code review 4378046: crypto/rsa: support > 3 primes.

crypto/rsa: support > 3 primes.

With full multi-prime support we can support version 1 PKCS#1 private
keys. This means exporting all the members of rsa.PrivateKey, thus
making the API a little messy. However there has already been another
request to export this so it seems to be something that's needed.

Over time, rsa.GenerateMultiPrimeKey will replace rsa.GenerateKey, but
I need to work on the prime balance first because we're no longer
generating primes which are a multiples of 8 bits.

Fixes issue 987.

[rsc, 2011-04-08 19:02:45]

I am starting to think that our insistence on not having Init was
a bad idea.

Where do PrivateKeys come from?
I see
   rsa.GenerateKey
   x509.CreateCertificate
   x509.ParsePKCS1PrivateKey

As long as those functions return fully initialized keys, I think
it might be okay to reintroduce Init for the special case code
that wants to create its own keys.

Then the rwMutex goes away along with all the uncertainty
about who can look at what.

http://codereview.appspot.com/4378046/diff/5001/src/pkg/crypto/rsa/rsa.go
File src/pkg/crypto/rsa/rsa.go (right):

http://codereview.appspot.com/4378046/diff/5001/src/pkg/crypto/rsa/rsa.go#new...
src/pkg/crypto/rsa/rsa.go:96: rwMutex          sync.RWMutex // protects the
following
If rwMutex must be held in order to read Dp, Dq, Qinv, AdditionalPrimes safely, 
it's strange to export them but not rwMutex.

Does it makes sense to split all the computed material out into 
a different structure so that the interface can be something like:

type PrivateKey struct {
    PublicKey
    D *big.Int
    P *big.Int
    Q *big.Int
    Additional []*big.Int
    rwMutex sync.RWMutex
    derived DerivedKey
}

func (k *PrivateKey) Derived() *DerivedKey

[agl1, 2011-04-22 15:26:02]

I'm very happy to get rid of the mutex. I've also split the precomputed values
into their own structure and disentangled the additional primes from the
precomputation data. Now, the obvious parts of the PrivateKey are at the
beginning and the precomputation details are somewhat hidden away. They are
still public, however, because x509 needs to be able to export them.

[rsc, 2011-04-22 18:43:14]

LGTM

http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go
File src/pkg/crypto/rsa/rsa.go (right):

http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go#ne...
src/pkg/crypto/rsa/rsa.go:108: // information that we mirror this.
s/information/important/?

[agl1, 2011-04-22 19:33:40]

Hello rsc (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://go.googlecode.com/hg/

[agl1, 2011-04-22 19:34:08]

*** Submitted as http://code.google.com/p/go/source/detail?r=351ee4100fc6 ***

crypto/rsa: support > 3 primes.

With full multi-prime support we can support version 1 PKCS#1 private
keys. This means exporting all the members of rsa.PrivateKey, thus
making the API a little messy. However there has already been another
request to export this so it seems to be something that's needed.

Over time, rsa.GenerateMultiPrimeKey will replace rsa.GenerateKey, but
I need to work on the prime balance first because we're no longer
generating primes which are a multiples of 8 bits.

Fixes issue 987.

R=rsc
CC=golang-dev
http://codereview.appspot.com/4378046

http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go
File src/pkg/crypto/rsa/rsa.go (right):

http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go#ne...
src/pkg/crypto/rsa/rsa.go:108: // information that we mirror this.
On 2011/04/22 18:43:14, rsc wrote:
> s/information/important/?

Done.

[rsc, 2011-04-22 19:41:25]

broke build

On Fri, Apr 22, 2011 at 15:34,  <agl@golang.org> wrote:
> *** Submitted as
> http://code.google.com/p/go/source/detail?r=351ee4100fc6 ***
>
> crypto/rsa: support > 3 primes.
>
> With full multi-prime support we can support version 1 PKCS#1 private
> keys. This means exporting all the members of rsa.PrivateKey, thus
> making the API a little messy. However there has already been another
> request to export this so it seems to be something that's needed.
>
> Over time, rsa.GenerateMultiPrimeKey will replace rsa.GenerateKey, but
> I need to work on the prime balance first because we're no longer
> generating primes which are a multiples of 8 bits.
>
> Fixes issue 987.
>
> R=rsc
> CC=golang-dev
> http://codereview.appspot.com/4378046
>
>
>
> http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go
> File src/pkg/crypto/rsa/rsa.go (right):
>
>
http://codereview.appspot.com/4378046/diff/11001/src/pkg/crypto/rsa/rsa.go#ne...
> src/pkg/crypto/rsa/rsa.go:108: // information that we mirror this.
> On 2011/04/22 18:43:14, rsc wrote:
>>
>> s/information/important/?
>
> Done.
>
> http://codereview.appspot.com/4378046/
>