#qi-hardware IRC log for Friday, 2015-06-26

mthapelete, larsc: the OHCI controller on the JZ4770 works with the ohci-platform driver; if that is the case for JZ4740 as well the ohci-jz4740 driver could be retired05:55
larscmth: the driver is using ohci-platform, just with some custom exensions for turning on a external regulator iirc11:05
larscwpwrak: It's interesting that IPv6 is as old as the Linux kernel11:13
mthlarsc: it is using a mostly generic driver, yes, but not the driver called "ohci-platform"15:48
larscmth: right15:48
mthI think all the things it does in those extensions can be done with DT or pdata and ohci-platform15:49
larscthe regulator stuff could be replaced by the generic phy15:49
larscpatches are welcome ;)15:49
mthwe haven't kept the Dingoo A320 kernel up-to-date, so testing for me would take a lot of time15:50
larscI don't have a board with the usb host anymore15:51
mthactualy, the A320 doesn't use the usb host either15:51
larscso just remove the driver?15:52
mthI do have a patch for that: https://github.com/gcwnow/linux/commit/f5bcf87415:52
mthbut it doesn't verify the assumption that everything the custom code did can be done with ohci-platform as well15:53
mthalthough a quick look at the code suggests that it can15:53
larscit probably can15:53
pcercueisomebody should test on jz4740...15:55
mthon the other hand, if there is no-one using the driver anymore, it is safe to remove it ;)15:56
pcercueino one is using jz4740 anymore, it is safe to remove it :)15:58
mthlarsc: would such an immediate removal patch be accepted or would there have to be a deprecation period?16:07
larscif you say there are no more users there should be no reason not to drop it16:07
larscjust make sure to also drop avt2 support from the qi_lb60 board driver16:07
larscsay that it was only a prototype platform of which only a view pcbs were made, none of which are in use anymore16:08
mthwell, I don't know any persons using it, that doesn't mean they can't exist somewhere16:08
mthok16:10
mthby the way, would this be useful to submit? https://github.com/gcwnow/linux/commit/e479f45316:11
larscralf said binutils needs the 0xfffffff....16:13
larscthere were only like 5 or so of the avt2 boards16:13
larscI had one, and wolfgang had some16:13
mthI've been building without the ffffffff for a long time and there are no problems here16:15
mthmaybe they're needed for 64-bit target systems?16:15
larsche said for all systems16:16
larscwhen you are building on a 64-bit host16:17
mththat's what I do most of the time16:17
larscjoin #mipslinix16:17
larscjoin #mipslinux16:17
larscand discuess it with him16:17
mthI should probably double-check the problem first; the patch is a cleaned up version of something I wrote long ago16:19
mthiirc the problem is that a "long" arg is declared using libconfuse and that lib will just parse as many digits as fit in the "long" type and call it a day16:20
mthlibconfuse used by mkimage16:20
larscyes16:21
larscso mkimage only breaks on a 32-bit host16:21
mthdoes the NanoNote load a uImage? we use vmlinuz.bin with hardcoded load address for the GCW Zero16:23
larscuImage16:25
--- Sat Jun 27 201500:00

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