#qi-hardware IRC log for Tuesday, 2012-09-04

qi-botThe build was successful: http://fidelio.qi-hardware.com/~xiangfu/build-nanonote/openwrt-xburst.minimal-20120903-0339 06:13
Fallenouwin 1910:11
qi-bot[commit] Xiangfu: new pacakge: mini-jtag (master) http://qi-hw.com/p/openwrt-packages/56f85bc13:20
wpwrakthey missed the obvious solution: draw it in n-space, then project on 2D space14:01
xiangfuwpwrak, I got your email. I am waiting the package now. then try to look into your source code. 14:03
wpwrakyou'll like it - it's a big mess :)14:05
xiangfuI like swimming in your source code. :-)14:08
qi-bot[commit] Xiangfu: mini-jtag: forget create dest folder (master) http://qi-hw.com/p/openwrt-packages/fb1e28f14:24
mthlarsc (or anyone else): I'm a bit confused by ioremap_nocache: is it ever desirable to have an I/O range mapped *with* cache?16:09
larscmaybe ;)16:10
larsce.g. if the memory is non volatile16:11
larscon mips ioremap and ioremap_nocache behave the same, but here is also ioremap_chached16:11
mththe i2s driver driver uses ioremap_nocache, while most other drivers just just ioremap16:12
mthbut they're actually doing the same thing then?16:12
larscioremap is older16:12
larscat somepoint ioremap_nocache and ioremap_cached were introduced16:12
mthfor the JZ4770 drivers I'm trying to use the devm_* functions as much as possible, since it really simplifies the probe/remove operations16:13
mthbut there is no devm_request_and_ioremap_uncached, just devm_request_and_ioremap16:13
larscthat one internally uses the noncached version by default16:14
larscand falls back to cachable if the ioresource has a CACHABLE flag set16:15
mthcurrently the remapping is just for show anyway, the Ingenic code just pokes using direct pointers without using writeb etc ;)16:18
mthbut it will at least log something if two driver try to claim the same address range16:18
--- Wed Sep 5 201200:00

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