[Spice-devel] [spice-server 1/3] test_vdagent: Pass proper types to spice_server_add_interface
Christophe Fergeau
cfergeau at redhat.com
Tue Jan 31 15:04:56 UTC 2017
On Tue, Jan 31, 2017 at 07:45:44AM -0500, Frediano Ziglio wrote:
> Title should be test-vdagent (with dash).
Thanks, fixed.
>
> >
> > This is a revert of 93b4f4050^ and 93b4f4050.
> > For a SpiceCharDeviceInstance, the base interface must be a
> > SpiceCharDeviceInterface instance, not a SpiceBaseInterface instance, or
> > spice-server code will end up reading out of bounds.
> >
> > vmc_state/vmc_write/vmc_read implementations also have to be provided.
> > ---
> > server/tests/test-vdagent.c | 65
> > ++++++++++++++++++++++++++++++++++++++++-----
> > 1 file changed, 59 insertions(+), 6 deletions(-)
> >
> > diff --git a/server/tests/test-vdagent.c b/server/tests/test-vdagent.c
> > index 7f905ad..e06229e 100644
> > --- a/server/tests/test-vdagent.c
> > +++ b/server/tests/test-vdagent.c
> > @@ -45,11 +45,64 @@ static void pinger(SPICE_GNUC_UNUSED void *opaque)
> > core->timer_start(ping_timer, ping_ms);
> > }
> >
> > -static SpiceBaseInterface base = {
> > - .type = SPICE_INTERFACE_CHAR_DEVICE,
> > - .description = "test spice virtual channel char device",
> > - .major_version = SPICE_INTERFACE_CHAR_DEVICE_MAJOR,
> > - .minor_version = SPICE_INTERFACE_CHAR_DEVICE_MINOR,
> > +static int vmc_write(SPICE_GNUC_UNUSED SpiceCharDeviceInstance *sin,
> > + SPICE_GNUC_UNUSED const uint8_t *buf,
> > + int len)
> > +{
> > + return len;
> > +}
> > +
> > +static int vmc_read(SPICE_GNUC_UNUSED SpiceCharDeviceInstance *sin,
> > + uint8_t *buf,
> > + int len)
> > +{
> > + static uint8_t c = 0;
> > + static uint8_t message[2048];
> > + static unsigned pos = 0;
> > + static unsigned message_size;
> > + int ret;
> > +
> > + if (pos == 0) {
> > + VDIChunkHeader *hdr = (VDIChunkHeader *)message;
> > + VDAgentMessage *msg = (VDAgentMessage *)&hdr[1];
> > + uint8_t *p = message;
> > + int size = sizeof(message);
> > + message_size = size;
> > + /* fill in message */
> > + hdr->port = VDP_SERVER_PORT;
> > + hdr->size = message_size - sizeof(VDIChunkHeader);
> > + msg->protocol = VD_AGENT_PROTOCOL;
> > + msg->type = VD_AGENT_END_MESSAGE;
> > + msg->opaque = 0;
> > + msg->size = message_size - sizeof(VDIChunkHeader) -
> > sizeof(VDAgentMessage);
> > + size -= sizeof(VDIChunkHeader) + sizeof(VDAgentMessage);
> > + p += sizeof(VDIChunkHeader) + sizeof(VDAgentMessage);
> > + for (; size; --size, ++p, ++c)
> > + *p = c;
> > + }
> > + ret = MIN(message_size - pos, len);
> > + memcpy(buf, &message[pos], ret);
> > + pos += ret;
> > + if (pos == message_size) {
> > + pos = 0;
> > + }
> > + //printf("vmc_read %d (ret %d)\n", len, ret);
> > + return ret;
> > +}
> > +
> > +static void vmc_state(SPICE_GNUC_UNUSED SpiceCharDeviceInstance *sin,
> > + SPICE_GNUC_UNUSED int connected)
> > +{
> > +}
> > +
> > +static SpiceCharDeviceInterface vmc_interface = {
> > + .base.type = SPICE_INTERFACE_CHAR_DEVICE,
> > + .base.description = "test spice virtual channel char device",
> > + .base.major_version = SPICE_INTERFACE_CHAR_DEVICE_MAJOR,
> > + .base.minor_version = SPICE_INTERFACE_CHAR_DEVICE_MINOR,
> > + .state = vmc_state,
> > + .write = vmc_write,
> > + .read = vmc_read,
> > };
> >
> > SpiceCharDeviceInstance vmc_instance = {
> > @@ -63,7 +116,7 @@ int main(void)
> > core = basic_event_loop_init();
> > test = test_new(core);
> >
> > - vmc_instance.base.sif = &base;
> > + vmc_instance.base.sif = &vmc_interface.base;
> > spice_server_add_interface(test->server, &vmc_instance.base);
> >
> > ping_timer = core->timer_add(pinger, NULL);
>
> How did you discover this problem?
> This test were supposed to be working even without vmc_read, why now
> is needed? It's not exactly a stub (like "return -1").
> Maybe NULLs are fine too?
s/SpiceBaseInstance/SpiceCharDeviceInstance/ was found because the test
was crashing at start. vmc_read implementation is the same as the one
which was present before 93b4f4050. Arguably, it's not really useful as
then the test just loops forever on vmc_read, but I'd say that's a
different issue (and actually I don't really know what the test is
supposed to be testing /o\ )
Without a vmc_read implementation, you get a crash when trying to read
from the device in vdi_port_read_one_msg_from_device(). There used to be
an assert when one of the vfuncs was missing, which alos caused me to
add this.
Christophe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 801 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/spice-devel/attachments/20170131/56c8b0fc/attachment.sig>
More information about the Spice-devel
mailing list