On Thu, Feb 26, 2009 at 2:00 AM, Matthew Toseland <toad at amphibian.dyndns.org> wrote: > On Wednesday 25 February 2009 09:57:25 Florent Daigniere wrote: >> Matthew Toseland wrote: >> > On Thursday 29 January 2009 17:12:57 nextgens at freenetproject.org wrote: >> >> Author: nextgens >> >> Date: 2009-01-29 17:12:57 +0000 (Thu, 29 Jan 2009) >> >> New Revision: 25364 >> >> >> >> Modified: >> >> ? ?trunk/freenet/src/freenet/node/FNPPacketMangler.java >> >> ? ?trunk/freenet/src/freenet/node/PacketSender.java >> >> ? ?trunk/freenet/src/freenet/node/PeerNode.java >> >> ? ?trunk/freenet/src/freenet/node/RequestSender.java >> >> ? ?trunk/freenet/src/freenet/support/Logger.java >> >> ? ?trunk/freenet/src/freenet/support/LoggerHook.java >> >> Log: >> >> some untested code trying to address the lock-contention problem in the >> > logger >> > >> > How expensive are the volatile's? >> >> "Free" to read on x86; and that's what matters. > > Ok so in java all it means is that the compiler will do a new memory fetch on > every access? >
and no instruction reordering before/after this, lessor loop-unrolling, etc. volatile on int/boolean/short are free *on its own*, it just suppress some optimization.
