Date: Thu, 21 Dec 2006 19:23:20 -0800 From: Peter Grehan <grehan@freebsd.org> To: Suleiman Souhlal <ssouhlal@FreeBSD.org> Cc: cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/powerpc/powerpc mmu_oea.c Message-ID: <458B4FA8.40408@freebsd.org> In-Reply-To: <458A48C9.40807@FreeBSD.org> References: <200612200110.kBK1AMeI091785@repoman.freebsd.org> <458A48C9.40807@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
> Could we just evict one random entry from either the primary or > secondary PTEG and use that spot to insert the current PTE? > (Isn't the whole point of pvo_entry to permit pages to be searched > quickly if their virtual->physical mapping has already been computed > before, basically making page faults to pages that have been evicted > from their PTEGs very cheap (because we don't need to walk the vm_map > entries etc again)?) Yes, though you have to make sure it isn't wired, and probably prefer discarding user PTEs over kernel PTEs. I'm also not sure if the code path on a DSI miss for a valid PTE works. But this shouldn't be too hard to test. later, Peter.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?458B4FA8.40408>