Date: Thu, 21 May 2020 13:09:53 -0400 From: Mark Johnston <markj@freebsd.org> To: Brandon Bergren <bdragon@freebsd.org> Cc: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: Re: svn commit: r361336 - head/sys/powerpc/aim Message-ID: <20200521170953.GF85681@raichu> In-Reply-To: <202005211553.04LFrH3m086488@repo.freebsd.org> References: <202005211553.04LFrH3m086488@repo.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, May 21, 2020 at 03:53:17PM +0000, Brandon Bergren wrote: > Author: bdragon > Date: Thu May 21 15:53:16 2020 > New Revision: 361336 > URL: https://svnweb.freebsd.org/changeset/base/361336 > > Log: > [PowerPC] Fix kernel boot on powerpc > > Recent changes have caused the vmspace objects to start coming from KVA > instead of direct-mapped memory on powerpc. As far as I can tell, this is > not actually a problem, so we should stop arbitrarily asserting that it is. > > I do not know why this was not being triggered before. UMA was recently changed in r357549 to use multi-page slabs if doing so would reduce internal fragmentation to a certain degree. In that case the slabs will be mapped into KVA. vmspace objects are quite large but smaller than a page, so they benefit from this. You can verify by checking the vm.uma.VMSPACE.keg.ppera sysctl, which gives the number of pages per slab.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20200521170953.GF85681>