Struct cortex_m::peripheral::SCB
source · pub struct SCB { /* private fields */ }
Expand description
System Control Block
Implementations§
source§impl SCB
impl SCB
sourcepub fn vect_active() -> VectActive
pub fn vect_active() -> VectActive
Returns the active exception number
source§impl SCB
impl SCB
sourcepub fn enable_icache(&mut self)
pub fn enable_icache(&mut self)
Enables I-cache if currently disabled.
This operation first invalidates the entire I-cache.
sourcepub fn disable_icache(&mut self)
pub fn disable_icache(&mut self)
Disables I-cache if currently enabled.
This operation invalidates the entire I-cache after disabling.
sourcepub fn icache_enabled() -> bool
pub fn icache_enabled() -> bool
Returns whether the I-cache is currently enabled.
sourcepub fn invalidate_icache(&mut self)
pub fn invalidate_icache(&mut self)
Invalidates the entire I-cache.
sourcepub fn enable_dcache(&mut self, cpuid: &mut CPUID)
pub fn enable_dcache(&mut self, cpuid: &mut CPUID)
Enables D-cache if currently disabled.
This operation first invalidates the entire D-cache, ensuring it does not contain stale values before being enabled.
sourcepub fn disable_dcache(&mut self, cpuid: &mut CPUID)
pub fn disable_dcache(&mut self, cpuid: &mut CPUID)
Disables D-cache if currently enabled.
This operation subsequently cleans and invalidates the entire D-cache, ensuring all contents are safely written back to main memory after disabling.
sourcepub fn dcache_enabled() -> bool
pub fn dcache_enabled() -> bool
Returns whether the D-cache is currently enabled.
sourcepub fn clean_dcache(&mut self, cpuid: &mut CPUID)
pub fn clean_dcache(&mut self, cpuid: &mut CPUID)
Cleans the entire D-cache.
This function causes everything in the D-cache to be written back to main memory, overwriting whatever is already there.
sourcepub fn clean_invalidate_dcache(&mut self, cpuid: &mut CPUID)
pub fn clean_invalidate_dcache(&mut self, cpuid: &mut CPUID)
Cleans and invalidates the entire D-cache.
This function causes everything in the D-cache to be written back to main memory, and then marks the entire D-cache as invalid, causing future reads to first fetch from main memory.
sourcepub unsafe fn invalidate_dcache_by_address(&mut self, addr: usize, size: usize)
pub unsafe fn invalidate_dcache_by_address(&mut self, addr: usize, size: usize)
Invalidates D-cache by address.
addr
: The address to invalidate, which must be cache-line aligned.size
: Number of bytes to invalidate, which must be a multiple of the cache line size.
Invalidates D-cache cache lines, starting from the first line containing addr
,
finishing once at least size
bytes have been invalidated.
Invalidation causes the next read access to memory to be fetched from main memory instead of the cache.
§Cache Line Sizes
Cache line sizes vary by core. For all Cortex-M7 cores, the cache line size is fixed
to 32 bytes, which means addr
must be 32-byte aligned and size
must be a multiple
of 32. At the time of writing, no other Cortex-M cores have data caches.
If addr
is not cache-line aligned, or size
is not a multiple of the cache line size,
other data before or after the desired memory would also be invalidated, which can very
easily cause memory corruption and undefined behaviour.
§Safety
After invalidating, the next read of invalidated data will be from main memory. This may cause recent writes to be lost, potentially including writes that initialized objects. Therefore, this method may cause uninitialized memory or invalid values to be read, resulting in undefined behaviour. You must ensure that main memory contains valid and initialized values before invalidating.
addr
must be aligned to the size of the cache lines, and size
must be a
multiple of the cache line size, otherwise this function will invalidate other memory,
easily leading to memory corruption and undefined behaviour. This precondition is checked
in debug builds using a debug_assert!()
, but not checked in release builds to avoid
a runtime-dependent panic!()
call.
sourcepub unsafe fn invalidate_dcache_by_ref<T>(&mut self, obj: &mut T)
pub unsafe fn invalidate_dcache_by_ref<T>(&mut self, obj: &mut T)
Invalidates an object from the D-cache.
obj
: The object to invalidate.
Invalidates D-cache starting from the first cache line containing obj
,
continuing to invalidate cache lines until all of obj
has been invalidated.
Invalidation causes the next read access to memory to be fetched from main memory instead of the cache.
§Cache Line Sizes
Cache line sizes vary by core. For all Cortex-M7 cores, the cache line size is fixed
to 32 bytes, which means obj
must be 32-byte aligned, and its size must be a multiple
of 32 bytes. At the time of writing, no other Cortex-M cores have data caches.
If obj
is not cache-line aligned, or its size is not a multiple of the cache line size,
other data before or after the desired memory would also be invalidated, which can very
easily cause memory corruption and undefined behaviour.
§Safety
After invalidating, obj
will be read from main memory on next access. This may cause
recent writes to obj
to be lost, potentially including the write that initialized it.
Therefore, this method may cause uninitialized memory or invalid values to be read,
resulting in undefined behaviour. You must ensure that main memory contains a valid and
initialized value for T before invalidating obj
.
obj
must be aligned to the size of the cache lines, and its size must be a
multiple of the cache line size, otherwise this function will invalidate other memory,
easily leading to memory corruption and undefined behaviour. This precondition is checked
in debug builds using a debug_assert!()
, but not checked in release builds to avoid
a runtime-dependent panic!()
call.
sourcepub unsafe fn invalidate_dcache_by_slice<T>(&mut self, slice: &mut [T])
pub unsafe fn invalidate_dcache_by_slice<T>(&mut self, slice: &mut [T])
Invalidates a slice from the D-cache.
slice
: The slice to invalidate.
Invalidates D-cache starting from the first cache line containing members of slice
,
continuing to invalidate cache lines until all of slice
has been invalidated.
Invalidation causes the next read access to memory to be fetched from main memory instead of the cache.
§Cache Line Sizes
Cache line sizes vary by core. For all Cortex-M7 cores, the cache line size is fixed
to 32 bytes, which means slice
must be 32-byte aligned, and its size must be a multiple
of 32 bytes. At the time of writing, no other Cortex-M cores have data caches.
If slice
is not cache-line aligned, or its size is not a multiple of the cache line size,
other data before or after the desired memory would also be invalidated, which can very
easily cause memory corruption and undefined behaviour.
§Safety
After invalidating, slice
will be read from main memory on next access. This may cause
recent writes to slice
to be lost, potentially including the write that initialized it.
Therefore, this method may cause uninitialized memory or invalid values to be read,
resulting in undefined behaviour. You must ensure that main memory contains valid and
initialized values for T before invalidating slice
.
slice
must be aligned to the size of the cache lines, and its size must be a
multiple of the cache line size, otherwise this function will invalidate other memory,
easily leading to memory corruption and undefined behaviour. This precondition is checked
in debug builds using a debug_assert!()
, but not checked in release builds to avoid
a runtime-dependent panic!()
call.
sourcepub fn clean_dcache_by_address(&mut self, addr: usize, size: usize)
pub fn clean_dcache_by_address(&mut self, addr: usize, size: usize)
Cleans D-cache by address.
addr
: The address to start cleaning at.size
: The number of bytes to clean.
Cleans D-cache cache lines, starting from the first line containing addr
,
finishing once at least size
bytes have been invalidated.
Cleaning the cache causes whatever data is present in the cache to be immediately written to main memory, overwriting whatever was in main memory.
§Cache Line Sizes
Cache line sizes vary by core. For all Cortex-M7 cores, the cache line size is fixed
to 32 bytes, which means addr
should generally be 32-byte aligned and size
should be a
multiple of 32. At the time of writing, no other Cortex-M cores have data caches.
If addr
is not cache-line aligned, or size
is not a multiple of the cache line size,
other data before or after the desired memory will also be cleaned. From the point of view
of the core executing this function, memory remains consistent, so this is not unsound,
but is worth knowing about.
sourcepub fn clean_dcache_by_ref<T>(&mut self, obj: &T)
pub fn clean_dcache_by_ref<T>(&mut self, obj: &T)
Cleans an object from the D-cache.
obj
: The object to clean.
Cleans D-cache starting from the first cache line containing obj
,
continuing to clean cache lines until all of obj
has been cleaned.
It is recommended that obj
is both aligned to the cache line size and a multiple of
the cache line size long, otherwise surrounding data will also be cleaned.
Cleaning the cache causes whatever data is present in the cache to be immediately written to main memory, overwriting whatever was in main memory.
sourcepub fn clean_dcache_by_slice<T>(&mut self, slice: &[T])
pub fn clean_dcache_by_slice<T>(&mut self, slice: &[T])
Cleans a slice from D-cache.
slice
: The slice to clean.
Cleans D-cache starting from the first cache line containing members of slice
,
continuing to clean cache lines until all of slice
has been cleaned.
It is recommended that slice
is both aligned to the cache line size and a multiple of
the cache line size long, otherwise surrounding data will also be cleaned.
Cleaning the cache causes whatever data is present in the cache to be immediately written to main memory, overwriting whatever was in main memory.
sourcepub fn clean_invalidate_dcache_by_address(&mut self, addr: usize, size: usize)
pub fn clean_invalidate_dcache_by_address(&mut self, addr: usize, size: usize)
Cleans and invalidates D-cache by address.
addr
: The address to clean and invalidate.size
: The number of bytes to clean and invalidate.
Cleans and invalidates D-cache starting from the first cache line containing addr
,
finishing once at least size
bytes have been cleaned and invalidated.
It is recommended that addr
is aligned to the cache line size and size
is a multiple of
the cache line size, otherwise surrounding data will also be cleaned.
Cleaning and invalidating causes data in the D-cache to be written back to main memory, and then marks that data in the D-cache as invalid, causing future reads to first fetch from main memory.
source§impl SCB
impl SCB
sourcepub fn set_sleepdeep(&mut self)
pub fn set_sleepdeep(&mut self)
Set the SLEEPDEEP bit in the SCR register
sourcepub fn clear_sleepdeep(&mut self)
pub fn clear_sleepdeep(&mut self)
Clear the SLEEPDEEP bit in the SCR register
source§impl SCB
impl SCB
sourcepub fn set_sleeponexit(&mut self)
pub fn set_sleeponexit(&mut self)
Set the SLEEPONEXIT bit in the SCR register
sourcepub fn clear_sleeponexit(&mut self)
pub fn clear_sleeponexit(&mut self)
Clear the SLEEPONEXIT bit in the SCR register
source§impl SCB
impl SCB
sourcepub fn set_pendsv()
pub fn set_pendsv()
Set the PENDSVSET bit in the ICSR register which will pend the PendSV interrupt
sourcepub fn is_pendsv_pending() -> bool
pub fn is_pendsv_pending() -> bool
Check if PENDSVSET bit in the ICSR register is set meaning PendSV interrupt is pending
sourcepub fn clear_pendsv()
pub fn clear_pendsv()
Set the PENDSVCLR bit in the ICSR register which will clear a pending PendSV interrupt
sourcepub fn set_pendst()
pub fn set_pendst()
Set the PENDSTSET bit in the ICSR register which will pend a SysTick interrupt
sourcepub fn is_pendst_pending() -> bool
pub fn is_pendst_pending() -> bool
Check if PENDSTSET bit in the ICSR register is set meaning SysTick interrupt is pending
sourcepub fn clear_pendst()
pub fn clear_pendst()
Set the PENDSTCLR bit in the ICSR register which will clear a pending SysTick interrupt
source§impl SCB
impl SCB
sourcepub fn get_priority(system_handler: SystemHandler) -> u8
pub fn get_priority(system_handler: SystemHandler) -> u8
Returns the hardware priority of system_handler
NOTE: Hardware priority does not exactly match logical priority levels. See
NVIC.get_priority
for more details.
sourcepub unsafe fn set_priority(&mut self, system_handler: SystemHandler, prio: u8)
pub unsafe fn set_priority(&mut self, system_handler: SystemHandler, prio: u8)
Sets the hardware priority of system_handler
to prio
NOTE: Hardware priority does not exactly match logical priority levels. See
NVIC.get_priority
for more details.
On ARMv6-M, updating a system handler priority requires a read-modify-write operation. On ARMv7-M, the operation is performed in a single, atomic write operation.
§Unsafety
Changing priority levels can break priority-based critical sections (see
register::basepri
) and compromise memory safety.
sourcepub fn enable(&mut self, exception: Exception)
pub fn enable(&mut self, exception: Exception)
Enable the exception
If the exception is enabled, when the exception is triggered, the exception handler will be executed instead of the HardFault handler. This function is only allowed on the following exceptions:
MemoryManagement
BusFault
UsageFault
SecureFault
(can only be enabled from Secure state)
Calling this function with any other exception will do nothing.
sourcepub fn disable(&mut self, exception: Exception)
pub fn disable(&mut self, exception: Exception)
Disable the exception
If the exception is disabled, when the exception is triggered, the HardFault handler will be executed instead of the exception handler. This function is only allowed on the following exceptions:
MemoryManagement
BusFault
UsageFault
SecureFault
(can not be changed from Non-secure state)
Calling this function with any other exception will do nothing.
sourcepub fn is_enabled(&self, exception: Exception) -> bool
pub fn is_enabled(&self, exception: Exception) -> bool
Check if an exception is enabled
This function is only allowed on the following exception:
MemoryManagement
BusFault
UsageFault
SecureFault
(can not be read from Non-secure state)
Calling this function with any other exception will read false
.