Struct MemFlags

pub struct MemFlags { /* private fields */ }

Flags for memory operations like load/store.

Each of these flags introduce a limited form of undefined behavior. The flags each enable certain optimizations that need to make additional assumptions. Generally, the semantics of a program does not change when a flag is removed, but adding a flag will.

In addition, the flags determine the endianness of the memory access. By default, any memory access uses the native endianness determined by the target ISA. This can be overridden for individual accesses by explicitly specifying little- or big-endian semantics via the flags.

Implementations

impl MemFlags

pub const fn new() -> Self

Create a new empty set of flags.

pub const fn trusted() -> Self

Create a set of flags representing an access from a “trusted” address, meaning it’s known to be aligned and non-trapping.

pub const fn alias_region(self) -> Option<AliasRegion>

Reads the alias region that this memory operation works with.

pub const fn with_alias_region(self, region: Option<AliasRegion>) -> Self

Sets the alias region that this works on to the specified region.

pub fn set_alias_region(&mut self, region: Option<AliasRegion>)

Sets the alias region that this works on to the specified region.

pub fn set_by_name(&mut self, name: &str) -> Result<bool, &'static str>

Set a flag bit by name.

Returns true if the flag was found and set, false for an unknown flag name.

Errors

Returns an error message if the name is known but couldn’t be applied due to it being a semantic error.

pub const fn endianness(self, native_endianness: Endianness) -> Endianness

Return endianness of the memory access. This will return the endianness explicitly specified by the flags if any, and will default to the native endianness otherwise. The native endianness has to be provided by the caller since it is not explicitly encoded in CLIF IR – this allows a front end to create IR without having to know the target endianness.

pub const fn explicit_endianness(self) -> Option<Endianness>

Return endianness of the memory access, if explicitly specified.

If the endianness is not explicitly specified, this will return None, which means “native endianness”.

pub fn set_endianness(&mut self, endianness: Endianness)

Set endianness of the memory access.

pub const fn with_endianness(self, endianness: Endianness) -> Self

Set endianness of the memory access, returning new flags.

pub const fn notrap(self) -> bool

Test if this memory operation cannot trap.

By default MemFlags will assume that any load/store can trap and is associated with a TrapCode::HeapOutOfBounds code. If the trap code is configured to None though then this method will return true and indicates that the memory operation will not trap.

If this returns true then the memory is accessible, which means that accesses will not trap. This makes it possible to delete an unused load or a dead store instruction.

pub fn set_notrap(&mut self)

Sets the trap code for this MemFlags to None.

pub const fn with_notrap(self) -> Self

Sets the trap code for this MemFlags to None, returning the new flags.

pub const fn aligned(self) -> bool

Test if the aligned flag is set.

By default, Cranelift memory instructions work with any unaligned effective address. If the aligned flag is set, the instruction is permitted to trap or return a wrong result if the effective address is misaligned.

pub fn set_aligned(&mut self)

Set the aligned flag.

pub const fn with_aligned(self) -> Self

Set the aligned flag, returning new flags.

pub const fn readonly(self) -> bool

Test if the readonly flag is set.

Loads with this flag have no memory dependencies. This results in undefined behavior if the dereferenced memory is mutated at any time between when the function is called and when it is exited.

pub fn set_readonly(&mut self)

Set the readonly flag.

pub const fn with_readonly(self) -> Self

Set the readonly flag, returning new flags.

pub const fn checked(self) -> bool

Test if the checked bit is set.

Loads and stores with this flag are verified to access pointers only with a validated PointsTo fact attached, and with that fact validated, when using the proof-carrying-code framework. If initial facts on program inputs are correct (i.e., correctly denote the shape and types of data structures in memory), and if PCC validates the compiled output, then all checked-marked memory accesses are guaranteed (up to the checker’s correctness) to access valid memory. This can be used to ensure memory safety and sandboxing.

pub fn set_checked(&mut self)

Set the checked bit.

pub const fn with_checked(self) -> Self

Set the checked bit, returning new flags.

pub const fn trap_code(self) -> Option<TrapCode>

Get the trap code to report if this memory access traps.

A None trap code indicates that this memory access does not trap.

pub const fn with_trap_code(self, code: Option<TrapCode>) -> Self

Configures these flags with the specified trap code code.

A trap code indicates that this memory operation cannot be optimized away and it must “stay where it is” in the programs. Traps are considered side effects, for example, and have meaning through the trap code that is communicated and which instruction trapped.

Trait Implementations

impl Clone for MemFlags

fn clone(&self) -> MemFlags

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for MemFlags

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for MemFlags

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Hash for MemFlags

fn hash<__H: Hasher>(&self, state: &mut __H)

Feeds this value into the given Hasher.

fn hash_slice<H>(data: &[Self], state: &mut H)

where H: Hasher, Self: Sized,

Feeds a slice of this type into the given Hasher.

impl PartialEq for MemFlags

fn eq(&self, other: &MemFlags) -> bool

Tests for self and other values to be equal, and is used by ==.

fn ne(&self, other: &Rhs) -> bool

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Copy for MemFlags

impl Eq for MemFlags

impl StructuralPartialEq for MemFlags