On Wed, Jan 31, 2018 at 07:25:21PM +0300, Klim Kireev wrote:
> On one of our client's node, due to trying to read from closed ioc,
> a segmentation fault occured. Corresponding backtrace:
>
> 0 object_get_class (obj=obj@entry=0x0)
> 1 qio_channel_readv_full (ioc=0x0, iov=0x7ffe55277180 ...
> 2 qio_channel_read (ioc=<optimized out> ...
> 3 vnc_client_read_buf (vs=vs@entry=0x55625f3c6000, ...
> 4 vnc_client_read_plain (vs=0x55625f3c6000)
> 5 vnc_client_read (vs=0x55625f3c6000)
> 6 vnc_client_io (ioc=<optimized out>, condition=G_IO_IN, ...
> 7 g_main_dispatch (context=0x556251568a50)
> 8 g_main_context_dispatch (context=context@entry=0x556251568a50)
> 9 glib_pollfds_poll ()
> 10 os_host_main_loop_wait (timeout=<optimized out>)
> 11 main_loop_wait (nonblocking=nonblocking@entry=0)
> 12 main_loop () at vl.c:1909
> 13 main (argc=<optimized out>, argv=<optimized out>, ...
>
> Having analyzed the coredump, I understood that the reason is that
> ioc_tag is reset on vnc_disconnect_start and ioc is cleaned
> in vnc_disconnect_finish. Between these two events due to some
> reasons the ioc_tag was set again and after vnc_disconnect_finish
> the handler is running with freed ioc,
> which led to the segmentation fault.
>
> The patch checks vs->disconnecting in places where we call
> qio_channel_add_watch to prevent such an occurrence.
>
> Signed-off-by: Klim Kireev <klim.kir...@virtuozzo.com>
> ---
> Changelog:
> v2: Attach the backtrace
>
> v3: Change checks
>
> ui/vnc.c | 18 ++++++++++++++----
> 1 file changed, 14 insertions(+), 4 deletions(-)
>
> diff --git a/ui/vnc.c b/ui/vnc.c
> index 33b087221f..708204fa7e 100644
> --- a/ui/vnc.c
> +++ b/ui/vnc.c
> @@ -1407,13 +1407,19 @@ static void vnc_client_write_locked(VncState *vs)
> } else
> #endif /* CONFIG_VNC_SASL */
> {
> - vnc_client_write_plain(vs);
> + if (vs->disconnecting == FALSE) {
> + vnc_client_write_plain(vs);
> + } else {
> + if (vs->ioc_tag != 0) {
> + g_source_remove(vs->ioc_tag);
> + vs->ioc_tag = 0;
> + }
> + }
> }
> }
I'm not sure it is safe to only do the check in the else{} branch
of this code. If this code is reachable, then I think the SASL
branch will cause the same crash problems too. I think probably
need to push the checks up a level or two in the caller stack...
>
> static void vnc_client_write(VncState *vs)
> {
> -
> vnc_lock_output(vs);
> if (vs->output.offset) {
> vnc_client_write_locked(vs);
> @@ -1421,8 +1427,12 @@ static void vnc_client_write(VncState *vs)
> if (vs->ioc_tag) {
> g_source_remove(vs->ioc_tag);
> }
> - vs->ioc_tag = qio_channel_add_watch(
> - vs->ioc, G_IO_IN, vnc_client_io, vs, NULL);
> + if (vs->disconnecting == FALSE) {
> + vs->ioc_tag = qio_channel_add_watch(
> + vs->ioc, G_IO_IN, vnc_client_io, vs, NULL);
> + } else {
> + vs->ioc_tag = 0;
> + }
> }
> vnc_unlock_output(vs);
> }
...I think perhaps we should do the check in the vnc_client_io()
method, and also in the vnc_flush() method.
I think we also need to put a check in the vnc_jobs_consume_buffer()
method, which can be triggered from a bottom-half.
Regards,
Daniel
--
|: 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 :|
