MIME-Version: 1.0
In-Reply-To: <C3ED56D2-CB1F-4DE5-AB43-F826705806FB@xbt.hk>
References: <9CCCE945-9432-41B9-8559-AFE7CF233603@xbt.hk>
	<CAMZUoKms85DhtS1mN70nq4LSY7QtXym6E4_yvQk5Q0tizkVwEQ@mail.gmail.com>
	<C3ED56D2-CB1F-4DE5-AB43-F826705806FB@xbt.hk>
From: "Russell O'Connor" <roconnor@blockstream.io>
Date: Fri, 1 Jun 2018 14:15:32 -0400
Message-ID: <CAMZUoKmqdT3fte0o-CSppMV125u9zmxheaP549=nqkeVGSryMA@mail.gmail.com>
To: Johnson Lau <jl2012@xbt.hk>
Content-Type: multipart/alternative; boundary="0000000000000589c8056d9893c8"
Cc: bitcoin-dev <bitcoin-dev@lists.linuxfoundation.org>
Subject: Re: [bitcoin-dev] SIGHASH2 for version 1 witness programme
Precedence: list

--0000000000000589c8056d9893c8
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Fri, Jun 1, 2018 at 1:03 PM, Johnson Lau <jl2012@xbt.hk> wrote:

> On 1 Jun 2018, at 11:03 PM, Russell O'Connor <roconnor@blockstream.io>
> wrote:
> On Thu, May 31, 2018 at 2:35 PM, Johnson Lau via bitcoin-dev <
> bitcoin-dev@lists.linuxfoundation.org> wrote:
>
>>
>>   Double SHA256 of the serialization of:
>>
>
> Should we replace the Double SHA256 with a Single SHA256?  There is no
> possible length extension attack here.  Or are we speculating that there =
is
> a robustness of Double SHA256 in the presence of SHA256 breaking?
>
> I suggest putting `sigversion` at the beginning instead of the end of the
> format.  Because its value is constant, the beginning of the SHA-256
> computation could be pre-computed in advance.  Furthermore, if we make th=
e
> `sigversion` exactly 64-bytes long then the entire first block of the
> SHA-256 compression function could be pre-computed.
>
> Can we add CHECKSIGFROMSTACK or do you think that would go into a separat=
e
> BIP?
>
>
> I think it=E2=80=99s just a tradition to use double SHA256. One reason we=
 might
> want to keep dSHA256 is a blind signature might be done by giving only th=
e
> single SHA256 hash to the signer. At the same time, a non-Bitcoin signatu=
re
> scheme might use SHA512-SHA256. So a blind signer could distinguish the
> message type without learning the message.
>
> sigversion is a response to Peter Todd=E2=80=99s comments on BIP143:
> https://petertodd.org/2016/segwit-consensus-critical-code-review#bip143-
> transaction-signature-verification
>
> I make it a 0x01000000 at the end of the message because the last 4 bytes
> has been the nHashType in the legacy/BIP143 protocol. Since the maximum
> legacy nHashType is 0xff, no collision could ever occur.
>
> Putting a 64-byte constant at the beginning should also work, since a
> collision means SHA256 is no longer preimage resistance. I don=E2=80=99t =
know much
> about SHA256 optimisation. How good it is as we put a 64-byte constant at
> the beginning, while we also make the message 64-byte longer?
>

In theory, having a fixed 64 byte constant at the beginning results in zero
overhead for those 64 bytes.  An implementation would just start the usual
SHA-256 algorithm with a different pre-computed and fixed initial value
than SHA-256's standard initial value.  The SHA-256 padding counter would
also need to start at 64*8 bits rather than starting at 0 bits.  In
practice, assuming a OpenSSL-like implementation of SHA-256, it should be
easy to implement this optimization. One would replace SHA256_Init call
with a variant that initializes the SHA256_CTX to this pre-computed value
and sets SHA256_CTX's num counter to the appropriate value.  Non-optimized
implementations can still just add the 64 byte prefix and use any SHA-256
implementation.

For CHECKSIGFROMSTACK (CSFS), I think the question is whether we want to
> make it as a separate opcode, or combine that with CHECKSIG. If it is a
> separate opcode, I think it should be a separate BIP. If it is combined
> with CHECKSIG, we could do something like this: If the bit 10 of SIGHASH2
> is set, CHECKSIG will pop one more item from stack, and serialize its
> content with the transaction digest. Any thought?
>

I prefer a different opcode for CHECKSIGFROMSTACK because I dislike opcodes
that pop a non-static number of elements off the stack.  Popping a dynamic
number of stack elements makes it more difficult to validate that a Script
pubkey doesn't allow any funny business.

--0000000000000589c8056d9893c8
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div class=3D"gmail_extra"><div class=3D"gmail_quote">On F=
ri, Jun 1, 2018 at 1:03 PM, Johnson Lau <span dir=3D"ltr">&lt;<a href=3D"ma=
ilto:jl2012@xbt.hk" target=3D"_blank">jl2012@xbt.hk</a>&gt;</span> wrote:<b=
r><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;borde=
r-left:1px solid rgb(204,204,204);padding-left:1ex"><div style=3D"overflow-=
wrap: break-word;"><div><div class=3D"gmail-h5"><div><blockquote type=3D"ci=
te"><div class=3D"gmail_extra">On 1 Jun 2018, at 11:03 PM, Russell O&#39;Co=
nnor &lt;<a href=3D"mailto:roconnor@blockstream.io" target=3D"_blank">rocon=
nor@blockstream.io</a>&gt; wrote:<br><div class=3D"gmail_quote">On Thu, May=
 31, 2018 at 2:35 PM, Johnson Lau via bitcoin-dev <span dir=3D"ltr">&lt;<a =
href=3D"mailto:bitcoin-dev@lists.linuxfoundation.org" target=3D"_blank">bit=
coin-dev@lists.linuxfounda<wbr>tion.org</a>&gt;</span> wrote:<br><blockquot=
e class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px s=
olid rgb(204,204,204);padding-left:1ex"><br>
=C2=A0 Double SHA256 of the serialization of:<br></blockquote><div><br></di=
v><div>Should we replace the Double SHA256 with a Single SHA256?=C2=A0 Ther=
e is no possible length extension attack here.=C2=A0 Or are we speculating =
that there is a robustness of Double SHA256 in the presence of SHA256 break=
ing?<br><br></div><div>I suggest putting `sigversion` at the beginning inst=
ead of the end of the format.=C2=A0 Because its value is constant, the begi=
nning of the SHA-256 computation could be pre-computed in advance.=C2=A0 Fu=
rthermore, if we make the `sigversion` exactly 64-bytes long then the entir=
e first block of the SHA-256 compression function could be pre-computed.<br=
></div><div><br></div><div>Can we add CHECKSIGFROMSTACK or do you think tha=
t would go into a separate BIP?<br></div></div></div><div>
</div></blockquote></div><br></div></div><div>I think it=E2=80=99s just a t=
radition to use double SHA256. One reason we might want to keep dSHA256 is =
a blind signature might be done by giving only the single SHA256 hash to th=
e signer. At the same time, a non-Bitcoin signature scheme might use SHA512=
-SHA256. So a blind signer could distinguish the message type without learn=
ing the message.</div><div><br></div><div>sigversion is a response to Peter=
 Todd=E2=80=99s comments on BIP143:=C2=A0<a href=3D"https://petertodd.org/2=
016/segwit-consensus-critical-code-review#bip143-transaction-signature-veri=
fication" target=3D"_blank">https://petertodd.org/<wbr>2016/segwit-consensu=
s-<wbr>critical-code-review#bip143-<wbr>transaction-signature-<wbr>verifica=
tion</a></div><div><br></div><div>I make it a 0x01000000 at the end of the =
message because the last 4 bytes has been the nHashType in the legacy/BIP14=
3 protocol. Since the maximum legacy nHashType is 0xff, no collision could =
ever occur.</div><div><br></div><div>Putting a 64-byte constant at the begi=
nning should also work, since a collision means SHA256 is no longer preimag=
e resistance. I don=E2=80=99t know much about SHA256 optimisation. How good=
 it is as we put a 64-byte constant at the beginning, while we also make th=
e message 64-byte longer?</div></div></blockquote><div><br></div><div>In th=
eory, having a fixed 64 byte constant at the beginning results in zero over=
head for those 64 bytes.=C2=A0 An implementation would just start the usual=
 SHA-256 algorithm with a different pre-computed and fixed initial value th=
an SHA-256&#39;s standard initial value.=C2=A0 The SHA-256 padding counter =
would also need to start at 64*8 bits rather than starting at 0 bits.=C2=A0=
 In practice, assuming a OpenSSL-like implementation of SHA-256, it should =
be easy to implement this optimization. One would replace SHA256_Init call =
with a variant that initializes the SHA256_CTX to this pre-computed value a=
nd sets SHA256_CTX&#39;s num counter to the appropriate value.=C2=A0 Non-op=
timized implementations can still just add the 64 byte prefix and use any S=
HA-256 implementation.<br><br></div><blockquote class=3D"gmail_quote" style=
=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding=
-left:1ex"><div style=3D"overflow-wrap: break-word;"><div></div><div>For CH=
ECKSIGFROMSTACK (CSFS), I think the question is whether we want to make it =
as a separate opcode, or combine that with CHECKSIG. If it is a separate op=
code, I think it should be a separate BIP. If it is combined with CHECKSIG,=
 we could do something like this: If the bit 10 of SIGHASH2 is set, CHECKSI=
G will pop one more item from stack, and serialize its content with the tra=
nsaction digest. Any thought?</div></div></blockquote><div><br></div><div>I=
 prefer a different opcode for CHECKSIGFROMSTACK because I dislike opcodes =
that pop a non-static number of elements off the stack.=C2=A0 Popping a dyn=
amic number of stack elements makes it more difficult to validate that a Sc=
ript pubkey doesn&#39;t allow any funny business.<br></div></div></div></di=
v>

--0000000000000589c8056d9893c8--