"Mark A. Greer" [off-list ref] writes:
On Fri, Apr 27, 2012 at 03:25:38PM -0700, Kevin Hilman wrote:
quoted
"Mark A. Greer" [off-list ref] writes:
quoted
On Thu, Apr 26, 2012 at 04:29:45PM -0700, Kevin Hilman wrote:
Hi Kevin.
quoted
This is a rebased version of this series which is ready for broader
testing. I'd especially appreciate testing from those of you with
AM35x platforms.
Currently, our SoC detection is based on SoC family detection
(using die ID) and the presence of specific IP blocks (or feature.)
This series begins the separation of the SoC family detection and
specific IP detection by completely removing IP detection from the SoC
family detection (cpu_is_*.)
Applies on top of v3.4-rc4, boot tested on AM3517 EVM.
Your series applied and booted without issue until the point that
pm_idle is called. At that point it hung because its doing a wfi
and not getting an interrupt to wake up back up. That's expected
behaviour so your patches are good (IMHO) but I need to finish and
submit a new version of my patches to fix that issue.
For the benefit of others wanting to test this:
You have to use 'nohlt' on the cmdline on AM35x in order to avoid WFI
and the wakeup problems that are still being worked on.
Oops, yes. I meant to add that and say that my am3517evm boots fine
with your patches and 'nohlt' on the cmdline. I'm using omap2plus_defconfig
with CONFIG_TI_DAVINCI_EMAC enabled.
Thanks for testing, will add a Tested-by from you.
Kevin