> In which case this kernel patch needs instead to speak the bootloader > wakeup protocol instead of speaking to the h/w directly like you've done > here, right? > > Or is it possible for the bootloader to set these things up and then put > the CPU back to sleep such that it both retains any settings and is > wakable by this patch? This code contains core resets and power control, > which makes me suspect not.
And I think secondary cpus remains setting after h/w boot. >> > Wouldn't it be better to do all this stuff in the bootloader and >> either >> > implement PSCI or have the bootloader do the traditional holding pen >> and >> > mbox address thing? Uboot doesnot support PSCI, it use traditional holding pen for sunxi platform now. >> > >> I have modified uboot to set cntfrq and cntvoff in all smp cpus,and it >> works well. I guess kernel should believe all cpu should be all same >> when kernel boot. Bootloader should do it to ensure that. > > Yes, I think all CPUs must be in the same state at boot. > > But if you've done all that then what is this patch for? > > Do you have links to your u-boot patches? > > Ian. Yes, This is my patch for uboot. My working uboot code is in https://github.com/linux-sunxi/u-boot-sunxi.git Note this is only test patch. I have not commit it formally for uboot. Fan
0001-Add-setting-CNTVOFF-and-CNTFRQ-for-all-smp-cpus.patch
Description: Binary data
