Date: Thu, 28 Feb 2002 14:37:29 -0800 (PST) From: <luoqi@FreeBSD.org> To: benno@freebsd.org Cc: freebsd-ppc@freebsd.org Subject: Re: cvs commit: src/sys/powerpc/powerpc pmap.c Message-ID: <200202282237.g1SMbTB46056@freefall.freebsd.org>
index | next in thread | raw e-mail
> > In FreeBSD, the zone allocator won't do this job, first, you are > > not guaranteed to get page below 256M, second, kva returned from > > zalloc does not necessarily equal to pa (you don't really need the > > kva). You'll have to write you own allocator. The tricky part is > > how to allocate page below 256M. We could probably do something > > similar to contigmalloc(), but my impression is this call could > > be quite expensive. Especially so if we adopt a merged pvo struct, > > because pv_entry allocation is much more frequent than overflow. > > Yeah, I was kind of avoiding thinking about this for the moment. =) > Hopefully Jeff's UMA stuff should help, as I believe that it > allows you to specify allocation routines which will allow this to work. > I wouldn't expect much help from UMA, in this customized allocation subroutine, you will still have to search for a page in low memory, that's not going to be cheap. Of course, we may follow NetBSD and have multiple freelists for different types of memory. In fact, this is not too hard to do and may benifit us in the long run. I'm just trying to figure out if there is enough justification for very rare events like hash table overflow. Last time I looked at Darwin's code, they'll simply panic if all entries in a particular pte group are all wired down. Not that I'm advocating this solution, but the fact that this hasn't become any problem shows how low the probability for an overflow is. > Hrm. I think I'd rather wait until Jeff's UMA goes in and reassess > then. At the moment I'm running on a simulator configured to use 32MB > of RAM so the BAT mapping will work. > You're absolutely right there's more urgent stuff to work on. Pmap module should be quite straightforward after this issue is resolved. I'm glad to see progresses are made in ppc port, it's a pity I don't have much time to work on it myself :( But I have my simulator ready, nd I can't wait to see the first shell prompt! -lq To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-ppc" in the body of the messagehelp
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200202282237.g1SMbTB46056>