You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Bugs in wasm code causing infinite loops and/or dry ran txs with custom wasm code may be a DoS vector, since wasm instructions have no associated gas cost (see get_gas_rules() in crates/namada/src/vm/wasm/run.rs). We should associate some non-zero cost to executing wasm opcodes, in order to avoid the aforementioned problems.
The text was updated successfully, but these errors were encountered:
This is a known issue due to the lack of a proper runtime gas meter but I believe two things should be considered:
For actual txs in storage we always have the whitelist. This attack can be exploited on loops that don't call host functions (since these are tracking gas). So as long as the whitelist is there and we check what gets whitelisted it should not happen
For dry runs this should be fixed in Fix allowlist checks #2819 since the whitelist check is moved inside the call to tx which is called from apply_wasm_tx
Still I'll try to see how hard it is to swap the current whitelist-based gas metering system with a proper runtime one
Bugs in wasm code causing infinite loops and/or dry ran txs with custom wasm code may be a DoS vector, since wasm instructions have no associated gas cost (see
get_gas_rules()
incrates/namada/src/vm/wasm/run.rs
). We should associate some non-zero cost to executing wasm opcodes, in order to avoid the aforementioned problems.The text was updated successfully, but these errors were encountered: