Gary Jennejohn <gary.jennejohn@freenet.de> wrote:
quoted text
> Worse is that there is a potential infinite loop with all interrupts
> blocked in cbm_raw_write().

Do you mean the one that was fixed in 0.4.3rc1?

<http://opencbm.cvs.sourceforge.net/viewvc/opencbm/cbm4win/sys/linux/cbm_module.c?r1=1.13.2.18&r2=1.13.2.19>

That was a sticking point for me too.  :)

(The loop was subsequently upped to 1ms, but it's still better than
infinite.)

quoted text
> It's a pity that Linux doesn't allowing blocking single interrupts, like
> that various BSDs out there.

Do you mean the ability to disable a specific IRQ?  I'm curious -- how
would that help?  The issue here is that the module needs to take a
certain action within a strict time interval, so any interrupt could
throw it off.

(I guess the real pity is the lack of real-time in the vanilla kernel.)


-- 
The world is coming to an end ... SAVE YOUR BUFFERS!!!

_______________________________________________
Kernel-mentors mailing list
Kernel-mentors@selenic.com
http://selenic.com/mailman/listinfo/kernel-mentors