chardev: convert leftover glib APIs to use dedicate gcontext
v3 cover chardev: convert leftover glib APIs to use dedicate gcontext
v3:
- rename function to qemu_chr_timeout_add_ms() [Stefan]
- add comment on return code [Stefan]
- add comment in commit message on why change GSource name [Stefan]
v2:
- add r-bs
- fix patch 3 on some s->ms conversion [Marc-André]
There were existing work that tried to allow chardev to be run in a dedicated gcontext rather than the default main context/thread. Basically that work passed in the correct gcontext during g_source_attach(). However, I found something missing along the way, that some legacy glib APIs are used by chardev code which take the main context as default:
g_timeout_add_seconds
g_timeout_add
g_idle_add
To fully allow the chardevs to be run in dedicated gcontext, we need to convert all these legacy APIs into g_source_attach() calls as well, with the correct gcontext passed in.
This series tries to clean the rest of things up.
I picked up patch 1 from monitor-oob series into this series (which is a missing of chardev frontend call fix for g_source_attach()), so that this series can be a complete fix.
Please review. Thanks,
Peter Xu (3):
chardev: use backend chr context when watch for fe
chardev: let g_idle_add() be with chardev gcontext
chardev: introduce qemu_chr_timeout_add_ms()
chardev/char-fe.c | 2 +- chardev/char-pty.c | 16 ++++++++-------- chardev/char-socket.c | 6 ++++-- chardev/char.c | 21 +++++++++++++++++++++ hw/char/terminal3270.c | 7 ++++--- include/chardev/char.h | 3 +++ 6 files changed, 41 insertions(+), 14 deletions(-)
Signature · 1 line
2.14.3
PATCH v3 1/3 chardev: use backend chr context when watch for fe
In commit 6bbb6c0644 ("chardev: use per-dev context for io_add_watch_poll", 2017-09-22) all the chardev watches are converted to use per-chardev gcontext to support chardev to be run outside default main thread. However that's still missing one call from the frontend code. Touch that up.
--- chardev/char-fe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
chardev/char-fe.c +1 -1
@@ -356,7 +356,7 @@ guint qemu_chr_fe_add_watch(CharBackend *be, GIOCondition cond, }g_source_set_callback (src, (GSourceFunc )func, user_data,NULL ); - tag =g_source_attach (src,NULL ); + tag =g_source_attach (src, s->gcontext);g_source_unref (src);return tag;
Signature · 1 line
2.14.3
PATCH v3 2/3 chardev: let g_idle_add() be with chardev gcontext
In reply to Peter XuThe idle task will be attached to main gcontext even if the chardev backend is running in another gcontext. Fix the only caller by extending the g_idle_add() logic into the more powerful g_source_attach(). It's basically g_idle_add_full() implementation, but with the chardev's gcontext passed in.
--- chardev/char-pty.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
chardev/char-pty.c +6 -1
@@ -210,9 +210,14 @@ static void pty_chr_state(Chardev *chr, int connected) s->timer_tag =0 ; }if (!s->connected) { +GSource *source =g_idle_source_new (); +g_assert (s->open_tag ==0 ); s->connected =1 ; - s->open_tag =g_idle_add (qemu_chr_be_generic_open_func, chr); +g_source_set_callback (source, qemu_chr_be_generic_open_func, + chr,NULL ); + s->open_tag =g_source_attach (source, chr->gcontext); +g_source_unref (source); }if (!chr->gsource) { chr->gsource = io_add_watch_poll(chr, s->ioc,
Signature · 1 line
2.14.3
PATCH v3 3/3 chardev: introduce qemu_chr_timeout_add_ms()
In reply to Peter XuIt's a replacement of g_timeout_add[_seconds]() for chardevs. Chardevs now can have dedicated gcontext, we should always bind chardev tasks onto those gcontext rather than the default main context. Since there are quite a few of g_timeout_add[_seconds]() callers, a new function qemu_chr_timeout_add_ms() is introduced.
One thing to mention is that, terminal3270 is still always running on main gcontext. However let's convert that as well since it's still part of chardev codes and in case one day we'll miss that when we move it out of main gcontext too.
Also, in pty_chr_rearm_timer() the GSource name is not user visible, and does not really matter much. Unify the two calls into one.
--- chardev/char-pty.c | 9 ++------- chardev/char-socket.c | 6 ++++-- chardev/char.c | 21 +++++++++++++++++++++ hw/char/terminal3270.c | 7 ++++--- include/chardev/char.h | 3 +++ 5 files changed, 34 insertions(+), 12 deletions(-)
chardev/char-pty.c +2 -7
@@ -78,13 +78,8 @@ static void pty_chr_rearm_timer(Chardev *chr, int ms) s->timer_tag =0 ; } -if (ms ==1000 ) { - name =g_strdup_printf ("pty-timer-secs-%s" , chr->label); - s->timer_tag =g_timeout_add_seconds (1 , pty_chr_timer, chr); - }else { - name =g_strdup_printf ("pty-timer-ms-%s" , chr->label); - s->timer_tag =g_timeout_add (ms, pty_chr_timer, chr); - } + name =g_strdup_printf ("pty-timer-ms-%s" , chr->label); + s->timer_tag =qemu_chr_timeout_add_ms (chr, ms, pty_chr_timer, chr);g_source_set_name_by_id (s->timer_tag, name);g_free (name); }
chardev/char-socket.c +4 -2
@@ -73,8 +73,10 @@ static void qemu_chr_socket_restart_timer(Chardev *chr)char *name;assert (s->connected ==0 ); - s->reconnect_timer =g_timeout_add_seconds (s->reconnect_time, - socket_reconnect_timeout, chr); + s->reconnect_timer =qemu_chr_timeout_add_ms (chr, + s->reconnect_time *1000 , + socket_reconnect_timeout, + chr); name =g_strdup_printf ("chardev-socket-reconnect-%s" , chr->label);g_source_set_name_by_id (s->reconnect_timer, name);g_free (name);
chardev/char.c +21 -0
@@ -1084,6 +1084,27 @@ void qmp_chardev_send_break(const char *id, Error **errp)qemu_chr_be_event (chr,CHR_EVENT_BREAK ); } +/* +* Add a timeout callback for the chardev (in milliseconds). Please +* use this to add timeout hook for chardev instead of g_timeout_add() +* and g_timeout_add_seconds(), to make sure the gcontext that the +* task bound to is correct. Returns the ID of the timeout GSource in +* gcontext of the chardev. +*/ +guint qemu_chr_timeout_add_ms (Chardev *chr,guint ms,GSourceFunc func, +void *private) +{ +GSource *source =g_timeout_source_new (ms); +guint id; + +assert (func); +g_source_set_callback (source, func, private,NULL ); + id =g_source_attach (source, chr->gcontext); +g_source_unref (source); + +return id; +} +void qemu_chr_cleanup (void ) {object_unparent (get_chardevs_root ());
hw/char/terminal3270.c +4 -3
@@ -94,8 +94,8 @@ static void terminal_read(void *opaque, const uint8_t *buf, int size)g_source_remove (t->timer_tag); t->timer_tag =0 ; } - t->timer_tag =g_timeout_add_seconds (600 , send_timing_mark_cb, t); - + t->timer_tag =qemu_chr_timeout_add_ms (t->chr.chr,600 *1000 , + send_timing_mark_cb, t);memcpy (&t->inv[t->in_len], buf, size); t->in_len += size;if (t->in_len <2 ) {
@@ -157,7 +157,8 @@ static void chr_event(void *opaque, int event)* char-socket.c. Once qemu receives the terminal-type of the * client, mark handshake done and trigger everything rolling again. */ - t->timer_tag =g_timeout_add_seconds (600 , send_timing_mark_cb, t); + t->timer_tag =qemu_chr_timeout_add_ms (t->chr.chr,600 *1000 , + send_timing_mark_cb, t);break ;case CHR_EVENT_CLOSED : sch->curr_status.scsw.dstat =SCSW_DSTAT_DEVICE_END ;
include/chardev/char.h +3 -0
@@ -256,6 +256,9 @@ Chardev *qemu_chardev_new(const char *id, const char *typename,extern int term_escape_char; +guint qemu_chr_timeout_add_ms (Chardev *chr,guint ms,GSourceFunc func, +void *private); +/* console.c */ void qemu_chr_parse_vc (QemuOpts *opts,ChardevBackend *backend,Error **errp);
Signature · 1 line
2.14.3
Quoted 16 lines · Peter Xu
v3:
- rename function toqemu_chr_timeout_add_ms()[Stefan]
- add comment on return code [Stefan]
- add comment in commit message on why change GSource name [Stefan]v2:
- add r-bs
- fix patch 3 on somes->msconversion [Marc-André]There were existing work that tried to allow chardev to be run in a dedicated gcontext rather than the default main context/thread. Basically that work passed in the correct gcontext during
g_source_attach(). However, I found something missing along the way, that some legacy glib APIs are used by chardev code which take the main context as default:
Quoted 3 lines
g_timeout_add_seconds
g_timeout_add
g_idle_add
Self NAK.
I just noticed something more critical: these calls are still managing gsources using IDs rather than GSource objects. That should not work, especially if we want to remove these events using g_source_remove(). That can be fairly dangerous after OOB series merged. I need to use GSource objects to replace the tag IDs.
I'll prepare another version tomorrow. Sorry for the troublesome.
Quoted 29 lines
To fully allow the chardevs to be run in dedicated gcontext, we need to convert all these legacy APIs into
g_source_attach()calls as well, with the correct gcontext passed in.This series tries to clean the rest of things up.
I picked up patch 1 from monitor-oob series into this series (which is a missing of chardev frontend call fix for
g_source_attach()), so that this series can be a complete fix.Please review. Thanks,
Peter Xu (3):
chardev: use backend chr context when watch for fe
chardev: letg_idle_add()be with chardev gcontext
chardev: introduceqemu_chr_timeout_add_ms()chardev/char-fe.c | 2 +- chardev/char-pty.c | 16 ++++++++-------- chardev/char-socket.c | 6 ++++-- chardev/char.c | 21 +++++++++++++++++++++ hw/char/terminal3270.c | 7 ++++--- include/chardev/char.h | 3 +++ 6 files changed, 41 insertions(+), 14 deletions(-)Signature · 1 line
2.14.3
Signature · 1 line
Peter Xu