On Sun, 2015-09-27 at 12:37 +0200, Hans de Goede wrote:
This is likely a problem with your kernel config, make sure that you've the axp209 mfd and regulator drivers enabled and loaded.
Thanks, Hans. I really should have figured that out on my own!
Do you think it's worth asking PeterR if he will change the default Fedora armv7 config to CONFIG_REGULATOR_AXP20X=y rather than the current CONFIG_REGULATOR_AXP20X=m? (I suspect that request might carry more weight if it comes from you rather than me. ;)
Regards
Clive
On Sunday, 27 September 2015 13:37:01 UTC+1, Clive Messer wrote:
On Sun, 2015-09-27 at 12:37 +0200, Hans de Goede wrote:
This is likely a problem with your kernel config, make sure that you've the axp209 mfd and regulator drivers enabled and loaded.
Thanks, Hans. I really should have figured that out on my own!
Do you think it's worth asking PeterR if he will change the default Fedora armv7 config to CONFIG_REGULATOR_AXP20X=y rather than the current CONFIG_REGULATOR_AXP20X=m? (I suspect that request might carry more weight if it comes from you rather than me. ;)
Hans,
Do you have an opinion, (or anyone else for that matter), on what would be the upstream preferred way of getting the axp20x regulator driver to auto-load, from a DT reference, assuming it is built as a module? Something like the attached patch? (Not tested, just thinking out loud.)
Regards
Clive
Hi,
On 27-09-15 22:14, Clive Messer wrote:
On Sunday, 27 September 2015 13:37:01 UTC+1, Clive Messer wrote:
On Sun, 2015-09-27 at 12:37 +0200, Hans de Goede wrote:
This is likely a problem with your kernel config, make sure that you've the axp209 mfd and regulator drivers enabled and loaded.
Thanks, Hans. I really should have figured that out on my own!
Do you think it's worth asking PeterR if he will change the default Fedora armv7 config to CONFIG_REGULATOR_AXP20X=y rather than the current CONFIG_REGULATOR_AXP20X=m? (I suspect that request might carry more weight if it comes from you rather than me. ;)
Hans,
Do you have an opinion, (or anyone else for that matter), on what would be the upstream preferred way of getting the axp20x regulator driver to auto-load, from a DT reference, assuming it is built as a module? Something like the attached patch? (Not tested, just thinking out loud.)
This is already fixed by this patch:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drive...
Regards,
Hans
On Sun, 2015-09-27 at 12:37 +0200, Hans de Goede wrote:
This is likely a problem with your kernel config, make sure that you've the axp209 mfd and regulator drivers enabled and loaded.
Thanks, Hans. I really should have figured that out on my own!
Do you think it's worth asking PeterR if he will change the default Fedora armv7 config to CONFIG_REGULATOR_AXP20X=y rather than the current CONFIG_REGULATOR_AXP20X=m? (I suspect that request might carry more weight if it comes from you rather than me. ;)
Hans,
Do you have an opinion, (or anyone else for that matter), on what would be the upstream preferred way of getting the axp20x regulator driver to auto-load, from a DT reference, assuming it is built as a module? Something like the attached patch? (Not tested, just thinking out loud.)
This is already fixed by this patch:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drive...
I've pulled this patch into the F-23/4.2.x kernel, it'll be in the 4.2.2 build, it's already in rawhide.
Peter
On Wed, 2015-09-30 at 04:29 +0100, Peter Robinson wrote:
This is already fixed by this patch:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/com mit/drivers/regulator/axp20x -regulator.c?id=d4ea7d86457a8d0ea40ce77bdeda1fc966cc35ec
I've pulled this patch into the F-23/4.2.x kernel, it'll be in the 4.2.2 build, it's already in rawhide.
Thanks, Peter.
One other thing to think about. Although the cpufreq-dt module will autoload from dt ref, it needs to be in the initramfs image created by dracut. This doesn't happen automatically. Currently have created a /etc/dracut.conf.d/sunxi.conf with...
------------ ## ## Need to make sure these are in the initramfs image and they ## will be automatically loaded via DT reference ## add_drivers+="cpufreq-dt sunxi-wdt" ------------
Regards
Clive
This is already fixed by this patch:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/com mit/drivers/regulator/axp20x -regulator.c?id=d4ea7d86457a8d0ea40ce77bdeda1fc966cc35ec
I've pulled this patch into the F-23/4.2.x kernel, it'll be in the 4.2.2 build, it's already in rawhide.
Thanks, Peter.
One other thing to think about. Although the cpufreq-dt module will autoload from dt ref, it needs to be in the initramfs image created by dracut. This doesn't happen automatically. Currently have created a /etc/dracut.conf.d/sunxi.conf with...
Weird, it appears to load and work post initrd with the BeagleBone.
In /sys/devices/system/cpu/cpu0/cpufreq I get all the bits I would expect with cur/max/mon freqs etc.
I'll have a look when I get a spare moment on a A20 device.
Peter
On Wed, 2015-09-30 at 11:17 +0100, Peter Robinson wrote:
Weird, it appears to load and work post initrd with the BeagleBone.
In /sys/devices/system/cpu/cpu0/cpufreq I get all the bits I would expect with cur/max/mon freqs etc.
I'll have a look when I get a spare moment on a A20 device.
One other random thought, although along the same lines....
Need the anatop-regulator to autoload for the imx boards. As you took the patch for the autoload of axp20x reg into your 4.2.x builds, can you also pull a similar patch for anatop reg. Think this should be in 4.3 already, but not in 4.2.
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/patch/drivers...
Clive
Hi,
On 27-09-15 14:36, Clive Messer wrote:
On Sun, 2015-09-27 at 12:37 +0200, Hans de Goede wrote:
This is likely a problem with your kernel config, make sure that you've the axp209 mfd and regulator drivers enabled and loaded.
Thanks, Hans. I really should have figured that out on my own!
Do you think it's worth asking PeterR if he will change the default Fedora armv7 config to CONFIG_REGULATOR_AXP20X=y rather than the current CONFIG_REGULATOR_AXP20X=m?
Peter tries to keep the Fedora ARM kernel config as modular as possible because of the wide variety of hardware it runs on + low amount of memory on some hardware. So this will have to be kept modular (I know this because I already asked :) ).
Regards,
Hans