#qi-hardware IRC log for Friday, 2013-05-31

mthlarsc: I see you've been doing work on regmap recently...05:05
mthI'd like a second opinion on this: in regcache_rbtree_sync in drivers/base/regmap/regcache-rbtree.c05:06
mthif (rbnode->base_reg < min) continue;05:06
mthit seems to me that this is overly strict, since a block could start before the minimum, but end after it05:06
mthif (rbnode->base_reg + rbnode->blklen < min) continue;05:07
mththat statement a few lines below seems to be the correct bounds check05:07
mthalso, I was about to suggest the JZ4740 could switch from an rbtree cache to a flat cache, but then I found out sync is not implemented for flat caches yet05:08
mthdo you think it makes sense to add that implementation?05:08
mthby the way, it seems wasteful to have a tree and then use linear search to find a value ;)05:14
larscmth: yea, that looks a bit strange07:06
mthhttp://www.treewalker.org/temp/regcache-flat-sync.diff07:08
mththat implementation works on the JZ4770, I didn't test it on the 474007:09
mththe 4770 doesn't have mmio mapped codec registers anymore07:10
mthhttp://www.treewalker.org/temp/regcache-rbtree-sync.diff07:21
mththat's the patch removing the first "if" for the rbtree sync07:21
mthlarsc: would you like to review these or shall I send them to the maintainers?07:23
larsclooks good to me07:24
larscmth: -ENOPATCH14:47
mthlarsc: which one, the flat or rbtree?14:49
larscrbtree14:49
mthfor some reason, git send-email asked to send twice14:50
mthand I'm not on the Cc: for the first time it asked14:50
larscit asks for each patch if you don't answer (a)all14:56
mthbut I ran it separately for each patch14:56
mthit asked twice for the second patch, but apparently the first time it ask the e-mail was not completely constructed yet14:57
mthI used --compose on the second patch, maybe that triggered it14:57
--- Sat Jun 1 201300:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!