Re: splxxx replacements?

[Gary Thorpe <gthorpe@xxxxxxxxxxxxx>]

Matthew Dillon wrote:
> :That's basically the equivalent of splhigh() and splx(): isn't that 
> :worse than blocking individual levels?
>
>     No, not really.   We lose a little handling devices without DMA, but
>     that's basically just serial ports and it isn't worth the aggravation of
>     trying to manage per-interrupt masking.  The per-interrupt (spl) masking
>     also doesn't work well in a SMP environment anyway.
>
> 					-Matt
> 					Matthew Dillon 
> 					<dillon@xxxxxxxxxxxxx>

Thanks for some explanation. However, it seems the crit_enter/exit only 
works on one CPU in an SMP environment too (looking back at previous 
posting in thread). Is there really a practical as well as conceptual 
difference?