Date: Wed, 24 Jul 2013 06:13:00 +0000 (UTC) From: Gleb Smirnoff <glebius@FreeBSD.org> To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-head@freebsd.org Subject: svn commit: r253591 - head/sys/vm Message-ID: <201307240613.r6O6D09Y034129@svn.freebsd.org>
next in thread | raw e-mail | index | archive | help
Author: glebius Date: Wed Jul 24 06:13:00 2013 New Revision: 253591 URL: http://svnweb.freebsd.org/changeset/base/253591 Log: Since r251709 a slab no longer use 8-bit indicies to manage items, thus remove a stale comment. Reviewed by: jeff Modified: head/sys/vm/uma_int.h Modified: head/sys/vm/uma_int.h ============================================================================== --- head/sys/vm/uma_int.h Wed Jul 24 04:24:21 2013 (r253590) +++ head/sys/vm/uma_int.h Wed Jul 24 06:13:00 2013 (r253591) @@ -49,14 +49,6 @@ * 10% memory waste we potentially allocate a separate uma_slab_t if this will * improve the number of items per slab that will fit. * - * Other potential space optimizations are storing the 8bit of linkage in space - * wasted between items due to alignment problems. This may yield a much better - * memory footprint for certain sizes of objects. Another alternative is to - * increase the UMA_SLAB_SIZE, or allow for dynamic slab sizes. I prefer - * dynamic slab sizes because we could stick with 8 bit indices and only use - * large slab sizes for zones with a lot of waste per slab. This may create - * inefficiencies in the vm subsystem due to fragmentation in the address space. - * * The only really gross cases, with regards to memory waste, are for those * items that are just over half the page size. You can get nearly 50% waste, * so you fall back to the memory footprint of the power of two allocator. I
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201307240613.r6O6D09Y034129>