Date: Sat, 17 Mar 2001 12:45:23 -0800 (PST) From: Matthew Jacob <mjacob@feral.com> To: John Baldwin <jhb@FreeBSD.ORG> Cc: Warner Losh <imp@harmony.village.org>, arch@FreeBSD.ORG Subject: Re: man pages Message-ID: <Pine.LNX.4.21.0103171243480.2328-100000@zeppo.feral.com> In-Reply-To: <XFMail.010317123150.jhb@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> >: critical region of control > > ... > >: The driver may not call any functions external to itself in between > >: the time it calls ddi_enter_critical() and the time it calls > >: ddi_exit_critical(). > > ... > > > > Hmmm. That's what I need, with the above exception being too > > restrictive. I'd need to call functions that frob the hardware. :-) > > This would allow us to get rid of most of the (ab)uses of *_intr as it is right > now. If we were to use a more abstract name for this than just > intr_restore/disable, would critical_enter() and critical_exit() work? If I do > this, I may not even mess with the current set of foo_intr() functions as they > may turn out to no longer be useful. Some things like the ACPI code would need > to be taught to not assume cli/sti though. (Intel's ACPICA code is > Linux-centric in some places, but I think they are very open to fixes, so I > don't expect that will be a problem.) Personally, I like the 'intr' aspect- this is, in fact, what we are doing. ddi_enter_critical/ddi_exit_critical as an attempt to be a bit more generic, and the restriction against calling other functions is too restrictive now (IMO)- you could interpret this as prohibiting calls to the PROM, for instance. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-arch" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.21.0103171243480.2328-100000>