Date: Fri, 23 Jan 1998 23:19:35 +1030 From: Mike Smith <mike@smith.net.au> To: Luigi Rizzo <luigi@labinfo.iet.unipi.it> Cc: daniel_sobral@voga.com.br, hackers@FreeBSD.ORG Subject: Re: uiomove() Message-ID: <199801231249.XAA00630@word.smith.net.au> In-Reply-To: Your message of "Fri, 23 Jan 1998 11:47:42 BST." <199801231047.LAA02217@labinfo.iet.unipi.it>
next in thread | previous in thread | raw e-mail | index | archive | help
> > Yeah (well, kind of yeah), but... What about SMP? What about a process in > > one processor calling read() while a process in another processor calls > > write()? > > not sure, but I think in our SMP implementation only one process > can be in the kernel at each time, isn't it ? Too much of the > existing code assumes that, and it would have been impossible to > fix everything to have SMP running. This is no longer true. The SMP people started with a single "monster lock" as you describe. This is the traditional starting point, as it reduces the amount of work required to begin with. >From there, there have been several rounds of lock "push downs", where parts of the kernel are made SMP-aware. At this point in time, drivers are still below a single lock, and are likely to remain there for some time. 8) -- \\ Sometimes you're ahead, \\ Mike Smith \\ sometimes you're behind. \\ mike@smith.net.au \\ The race is long, and in the \\ msmith@freebsd.org \\ end it's only with yourself. \\
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199801231249.XAA00630>