Fwd: Fwd: Re: Cross Compile Module for armv7 - Booting 3.15 rc5 on Allwinner A10
Proceeded to compile kernel and modules - without error. Followed steps in attached file. Fails to boot and because I have no access to u-boot prompt am unable to troubleshoot. Anyone have any comments on my compile process? Thanks ................................. I didn't get far: make -j5 ARCH=arm CROSS_COMPILE=/usr/bin/arm-linux-gnueabihf- defconfig sunxi_defconfig HOSTCC scripts/basic/fixdep HOSTCC scripts/kconfig/conf.o SHIPPED scripts/kconfig/zconf.tab.c SHIPPED scripts/kconfig/zconf.lex.c SHIPPED scripts/kconfig/zconf.hash.c HOSTCC scripts/kconfig/conf.o HOSTCC scripts/kconfig/zconf.tab.o HOSTCC scripts/kconfig/zconf.tab.o HOSTLD scripts/kconfig/conf fixdep: error opening depfile: scripts/kconfig/.zconf.tab.o.d: No such file or directory make[1]: *** [scripts/kconfig/zconf.tab.o] Error 2 make: *** [defconfig] Error 2 make: *** Waiting for unfinished jobs.... # # configuration written to .config I have scripts/kconfig/.zconf.tab.o.cmd and scripts/kconfig/.zconf.tab.o.tmp How do I report an issue with this? I still got a .config so I will proceed and see what happens. Thanks Rob -------- Original Message -------- Subject: Re: Cross Compile Module for armv7 Date: Mon, 19 May 2014 22:39:51 +1000 From: Robert Brown <rebrown@exemail.com.au> To: Chris Samuel <chris@csamuel.org> Wow thanks I would never have found that. I will give it a go over the next day or two. Can I expect most things to be working yet e.g. usb, network, HDMI display,wireless? Regards Rob On 19/05/14 21:34, Chris Samuel wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Hi Rob,
On Mon, 19 May 2014 06:36:48 PM Robert Brown wrote:
Thanks for the info. Pleasure!
Looking here: https://github.com/linux-sunxi/linux-sunxi/tree/sunxi-devel
it looks like 3.15 rc1. Should I be checking somewhere else? If the architecture is supported in the mainline kernel then you can just use the standard place, which is:
The latest tarball is here:
https://www.kernel.org/pub/linux/kernel/v3.x/testing/linux-3.15-rc5.tar.xz
but there are currently 77 commits beyond that (not a lot, but that's because Linus's travelling and is off-net), including some for Allwinner chips, so you may want to use git to clone his repo and build from there.
Browse here:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/
Clone with this:
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
cheers! Chris - -- Chris Samuel :http://www.csamuel.org/ : Melbourne, VIC
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Tue, 20 May 2014 07:52:52 PM Robert Brown wrote:
Proceeded to compile kernel and modules - without error. Followed steps in attached file. Fails to boot and because I have no access to u-boot prompt am unable to troubleshoot. Anyone have any comments on my compile process?
I suspect that's related to the fact that the boards assume FEX support whereas the mainline kernel universe uses Device Tree instead. The LWN article[1] I've just sent you off-list comments that: # The solution was a runtime FEX-to-DT translator called sunxi-babelfish. # It behaves like a normal U-Boot kernel image, but instead reads the # FEX file and constructs a DT file that gets passed to the DT-based # kernel that is next in the boot progression. So mainline kernels can # now boot on most available boards and tablets. Why not support FEX? Well basically because it's a reinvention of Device Tree, but badly designed. # Leighton showed why FEX was unsuitable for more than just a single # SoC—it was not possible to support a driver for new hardware in # Allwinner's FEX-only kernels. But as the article goes on to point out the mainline support for these devices has a long way to go, so it looks for the near future you're really stuck with the old, hardware specific, kernels from the sunxi people. :-( [1] - https://lwn.net/Articles/597916/ for other subscribers, free from Thursday when the next LWN comes out. Best of luck, Chris - -- Chris Samuel : http://www.csamuel.org/ : Melbourne, VIC -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iQEVAwUBU3swl41yjaOTJg85AQKD7ggAqeSvL8qwXrZSuq7uqvYf2vEwJ4BOVjO9 Luhyi9joQyTJ6bewnkmX8kBqtqlZ6CRMX4UI/XmMvN2ww8xpU8mpR3htMf2ZViq8 HRrkSpm/daKlCLNUOea5+VK3v3TbyuDzIEinhQQuyq0dqXV6ofhZ7zFKoPPj0eiT EwWcA0bk/LhrZBS9dAVwfLxput0h6NV9rBjDhnYr63WLMKmikk2ONFy0IPwe73lz AudzviBQlVEU4rZoFjwL4BJwDrTaswJc5/CEeq3n4UKhysuY0QCFQiS1/OD/5stz 03E87kJcEWrMRJsYuPQAKI6bAELwTuRGlMOeXt8B25I7XN/Oqu6udw== =aBoY -----END PGP SIGNATURE-----
On Tue, May 20, 2014 at 08:38:12PM +1000, Chris Samuel wrote:
Why not support FEX? Well basically because it's a reinvention of Device Tree, but badly designed.
to be fair, Allwinner's FEX was around before anyone decided that DT for Arm was the right way to go. IIRC DT in mainline was only for PPC at that time. the lwn article doesn't really point that out. BTW, from following sunxi-devel it seems DT has had to change a lot (and is still evolving) to support all the Arm SoCs out there. generalising is hard and slow work. I suspect FEX was probably the right solution for the company at the time, and it's certainly better than hardcoding everything like older Android kernels used to do. (as a data point my new qualcomm based phone runs a 3.4 kernel with DT) I'm reasonably hopeful my a10 DAC + backup, and a20 usb video cubies will be running mainline kernels in my lifetime :) cheers, robin
[1] - https://lwn.net/Articles/597916/ for other subscribers, free from Thursday when the next LWN comes out.
On 21/05/14 02:01, Robin Humble wrote:
BTW, from following sunxi-devel it seems DT has had to change a lot (and is still evolving) to support all the Arm SoCs out there. generalising is hard and slow work. I suspect FEX was probably the right solution for the company at the time, and it's certainly better than hardcoding everything like older Android kernels used to do. (as a data point my new qualcomm based phone runs a 3.4 kernel with DT) I'm reasonably hopeful my a10 DAC + backup, and a20 usb video cubies will be running mainline kernels in my lifetime :) cheers, robin
Gee ..... I hope you're about 105 years old when you say that. :>)
[1] - https://lwn.net/Articles/597916/ for other subscribers, free from Thursday when the next LWN comes out.
luv-main mailing list luv-main@luv.asn.au http://lists.luv.asn.au/listinfo/luv-main
Just to clarify something in the article: On 20/05/14 20:38, Chris Samuel wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On Tue, 20 May 2014 07:52:52 PM Robert Brown wrote:
Proceeded to compile kernel and modules - without error. Followed steps in attached file. Fails to boot and because I have no access to u-boot prompt am unable to troubleshoot. Anyone have any comments on my compile process? I suspect that's related to the fact that the boards assume FEX support whereas the mainline kernel universe uses Device Tree instead.
The LWN article[1] I've just sent you off-list comments that:
# The solution was a runtime FEX-to-DT translator called sunxi-babelfish. # It behaves like a normal U-Boot kernel image, but instead reads the # FEX file and constructs a DT file that gets passed to the DT-based # kernel that is next in the boot progression. So mainline kernels can # now boot on most available boards and tablets. The A10 is presumably in this list so should at least boot on 3.15 rc5? and also from the article: "Most everything needed for a headless device is available" I am ok with it being headless but can any boot problems be diagnosed?
Why not support FEX? Well basically because it's a reinvention of Device Tree, but badly designed.
# Leighton showed why FEX was unsuitable for more than just a single # SoC—it was not possible to support a driver for new hardware in # Allwinner's FEX-only kernels.
But as the article goes on to point out the mainline support for these devices has a long way to go, so it looks for the near future you're really stuck with the old, hardware specific, kernels from the sunxi people. :-(
[1] - https://lwn.net/Articles/597916/ for other subscribers, free from Thursday when the next LWN comes out.
Best of luck, Chris - -- Chris Samuel : http://www.csamuel.org/ : Melbourne, VIC
-----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux)
iQEVAwUBU3swl41yjaOTJg85AQKD7ggAqeSvL8qwXrZSuq7uqvYf2vEwJ4BOVjO9 Luhyi9joQyTJ6bewnkmX8kBqtqlZ6CRMX4UI/XmMvN2ww8xpU8mpR3htMf2ZViq8 HRrkSpm/daKlCLNUOea5+VK3v3TbyuDzIEinhQQuyq0dqXV6ofhZ7zFKoPPj0eiT EwWcA0bk/LhrZBS9dAVwfLxput0h6NV9rBjDhnYr63WLMKmikk2ONFy0IPwe73lz AudzviBQlVEU4rZoFjwL4BJwDrTaswJc5/CEeq3n4UKhysuY0QCFQiS1/OD/5stz 03E87kJcEWrMRJsYuPQAKI6bAELwTuRGlMOeXt8B25I7XN/Oqu6udw== =aBoY -----END PGP SIGNATURE-----
_______________________________________________ luv-main mailing list luv-main@luv.asn.au http://lists.luv.asn.au/listinfo/luv-main
participants (3)
-
Chris Samuel -
Robert Brown -
Robin Humble