[okl4-developer] L4 > scheduler

[Harry Vennik]

Hi Antonin,

Personally I do not understand all of your questions, but I can help  
you a little on these:

>
> - what about thread needing lesser-priority threads ? If A is an  
> high prio thread and need a message to/from B, a low prio thread,  
> with all the IPC stuff, will A be stopped and resumed after B is  
> done (maybe after C,D,E,F...) , or will A give temporarly a high  
> prio for B to execute and resume? (B is not created by A, otherwise  
> the problem is solved)
>
If A waits for an IPC message from B, then A is thus blocked, so B  
may get a chance to execute. But generally I think that high-pri  
threads are often called by lower-pri threads, and thus not 'need'  
the low-pri thread, but the low-pri thread will 'need' the high-pri  
thread. (However, this still involves a high-pri thread waiting for  
lower-pri threads, but then because the high-pri server thread just  
has nothing to do (i.e. is idle) until it gets called.)

> - Is it possible for some "malware" to change the priority of  
> threads and then not realase proc units? Time quanta seems to avoid  
> unrealasing, yeti'm not sure in a case of malware.
>
Only privileged threads may change thread priorities, so the malware  
would have to execute as a privileged thread (i.e. in the root  
address space!). So I think this is possible, but very unlikely. If  
what you're running on top of OKL4 is not very badly done, the  
malware would have to enter your system in a trojan-horse like  
manner, i.e. be manually installed to run in the root address space.

>
>
> Thanks
>
> -- 
> Antonin Subtil,
> 35 rue Donissan
> FR-33000 Bordeaux
> +33 (0)6.77.40.17.69
> _______________________________________________
> Developer mailing list
> Developer at okl4.org
> https://lists.okl4.org/mailman/listinfo/developer

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.okl4.org/pipermail/developer/attachments/20080325/c1ac23e8/attachment.htm