Hello Martin,

> a few comments on your current code (please don't take it as criticism,
> but rather as TODOs :)):

No problem, because it is a prototype.

> * I agree with the comments by Jiri Svoboda. Try to split the IPC code
>    from compctl into a module in libc, to make compctl more abstract and
>    the IPC stubs reusable.

OK. I was planning to do so even without your comments.

> * Currently you have disabled kfb.

Only temporarily.

>>    a) kfb is used on non-ia32, non-amd64 platforms.
>    b) kfb should be still usable in case vesafb fails to initialize (for
>       any reason).
>
>    There needs to be some kind of hand-off between kfb and vesafb, both
>    drivers need to coexist side-by-side in a limited way.

To be honest, I don't have any idea how to do it in an elegant way.
May be we should disable autoloading and load drivers explicitly by
compositor? Is there a mechanism to load drivers on demand?

> * I know that this is not your idea, but the dual naming of the
>    compositor nodes (comp:0/* vs. comp/:0) is somewhat confusing. I
>    would really prefer creating the control node as for example
>    comp:0/ctrl.

At first I planned to add a a /ctl node to comp:0/ directory, but then
decided that it will be rather natural to control compositor through
it main node. But if you want the former way, I'll do.

> * The solution with the separate thread in vesafb is not very elegant.
>    You should first initialize the hardware (port_init()) and only after
>    the initialization succeeds register the DDF function.

Well, then what to do with compositor which starts before driver?

> *  CamelCase in libvbe

It was done on purpose: to name structures like in VBE standard. Could
naming convention be weaken here?

_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel

Reply via email to