aya: Support loading programs with ksyms

Author: altugbozkurt07

aya-obj/src/btf/btf.rs

@@ -24,6 +24,7 @@ use crate::{
         info::{FuncSecInfo, LineSecInfo},
         relocation::Relocation,
     },
+    extern_types::ExternCollection,
     generated::{btf_ext_header, btf_header},
     util::{HashMap, bytes_of},
 };
@@ -264,6 +265,8 @@ pub struct Btf {
     strings: Vec<u8>,
     types: BtfTypes,
     _endianness: Endianness,
+    /// Extern functions parsed from ksyms section
+    pub(crate) externs: ExternCollection,
 }
 
 fn add_type(header: &mut btf_header, types: &mut BtfTypes, btf_type: BtfType) -> u32 {
@@ -292,6 +295,7 @@ impl Btf {
             strings: vec![0],
             types: BtfTypes::default(),
             _endianness: Endianness::default(),
+            externs: ExternCollection::new(),
         }
     }
 
@@ -364,6 +368,7 @@ impl Btf {
             strings,
             types,
             _endianness: endianness,
+            externs: ExternCollection::new(),
         })
     }
 
@@ -501,6 +506,9 @@ impl Btf {
         symbol_offsets: &HashMap<String, u64>,
         features: &BtfFeatures,
     ) -> Result<(), BtfError> {
+        if !self.externs.is_empty() {
+            self.fixup_ksyms_datasec(self.externs.datasec_id, self.externs.dummy_ksym_var_id)?;
+        }
         let enum64_placeholder_id = OnceCell::new();
         let filler_var_id = OnceCell::new();
         let mut types = mem::take(&mut self.types);
@@ -812,6 +820,128 @@ impl Btf {
         self.types = types;
         Ok(())
     }
+
+    /// Fixes up BTF for `.ksyms` datasec entries containing extern kernel symbol and
+    /// makes it acceptable by the kernel:
+    ///
+    /// * Changes linkage of extern functions to `GLOBAL`, fixes parameter names, injects
+    ///   a dummy variable representing them in datasec.
+    /// * Changes linkage of extern variables to `GLOBAL_ALLOCATED`, replaces their type
+    ///   with `int`.
+    pub(crate) fn fixup_ksyms_datasec(
+        &mut self,
+        datasec_id: Option<u32>,
+        dummy_var_id: Option<u32>,
+    ) -> Result<(), BtfError> {
+        // Extract dummy variable's name offset and type ID for patching func_proto names and datasec entries.
+        // If dummy var exists: use its name_offset (for string table patching) and btf_type (underlying int type).
+        // If no dummy var (fallback): search for a 4-byte int type directly, with no name_offset.
+        // Both paths provide an int_btf_id to ensure type consistency in datasec variable entries.
+        let (dummy_var_name_offset, int_btf_id) = if let Some(dummy_id) = dummy_var_id {
+            let dummy_type = &self.types.types[dummy_id as usize];
+            if let BtfType::Var(v) = dummy_type {
+                (Some(v.name_offset), v.btf_type)
+            } else {
+                return Err(BtfError::InvalidDatasec);
+            }
+        } else {
+            let int_id = self
+                .types
+                .types
+                .iter()
+                .enumerate()
+                .find_map(|(idx, t)| {
+                    if let BtfType::Int(int_type) = t {
+                        (int_type.size == 4).then_some(idx as u32)
+                    } else {
+                        None
+                    }
+                })
+                .ok_or(BtfError::InvalidDatasec)?;
+
+            (None, int_id)
+        };
+
+        let datasec_id = datasec_id.ok_or(BtfError::InvalidDatasec)?;
+
+        let datasec_name = {
+            let datasec = &self.types.types[datasec_id as usize];
+            let BtfType::DataSec(d) = datasec else {
+                return Err(BtfError::InvalidDatasec);
+            };
+            self.string_at(d.name_offset)?.into_owned()
+        };
+
+        debug!("DATASEC {datasec_name}: fixing up extern ksyms");
+
+        let entry_type_ids: Vec<u32> = {
+            let BtfType::DataSec(d) = &self.types.types[datasec_id as usize] else {
+                return Err(BtfError::InvalidDatasec);
+            };
+            d.entries.iter().map(|e| e.btf_type).collect()
+        };
+
+        let mut offset = 0u32;
+        let size = mem::size_of::<i32>() as u32;
+
+        for (i, &type_id) in entry_type_ids.iter().enumerate() {
+            match &self.types.types[type_id as usize] {
+                BtfType::Func(f) => {
+                    let (func_name, proto_id) =
+                        { (self.string_at(f.name_offset)?.into_owned(), f.btf_type) };
+
+                    if let BtfType::Func(f) = &mut self.types.types[type_id as usize] {
+                        f.set_linkage(FuncLinkage::Global);
+                    }
+
+                    if let Some(dummy_name_off) = dummy_var_name_offset {
+                        if let BtfType::FuncProto(func_proto) =
+                            &mut self.types.types[proto_id as usize]
+                        {
+                            for param in &mut func_proto.params {
+                                if param.btf_type != 0 && param.name_offset == 0 {
+                                    param.name_offset = dummy_name_off;
+                                }
+                            }
+                        }
+                    }
+
+                    if let (Some(dummy_id), BtfType::DataSec(d)) =
+                        (dummy_var_id, &mut self.types.types[datasec_id as usize])
+                    {
+                        d.entries[i].btf_type = dummy_id;
+                    }
+
+                    debug!("DATASEC {datasec_name}: FUNC {func_name}: fixup offset {offset}");
+                }
+                BtfType::Var(v) => {
+                    let var_name = { self.string_at(v.name_offset)?.into_owned() };
+
+                    if let BtfType::Var(v) = &mut self.types.types[type_id as usize] {
+                        v.linkage = VarLinkage::Global;
+                        v.btf_type = int_btf_id;
+                    }
+
+                    debug!("DATASEC {datasec_name}: VAR {var_name}: fixup offset {offset}");
+                }
+                _ => unreachable!(),
+            }
+
+            if let BtfType::DataSec(d) = &mut self.types.types[datasec_id as usize] {
+                d.entries[i].offset = offset;
+                d.entries[i].size = size;
+            }
+
+            offset += size;
+        }
+
+        if let BtfType::DataSec(d) = &mut self.types.types[datasec_id as usize] {
+            d.size = offset;
+            debug!("DATASEC {datasec_name}: fixup size to {offset}");
+        }
+
+        Ok(())
+    }
 }
 
 impl Default for Btf {

aya-obj/src/btf/extern_types.rs

@@ -0,0 +1,187 @@
+use alloc::vec::Vec;
+
+use log::debug;
+
+use crate::{
+    KsymsError, Object,
+    btf::{Btf, BtfError, BtfType, DataSec, DataSecEntry},
+    extern_types::{ExternDesc, ExternType},
+    relocation::Symbol,
+    util::HashMap,
+};
+impl Btf {
+    /// Creates a dummy global variable named "dummy_ksym" with a 4-byte int type for unresolved kernel symbols.
+    pub(crate) fn create_dummy_ksym_var(&mut self) -> Result<u32, BtfError> {
+        let int_type_id = {
+            let mut found_id = None;
+            for (idx, t) in self.types().enumerate() {
+                if let BtfType::Int(int) = t {
+                    if int.size == 4 {
+                        found_id = Some((idx) as u32);
+                        break;
+                    }
+                }
+            }
+            found_id
+        };
+
+        let int_type_id = if let Some(id) = int_type_id {
+            id
+        } else {
+            let name_offset = self.add_string("int");
+            self.add_type(BtfType::Int(crate::btf::Int::new(
+                name_offset,
+                4,
+                crate::btf::IntEncoding::Signed,
+                0,
+            )))
+        };
+
+        debug!("Found/created int type_id: {}", int_type_id);
+        if let Ok(BtfType::Int(int)) = self.type_by_id(int_type_id) {
+            debug!(
+                "Int type size: {}, encoding: {:?}",
+                int.size,
+                int.encoding()
+            );
+        }
+
+        let name_offset = self.add_string("dummy_ksym");
+        let dummy_var_id = self.add_type(BtfType::Var(crate::btf::Var::new(
+            name_offset,
+            int_type_id,
+            crate::btf::VarLinkage::Global,
+        )));
+
+        debug!("Created dummy_var type_id: {}", dummy_var_id);
+        if let Ok(BtfType::Var(var)) = self.type_by_id(dummy_var_id) {
+            debug!("Dummy var points to type_id: {}", var.btf_type);
+        }
+
+        Ok(dummy_var_id)
+    }
+
+    /// Searches for the `.ksyms` datasec in BTF, returns it if found.
+    fn find_ksyms_datasec(&self) -> Result<Option<(u32, DataSec)>, BtfError> {
+        for (idx, btf_type) in self.types().enumerate() {
+            if let BtfType::DataSec(datasec) = btf_type {
+                let name = self.type_name(btf_type)?;
+                if name == ".ksyms" {
+                    return Ok(Some((idx as u32, datasec.clone())));
+                }
+            }
+        }
+
+        Ok(None)
+    }
+
+    /// Checks if datasec contains any functions.
+    pub(crate) fn datasec_has_functions(&self, datasec: &DataSec) -> bool {
+        datasec.entries.iter().any(|entry| {
+            self.type_by_id(entry.btf_type)
+                .map(|t| matches!(t, BtfType::Func(_)))
+                .unwrap_or(false)
+        })
+    }
+
+    /// Collects extern descriptors from datasec entries.
+    pub(crate) fn collect_extern_entries(
+        &self,
+        datasec: &DataSec,
+        symbol_table: &HashMap<usize, Symbol>,
+    ) -> Result<Vec<ExternDesc>, BtfError> {
+        let mut result = Vec::new();
+
+        for entry in &datasec.entries {
+            let Some(extern_desc) = self.process_datasec_entry(entry, symbol_table)? else {
+                continue;
+            };
+
+            result.push(extern_desc);
+        }
+
+        Ok(result)
+    }
+
+    /// Processes a single datasec entry, returns [`ExternDesc`] if it's an extern.
+    fn process_datasec_entry(
+        &self,
+        entry: &DataSecEntry,
+        symbol_table: &HashMap<usize, Symbol>,
+    ) -> Result<Option<ExternDesc>, BtfError> {
+        let btf_type = self.type_by_id(entry.btf_type)?;
+
+        let (name, is_func, var_btf_type) = match btf_type {
+            BtfType::Func(func) => {
+                let name = self.string_at(func.name_offset)?.into_owned();
+                (name, true, func.btf_type)
+            }
+            BtfType::Var(var) => {
+                let name = self.string_at(var.name_offset)?.into_owned();
+                (name, false, var.btf_type)
+            }
+            _ => return Ok(None),
+        };
+
+        let symbol = find_symbol_by_name(symbol_table, &name).ok_or(BtfError::InvalidSymbolName)?;
+
+        // Resolve through modifiers (const, volatile, typedef, etc.)
+        // Type ID 0 represents void in BTF
+        let resolved_type_id = self.resolve_type(var_btf_type).unwrap_or(var_btf_type);
+
+        // Typeless ksyms are declared as `extern const void symbol __ksym`
+        // They resolve to void (type_id 0) and are resolved via /proc/kallsyms
+        let is_typeless = !is_func && resolved_type_id == 0;
+
+        let mut extern_desc = ExternDesc::new(
+            name,
+            ExternType::Ksym,
+            entry.btf_type,
+            symbol.is_weak,
+            is_func,
+        );
+
+        // For typeless ksyms, don't set type_id so they skip kernel BTF resolution
+        if !is_typeless {
+            extern_desc.type_id = Some(resolved_type_id);
+        }
+
+        Ok(Some(extern_desc))
+    }
+}
+
+fn find_symbol_by_name<'a>(
+    symbol_table: &'a HashMap<usize, Symbol>,
+    name: &str,
+) -> Option<&'a Symbol> {
+    symbol_table
+        .values()
+        .find(|sym| sym.name.as_deref() == Some(name))
+}
+
+impl Object {
+    /// Collects extern kernel symbols from BTF datasec entries.
+    pub fn collect_ksyms_from_btf(&mut self) -> Result<(), KsymsError> {
+        let btf = self.btf.as_mut().ok_or(KsymsError::NoBtf)?;
+        let Some((datasec_id, datasec)) = btf.find_ksyms_datasec()? else {
+            return Ok(());
+        };
+
+        if btf.datasec_has_functions(&datasec) {
+            let dummy_var_id = btf.create_dummy_ksym_var()?;
+            btf.externs.set_dummy_var_id(dummy_var_id);
+        }
+
+        let collected = btf.collect_extern_entries(&datasec, &self.symbol_table)?;
+
+        for extern_desc in collected {
+            btf.externs.insert(extern_desc.name.clone(), extern_desc);
+        }
+
+        if !btf.externs.is_empty() {
+            btf.externs.datasec_id = Some(datasec_id);
+        }
+
+        Ok(())
+    }
+}

aya-obj/src/btf/mod.rs

@@ -2,6 +2,7 @@
 
 #[expect(clippy::module_inception)]
 mod btf;
+mod extern_types;
 mod info;
 mod relocation;
 mod types;