On 08/22/13 11:57, Gerd Hoffmann wrote: > pty_chr_timer first calls pty_chr_update_read_handler(), then clears > timer_tag (because it is a one-shot timer). This is the wrong order > though. pty_chr_update_read_handler might re-arm time timer, and the > new timer_tag gets overwitten in that case. > > This leads to crashes when unplugging a pty chardev: pty_chr_close > thinks no timer is running -> timer isn't canceled -> pty_chr_timer gets > called with stale CharDevState -> BOOM. > > This patch fixes the ordering. > Kill the pointless goto while being at it. > > https://bugzilla.redhat.com/show_bug.cgi?id=994414 > > Signed-off-by: Gerd Hoffmann <kra...@redhat.com> > --- > qemu-char.c | 12 ++++-------- > 1 file changed, 4 insertions(+), 8 deletions(-) > > diff --git a/qemu-char.c b/qemu-char.c > index 1be1cf6..1621fbd 100644 > --- a/qemu-char.c > +++ b/qemu-char.c > @@ -1026,15 +1026,11 @@ static gboolean pty_chr_timer(gpointer opaque) > struct CharDriverState *chr = opaque; > PtyCharDriver *s = chr->opaque; > > - if (s->connected) { > - goto out; > - } > - > - /* Next poll ... */ > - pty_chr_update_read_handler(chr); > - > -out: > s->timer_tag = 0; > + if (!s->connected) { > + /* Next poll ... */ > + pty_chr_update_read_handler(chr); > + } > return FALSE; > }
pty_chr_timer() s->timer_tag = 0; pty_chr_update_read_handler() // s->connected == 0 pty_chr_state(..., 1) // G_IO_HUP is clear not calling: g_source_remove(s->timer_tag) s->connected = 1; qemu_chr_be_generic_open() s->fd_tag = io_add_watch_poll() So, in this order, s->timer_tag is not removed in pty_chr_state(). But that shouldn't be necessary anyway, since pty_chr_timer() returns FALSE, and its associated tag (s->timer_tag, see pty_chr_rearm_timer()) is removed anyway. Seems OK to me. Reviewed-by: Laszlo Ersek <ler...@redhat.com>