[RFC RESEND 2/4] ARM: OMAP3: Dynamically disable secure timer nodes for secure devices

[Jon Hunter]

On 08/17/2012 12:32 AM, Hiremath, Vaibhav wrote:
> On Thu, Aug 16, 2012 at 22:27:42, Hunter, Jon wrote:
>>
>> On 08/15/2012 04:13 AM, Vaibhav Hiremath wrote:
>>>
>>>
>>> On 7/14/2012 3:56 AM, Jon Hunter wrote:
>>>> OMAP3 devices may or may not have security features enabled. Security enabled
>>>> devices are known as high-secure (HS) and devices without security are known as
>>>> general purpose (GP).
>>>>
>>>> For OMAP3 devices there are 12 general purpose timers available. On secure
>>>> devices the 12th timer is reserved for secure usage and so cannot be used by
>>>> the kernel, where as for a GP device it is available. We can detect the OMAP
>>>> device type, secure or GP, at runtime via an on-chip register. Today, when not
>>>> using DT, we do not register the 12th timer as a linux device if the device is
>>>> secure.
>>>>
>>>> When using device tree, device tree is going to register all the timer devices
>>>> it finds in the device tree blob. To prevent device tree from registering 12th
>>>> timer on a secure OMAP3 device we can add a status property to the timer
>>>> binding with the value "disabled" at boot time. Note that timer 12 on a OMAP3
>>>> device has a property "ti,timer-secure" to indicate that it will not be
>>>> available on a secure device and so for secure OMAP3 devices, we search for
>>>> timers with this property and then disable them. Using the prom_add_property()
>>>> function to dynamically add a property was a recommended approach suggested by
>>>> Rob Herring [1].
>>>>
>>>> I have tested this on an OMAP3 GP device and faking it to pretend to be a
>>>> secure device to ensure that any timers marked with "ti,timer-secure" are not
>>>> registered on boot. I have also made sure that all timers are registered as
>>>> expected on a GP device by default.
>>>>
>>>> [1] http://comments.gmane.org/gmane.linux.ports.arm.omap/79203
>>>>
>>>> Signed-off-by: Jon Hunter <jon-hunter at ti.com>
>>>> ---
>>>>  arch/arm/mach-omap2/board-generic.c |    1 +
>>>>  arch/arm/mach-omap2/common.h        |    1 +
>>>>  arch/arm/mach-omap2/timer.c         |   36 +++++++++++++++++++++++++++++++++++
>>>>  3 files changed, 38 insertions(+)
>>>>
>>>> diff --git a/arch/arm/mach-omap2/board-generic.c b/arch/arm/mach-omap2/board-generic.c
>>>> index 6f93a20..20124d7 100644
>>>> --- a/arch/arm/mach-omap2/board-generic.c
>>>> +++ b/arch/arm/mach-omap2/board-generic.c
>>>> @@ -40,6 +40,7 @@ static struct of_device_id omap_dt_match_table[] __initdata = {
>>>>  static void __init omap_generic_init(void)
>>>>  {
>>>>  	omap_sdrc_init(NULL, NULL);
>>>> +	omap_dmtimer_init();
>>>>  
>>>>  	of_platform_populate(NULL, omap_dt_match_table, NULL, NULL);
>>>>  }
>>>> diff --git a/arch/arm/mach-omap2/common.h b/arch/arm/mach-omap2/common.h
>>>> index 1f65b18..d6a4875 100644
>>>> --- a/arch/arm/mach-omap2/common.h
>>>> +++ b/arch/arm/mach-omap2/common.h
>>>> @@ -326,6 +326,7 @@ extern void omap_sdrc_init(struct omap_sdrc_params *sdrc_cs0,
>>>>  				      struct omap_sdrc_params *sdrc_cs1);
>>>>  struct omap2_hsmmc_info;
>>>>  extern int omap4_twl6030_hsmmc_init(struct omap2_hsmmc_info *controllers);
>>>> +extern void omap_dmtimer_init(void);
>>>>  
>>>>  #endif /* __ASSEMBLER__ */
>>>>  #endif /* __ARCH_ARM_MACH_OMAP2PLUS_COMMON_H */
>>>> diff --git a/arch/arm/mach-omap2/timer.c b/arch/arm/mach-omap2/timer.c
>>>> index 13d20c8..e3b9931 100644
>>>> --- a/arch/arm/mach-omap2/timer.c
>>>> +++ b/arch/arm/mach-omap2/timer.c
>>>> @@ -36,6 +36,7 @@
>>>>  #include <linux/clocksource.h>
>>>>  #include <linux/clockchips.h>
>>>>  #include <linux/slab.h>
>>>> +#include <linux/of.h>
>>>>  
>>>>  #include <asm/mach/time.h>
>>>>  #include <plat/dmtimer.h>
>>>> @@ -482,6 +483,41 @@ static int __init omap2_dm_timer_init(void)
>>>>  }
>>>>  arch_initcall(omap2_dm_timer_init);
>>>>  
>>>> +static struct property timer_disabled = {
>>>> +	.name = "status",
>>>> +	.length = sizeof("disabled"),
>>>> +	.value = "disabled",
>>>> +};
>>>> +
>>>> +static struct of_device_id omap3_timer_match[] __initdata = {
>>>> +	{ .compatible = "ti,omap3-timer", },
>>>> +	{ }
>>>> +};
>>>> +
>>>> +/**
>>>> + * omap_dmtimer_init - initialisation function when device tree is used
>>>> + *
>>>> + * For secure OMAP3 devices, timers with device type "timer-secure" cannot
>>>> + * be used by the kernel as they are reserved. Therefore, to prevent the
>>>> + * kernel registering these devices remove them dynamically from the device
>>>> + * tree on boot.
>>>> + */
>>>> +void __init omap_dmtimer_init(void)
>>>> +{
>>>> +	struct device_node *np;
>>>> +
>>>> +	if (!cpu_is_omap34xx())
>>>> +		return;
>>>> +
>>>
>>> Sorry for responding so late, but why only omap34xx check here?
>>> Isn't this applicable to all omap & non-omap devices?
>>
>> It is only applicable to omap3 devices as far as omap is concerned.
>>
>> By non-omap, you are referring to the AMxxx stuff?
>>
>> Do AMxxx devices even support security (ie. secure boot and have secure
>> peripherals)? If not then this will work for AMxxx devices too.
>>
> 
> Yes it does. 
> 
> AM33xx has 8 timers and Timer-0 is a secure timer. As in case of OMAP3, we 
> do not even register timer-0 to kernel.

Ok, however, please note that the omap3 timer 12 is a special case,
because on a non-secure device it is available for general purpose use
and only reserved for secure use on a secure device.

For omap4/5 timer 12 is always a secure timer and so we never register
it for any device (secure or non-secure) by using the "ti,timer-secure"
property in DT.

So, for AM33xx, is timer0 always reserved for secure used on non-secure
and secure devices? If so, then timer0 just needs to add the
"ti,timer-secure" property in device-tree file and it will not be
registered and no change is needed in the above code.

>> Let me know if the changelog is not clear why this is needed for an
>> omap3 device.
>>
> 
> The changelog description is clear, but it is not only restricted to OMAP3.

Not exactly, per the above it depends on whether the timer is always
secure for all device types or not.

Cheers
Jon