virtio-console: notify about the terminal size
v2 cover virtio-console: notify about the terminal size
The goal of this series is to have a resizable terminal into a guest without having to set up networking and using, e.g. ssh.
The virtio spec allows a virtio-console device to notify the guest about terminal resizes in the host. Linux Kernel implements the driver part of the spec. This series implement the device part in QEMU.
In this series resize notifications are only supported for the stdio backend but I think it should be easy to add support for the vc backend. Support for tty/serial backends is complicated by the fact that there is no clean way to detect resizes of the underlying terminal.
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in virtio_console_resize struct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.
v2:
fix adding a new virtio feature bit to the virtio console device
Szymon Lukasz (6):
main-loop: change the handling of SIGWINCH
chardev: add support for retrieving the terminal size
chardev: add support for notifying about terminal resizes
char-stdio: add support for the terminal size
virtio-serial-bus: add terminal resize messages
virtio-console: notify the guest about terminal resizes
backends/cryptodev-vhost-user.c | 1 + chardev/char-fe.c | 11 ++++++ chardev/char-mux.c | 7 ++++ chardev/char-stdio.c | 35 +++++++++++++++++ chardev/char.c | 1 + hw/block/vhost-user-blk.c | 1 + hw/char/terminal3270.c | 1 + hw/char/trace-events | 1 + hw/char/virtio-console.c | 65 +++++++++++++++++++++++++++++-- hw/char/virtio-serial-bus.c | 42 +++++++++++++++++++- hw/core/machine.c | 1 + hw/ipmi/ipmi_bmc_extern.c | 1 + hw/usb/ccid-card-passthru.c | 1 + hw/usb/dev-serial.c | 1 + hw/usb/redirect.c | 1 + include/chardev/char-fe.h | 11 ++++++ include/chardev/char.h | 2 + include/hw/virtio/virtio-serial.h | 5 +++ include/qemu/main-loop.h | 4 ++ monitor/hmp.c | 1 + monitor/qmp.c | 1 + net/vhost-user.c | 1 + ui/curses.c | 11 +++--- util/main-loop.c | 21 ++++++++++ 24 files changed, 216 insertions(+), 11 deletions(-)
Signature · 1 line
2.27.0
PATCH v2 1/6 main-loop: change the handling of SIGWINCH
Block SIGWINCH, so it is delivered only via signalfd. Install a handler that uses NotifierList to tell interested parties about SIGWINCH delivery.
--- include/qemu/main-loop.h | 4 ++++ ui/curses.c | 11 ++++++----- util/main-loop.c | 21 +++++++++++++++++++++ 3 files changed, 31 insertions(+), 5 deletions(-)
include/qemu/main-loop.h +4 -0
@@ -325,4 +325,8 @@ typedef struct MainLoopPoll {void main_loop_poll_add_notifier (Notifier *notify);void main_loop_poll_remove_notifier (Notifier *notify); +#ifndef _WIN32 +void sigwinch_add_notifier (Notifier *n); +#endif +#endif
ui/curses.c +6 -5
@@ -34,6 +34,7 @@#include <iconv.h> #include "qapi/error.h" +#include "qemu/main-loop.h" #include "qemu/module.h" #include "ui/console.h" #include "ui/input.h"
@@ -146,7 +147,7 @@ static void curses_resize(DisplayChangeListener *dcl, } #if !defined(_WIN32) && defined(SIGWINCH) && defined(KEY_RESIZE) -static volatile sig_atomic_t got_sigwinch; +static bool got_sigwinch; static void curses_winch_check(void) { struct winsize {
@@ -169,17 +170,17 @@ static void curses_winch_check(void) invalidate = 1; } -static void curses_winch_handler (int signum) +static void curses_winch_handler (Notifier *n,void *data) { got_sigwinch = true; }static void curses_winch_init (void ) { -struct sigaction old, winch = { - .sa_handler = curses_winch_handler, +static Notifier n = { + .notify = curses_winch_handler }; -sigaction (SIGWINCH , &winch, &old); +sigwinch_add_notifier (&n); }#else static void curses_winch_check (void ) {}
util/main-loop.c +21 -0
@@ -90,6 +90,7 @@ static int qemu_signal_init(Error **errp)sigaddset (&set,SIGIO );sigaddset (&set,SIGALRM );sigaddset (&set,SIGBUS ); +sigaddset (&set,SIGWINCH );/* SIGINT cannot be handled via signalfd, so that ^C can be used * to interrupt QEMU when it is being run under gdb. SIGHUP and * SIGTERM are also handled asynchronously, even though it is not
@@ -111,6 +112,26 @@ static int qemu_signal_init(Error **errp) return 0; } +static NotifierList sigwinch_notifiers = +NOTIFIER_LIST_INITIALIZER (sigwinch_notifiers); + +static void sigwinch_handler (int signum) +{ +notifier_list_notify (&sigwinch_notifiers,NULL ); +} + +void sigwinch_add_notifier (Notifier *n) +{ +if (notifier_list_empty (&sigwinch_notifiers)) { +struct sigaction action = { + .sa_handler = sigwinch_handler, + }; +sigaction (SIGWINCH , &action,NULL ); + } + +notifier_list_add (&sigwinch_notifiers, n); +} +#else /* _WIN32 */ static int qemu_signal_init (Error **errp)
Signature · 1 line
2.27.0
PATCH v2 2/6 chardev: add support for retrieving the terminal size
In reply to Szymon LukaszExtend the class of chardevs with a new function - chr_get_winsize. A chardev backend should implement if it is able to get the size of the connected terminal and can detect changes in the terminal size, i.e. if the backend cannot detect resizes it must not implement this (e.g. if we have a tty backend connected to some (pseudo)terminal there is no clean way to detect resizes since SIGWINCH is sent only for the controlling terminal).
--- chardev/char-fe.c | 11 +++++++++++ chardev/char-mux.c | 7 +++++++ include/chardev/char-fe.h | 11 +++++++++++ include/chardev/char.h | 1 + 4 files changed, 30 insertions(+)
chardev/char-fe.c +11 -0
@@ -336,6 +336,17 @@ void qemu_chr_fe_set_echo(CharBackend *be, bool echo) } } +int qemu_chr_fe_get_winsize (CharBackend *be,uint16_t *cols,uint16_t *rows) +{ +Chardev *chr = be->chr; + +if (chr &&CHARDEV_GET_CLASS (chr)->chr_get_winsize) { +return CHARDEV_GET_CLASS (chr)->chr_get_winsize (chr, cols, rows); + } + +return -1 ; +} +void qemu_chr_fe_set_open (CharBackend *be,int fe_open) {Chardev *chr = be->chr;
chardev/char-mux.c +7 -0
@@ -293,6 +293,12 @@ static void mux_chr_update_read_handlers(Chardev *chr) chr->gcontext, true, false); } +static int mux_chr_get_winsize (Chardev *chr,uint16_t *cols,uint16_t *rows) +{ +MuxChardev *d =MUX_CHARDEV (chr); +return qemu_chr_fe_get_winsize (&d->chr, cols, rows); +} +void mux_set_focus (Chardev *chr,int focus) {MuxChardev *d =MUX_CHARDEV (chr);
@@ -385,6 +391,7 @@ static void char_mux_class_init(ObjectClass *oc, void *data) cc->chr_be_event = mux_chr_be_event; cc->chr_machine_done = open_muxes; cc->chr_update_read_handler = mux_chr_update_read_handlers; + cc->chr_get_winsize = mux_chr_get_winsize; }static const TypeInfo char_mux_type_info = {
include/chardev/char-fe.h +11 -0
@@ -154,6 +154,17 @@ int qemu_chr_fe_wait_connected(CharBackend *be, Error **errp);*/ void qemu_chr_fe_set_echo (CharBackend *be,bool echo); +/** +* qemu_chr_fe_get_winsize: +* @cols: the address for storing columns +* @rows: the address for storing rows +* +* Get the terminal size of the backend. +* +* Returns: 0 on success and < 0 on error +*/ +int qemu_chr_fe_get_winsize (CharBackend *be,uint16_t *cols,uint16_t *rows); +/** * qemu_chr_fe_set_open: *
include/chardev/char.h +1 -0
@@ -276,6 +276,7 @@ typedef struct ChardevClass {void (*chr_be_event)(Chardev *s,QEMUChrEvent event);/* Return 0 if succeeded, 1 if failed */ int (*chr_machine_done)(Chardev *chr); +int (*chr_get_winsize)(Chardev *chr,uint16_t *cols,uint16_t *rows); } ChardevClass; Chardev *qemu_chardev_new(const char *id, const char *typename,
Signature · 1 line
2.27.0
PATCH v2 3/6 chardev: add support for notifying about terminal resizes
In reply to Szymon LukaszAdd a new chardev event, CHR_EVENT_RESIZE, which a backend should trigger if detects the size of the connected terminal changed.
--- backends/cryptodev-vhost-user.c | 1 + chardev/char.c | 1 + hw/block/vhost-user-blk.c | 1 + hw/char/terminal3270.c | 1 + hw/char/virtio-console.c | 1 + hw/ipmi/ipmi_bmc_extern.c | 1 + hw/usb/ccid-card-passthru.c | 1 + hw/usb/dev-serial.c | 1 + hw/usb/redirect.c | 1 + include/chardev/char.h | 1 + monitor/hmp.c | 1 + monitor/qmp.c | 1 + net/vhost-user.c | 1 + 13 files changed, 13 insertions(+)
backends/cryptodev-vhost-user.c +1 -0
@@ -174,6 +174,7 @@ static void cryptodev_vhost_user_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
chardev/char.c +1 -0
@@ -74,6 +74,7 @@ void qemu_chr_be_event(Chardev *s, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/block/vhost-user-blk.c +1 -0
@@ -403,6 +403,7 @@ static void vhost_user_blk_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/char/terminal3270.c +1 -0
@@ -169,6 +169,7 @@ static void chr_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/char/virtio-console.c +1 -0
@@ -165,6 +165,7 @@ static void chr_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/ipmi/ipmi_bmc_extern.c +1 -0
@@ -439,6 +439,7 @@ static void chr_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/usb/ccid-card-passthru.c +1 -0
@@ -321,6 +321,7 @@ static void ccid_card_vscard_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT :case CHR_EVENT_CLOSED : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/usb/dev-serial.c +1 -0
@@ -507,6 +507,7 @@ static void usb_serial_event(void *opaque, QEMUChrEvent event)break ;case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
hw/usb/redirect.c +1 -0
@@ -1383,6 +1383,7 @@ static void usbredir_chardev_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
include/chardev/char.h +1 -0
@@ -22,6 +22,7 @@ typedef enum {CHR_EVENT_OPENED ,/* new connection established */ CHR_EVENT_MUX_IN ,/* mux-focus was set to this terminal */ CHR_EVENT_MUX_OUT ,/* mux-focus will move on */ +CHR_EVENT_RESIZE ,/* the terminal size of the chardev changed */ CHR_EVENT_CLOSED /* connection closed. NOTE: currently this event * is only bound to the read port of the chardev. * Normally the read port and write port of a
monitor/hmp.c +1 -0
@@ -1373,6 +1373,7 @@ static void monitor_event(void *opaque, QEMUChrEvent event)break ;case CHR_EVENT_BREAK : +case CHR_EVENT_RESIZE :/* Ignored */ break ; }
monitor/qmp.c +1 -0
@@ -371,6 +371,7 @@ static void monitor_qmp_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
net/vhost-user.c +1 -0
@@ -297,6 +297,7 @@ static void net_vhost_user_event(void *opaque, QEMUChrEvent event)case CHR_EVENT_BREAK :case CHR_EVENT_MUX_IN :case CHR_EVENT_MUX_OUT : +case CHR_EVENT_RESIZE :/* Ignore */ break ; }
Signature · 1 line
2.27.0
PATCH v2 4/6 char-stdio: add support for the terminal size
In reply to Szymon LukaszImplement chr_get_winsize for the stdio backend and trigger CHR_EVENT_RESIZE upon SIGWINCH delivery.
--- chardev/char-stdio.c | 35 +++++++++++++++++++++++++++++++++++ 1 file changed, 35 insertions(+)
chardev/char-stdio.c +35 -0
@@ -34,7 +34,9 @@#include "chardev/char-win-stdio.h" #else #include <termios.h> +#include <sys/ioctl.h> #include "chardev/char-fd.h" +#include "qemu/main-loop.h" #endif #ifndef _WIN32
@@ -45,6 +47,13 @@ static bool stdio_in_use;static bool stdio_allow_signal;static bool stdio_echo_state; +typedef struct { +FDChardev parent; +Notifier resize_notifier; +}StdioChardev ; + +#define STDIO_CHARDEV (obj) OBJECT_CHECK(StdioChardev, (obj), TYPE_CHARDEV_STDIO) +static void term_exit (void ) {if (stdio_in_use) {
@@ -82,11 +91,32 @@ static void term_stdio_handler(int sig)qemu_chr_set_echo_stdio (NULL , stdio_echo_state); } +static int qemu_chr_get_winsize_stdio (Chardev *chr,uint16_t *cols, +uint16_t *rows) +{ +struct winsize ws; + +if (ioctl (1 ,TIOCGWINSZ , &ws) <0 ) { +return -1 ; + } + + *cols = ws.ws_col; + *rows = ws.ws_row; +return 0 ; +} + +static void term_resize_notify (Notifier *n,void *data) +{ +StdioChardev *s =container_of (n, StdioChardev, resize_notifier); +qemu_chr_be_event (CHARDEV (s),CHR_EVENT_RESIZE ); +} +static void qemu_chr_open_stdio (Chardev *chr,ChardevBackend *backend,bool *be_opened,Error **errp) { +StdioChardev *s =STDIO_CHARDEV (chr);ChardevStdio *opts = backend->u.stdio.data;struct sigaction act;
@@ -116,6 +146,9 @@ static void qemu_chr_open_stdio(Chardev *chr, stdio_allow_signal = opts->signal; }qemu_chr_set_echo_stdio (chr, false); + + s->resize_notifier.notify = term_resize_notify; +sigwinch_add_notifier (&s->resize_notifier); }#endif
@@ -139,6 +172,7 @@ static void char_stdio_class_init(ObjectClass *oc, void *data)#ifndef _WIN32 cc->open = qemu_chr_open_stdio; cc->chr_set_echo = qemu_chr_set_echo_stdio; + cc->chr_get_winsize = qemu_chr_get_winsize_stdio;#endif }
@@ -155,6 +189,7 @@ static const TypeInfo char_stdio_type_info = { .parent = TYPE_CHARDEV_WIN_STDIO, #else .parent = TYPE_CHARDEV_FD, + .instance_size = sizeof(StdioChardev), #endif .instance_finalize = char_stdio_finalize, .class_init = char_stdio_class_init,
Signature · 1 line
2.27.0
PATCH v2 5/6 virtio-serial-bus: add terminal resize messages
In reply to Szymon LukaszImplement the part of the virtio spec that allows to notify the virtio driver about terminal resizes. The virtio spec contains two methods to achieve that:
For legacy drivers, we have only one port and we put the terminal size in the config space and inject the config changed interrupt.
For multiport devices, we use the control virtqueue to send a packet containing the terminal size. Note that the Linux kernel expects the fields indicating the number of rows and columns in a packet to be in a different order than the one specified in the current version of the virtio spec. We follow the Linux implementation, so hopefully there is no implementation of this functionality conforming to the spec.
--- hw/char/trace-events | 1 + hw/char/virtio-serial-bus.c | 42 +++++++++++++++++++++++++++++-- hw/core/machine.c | 1 + include/hw/virtio/virtio-serial.h | 5 ++++ 4 files changed, 47 insertions(+), 2 deletions(-)
hw/char/trace-events +1 -0
@@ -10,6 +10,7 @@ serial_ioport_write(uint16_t addr, uint8_t value) "write addr 0x%02x val 0x%02x" # virtio-serial-bus.c virtio_serial_send_control_event(unsigned int port, uint16_t event, uint16_t value) "port %u, event %u, value %u" +virtio_serial_send_console_resize(unsigned int port, uint16_t cols, uint16_t rows) "port %u, cols %u, rows %u" virtio_serial_throttle_port(unsigned int port, bool throttle) "port %u, throttle %d" virtio_serial_handle_control_message(uint16_t event, uint16_t value) "event %u, value %u" virtio_serial_handle_control_message_port(unsigned int port) "port %u"
hw/char/virtio-serial-bus.c +40 -2
60 lines not on this page · raw message
hw/core/machine.c +1 -0
@@ -30,6 +30,7 @@GlobalProperty hw_compat_5_0[] = { {"virtio-balloon-device" ,"page-poison" ,"false" }, + {"virtio-serial-device" ,"console-size" ,"off" }, };const size_t hw_compat_5_0_len =G_N_ELEMENTS (hw_compat_5_0);
include/hw/virtio/virtio-serial.h +5 -0
@@ -188,6 +188,8 @@ struct VirtIOSerial {virtio_serial_conf serial;uint64_t host_features; + +uint16_t port0_cols, port0_rows; };/* Interface to the virtio-serial bus */
@@ -222,6 +224,9 @@ size_t virtio_serial_guest_ready(VirtIOSerialPort *port);*/ void virtio_serial_throttle_port (VirtIOSerialPort *port,bool throttle); +void virtio_serial_send_console_resize (VirtIOSerialPort *port, +uint16_t cols,uint16_t rows); +#define TYPE_VIRTIO_SERIAL "virtio-serial-device"#define VIRTIO_SERIAL (obj) \ OBJECT_CHECK(VirtIOSerial, (obj), TYPE_VIRTIO_SERIAL)
Signature · 1 line
2.27.0
Re: [PATCH v2 5/6] virtio-serial-bus: add terminal resize messages
Quoted 7 lines · Szymon Lukasz
Implement the part of the virtio spec that allows to notify the virtio driver about terminal resizes. The virtio spec contains two methods to achieve that:
For legacy drivers, we have only one port and we put the terminal size in the config space and inject the config changed interrupt.
Quoted 8 lines
For multiport devices, we use the control virtqueue to send a packet containing the terminal size. Note that the Linux kernel expects the fields indicating the number of rows and columns in a packet to be in a different order than the one specified in the current version of the virtio spec. We follow the Linux implementation, so hopefully there is no implementation of this functionality conforming to the spec.
Signed-off-by Szymon Lukasz <[email protected]>
Looks right to me:
Quoted 123 lines
--- hw/char/trace-events | 1 + hw/char/virtio-serial-bus.c | 42 +++++++++++++++++++++++++++++-- hw/core/machine.c | 1 + include/hw/virtio/virtio-serial.h | 5 ++++ 4 files changed, 47 insertions(+), 2 deletions(-)hw/char/trace-events +1
-0@@ -10,6 +10,7 @@ serial_ioport_write(uint16_t addr, uint8_t value) "write addr 0x%02x val 0x%02x" # virtio-serial-bus.c virtio_serial_send_control_event(unsigned int port, uint16_t event, uint16_t value) "port %u, event %u, value %u" +virtio_serial_send_console_resize(unsigned int port, uint16_t cols, uint16_t rows) "port %u, cols %u, rows %u" virtio_serial_throttle_port(unsigned int port, bool throttle) "port %u, throttle %d" virtio_serial_handle_control_message(uint16_t event, uint16_t value) "event %u, value %u" virtio_serial_handle_control_message_port(unsigned int port) "port %u"hw/char/virtio-serial-bus.c +40
-2@@ -261,6 +261,42 @@ static size_t send_control_event(VirtIOSerial *vser, uint32_t port_id,return send_control_msg (vser, &cpkt,sizeof (cpkt)); } +/* +* This struct should be added to the Linux kernel uapi headers +* and later imported to standard-headers/linux/virtio_console.h +*/ +struct virtio_console_resize { +__virtio16 rows; +__virtio16 cols; +}; + +void virtio_serial_send_console_resize (VirtIOSerialPort *port, +uint16_t cols,uint16_t rows) +{ +VirtIOSerial *vser = port->vser; +VirtIODevice *vdev =VIRTIO_DEVICE (vser); + +if (virtio_vdev_has_feature (vdev,VIRTIO_CONSOLE_F_MULTIPORT )) { +struct { +struct virtio_console_control control; +struct virtio_console_resize resize; + } buffer; + +virtio_stl_p (vdev, &buffer.control.id, port->id); +virtio_stw_p (vdev, &buffer.control.event,VIRTIO_CONSOLE_RESIZE ); +virtio_stw_p (vdev, &buffer.resize.cols, cols); +virtio_stw_p (vdev, &buffer.resize.rows, rows); + +trace_virtio_serial_send_console_resize (port->id, cols, rows); +send_control_msg (vser, &buffer,sizeof (buffer)); + + }else if (virtio_vdev_has_feature (vdev,VIRTIO_CONSOLE_F_SIZE )) { + vser->port0_cols = cols; + vser->port0_rows = rows; +virtio_notify_config (vdev); + } +} +/* Functions for use inside qemu to open and read from/write to ports */ int virtio_serial_open (VirtIOSerialPort *port) {@@ -572,8 +608,8 @@ static void get_config(VirtIODevice *vdev, uint8_t *config_data)struct virtio_console_config *config = (struct virtio_console_config *)config_data; - config->cols =0 ; - config->rows =0 ; + config->cols =virtio_tswap16 (vdev, vser->port0_cols); + config->rows =virtio_tswap16 (vdev, vser->port0_rows); config->max_nr_ports =virtio_tswap32 (vdev, vser->serial.max_virtserial_ports); }@@ -1168,6 +1204,8 @@ static Property virtio_serial_properties[] = { 31),DEFINE_PROP_BIT64 ("emergency-write" , VirtIOSerial, host_features,VIRTIO_CONSOLE_F_EMERG_WRITE , true), +DEFINE_PROP_BIT64 ("console-size" , VirtIOSerial, host_features, +VIRTIO_CONSOLE_F_SIZE , true), DEFINE_PROP_END_OF_LIST(), };hw/core/machine.c +1
-0@@ -30,6 +30,7 @@GlobalProperty hw_compat_5_0[] = { {"virtio-balloon-device" ,"page-poison" ,"false" }, + {"virtio-serial-device" ,"console-size" ,"off" }, };const size_t hw_compat_5_0_len =G_N_ELEMENTS (hw_compat_5_0);include/hw/virtio/virtio-serial.h +5
-0@@ -188,6 +188,8 @@ struct VirtIOSerial {virtio_serial_conf serial;uint64_t host_features; + +uint16_t port0_cols, port0_rows; };/* Interface to the virtio-serial bus */ @@ -222,6 +224,9 @@ size_t virtio_serial_guest_ready(VirtIOSerialPort *port);*/ void virtio_serial_throttle_port (VirtIOSerialPort *port,bool throttle); +void virtio_serial_send_console_resize (VirtIOSerialPort *port, +uint16_t cols,uint16_t rows); +#define TYPE_VIRTIO_SERIAL "virtio-serial-device"#define VIRTIO_SERIAL (obj) \ OBJECT_CHECK(VirtIOSerial, (obj), TYPE_VIRTIO_SERIAL)Signature · 1 line
2.27.0
PATCH v2 6/6 virtio-console: notify the guest about terminal resizes
In reply to Szymon LukaszIf a virtio serial port is a console port forward terminal resize messages from the chardev backend to the guest.
--- hw/char/virtio-console.c | 64 +++++++++++++++++++++++++++++++++++++--- 1 file changed, 60 insertions(+), 4 deletions(-)
hw/char/virtio-console.c +60 -4
@@ -29,6 +29,7 @@ typedef struct VirtConsole {CharBackend chr;guint watch; +uint16_t cols, rows; } VirtConsole;/*
@@ -104,6 +105,36 @@ static ssize_t flush_buf(VirtIOSerialPort *port,return ret; } +static void virtconsole_send_resize (VirtIOSerialPort *port) +{ +uint16_t cols, rows; +VirtConsole *vcon =VIRTIO_CONSOLE (port); + +/* +* We probably shouldn't send these messages before +* we told the guest it is a console port (which we do +* by sending VIRTIO_CONSOLE_CONSOLE_PORT message). +* Instead of adding a new field to the device state +* lets just use the guest_connected field for that purpose +* since the guest should not care about the terminal size +* before opening the port. +*/ +if (!port->guest_connected) { +return ; + } + +if (qemu_chr_fe_get_winsize (&vcon->chr, &cols, &rows) <0 ) { + cols =0 ; + rows =0 ; + } + +if (cols != vcon->cols || rows != vcon->rows) { + vcon->cols = cols; + vcon->rows = rows; +virtio_serial_send_console_resize (port, cols, rows); + } +} +/* Callback function that's called when the guest opens/closes the port */ static void set_guest_connected (VirtIOSerialPort *port,int guest_connected) {
@@ -111,7 +142,9 @@ static void set_guest_connected(VirtIOSerialPort *port, int guest_connected)DeviceState *dev =DEVICE (port);VirtIOSerialPortClass *k =VIRTIO_SERIAL_PORT_GET_CLASS (port); -if (!k->is_console) { +if (k->is_console) { +virtconsole_send_resize (port); + }else {qemu_chr_fe_set_open (&vcon->chr, guest_connected); }
@@ -171,6 +204,22 @@ static void chr_event(void *opaque, QEMUChrEvent event) } } +static void chr_event_console (void *opaque,QEMUChrEvent event) +{ +VirtConsole *vcon = opaque; +VirtIOSerialPort *port =VIRTIO_SERIAL_PORT (vcon); + +switch (event) { +case CHR_EVENT_OPENED : +case CHR_EVENT_RESIZE : +trace_virtio_console_chr_event (port->id, event); +virtconsole_send_resize (port); +break ; +default : +break ; + } +} +static int chr_be_change (void *opaque) {VirtConsole *vcon = opaque;
@@ -179,7 +228,9 @@ static int chr_be_change(void *opaque)if (k->is_console) {qemu_chr_fe_set_handlers (&vcon->chr, chr_can_read, chr_read, -NULL , chr_be_change, vcon,NULL , true); + chr_event_console, chr_be_change, + vcon,NULL , true); +virtconsole_send_resize (port); }else {qemu_chr_fe_set_handlers (&vcon->chr, chr_can_read, chr_read, chr_event, chr_be_change, vcon,NULL , false);
@@ -207,7 +258,7 @@ static void virtconsole_enable_backend(VirtIOSerialPort *port, bool 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, + k->is_console ? chr_event_console : chr_event, chr_be_change, vcon,NULL , false); }else { qemu_chr_fe_set_handlers(&vcon->chr, NULL, NULL, NULL,
@@ -227,6 +278,11 @@ static void virtconsole_realize(DeviceState *dev, Error **errp)return ; } +if (k->is_console) { + vcon->cols = (uint16_t )-1 ; + vcon->rows = (uint16_t )-1 ; + } +if (qemu_chr_fe_backend_connected (&vcon->chr)) {/* * For consoles we don't block guest data transfer just
@@ -239,7 +295,7 @@ static void virtconsole_realize(DeviceState *dev, Error **errp)*/ if (k->is_console) {qemu_chr_fe_set_handlers (&vcon->chr, chr_can_read, chr_read, -NULL , chr_be_change, + chr_event_console, chr_be_change, vcon,NULL , true);virtio_serial_open (port); }else {
Signature · 1 line
2.27.0
Quoted 5 lines · Szymon Lukasz
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in
virtio_console_resizestruct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.
Well this is a bit of a mess :-(
The main virtio_console_config struct has cols, then rows.
The Linux impl of resizing appears to have arrived in 2010, and created a new struct with rows, then cols.
commit 8345adbf96fc1bde7d9846aadbe5af9b2ae90882
Author: Amit Shah <[email protected]> Date: Thu May 6 02:05:09 2010 +0530
virtio: console: Accept console size along with resize control message
The VIRTIO_CONSOLE_RESIZE control message sent to us by the host now contains the new {rows, cols} values for the console. This ensures each console port gets its own size, and we don't depend on the config-space rows and cols values at all now.
The virtio spec documenting this came 4 years later in 2014 and documented the resize struct with cols, then rows, which differs from Linux impl, but matches ordering of the main virtio_console_config:
commit 908cfaa782e950d6656d947599d7a6c9fb16cad1
Author: rusty <rusty@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652> Date: Wed Feb 12 03:15:57 2014 +0000
Feedback #6: Applied
As per minutes:
https://lists.oasis-open.org/archives/virtio/201402/msg00121.html
git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio@237 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652
I can understand why it is desirable for the resize struct to match the order of the initial config struct. I'm guessing it just wasn't realized that the Linux impl was inverted for resize
The FreeBSD impl of virtio-console doesn't do resize:
https://github.com/freebsd/freebsd/blob/master/sys/dev/virtio/console/virtio_console.c#L874
Not sure what other impls are going to be around, but I feel like Linux is going to be the most commonly deployed by orders of magnitude.
So I'd say QEMU should match Linux, and the spec should be fixed.
Have you reported this bug to the virtio spec people directly yet ?
I don't see an issue open at
https://github.com/oasis-tcs/virtio-spec/issues/
so I think one should be filed there
Regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 53 lines · Daniel P. Berrangé
Quoted 5 lines · Szymon Lukasz
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in
virtio_console_resizestruct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.Well this is a bit of a mess :-(
The main
virtio_console_configstruct has cols, then rows.The Linux impl of resizing appears to have arrived in 2010, and created a new struct with rows, then cols.
commit
8345adbf96fc1bde7d9846aadbe5af9b2ae90882Author: Amit Shah <[email protected]> Date: Thu May 6 02:05:09 2010 +0530virtio: console: Accept console size along with resize control message
The
VIRTIO_CONSOLE_RESIZEcontrol message sent to us by the host now contains the new {rows, cols} values for the console. This ensures each console port gets its own size, and we don't depend on the config-space rows and cols values at all now.Signed-off-by Amit Shah <[email protected]>CC Christian Borntraeger <[email protected]>CC Kusanagi Kouichi <[email protected]>Signed-off-by Rusty Russell <[email protected]>The virtio spec documenting this came 4 years later in 2014 and documented the resize struct with cols, then rows, which differs from Linux impl, but matches ordering of the main
virtio_console_config:commit
908cfaa782e950d6656d947599d7a6c9fb16cad1Author: rusty <rusty@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652> Date: Wed Feb 12 03:15:57 2014 +0000Feedback #6: Applied
As per minutes:
https://lists.oasis-open.org/archives/virtio/201402/msg00121.htmlSigned-off-by Rusty Russell <[email protected]>git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio@237 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652
I can understand why it is desirable for the resize struct to match the order of the initial config struct. I'm guessing it just wasn't realized that the Linux impl was inverted for resize
Quoted 8 lines
The FreeBSD impl of virtio-console doesn't do resize:
https://github.com/freebsd/freebsd/blob/master/sys/dev/virtio/console/virtio_console.c#L874
Not sure what other impls are going to be around, but I feel like Linux is going to be the most commonly deployed by orders of magnitude.
So I'd say QEMU should match Linux, and the spec should be fixed.
I had the same thoughts. I will ask for a change in the spec.
Quoted 17 lines
Have you reported this bug to the virtio spec people directly yet ?
I don't see an issue open at
https://github.com/oasis-tcs/virtio-spec/issues/
so I think one should be filed there
Regards,
DanielSignature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 65 lines · Daniel P. Berrangé
Quoted 5 lines · Szymon Lukasz
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in
virtio_console_resizestruct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.Well this is a bit of a mess :-(
The main
virtio_console_configstruct has cols, then rows.The Linux impl of resizing appears to have arrived in 2010, and created a new struct with rows, then cols.
commit
8345adbf96fc1bde7d9846aadbe5af9b2ae90882Author: Amit Shah <[email protected]> Date: Thu May 6 02:05:09 2010 +0530virtio: console: Accept console size along with resize control message
The
VIRTIO_CONSOLE_RESIZEcontrol message sent to us by the host now contains the new {rows, cols} values for the console. This ensures each console port gets its own size, and we don't depend on the config-space rows and cols values at all now.Signed-off-by Amit Shah <[email protected]>CC Christian Borntraeger <[email protected]>CC Kusanagi Kouichi <[email protected]>Signed-off-by Rusty Russell <[email protected]>The virtio spec documenting this came 4 years later in 2014 and documented the resize struct with cols, then rows, which differs from Linux impl, but matches ordering of the main
virtio_console_config:commit
908cfaa782e950d6656d947599d7a6c9fb16cad1Author: rusty <rusty@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652> Date: Wed Feb 12 03:15:57 2014 +0000Feedback #6: Applied
As per minutes:
https://lists.oasis-open.org/archives/virtio/201402/msg00121.htmlSigned-off-by Rusty Russell <[email protected]>git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio@237 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652
I can understand why it is desirable for the resize struct to match the order of the initial config struct. I'm guessing it just wasn't realized that the Linux impl was inverted for resize
The FreeBSD impl of virtio-console doesn't do resize:
https://github.com/freebsd/freebsd/blob/master/sys/dev/virtio/console/virtio_console.c#L874
Not sure what other impls are going to be around, but I feel like Linux is going to be the most commonly deployed by orders of magnitude.
So I'd say QEMU should match Linux, and the spec should be fixed.
Have you reported this bug to the virtio spec people directly yet ?
Quoted 8 lines
I don't see an issue open at
https://github.com/oasis-tcs/virtio-spec/issues/
so I think one should be filed there
Regards,
Daniel
One reports defects on the virtio-comments mailing list, issue tracker is just for tracking spec changes.
Quoted 4 lines
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 70 lines · Michael S. Tsirkin
Quoted 69 lines · Daniel P. Berrangé
Quoted 5 lines · Szymon Lukasz
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in
virtio_console_resizestruct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.Well this is a bit of a mess :-(
The main
virtio_console_configstruct has cols, then rows.The Linux impl of resizing appears to have arrived in 2010, and created a new struct with rows, then cols.
commit
8345adbf96fc1bde7d9846aadbe5af9b2ae90882Author: Amit Shah <[email protected]> Date: Thu May 6 02:05:09 2010 +0530virtio: console: Accept console size along with resize control message
The
VIRTIO_CONSOLE_RESIZEcontrol message sent to us by the host now contains the new {rows, cols} values for the console. This ensures each console port gets its own size, and we don't depend on the config-space rows and cols values at all now.Signed-off-by Amit Shah <[email protected]>CC Christian Borntraeger <[email protected]>CC Kusanagi Kouichi <[email protected]>Signed-off-by Rusty Russell <[email protected]>The virtio spec documenting this came 4 years later in 2014 and documented the resize struct with cols, then rows, which differs from Linux impl, but matches ordering of the main
virtio_console_config:commit
908cfaa782e950d6656d947599d7a6c9fb16cad1Author: rusty <rusty@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652> Date: Wed Feb 12 03:15:57 2014 +0000Feedback #6: Applied
As per minutes:
https://lists.oasis-open.org/archives/virtio/201402/msg00121.htmlSigned-off-by Rusty Russell <[email protected]>git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio@237 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652
I can understand why it is desirable for the resize struct to match the order of the initial config struct. I'm guessing it just wasn't realized that the Linux impl was inverted for resize
The FreeBSD impl of virtio-console doesn't do resize:
https://github.com/freebsd/freebsd/blob/master/sys/dev/virtio/console/virtio_console.c#L874
Not sure what other impls are going to be around, but I feel like Linux is going to be the most commonly deployed by orders of magnitude.
So I'd say QEMU should match Linux, and the spec should be fixed.
Have you reported this bug to the virtio spec people directly yet ?
I don't see an issue open at
Quoted 8 lines
Quoted 4 lines
so I think one should be filed there
Regards,
DanielOne reports defects on the virtio-comments mailing list, issue tracker is just for tracking spec changes.
NB That contradicts what the CONTRIBUTING.md file in virtio-spec says, which welcomes use of the issue tracker:
"Persons who are not TC members are invited to open issues and
provide comments using this repository's GitHub Issues tracking facility or using the TC's comment list. "
https://github.com/oasis-tcs/virtio-spec/blob/master/CONTRIBUTING.md
Regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 80 lines · Daniel P. Berrangé
Quoted 70 lines · Michael S. Tsirkin
Quoted 69 lines · Daniel P. Berrangé
Quoted 5 lines · Szymon Lukasz
Also there is a problem with the virtio spec and Linux Kernel implementation, the order of fields in
virtio_console_resizestruct differs between the kernel and the spec. I do not know if there is any implementation of the virtio-console driver that handles resize messages and uses a different order than Linux.Well this is a bit of a mess :-(
The main
virtio_console_configstruct has cols, then rows.The Linux impl of resizing appears to have arrived in 2010, and created a new struct with rows, then cols.
commit
8345adbf96fc1bde7d9846aadbe5af9b2ae90882Author: Amit Shah <[email protected]> Date: Thu May 6 02:05:09 2010 +0530virtio: console: Accept console size along with resize control message
The
VIRTIO_CONSOLE_RESIZEcontrol message sent to us by the host now contains the new {rows, cols} values for the console. This ensures each console port gets its own size, and we don't depend on the config-space rows and cols values at all now.Signed-off-by Amit Shah <[email protected]>CC Christian Borntraeger <[email protected]>CC Kusanagi Kouichi <[email protected]>Signed-off-by Rusty Russell <[email protected]>The virtio spec documenting this came 4 years later in 2014 and documented the resize struct with cols, then rows, which differs from Linux impl, but matches ordering of the main
virtio_console_config:commit
908cfaa782e950d6656d947599d7a6c9fb16cad1Author: rusty <rusty@0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652> Date: Wed Feb 12 03:15:57 2014 +0000Feedback #6: Applied
As per minutes:
https://lists.oasis-open.org/archives/virtio/201402/msg00121.htmlSigned-off-by Rusty Russell <[email protected]>git-svn-id: https://tools.oasis-open.org/version-control/svn/virtio@237 0c8fb4dd-22a2-4bb5-bc14-6c75a5f43652
I can understand why it is desirable for the resize struct to match the order of the initial config struct. I'm guessing it just wasn't realized that the Linux impl was inverted for resize
The FreeBSD impl of virtio-console doesn't do resize:
https://github.com/freebsd/freebsd/blob/master/sys/dev/virtio/console/virtio_console.c#L874
Not sure what other impls are going to be around, but I feel like Linux is going to be the most commonly deployed by orders of magnitude.
So I'd say QEMU should match Linux, and the spec should be fixed.
Have you reported this bug to the virtio spec people directly yet ?
I don't see an issue open at
Quoted 8 lines
Quoted 4 lines
so I think one should be filed there
Regards,
DanielOne reports defects on the virtio-comments mailing list, issue tracker is just for tracking spec changes.
Quoted 8 lines
NB That contradicts what the CONTRIBUTING.md file in virtio-spec says, which welcomes use of the issue tracker:
"Persons who are not TC members are invited to open issues and
provide comments using this repository's GitHub Issues tracking facility or using the TC's comment list. "https://github.com/oasis-tcs/virtio-spec/blob/master/CONTRIBUTING.md
I know, it's confusing. See
https://github.com/oasis-tcs/virtio-spec#use-of-github-issues
OASIS admins asked us not to change CONTRIBUTING.md so we had to resort to adding clarifications in a separate file like that.
Quoted 7 lines
Regards,
DanielSignature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 4 lines · Szymon Lukasz
In this series resize notifications are only supported for the stdio backend but I think it should be easy to add support for the vc backend. Support for tty/serial backends is complicated by the fact that there is no clean way to detect resizes of the underlying terminal.
In a libvirt managed scenario it is typical to have the virtio console connected to a UNIX socket. It would be desirable to have a way to deal with resizes there.
QEMU socket chardev (TCP & UNIX socket) supports a "telnet" protocol addition. Currently it doesn't almost nothing useful, but in theory we could wire up support for the telnet resize message:
https://tools.ietf.org/html/rfc1073
Another option is to allow dealing with resizes out of band, via the QMP monitor. ie we can introduce a qmp_chardev_winsize command that a client app can use to trigger the resize message to the guest OS. From libvirt's POV, this would be quite easy to support & useful to have.
Regards,
Daniel
Signature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|
Quoted 14 lines · Daniel P. Berrangé
Quoted 4 lines · Szymon Lukasz
In this series resize notifications are only supported for the stdio backend but I think it should be easy to add support for the vc backend. Support for tty/serial backends is complicated by the fact that there is no clean way to detect resizes of the underlying terminal.
In a libvirt managed scenario it is typical to have the virtio console connected to a UNIX socket. It would be desirable to have a way to deal with resizes there.
QEMU socket chardev (TCP & UNIX socket) supports a "telnet" protocol addition. Currently it doesn't almost nothing useful, but in theory we could wire up support for the telnet resize message:
Quoted 8 lines
https://tools.ietf.org/html/rfc1073Another option is to allow dealing with resizes out of band, via the QMP monitor. ie we can introduce a
qmp_chardev_winsizecommand that a client app can use to trigger the resize message to the guest OS. From libvirt's POV, this would be quite easy to support & useful to have.
I will look into that.
Quoted 9 lines
Regards,
DanielSignature · 3 lines
|: https://berrange.com -o- https://www.flickr.com/photos/dberrange :| |: https://libvirt.org -o- https://fstop138.berrange.com :| |: https://entangle-photo.org -o- https://www.instagram.com/dberrange :|