Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

elf.parse: resolve overflow issue in hash sections #237

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

elf.parse: resolve overflow issue in hash sections #237

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
14 changes: 7 additions & 7 deletions src/elf/mod.rs
View file
Open in desktop
Expand Up @@ -348,18 +348,18 @@ if_sylvan! {

fn gnu_hash_len(bytes: &[u8], offset: usize, ctx: Ctx) -> error::Result<usize> {
let buckets_num = bytes.pread_with::<u32>(offset, ctx.le)? as usize;
let min_chain = bytes.pread_with::<u32>(offset + 4, ctx.le)? as usize;
let bloom_size = bytes.pread_with::<u32>(offset + 8, ctx.le)? as usize;
let min_chain = bytes.pread_with::<u32>(offset.wrapping_add(4), ctx.le)? as usize;
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok back to this PR; so what are the semantics if this overflows? This seems like a subtle bug now (I guess it was always present) Also, won't this only crash in debug mode, but will wrap overflow in release? Ditto for the other explicit overflow wraps

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm using wrapping_xxx here because I've seen other code do it too, for example,

goblin/src/mach/imports.rs

Line 234 in 2188699

let seg_offset = bind_info.seg_offset.wrapping_add(addr);
, semantically, maybe overflowing_xxx is better?

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes, i think your intuition is right, was kind of just wondering out loud what pros/cons are of the explict wrap, and what it means for the gnu hash table calculations to overflow (i just assume it means the binary is bad, better not to crash then i suppose?)

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Does gnu_hash_len calculate total size of GNU_HASH table?
Then performing some checks like

goblin/src/elf/gnu_hash.rs

Lines 116 to 135 in ece3ce6

let hashtab = &hashtab[16..];
{
// SAFETY: Condition to check for an overflow
// size_of(chains) + size_of(buckets) + size_of(bloom_filter) == size_of(hashtab)
if dynsyms.len() <= symindex as usize {
return Err("symindex must be smaller than dynsyms.len()");
}
let chains_size = (dynsyms.len() - symindex as usize).checked_mul(U32_SIZE);
let buckets_size = (nbuckets as usize).checked_mul(U32_SIZE);
let bloom_size = (maskwords as usize).checked_mul(INT_SIZE);
let total_size = match (chains_size, buckets_size, bloom_size) {
(Some(a), Some(b), Some(c)) => {
a.checked_add(b).and_then(|t| t.checked_add(c))
}
_ => None,
};
match total_size {
Some(size) if size == hashtab.len() => {}
is Ok.

let bloom_size = bytes.pread_with::<u32>(offset.wrapping_add(8), ctx.le)? as usize;
// We could handle min_chain==0 if we really had to, but it shouldn't happen.
if buckets_num == 0 || min_chain == 0 || bloom_size == 0 {
return Err(error::Error::Malformed(format!("Invalid DT_GNU_HASH: buckets_num={} min_chain={} bloom_size={}",
buckets_num, min_chain, bloom_size)));
}
// Find the last bucket.
let buckets_offset = offset + 16 + bloom_size * if ctx.container.is_big() { 8 } else { 4 };
let buckets_offset = offset.wrapping_add(16).wrapping_add(bloom_size.wrapping_mul(if ctx.container.is_big() { 8 } else { 4 }));
let mut max_chain = 0;
for bucket in 0..buckets_num {
let chain = bytes.pread_with::<u32>(buckets_offset + bucket * 4, ctx.le)? as usize;
let chain = bytes.pread_with::<u32>(buckets_offset.wrapping_add(bucket.wrapping_mul(4)), ctx.le)? as usize;
if max_chain < chain {
max_chain = chain;
}
Expand All @@ -368,7 +368,7 @@ if_sylvan! {
return Ok(0);
}
// Find the last chain within the bucket.
let mut chain_offset = buckets_offset + buckets_num * 4 + (max_chain - min_chain) * 4;
let mut chain_offset = buckets_offset.wrapping_add(buckets_num.wrapping_mul(4)).wrapping_add((max_chain - min_chain).wrapping_mul(4));
loop {
let hash = bytes.pread_with::<u32>(chain_offset, ctx.le)?;
max_chain += 1;
Expand All @@ -382,9 +382,9 @@ if_sylvan! {
fn hash_len(bytes: &[u8], offset: usize, machine: u16, ctx: Ctx) -> error::Result<usize> {
// Based on readelf code.
let nchain = if (machine == header::EM_FAKE_ALPHA || machine == header::EM_S390) && ctx.container.is_big() {
bytes.pread_with::<u64>(offset + 4, ctx.le)? as usize
bytes.pread_with::<u64>(offset.wrapping_add(4), ctx.le)? as usize
} else {
bytes.pread_with::<u32>(offset + 4, ctx.le)? as usize
bytes.pread_with::<u32>(offset.wrapping_add(4), ctx.le)? as usize
};
Ok(nchain)
}
Expand Down