Re: [patch 08/19] mutex subsystem, core

[Ingo Molnar]

* Linus Torvalds <torvalds@xxxxxxxx> wrote:

> On Tue, 3 Jan 2006, Ingo Molnar wrote:
> > > 
> > > Is this an interrupt deadlock, or do you not allow interrupt contexts 
> > > to even trylock a mutex?
> > 
> > correct, no irq contexts are allowed. This is also checked for if 
> > CONFIG_DEBUG_MUTEXES is enabled.
> 
> Note that semaphores are definitely used from interrupt context, and 
> as such you can't replace them with mutexes if you do this.
>
> The prime example is the console semaphore. See kernel/printk.c, look 
> for "down_trylock()", and realize that they are all about interrupts.

yeah, i know - and i'm not trying to replace _all_ semaphore uses.  
down_trylock() is used only 54 times in drivers/*, while down() is used 
2986 times. And even of those 54 cases, only a small minority does it 
from an IRQ context.

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/