[Spice-devel] Purpose of interface version checks in server/reds.c

[Gerd Hoffmann]

On 07/01/11 15:58, Christophe Fergeau wrote:
> Hi,
>
> While looking at a crash in qemu when SPICE_INTERFACE_PLAYBACK_MINOR was
> increased, I came across the version checking code in spice_server_add_interface
> in server/reds.c which does stuff like
>
> if (strcmp(interface->type, SPICE_INTERFACE_RECORD) == 0) {
>      red_printf("SPICE_INTERFACE_RECORD");
>      if (interface->major_version != SPICE_INTERFACE_RECORD_MAJOR ||
>          interface->minor_version<  SPICE_INTERFACE_RECORD_MINOR) {
>          red_printf("unsupported record interface");
>          return -1;
>      }
>      snd_attach_record(SPICE_CONTAINEROF(sin, SpiceRecordInstance, base));
> }

That looks bogous.

> interface->major_version/interface->minor_version are filled in
> qemu/audio/spiceaudio.c using compile time constants from spice-server.
> It seems to be useful to detect when qemu was compiled against an older
> spice-server than the one that is currently installed.

Indeed.

> I'm a bit surprised by this mechanism, if we break these interfaces in such
> a way that qemu needs to be recompiled, shouldn't we use different library
> sonames to make sure things are recompiled?

No.  The purpose of the versions is exactly to avoid the need for a new 
soname.  The rules are basically:

   (1) You add stuff to the interface, strictly append-only to not break
       binary compatibility.
   (2) You bump the minor version of the interface.
   (3) You check the minor version at runtime to figure whenever the
       added fields contain valid stuff or not.

An example is here (core interface, minor goes from 2 to 3, new 
channel_event callback):

http://cgit.freedesktop.org/spice/spice/commit/?id=97f33fa86aa6edd25111b173dc0d9599ac29f879

cheers,
   Gerd