Message ID | 1356298529-7334-1-git-send-email-alevy@redhat.com |
---|---|
State | New |
Headers | show |
On (Sun) 23 Dec 2012 [23:35:29], Alon Levy wrote: > The target has not seen the guest_connected event via > spice_chr_guest_open or spice_chr_write, and so spice server wrongly > assumes there is no agent active, while the client continues to send > motion events only by the agent channel, which the server ignores. The > net effect is that the mouse is static in the guest. > > By registering the interface on post load spice server will pass on the > agent messages fixing the mouse behavior after migration. > > RHBZ #725965 > > Signed-off-by: Alon Levy <alevy@redhat.com> > --- > spice-qemu-char.c | 42 ++++++++++++++++++++++++++++++++++++++++++ > 1 file changed, 42 insertions(+) I suppose Gerd should pick this up in his tree? I have a couple of questions below, but I ack this approach. > static QLIST_HEAD(, SpiceCharDriver) spice_chars = > @@ -185,18 +190,23 @@ static void spice_chr_close(struct CharDriverState *chr) > printf("%s\n", __func__); > vmc_unregister_interface(s); > QLIST_REMOVE(s, next); > + qemu_free_timer(s->post_load.timer); Also vmstate_unregister()? I'm wondering if there can be a case where this function is called before the timer has had a chance to fire. It can happen if the spice port is hot-unplugged before the guest has had a chance to run on the target. In that case, qemu_del_timer() should be called as well, to ensure the timer doesn't fire with invalid args later. Amit
On (Mon) 24 Dec 2012 [13:09:20], Amit Shah wrote: > On (Sun) 23 Dec 2012 [23:35:29], Alon Levy wrote: > > The target has not seen the guest_connected event via > > spice_chr_guest_open or spice_chr_write, and so spice server wrongly > > assumes there is no agent active, while the client continues to send > > motion events only by the agent channel, which the server ignores. The > > net effect is that the mouse is static in the guest. > > > > By registering the interface on post load spice server will pass on the > > agent messages fixing the mouse behavior after migration. > > > > RHBZ #725965 > > > > Signed-off-by: Alon Levy <alevy@redhat.com> > > --- > > spice-qemu-char.c | 42 ++++++++++++++++++++++++++++++++++++++++++ > > 1 file changed, 42 insertions(+) > > I suppose Gerd should pick this up in his tree? > > I have a couple of questions below, but I ack this approach. > > > static QLIST_HEAD(, SpiceCharDriver) spice_chars = > > @@ -185,18 +190,23 @@ static void spice_chr_close(struct CharDriverState *chr) > > printf("%s\n", __func__); > > vmc_unregister_interface(s); > > QLIST_REMOVE(s, next); > > + qemu_free_timer(s->post_load.timer); > > Also vmstate_unregister()? > > I'm wondering if there can be a case where this function is called > before the timer has had a chance to fire. It can happen if the spice > port is hot-unplugged before the guest has had a chance to run on the > target. In that case, qemu_del_timer() should be called as well, to > ensure the timer doesn't fire with invalid args later. Alon, ping? Amit
diff --git a/spice-qemu-char.c b/spice-qemu-char.c index a4d7de8..e6eb523 100644 --- a/spice-qemu-char.c +++ b/spice-qemu-char.c @@ -2,6 +2,7 @@ #include "trace.h" #include "ui/qemu-spice.h" #include "char/char.h" +#include "migration/vmstate.h" #include <spice.h> #include <spice-experimental.h> #include <spice/protocol.h> @@ -26,6 +27,10 @@ typedef struct SpiceCharDriver { ssize_t bufsize, datalen; uint32_t debug; QLIST_ENTRY(SpiceCharDriver) next; + uint32_t guest_open; + struct { + QEMUTimer *timer; + } post_load; } SpiceCharDriver; static QLIST_HEAD(, SpiceCharDriver) spice_chars = @@ -185,18 +190,23 @@ static void spice_chr_close(struct CharDriverState *chr) printf("%s\n", __func__); vmc_unregister_interface(s); QLIST_REMOVE(s, next); + qemu_free_timer(s->post_load.timer); g_free(s); } static void spice_chr_guest_open(struct CharDriverState *chr) { SpiceCharDriver *s = chr->opaque; + + s->guest_open = 1; vmc_register_interface(s); } static void spice_chr_guest_close(struct CharDriverState *chr) { SpiceCharDriver *s = chr->opaque; + + s->guest_open = 0; vmc_unregister_interface(s); } @@ -217,6 +227,34 @@ static void print_allowed_subtypes(void) fprintf(stderr, "\n"); } +static void spice_chr_post_load_cb(void *opaque) +{ + SpiceCharDriver *s = opaque; + + spice_chr_guest_open(s->chr); +} + +static int spice_chr_post_load(void *opaque, int version_id) +{ + SpiceCharDriver *s = opaque; + + if (s && s->chr && s->guest_open) { + qemu_mod_timer(s->post_load.timer, 1); + } + return 0; +} + +static VMStateDescription spice_chr_vmstate = { + .name = "spice-chr", + .version_id = 1, + .minimum_version_id = 1, + .post_load = spice_chr_post_load, + .fields = (VMStateField[]) { + VMSTATE_UINT32(guest_open, SpiceCharDriver), + VMSTATE_END_OF_LIST() + }, +}; + static CharDriverState *chr_open(QemuOpts *opts, const char *subtype) { CharDriverState *chr; @@ -229,12 +267,16 @@ static CharDriverState *chr_open(QemuOpts *opts, const char *subtype) s->debug = debug; s->active = false; s->sin.subtype = subtype; + s->post_load.timer = qemu_new_timer_ns(vm_clock, + spice_chr_post_load_cb, s); chr->opaque = s; chr->chr_write = spice_chr_write; chr->chr_close = spice_chr_close; chr->chr_guest_open = spice_chr_guest_open; chr->chr_guest_close = spice_chr_guest_close; + vmstate_register(NULL, -1, &spice_chr_vmstate, s); + QLIST_INSERT_HEAD(&spice_chars, s, next); return chr;
The target has not seen the guest_connected event via spice_chr_guest_open or spice_chr_write, and so spice server wrongly assumes there is no agent active, while the client continues to send motion events only by the agent channel, which the server ignores. The net effect is that the mouse is static in the guest. By registering the interface on post load spice server will pass on the agent messages fixing the mouse behavior after migration. RHBZ #725965 Signed-off-by: Alon Levy <alevy@redhat.com> --- spice-qemu-char.c | 42 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 42 insertions(+)