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.

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

sp-trie: minor fix to avoid possible panic during node decoding #6486

Merged
merged 7 commits into from
Nov 15, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 10 additions & 0 deletions prdoc/pr_6486.prdoc
Original file line number Diff line number Diff line change
@@ -0,0 +1,10 @@
title: "sp-trie: minor fix to avoid panic on badly-constructed proof"

doc:
- audience: ["Runtime Dev", "Runtime User"]
description: |
"Added a check when decoding encoded proof nodes in `sp-trie` to avoid panicking when receiving a badly constructed proof, instead erroring out."

crates:
- name: sp-trie
bump: patch
8 changes: 8 additions & 0 deletions substrate/primitives/trie/src/node_codec.rs
Original file line number Diff line number Diff line change
Expand Up @@ -110,6 +110,10 @@ where
NodeHeader::Null => Ok(NodePlan::Empty),
NodeHeader::HashedValueBranch(nibble_count) | NodeHeader::Branch(_, nibble_count) => {
let padding = nibble_count % nibble_ops::NIBBLE_PER_BYTE != 0;
// data should be at least the size of the offset
if data.len() < input.offset {
return Err(Error::BadFormat)
}
Copy link
Contributor

Choose a reason for hiding this comment

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

Alternative would be to do the check with let offset = input.take(1)?.start, then use offset instead of input.offset. (not exposing input.offset).

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Hey @cheme! Thanks for the feedback. I like your alternative solution but the reason I went for a simple if clause was because take increments the offset of input, so a big part of the code would have to be refactored to adapt to this which might not be worth it for a simple fix like this.

Nevertheless, I made a mistake in this PR since the correct check would be data.len() < input.offset + 1 to make sure data[input.offset] does not panic 😅 I realized after testing and this PR was sadly already merged so I'm a bit ashamed, sorry about that. I already opened another PR (#6502) that fixes this

Copy link
Contributor

Choose a reason for hiding this comment

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

yeah, would have needed a peak method (reader is implemented to be usable with scale api, iirc there another std trait that expose peak), but really not worth it.

// check that the padding is valid (if any)
if padding && nibble_ops::pad_left(data[input.offset]) != 0 {
return Err(Error::BadFormat)
Expand Down Expand Up @@ -154,6 +158,10 @@ where
},
NodeHeader::HashedValueLeaf(nibble_count) | NodeHeader::Leaf(nibble_count) => {
let padding = nibble_count % nibble_ops::NIBBLE_PER_BYTE != 0;
// data should be at least the size of the offset
if data.len() < input.offset {
return Err(Error::BadFormat)
}
// check that the padding is valid (if any)
if padding && nibble_ops::pad_left(data[input.offset]) != 0 {
return Err(Error::BadFormat)
Expand Down
Loading