var_int ambiguous serialization consequences
4 messages, 1 Feb 2015. Started by Tamas Blummer.
Tamas Blummer (2), Wladimir J. van der Laan, Pieter Wuille
I wonder of consequences if var_int is used in its longer than necessary forms (e.g encoding 1 as 0xfd0100 instead of 0x01)
This is already of interest if applying size limit to a block, since transaction count is var_int but is not part of the hashed header or the merkle tree.
It could also be used to create variants of the same transaction message by altered representation of txIn and txout counts, that would remain valid provided signatures validate with the shortest form, as that is created while re-serializing for signature hashing. An implementation that holds mempool by raw message hashes could be tricked to believe that a modified encoded version of the same transaction is a real double spend. One could also mine a valid block with transactions that have a different hash if regularly parsed and re-serialized. An SPV client could be confused by such a transaction as it was present in the merkle tree proof with a different hash than it gets for the tx with its own serialization or from the raw message.
Tamas Blummer Bits of Proof
On Sun, 1 Feb 2015, Tamas Blummer wrote:
Quoted text (1 line)
> I wonder of consequences if var_int is used in its longer than necessary forms (e.g encoding 1 as 0xfd0100 instead of 0x01)
In serialize.h lingo you are talking about CompactSize, not VarInt.
CompactSizes indeed have redundancy in their representation, i.e. the same number can be represented as up to four different byte sequences.
VARINTs have a different format that (AFAIK) isn't used anywhere in the block chain. See WriteVarInt / ReadVarInt. These were designed to not have any redundancy in their representation.
Quoted text (2 lines)
> This is already of interest if applying size limit to a block, since transaction count is var_int but is not part of the hashed header or the > merkle tree.
Are you sure that this is a current concern? Non-canonical CompactSizes are forbidden - in serialize.h this is flagged in ReadCompactSize.
Wladimir
Thanks for the clarification. Yes, I referred to CompactSize using the lingo of https://en.bitcoin.it/wiki/Protocol_documentation
I am not sure if it is current concern. Apparently an exception is thrown if non-canonical CompactSize in a transaction s parsed. Is it ensured that transactions are always parsed before computing their hash?
Tamas Blummer
On Feb 1, 2015, at 11:44 AM, Wladimir <laanwj@gmail•com> wrote:
Quoted text (19 lines)
> > On Sun, 1 Feb 2015, Tamas Blummer wrote: > >> I wonder of consequences if var_int is used in its longer than necessary forms (e.g encoding 1 as 0xfd0100 instead of 0x01) > > In serialize.h lingo you are talking about CompactSize, not VarInt. > > CompactSizes indeed have redundancy in their representation, i.e. the same number can be represented as up to four different byte sequences. > > VARINTs have a different format that (AFAIK) isn't used anywhere in the block chain. See WriteVarInt / ReadVarInt. These were designed to not have any redundancy in their representation. > >> This is already of interest if applying size limit to a block, since transaction count is var_int but is not part of the hashed header or the >> merkle tree. > > Are you sure that this is a current concern? Non-canonical CompactSizes are forbidden - in serialize.h this is flagged in ReadCompactSize. > > Wladimir > >
Hashes are always computed by reserializing data structures, never by hashing wire data directly. This has been the case in every version of the reference client's code that I know of.
This even meant that for example a block of 999999 bytes with non-shortest length for the transaction count could be over the mazimum block size, but still be valid.
As Wladimir says, more recently we switched to just failing to deserialize (by throwing an exception) whenever a non-shortest form is used. On Feb 1, 2015 1:34 AM, "Tamas Blummer" <tamas@bitsofproof•com> wrote:
Quoted text (37 lines)
> I wonder of consequences if var_int is used in its longer than necessary > forms (e.g encoding 1 as 0xfd0100 instead of 0x01) > > This is already of interest if applying size limit to a block, since > transaction count is var_int but is not part of the hashed header or the > merkle tree. > > It could also be used to create variants of the same transaction message > by altered representation of txIn and txout counts, that would remain valid > provided signatures validate with the shortest form, as that is created > while re-serializing for signature hashing. An implementation that holds > mempool by raw message hashes could be tricked to believe that a modified > encoded version of the same transaction is a real double spend. One could > also mine a valid block with transactions that have a different hash if > regularly parsed and re-serialized. An SPV client could be confused by such > a transaction as it was present in the merkle tree proof with a different > hash than it gets for the tx with its own serialization or from the raw > message. > > Tamas Blummer > Bits of Proof > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming. The Go Parallel Website, > sponsored by Intel and developed in partnership with Slashdot Media, is > your > hub for all things parallel software development, from weekly thought > leadership blogs to news, videos, case studies, tutorials and more. Take a > look and join the conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Bitcoin-development mailing list > Bitcoin-development@lists•sourceforge•net > https://lists.sourceforge.net/lists/listinfo/bitcoin-development > >