| Message ID | 1398418275-9671-7-git-send-email-marc.zyngier@arm.com |
|---|---|
| State | New |
| Headers | show |
On Fri, Apr 25, 2014 at 10:31:14AM +0100, Marc Zyngier wrote: > So far, the CMOS RTC (PC-style mc146818) has always been used > through an ISA I/O port of some sort, with each platform configuring > the address in a static way. > > Some platforms (KVM with kvmtools, for example) offer a MMIO version > of the CMOS RTC, which requires a different access method. > > This patch select the RTC_DRV_CMOS_MMIO configuration in order to > support MMIO accesses as well as the older IO port method. > > While we're at it, switch to RTC_DRV_CMOS_PRIV_LOCK and remove the > ancient rtc_lock. rtc_lock is there so that if you build the RTC driver and nvram driver in, and load them, they will both want to access the RTC via its indexed registers, and they need to share the lock to avoid trampling on each others toes. Yes, it should be handled in a more modern way, but I don't think you can simply get rid of this in this manner. We can have both these drivers loaded on ARM platforms.
Hi Russell, On 25/04/14 11:27, Russell King - ARM Linux wrote: > On Fri, Apr 25, 2014 at 10:31:14AM +0100, Marc Zyngier wrote: >> So far, the CMOS RTC (PC-style mc146818) has always been used >> through an ISA I/O port of some sort, with each platform configuring >> the address in a static way. >> >> Some platforms (KVM with kvmtools, for example) offer a MMIO version >> of the CMOS RTC, which requires a different access method. >> >> This patch select the RTC_DRV_CMOS_MMIO configuration in order to >> support MMIO accesses as well as the older IO port method. >> >> While we're at it, switch to RTC_DRV_CMOS_PRIV_LOCK and remove the >> ancient rtc_lock. > > rtc_lock is there so that if you build the RTC driver and nvram driver > in, and load them, they will both want to access the RTC via its > indexed registers, and they need to share the lock to avoid trampling > on each others toes. Ah, I completely missed that one. Note to self: never look at that kind of code just after lunch... > Yes, it should be handled in a more modern way, but I don't think you > can simply get rid of this in this manner. We can have both these > drivers loaded on ARM platforms. Agreed. I suppose that would have to be handled by some kind of module dependency, but it is starting to look even uglier than I initially thought. Back to the drawing board... Thanks, M.
diff --git a/arch/arm/kernel/time.c b/arch/arm/kernel/time.c index 829a96d..38c8718 100644 --- a/arch/arm/kernel/time.c +++ b/arch/arm/kernel/time.c @@ -32,13 +32,6 @@ #include <asm/stacktrace.h> #include <asm/thread_info.h> -#if defined(CONFIG_RTC_DRV_CMOS) || defined(CONFIG_RTC_DRV_CMOS_MODULE) || \ - defined(CONFIG_NVRAM) || defined(CONFIG_NVRAM_MODULE) -/* this needs a better home */ -DEFINE_SPINLOCK(rtc_lock); -EXPORT_SYMBOL(rtc_lock); -#endif /* pc-style 'CMOS' RTC support */ - /* change this if you have some constant time drift */ #define USECS_PER_JIFFY (1000000/HZ) diff --git a/drivers/rtc/Kconfig b/drivers/rtc/Kconfig index 12bc27d..b2a512e 100644 --- a/drivers/rtc/Kconfig +++ b/drivers/rtc/Kconfig @@ -662,6 +662,8 @@ config RTC_DRV_CMOS tristate "PC-style 'CMOS'" depends on X86 || ARM || M32R || PPC || MIPS || SPARC64 default y if X86 + select RTC_DRV_CMOS_MMIO if ARM + select RTC_DRV_CMOS_PRIV_LOCK if ARM help Say "yes" here to get direct support for the real time clock found in every PC or ACPI-based system, and some other boards.
So far, the CMOS RTC (PC-style mc146818) has always been used through an ISA I/O port of some sort, with each platform configuring the address in a static way. Some platforms (KVM with kvmtools, for example) offer a MMIO version of the CMOS RTC, which requires a different access method. This patch select the RTC_DRV_CMOS_MMIO configuration in order to support MMIO accesses as well as the older IO port method. While we're at it, switch to RTC_DRV_CMOS_PRIV_LOCK and remove the ancient rtc_lock. Tested on Cortex-A7 with KVM and kvmtools. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> --- arch/arm/kernel/time.c | 7 ------- drivers/rtc/Kconfig | 2 ++ 2 files changed, 2 insertions(+), 7 deletions(-)