 1dc8a6695c
Author: Marc-André Lureau <marcandre.lureau@redhat.com>
Date: Tue Aug 16 12:33:32 2016 +0400
char: fix waiting for TLS and telnet connection
the tcp_chr_wait_connected() method was changed to check for a non-NULL
's->ioc' as a sign that there is already a connection present, as
opposed to checking the "connected" flag to supposedly fix handling of
TLS/telnet connections.
The original code would repeatedly call tcp_chr_wait_connected creating
many connections as 'connected' would never become true. The changed
code would still repeatedly call tcp_chr_wait_connected busy waiting
because s->ioc is set but the chardev will never see CHR_EVENT_OPENED.
IOW, the code is still broken with TLS/telnet, but in a different way.
Checking for a non-NULL 's->ioc' does not mean that a CHR_EVENT_OPENED
will be ready for a TLS/telnet connection. These protocols (and the
websocket protocol) all require the main loop to be running in order
to complete the protocol handshake before emitting CHR_EVENT_OPENED.
The tcp_chr_wait_connected() method is only used during early startup
before a main loop is running, so TLS/telnet/websock connections can
never complete initialization.
Making this work would require changing tcp_chr_wait_connected to run
a main loop. This is quite complex since we must not allow GSource's
that other parts of QEMU have registered to run yet. The current callers
of tcp_chr_wait_connected do not require use of the TLS/telnet/websocket
protocols, so the simplest option is to just forbid this combination
completely for now.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-14-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
If establishing a client connection fails, the tcp_chr_wait_connected
method should sleep for the reconnect timeout and then retry the
attempt. This ensures the callers don't immediately abort with an
error when the initial connection fails.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-13-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The socket connection state is indicated via the 'bool connected' field
in the SocketChardev struct. This variable is somewhat misleading
though, as it is only set to true once the connection has completed all
required handshakes (eg for TLS, telnet or websockets). IOW there is a
period of time in which the socket is connected, but the "connected"
flag is still false.
The socket chardev really has three states that it can be in,
disconnected, connecting and connected and those should be tracked
explicitly.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-12-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
In qmp_chardev_open_socket the code for connecting client chardevs is
split across two conditionals far apart with some server chardev code in
the middle. Split up the method so that code for client connection setup
is separate from code for server connection setup.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-11-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The tcp_chr_wait_connected method can deal with either server or client
chardevs, but some callers only care about one of these possibilities.
The tcp_chr_wait_connected method will also need some refactoring to
reliably deal with its primary goal of allowing a device frontend to
wait for an established connection, which will interfere with other
callers.
Split it into two methods, one responsible for server initiated
connections, the other responsible for client initiated connections.
In doing this split the tcp_char_connect_async() method is renamed
to become consistent with naming of the new methods.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-10-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The 'sioc' variable in qmp_chardev_open_socket was unused since
commit 3e7d4d20d3
Author: Peter Xu <peterx@redhat.com>
Date: Tue Mar 6 13:33:17 2018 +0800
chardev: use chardev's gcontext for async connect
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-9-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Now that all validation is separated off into a separate method,
we can directly populate the ChardevSocket struct from the
QemuOpts values, avoiding many local variables.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-7-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The 'wait'/'nowait' parameter is used to tell server sockets whether to
block until a client is accepted during initialization. Client chardevs
have always silently ignored this option. Various tests were mistakenly
passing this option for their client chardevs.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-6-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The 'reconnect' option is used to give the sleep time, in seconds,
before a client socket attempts to re-establish a connection to the
server. It does not make sense to set this for server sockets, as they
will always accept a new client connection immediately after the
previous one went away.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-5-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
The TLS creds option is not valid with certain address types. The user
config was only checked for errors when parsing legacy QemuOpts, thus
the user could pass unsupported values via QMP.
Pull all code for validating options out into a new method
qmp_chardev_validate_socket, that is called from the main
qmp_chardev_open_socket method. This adds a missing check for rejecting
TLS creds with the vsock address type.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20190211182442.8542-4-berrange@redhat.com>
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
New option "websocket" added to allow using WebSocket protocol for
chardev socket backend.
Example:
-chardev socket,websocket,server,id=...
Signed-off-by: Julia Suvorova <jusual@mail.ru>
Message-Id: <20181018223501.21683-3-jusual@mail.ru>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Upcoming websocket support requires additional parameters in function
headers that are already overloaded. This patch replaces the bunch of
parameters with a single structure pointer.
Signed-off-by: Julia Suvorova <jusual@mail.ru>
Message-Id: <20181018223501.21683-2-jusual@mail.ru>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
A chardev socket created with the 'fd=' argument is not going to
handle reconnection properly by recycling the same fd (or not in a
supported way). Let's forbid this case.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
qemu_chr_parse_socket() fills all ChardevSocket fields, but that
doesn't reflect correctly the arguments given with the options / on
the command line. "reconnect" takes a number as argument, and the
default value is 0, which doesn't help to identify the missing
option. The other arguments have default values that are less
problematic, leave them set by default for now.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
A socket chardev may not have associated address (when adding client
fd manually for example). But on disconnect, updating socket filename
expects an address and may lead to this crash:
Thread 1 "qemu-system-x86" received signal SIGSEGV, Segmentation fault.
0x0000555555d8c70c in SocketAddress_to_str (prefix=0x555556043062 "disconnected:", addr=0x0, is_listen=false, is_telnet=false) at /home/elmarco/src/qq/chardev/char-socket.c:388
388 switch (addr->type) {
(gdb) bt
#0 0x0000555555d8c70c in SocketAddress_to_str (prefix=0x555556043062 "disconnected:", addr=0x0, is_listen=false, is_telnet=false) at /home/elmarco/src/qq/chardev/char-socket.c:388
#1 0x0000555555d8c8aa in update_disconnected_filename (s=0x555556b1ed00) at /home/elmarco/src/qq/chardev/char-socket.c:419
#2 0x0000555555d8c959 in tcp_chr_disconnect (chr=0x555556b1ed00) at /home/elmarco/src/qq/chardev/char-socket.c:438
#3 0x0000555555d8cba1 in tcp_chr_hup (channel=0x555556b75690, cond=G_IO_HUP, opaque=0x555556b1ed00) at /home/elmarco/src/qq/chardev/char-socket.c:482
#4 0x0000555555da596e in qio_channel_fd_source_dispatch (source=0x555556bb68b0, callback=0x555555d8cb58 <tcp_chr_hup>, user_data=0x555556b1ed00) at /home/elmarco/src/qq/io/channel-watch.c:84
Replace filename with a generic "disconnected:socket" in this case.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
So far, tcp_chr_update_read_handler() only updated the read
handler. Let's also update the hup handler.
Factorize the code while at it. (note that s->ioc != NULL when
s->connected)
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180817135224.22971-4-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This reverts commit 25679e5d58.
This commit broke "reconnect socket" chardev that are created after
"machine_done": they no longer try to connect. It broke also
vhost-user-test that uses chardev while there is no "machine_done"
event.
The goal of this patch was to move the "connect" source to the
frontend context. chr->gcontext is set with
qemu_chr_fe_set_handlers(). But there is no guarantee that it will be
called, so we can't delay connection until then: the chardev should
still attempt to connect during open(). qemu_chr_fe_set_handlers() is
eventually called later and will update the context.
Unless there is a good reason to not use initially the default
context, I think we should revert to the previous state to fix the
regressions.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180817135224.22971-3-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This reverts commit 99f2f54174.
See next commit reverting 25679e5d58 as
well for rationale.
Signed-off-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Message-Id: <20180817135224.22971-2-marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
In the tcp_chr_write function, we checked errno,
but errno was not reset before a read or write operation.
Therefore, this check of errno's actions is often
incorrect after EAGAIN has occurred.
we need check errno together with ret < 0.
Signed-off-by: xinhua.Cao <caoxinhua@huawei.com>
Message-Id: <20180704033642.15996-1-caoxinhua@huawei.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Fixes: 9fc53a10f8
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
If we see EAGAIN, no data was sent over the socket, so we still have to
retry sending of msgfds next time.
Signed-off-by: linzhecheng <linzhecheng@huawei.com>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
This trips Coverity, which believes the subsequent qio_channel_create_watch
can dereference a NULL pointer. In reality, tcp_chr_connect's callers
all have s->ioc properly initialized, since they are all rooted at
tcp_chr_new_client.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
TLS handshake may create background GSource tasks, while we won't know
the correct GMainContext until the whole chardev (including frontend)
inited. Let's postpone the initial TLS handshake until machine done.
For dynamically created tcp chardev, we don't postpone that by checking
the init_machine_done variable.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
[peterx: add missing include line, do unit test]
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180308140714.28906-1-peterx@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Eric Blake <eblake@redhat.com>
* SCSI fix to pass maximum transfer size (Daniel Barboza)
* chardev fixes and improved iothread support (Daniel Berrangé, Peter)
* checkpatch tweak (Eric)
* make help tweak (Marc-André)
* make more PCI NICs available with -net or -nic (myself)
* change default q35 NIC to e1000e (myself)
* SCSI support for NDOB bit (myself)
* membarrier system call support (myself)
* SuperIO refactoring (Philippe)
* miscellaneous cleanups and fixes (Thomas)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (GNU/Linux)
iQEcBAABAgAGBQJapqaMAAoJEL/70l94x66DQoUH/Rvg+a8giz/SrEA4P8D3Cb2z
4GNbNUUoy4oU0ltD5IAMskMwpOsvl1batE0D+pKIlfO9NV4+Cj2kpgo0p9TxoYqM
VCby3wRtx27zb5nVytC6M++iIKXmeEMqXmFw61I6umddNPSl4IR3hiHEE0DM+7dV
UPIOvJeEiazyQaw3Iw+ZctNn8dDBKc/+6oxP9xRcYTaZ6hB4G9RZkqGNNSLcJkk7
R0UotdjzIZhyWMOkjIwlpTF4sWv8gsYUV4bPYKMYho5B0Obda2dBM3I1kpA8yDa/
xZ5lheOaAVBZvM5aMIcaQPa65MO9hLyXFmhMOgyfpJhLBBz6Qpa4OLLI6DeTN+0=
=UAgA
-----END PGP SIGNATURE-----
Merge remote-tracking branch 'remotes/bonzini/tags/for-upstream' into staging
* Record-replay lockstep execution, log dumper and fixes (Alex, Pavel)
* SCSI fix to pass maximum transfer size (Daniel Barboza)
* chardev fixes and improved iothread support (Daniel Berrangé, Peter)
* checkpatch tweak (Eric)
* make help tweak (Marc-André)
* make more PCI NICs available with -net or -nic (myself)
* change default q35 NIC to e1000e (myself)
* SCSI support for NDOB bit (myself)
* membarrier system call support (myself)
* SuperIO refactoring (Philippe)
* miscellaneous cleanups and fixes (Thomas)
# gpg: Signature made Mon 12 Mar 2018 16:10:52 GMT
# gpg: using RSA key BFFBD25F78C7AE83
# gpg: Good signature from "Paolo Bonzini <bonzini@gnu.org>"
# gpg: aka "Paolo Bonzini <pbonzini@redhat.com>"
# Primary key fingerprint: 46F5 9FBD 57D6 12E7 BFD4 E2F7 7E15 100C CD36 69B1
# Subkey fingerprint: F133 3857 4B66 2389 866C 7682 BFFB D25F 78C7 AE83
* remotes/bonzini/tags/for-upstream: (69 commits)
tcg: fix cpu_io_recompile
replay: update documentation
replay: save vmstate of the asynchronous events
replay: don't process async events when warping the clock
scripts/replay-dump.py: replay log dumper
replay: avoid recursive call of checkpoints
replay: check return values of fwrite
replay: push replay_mutex_lock up the call tree
replay: don't destroy mutex at exit
replay: make locking visible outside replay code
replay/replay-internal.c: track holding of replay_lock
replay/replay.c: bump REPLAY_VERSION again
replay: save prior value of the host clock
replay: added replay log format description
replay: fix save/load vm for non-empty queue
replay: fixed replay_enable_events
replay: fix processing async events
cpu-exec: fix exception_index handling
hw/i386/pc: Factor out the superio code
hw/alpha/dp264: Use the TYPE_SMC37C669_SUPERIO
...
Signed-off-by: Peter Maydell <peter.maydell@linaro.org>
# Conflicts:
# default-configs/i386-softmmu.mak
# default-configs/x86_64-softmmu.mak
When starting QEMU management apps will usually setup a monitor socket, and
then open it immediately after startup. If not using QEMU's own -daemonize
arg, this process can be troublesome to handle correctly. The mgmt app will
need to repeatedly call connect() until it succeeds, because it does not
know when QEMU has created the listener socket. If can't retry connect()
forever though, because an error might have caused QEMU to exit before it
even creates the monitor.
The obvious way to fix this kind of problem is to just pass in a pre-opened
socket file descriptor for the QEMU monitor to listen on. The management
app can now immediately call connect() just once. If connect() fails it
knows that QEMU has exited with an error.
The SocketAddress(Legacy) structs allow for FD passing via the monitor, and
now via inherited file descriptors from the process that spawned QEMU. The
final missing piece is adding a 'fd' parameter in the socket chardev
options.
This allows both HMP usage, pass any FD number with SCM_RIGHTS, then
running HMP commands:
getfd myfd
chardev-add socket,fd=myfd
Note that numeric FDs cannot be referenced directly in HMP, only named FDs.
And also CLI usage, by leak FD 3 from parent by clearing O_CLOEXEC, then
spawning QEMU with
-chardev socket,fd=3,id=mon
-mon chardev=mon,mode=control
Note that named FDs cannot be referenced in CLI args, only numeric FDs.
We do not wire this up in the legacy chardev syntax, so you cannot use FD
passing with '-qmp', you must use the modern '-mon' + '-chardev' pair.
When passing pre-opened FDs there is a restriction on use of TLS encryption.
It can be used on a server socket chardev, but cannot be used for a client
socket chardev. This is because when validating a server's certificate, the
client needs to have a hostname available to match against the certificate
identity.
An illustrative example of usage is:
#!/usr/bin/perl
use IO::Socket::UNIX;
use Fcntl;
unlink "/tmp/qmp";
my $srv = IO::Socket::UNIX->new(
Type => SOCK_STREAM(),
Local => "/tmp/qmp",
Listen => 1,
);
my $flags = fcntl $srv, F_GETFD, 0;
fcntl $srv, F_SETFD, $flags & ~FD_CLOEXEC;
my $fd = $srv->fileno();
exec "qemu-system-x86_64", \
"-chardev", "socket,fd=$fd,server,nowait,id=mon", \
"-mon", "chardev=mon,mode=control";
Reviewed-by: Eric Blake <eblake@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
To prepare for handling more address types, refactor the parsing of
socket address information to make it more robust and extensible.
Reviewed-by: Eric Blake <eblake@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Even if common tn3270 implementations do not support TLS, it is trivial to
have them proxied over a proxy like stunnel which adds TLS at the sockets
layer. We should thus not silently skip tn3270 protocol initialization
when TLS is enabled.
Reviewed-by: Eric Blake <eblake@redhat.com>
Acked-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
Now qio_channel_tls_handshake() is ready to receive the context. Let
socket chardev use it, then the TLS handshake of chardev will always be
with the chardev's context.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180306053320.15401-9-peterx@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This patch allows the socket chardev async connection be setup with
non-default gcontext. We do it by postponing the setup to machine done,
since until then we can know which context we should run the async
operation on.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180306053320.15401-8-peterx@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Generalize the function to create the async QIO task connection. Also,
fix the context pointer to use the chardev's gcontext.
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180306053320.15401-7-peterx@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
It was originally created by qio_channel_add_watch() so it's always
assigning the task to main context. Now we use the new API called
qio_channel_add_watch_source() so that we get the GSource handle rather
than the tag ID.
Meanwhile, caching the gsource and TCPChardevTelnetInit (which holds the
handshake data) in SocketChardev.telnet_source so that we can also do
dynamic context switch when update read handlers.
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180306053320.15401-5-peterx@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
TCP chardevs can be using QIO network listeners working in the
background when in listening mode. However the network listeners are
always running in main context. This can race with chardevs that are
running in non-main contexts.
To solve this, we need to re-setup the net listeners in
tcp_chr_update_read_handler() with the newly cached gcontext.
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Message-Id: <20180306053320.15401-4-peterx@redhat.com>
Acked-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Daniel P. Berrangé <berrange@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When this commit was applied
commit 9894dc0cdc
Author: Daniel P. Berrange <berrange@redhat.com>
Date: Tue Jan 19 11:14:29 2016 +0000
char: convert from GIOChannel to QIOChannel
The tcp_chr_recv() function was changed to return QIO_CHANNEL_ERR_BLOCK
which corresonds to -2. As such the handling for EAGAIN was able to be
removed from tcp_chr_read(). Unfortunately in a later commit:
commit b6572b4f97
Author: Marc-André Lureau <marcandre.lureau@redhat.com>
Date: Fri Mar 11 18:55:24 2016 +0100
char: translate from QIOChannel error to errno
The tcp_chr_recv() function was changed back to return -1, with errno
set to EAGAIN, without also re-addding support for this to tcp_chr_read()
Reported-by: Aleksey Kuleshov <rndfax@yandex.ru>
Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
Message-Id: <20180222121351.26191-1-berrange@redhat.com>
Reviewed-by: Marc-André Lureau <marcandre.lureau@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
