I'll take a look. I was thinking we should make this move. Glad to see it. On Mar 7, 2013 9:13 PM, "Brian Geffon" <bri...@apache.org> wrote:
> Hello all, > > So to those of you familiar with the freelists you know that it works this > way the head pointer uses the upper 16 bits for a version to prevent the > ABA problem. The big drawback to this is that it requires the following > NASTY macros to get at the pointer or the version: > > #define FREELIST_POINTER(_x) ((void*)(((((intptr_t)(_x).data)<<16)>>16) | \ > (((~((((intptr_t)(_x).data)<<16>>63)-1))>>48)<<48))) // sign extend > #define FREELIST_VERSION(_x) (((intptr_t)(_x).data)>>48) > #define SET_FREELIST_POINTER_VERSION(_x,_p,_v) \ > (_x).data = ((((intptr_t)(_p))&0x0000FFFFFFFFFFFFULL) | (((_v)&0xFFFFULL) > << 48)) > > Additionally, since this only leaves 16 bits it limits the number of > versions you can have. Well more and more x86_64 processors support DCAS > (double word compare and swap / 128bit CAS). This means that we can use > 64bits for a version which basically makes the versions unlimited but more > importantly it takes those macros above and simplifies them to: > > #define FREELIST_POINTER(_x) (_x).s.pointer > #define FREELIST_VERSION(_x) (_x).s.version > #define SET_FREELIST_POINTER_VERSION(_x,_p,_v) \ > (_x).s.pointer = _p; (_x).s.version = _v > > As you can imagine this will have a performance improvement, in my simple > tests I measured a performance improvement of around 6%. As of now I > haven't tried doing this with reclaimable freelists, but if this proves > fruitful maybe I can work with the taobao guys to make sure it wont have > problems with their stuff. > > This freelist stuff can be very tricky, so I'll ask anyone who is > interested or understands this code to please review these changes and most > importantly performance test it to make sure this makes. Patch is uploaded > to the ticket: https://issues.apache.org/jira/browse/TS-1742 > > Thanks > Brian >
