Struct addr2line::Context

pub struct Context<R: Reader> { /* private fields */ }

The state necessary to perform address to line translation.

Constructing a Context is somewhat costly, so users should aim to reuse Contexts when performing lookups for many addresses in the same executable.

Implementations

impl Context<EndianRcSlice<RunTimeEndian>>

pub fn new<'data, O: Object<'data>>(file: &O) -> Result<Self, Error>

Construct a new Context.

The resulting Context uses gimli::EndianRcSlice<gimli::RunTimeEndian>. This means it is not thread safe, has no lifetime constraints (since it copies the input data), and works for any endianity.

Performance sensitive applications may want to use Context::from_dwarf with a more specialised gimli::Reader implementation.

pub fn new_with_sup<'data, O: Object<'data>>( file: &O, sup_file: Option<&O> ) -> Result<Self, Error>

Construct a new Context.

Optionally also use a supplementary object file.

The resulting Context uses gimli::EndianRcSlice<gimli::RunTimeEndian>. This means it is not thread safe, has no lifetime constraints (since it copies the input data), and works for any endianity.

Performance sensitive applications may want to use Context::from_dwarf with a more specialised gimli::Reader implementation.

impl<R: Reader> Context<R>

pub fn from_sections( debug_abbrev: DebugAbbrev<R>, debug_addr: DebugAddr<R>, debug_aranges: DebugAranges<R>, debug_info: DebugInfo<R>, debug_line: DebugLine<R>, debug_line_str: DebugLineStr<R>, debug_ranges: DebugRanges<R>, debug_rnglists: DebugRngLists<R>, debug_str: DebugStr<R>, debug_str_offsets: DebugStrOffsets<R>, default_section: R ) -> Result<Self, Error>

Construct a new Context from DWARF sections.

This method does not support using a supplementary object file.

pub fn from_dwarf(sections: Dwarf<R>) -> Result<Context<R>, Error>

Construct a new Context from an existing gimli::Dwarf object.

pub fn find_dwarf_and_unit( &self, probe: u64 ) -> LookupResult<impl LookupContinuation<Output = Option<(&Dwarf<R>, &Unit<R>)>, Buf = R>>

Find the DWARF unit corresponding to the given virtual memory address.

pub fn find_location(&self, probe: u64) -> Result<Option<Location<'_>>, Error>

Find the source file and line corresponding to the given virtual memory address.

pub fn find_location_range( &self, probe_low: u64, probe_high: u64 ) -> Result<LocationRangeIter<'_, R>, Error>

Return source file and lines for a range of addresses. For each location it also returns the address and size of the range of the underlying instructions.

pub fn find_frames( &self, probe: u64 ) -> LookupResult<impl LookupContinuation<Output = Result<FrameIter<'_, R>, Error>, Buf = R>>

Return an iterator for the function frames corresponding to the given virtual memory address.

If the probe address is not for an inline function then only one frame is returned.

If the probe address is for an inline function then the first frame corresponds to the innermost inline function. Subsequent frames contain the caller and call location, until an non-inline caller is reached.

pub fn preload_units( &self, probe: u64 ) -> impl Iterator<Item = (SplitDwarfLoad<R>, impl FnOnce(Option<Arc<Dwarf<R>>>) -> Result<(), Error> + '_)>

Preload units for probe.

The iterator returns pairs of SplitDwarfLoads containing the information needed to locate and load split DWARF for probe and a matching callback to invoke once that data is available.

If this method is called, and all of the returned closures are invoked, addr2line guarantees that any future API call for the address probe will not require the loading of any split DWARF.

  const ADDRESS: u64 = 0xdeadbeef;
  ctx.preload_units(ADDRESS).for_each(|(load, callback)| {
    let dwo = do_split_dwarf_load(load);
    callback(dwo);
  });

  let frames_iter = match ctx.find_frames(ADDRESS) {
    LookupResult::Output(result) => result,
    LookupResult::Load { .. } => unreachable!("addr2line promised we wouldn't get here"),
  };

  // ...