Hi Peter, Andre,
On 4/24/24 12:29, Andre Przywara wrote:
On Tue, 23 Apr 2024 12:55:55 +0200 Quentin Schulz quentin.schulz@theobroma-systems.com wrote:
Hi Peter,
On 4/23/24 10:10, Peter Hoyes wrote:
From: Peter Hoyes Peter.Hoyes@arm.com
Polling cntpct_el0 in a tight loop for delays is inefficient. This is particularly apparent on Arm FVPs, which do not simulate real time, meaning that a 1s sleep can take a couple of orders of magnitude longer to execute in wall time.
If running at EL2 or above (where CNTHCTL_EL2 is available), enable the cntpct_el0 event stream temporarily and use wfe to implement the delay more efficiently. The event period is chosen as a trade-off between efficiency and the fact that Arm FVPs do not typically simulate real time.
This is only implemented for Armv8 boards, where an architectural timer exists.
Some mach-socfpga AArch64 boards already override __udelay to make it always inline, so guard the functionality with a new ARMV8_UDELAY_EVENT_STREAM Kconfig, enabled by default.
Signed-off-by: Peter Hoyes Peter.Hoyes@arm.com
arch/arm/cpu/armv8/Kconfig | 8 ++++++++ arch/arm/cpu/armv8/generic_timer.c | 27 +++++++++++++++++++++++++++ arch/arm/include/asm/system.h | 6 ++++-- 3 files changed, 39 insertions(+), 2 deletions(-)
diff --git a/arch/arm/cpu/armv8/Kconfig b/arch/arm/cpu/armv8/Kconfig index 9f0fb369f7..544c5e2d74 100644 --- a/arch/arm/cpu/armv8/Kconfig +++ b/arch/arm/cpu/armv8/Kconfig @@ -191,6 +191,14 @@ config ARMV8_EA_EL3_FIRST Exception handling at all exception levels for External Abort and SError interrupt exception are taken in EL3.
+config ARMV8_UDELAY_EVENT_STREAM
- bool "Use the event stream for udelay"
- default y if !ARCH_SOCFPGA
- help
Use the event stream provided by the AArch64 architectural timer fordelays. This is more efficient than the default pollingimplementation.- menuconfig ARMV8_CRYPTO bool "ARM64 Accelerated Cryptographic Algorithms"
diff --git a/arch/arm/cpu/armv8/generic_timer.c b/arch/arm/cpu/armv8/generic_timer.c index 8f83372cbc..e18b5c8187 100644 --- a/arch/arm/cpu/armv8/generic_timer.c +++ b/arch/arm/cpu/armv8/generic_timer.c @@ -115,3 +115,30 @@ ulong timer_get_boot_us(void)
return val / get_tbclk();}
+#if CONFIG_IS_ENABLED(ARMV8_UDELAY_EVENT_STREAM) +void __udelay(unsigned long usec) +{
- u64 target = get_ticks() + usec_to_tick(usec);
This can theoretically overflow, do we have any guarantee this cannot happen in real life, like... we would need U-Boot to be running for 100 years without being powered-down/reset or something like that? Can we document this assumption? Does this make sense?
The Arm ARM guarantees a "Roll-over time of not less than 40 years." (Armv8 ARM 0487K.a D12.1.2 "The system counter"). So that's not the 100 years you are asking for, but I guess still good enough?
I guess it is, since it is also stored in a u64 and is reset to 0 upon start-up according to the ARM. I also assume nobody is going to add a udelay of years in their code :) (and if they do, they would probably figure out something's wrong before it reaches the final products :) ).
Thanks all for the pointers to reference manuals and current implementations.
Cheers, Quentin