Subject: Re: misc/4543: spl(9) manpage in conflict with reality
To: Mike Long <firstname.lastname@example.org>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
Date: 11/19/1997 09:34:03
Mike Long writes:
>>Date: Wed, 19 Nov 1997 06:31:39 -0800 (PST)
>>From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
>>spl(9) describes the synchronization protocol used inside the kernel.
>>But the description is not correct. The man page begins
>>where historically, and on at least hte mips and vax ports the
>>converse is actually true: sptty() blocks tty interrupts and _lower_
>>spl levels. For example, spltty() does _not_ block out splsched().
>You are correct, I biffed this.
Thanks for the reply. The current manpage had gotten me quie seriously
I'd like to generate some text that clarifies both that, and the
requirements/invariants that (as far as I understand)
a) MI callers of the spl* code should obey
b) the semantics MD code must provide
in order to guarantee correct synchronization -- somewhat like the
manpage I originally counter-proposed to your draft, except describing
semantics rather than specific implentations. I think that would really help.
Does that sound appropriate for this manpage?