--- Log opened Sun May 17 00:00:33 2009
09:03 -!- tgrabiec [n=tomekg@rev-189-107.ramtel.pl] has joined #jato
09:42 < vegard> hm. isn't this weird: the _quick opcodes listed in the first edition of JVM spec are completely gone in the second edition
10:24 -!- penberg_home [n=penberg@cs149038.pp.htv.fi] has joined #jato
10:24 < penberg_home> tgrabiec: hey
10:25 < penberg_home> tgrabiec: radix tree, nice!
10:25 <@ahuillet> hi
10:25 < penberg_home> tgrabiec: can you separate the radix tree part from the cu mapping part, though?
10:25 < penberg_home> hiya
10:25 < vegard> heh. there's also a typo in the first edition that says idiv has value 100.
10:26 < penberg_home> tgrabiec: put it in vm/radix-tree.c, ok?
10:26 < penberg_home> tgrabiec: can be an incremental patch on top of the series too.
10:27 < penberg_home> ok, gotta run! :)
10:27 < penberg_home> later
10:27 -!- penberg_home [n=penberg@cs149038.pp.htv.fi] has quit [Client Quit]
10:28 -!- tgrabiec [n=tomekg@rev-189-107.ramtel.pl] has quit [Read error: 113 (No route to host)]
15:55 -!- wekt [n=kvirc@g228222001.adsl.alicedsl.de] has joined #jato
15:56 -!- wekt [n=kvirc@g228222001.adsl.alicedsl.de] has left #jato []
17:24 -!- tgrabiec [n=tomekg@rev-189-107.ramtel.pl] has joined #jato
20:42 -!- penberg_home [n=penberg@cs149038.pp.htv.fi] has joined #jato
20:43 < penberg_home> ahuillet: I can rebase but I'd rather just get rid of the branch
20:43 < penberg_home> is there something you want to keep?
20:43 <@ahuillet> your modifications
20:43 <@ahuillet> why would you throw them away?
20:43 < penberg_home> ahuillet: they don't work :)
20:44 < penberg_home> tgrabiec: hey, btw, please pay attention to the "why is this change needed" part of your patch descriptions.
20:44 <@ahuillet> they don't handle one of the two cases,but apart  from that...
20:44 < penberg_home> tgrabiec: it's easier for others to review them that way.
20:44 < penberg_home> ahuillet: then do you want me to merge them to master or...?
20:45 < tgrabiec> penberg_home: ok
20:45 <@ahuillet> penberg: if we both agree that we keep them then yes
20:45 < penberg_home> tgrabiec: did you see my radix tree comment?
20:45 <@ahuillet> I'll have to patch block predecessors for the second case
20:45 < penberg_home> tgrabiec: i think i'm going to go ahead and merge your patches as-is
20:45 <@ahuillet> tgrabiec : yeah, saying why a change is needed is important
20:45 <@ahuillet> your INSN_RET thing for example looked good but I did not know *why* you wanted it
20:46 < penberg_home> yeah, that one was bit strange
20:46 < penberg_home> especially as it's not part of the patch series numbering
20:46 < tgrabiec> penberg_home: where is the comment wbout radix tree ?
20:46 < penberg_home> tgrabiec: http://folk.uio.no/vegardno/jato-irc-logs/2009-05-17.txt
20:47 < penberg_home> oh, it seems vegard's logs are real-time.
20:47 < penberg_home> nice.
20:48 < tgrabiec> penberg_home: about the radix tree separation - no problem.
20:48 < penberg_home> tgrabiec: did you implement it from scratch?
20:49 < tgrabiec> penberg_home: yes...
20:49 < penberg_home> tgrabiec: ok, cool.
20:49 < tgrabiec> penberg_home: probably a test in test/jit would be appriciated for it ?
20:50 < penberg_home> tgrabiec: test/vm
20:50 < penberg_home> tgrabiec: but yes :)
20:50 < tgrabiec> right
20:51 < penberg_home> tgrabiec: so
20:51 < penberg_home> tgrabiec: we only insert starting address of a compiled method?
20:52 < tgrabiec> penberg_home: yes
20:52 < tgrabiec> penberg_home: oh crap, i forgot about locking... can it go in the next patch ?
20:52 < penberg_home> tgrabiec: yeah, i think so.
20:53 < penberg_home> buffer alignment size
20:53 < penberg_home> 4K
20:53 < penberg_home> really?
20:54 < tgrabiec> penberg_home: AFAIK in vm/buffer.c buffers are allocated aligned to the page size
20:55 < penberg_home> well, that's obviously something we want to fix
20:55 < penberg_home> IIRC, Intel recommends 16 byte alignment.
20:55 < penberg_home> tgrabiec: does that change things wrt. how we use the radix tree?
20:56 < tgrabiec> penberg_home: well, changing alignment to 16 bytes is changing just init_cu_mapping(), doesn't affect how we use it
20:56 < penberg_home> ok
21:01 < penberg_home> ahuillet: ok, so, I'm inclined to just merge the topic branch to master then
21:01 < penberg_home> I'll double check the commits tomorrow
21:01 <@ahuillet> k
21:24 -!- penberg_home [n=penberg@cs149038.pp.htv.fi] has quit []
--- Log closed Mon May 18 00:00:33 2009