AW: Rottserver UTCB Bug?
"Christian, Martin, OPEE45" <[email protected]>
| Newsgroups | gmane.comp.micro-kernel.l4.l4ka.general |
|---|---|
| Message-ID | <[email protected]> |
Allright! I've overlooked the line "space = allocate_space()" where I assume the new address space is created. Thanks, Martin. -----Ursprüngliche Nachricht----- Von: Espen Skoglund [mailto:[email protected]] Im Auftrag von Espen Skoglund Gesendet: Montag, 19. März 2007 03:57 An: Christian, Martin, OPEE45 Cc: [email protected] Betreff: Re: Rottserver UTCB Bug? F[Martin Christian] > Hi! > I was wondering why Sigma0/1 and Roottask share the same UTCB? Is this > on purpose? > I'm talking about: > http://l4hq.org/viewcvs/pistachio/kernel/src/api/v4/thread.cc?revision=1 > .65.4.15 > In init_root_servers() the function create_root_server() is called > with ROOT_UTCB_START for any rootserver. Thus every rootserver uses > the same TCRs - e. g. message registers. Wouldn't it be possible > that one rootserver overwrites values in the message registers, > which another rootserver still needs? E. g. sigma0 gets interrupted > during pagefault handling and roottask sends some IPC in between. No, the address you're referring to is simply the virtual address of the UTCB. There are conflicts since sigma0, sigma1, and roottask occupy different address spaces. eSk