On Wed, 21 Mar 2001 00:04:56 -0800,
george anzinger <[EMAIL PROTECTED]> wrote:
>Exactly so. The method does not depend on the sum of preemption being
>zip, but on each potential reader (writers take locks) passing thru a
>"sync point". Your notion of waiting for each task to arrive
>"naturally" at schedule() would work. It is, in fact, over kill as you
>could also add arrival at sys call exit as a (the) "sync point". In
>fact, for module unload, isn't this the real "sync point"? After all, a
>module can call schedule, or did I miss a usage counter somewhere?
A module can call schedule but it must do MOD_INC_USE_COUNT first.
Sleeping in module code without incrementing the module use count first
is a shooting offence. It is so full of races that you may as well
call it Daytona.
>By the way, there is a paper on this somewhere on the web. Anyone
>remember where?
http://www.rdrop.com/users/paulmck/paper/rclockpdcsproof.pdf
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
