This change set makes sure that if there is a tx with the same txid as a
previous tx in the block, the tx is marked as a duplicate and invalidated.
This change set is not backward compatible and so a new application
capability is introduced.

Change-Id: Ia1490079d716c5d51b23af0765bffed27b9b843c
Signed-off-by: Alessandro Sorniotti <ale.linux@sopit.net>

* changes:
  [FAB-6567] Move ACL resources deeper in tree
  [FAB-6566] mv config/resources -> resourcesconfig

The default MSP implementation can be found in mspimpl.go.
The file is gigantic and it is getting complex to modify it.
In order to simply the picture, mspimpl.go will be split
in multiple files:
- mspimplsetup.go to handle the setup related functions
- mspimplvalidate.go to handle the validation related functions

No new code has been introduced, therefore no tests
have been added.

This change-set is propaedeutic to the implementation
of FAB-5668

Change-Id: I9b851bf919e185d79f59b66be11e40e9822033f7
Signed-off-by: Angelo De Caro <adc@zurich.ibm.com>

* changes:
  [FAB-5091] Purge from transient store by block height
  [FAB-6433] Pass ledger height at Persist()

This change set adds a core.yaml config that makes the coordinator
purge from the transient store private data that has been persisted
in blocks too far from the current ledger height.

The purge happens in every multiple of the configuration value upon commit.

Change-Id: I3a734993d17c8e83b6171510574f3923d682ddaa
Signed-off-by: yacovm <yacovm@il.ibm.com>

There are currently two types of BCCSP
providers:
- SW
- PKCS11

This CR addes a new "PLUGIN" type support
BCCSP providers implemented as Go plugins.
This is useful for providers who might
not be able to use PKCS11 as well as for
providers who want to use different
cryptographic algorithms that are not
supported via PKCS11 or the SW BCCSP.

Change-Id: I908634cbd4e6215ad682c97f5acc7a064bca643f
Signed-off-by: Gari Singh <gari.r.singh@gmail.com>

This change set makes the coordinator pass the ledger height when
it persists private data into the transient store.

Change-Id: Id890b4ce6043b841e0d9c9db7959e01e393b2f78
Signed-off-by: yacovm <yacovm@il.ibm.com>

This CR adds a check for the length of a config item's path and
triggers a panic if it is zero. The decision to still trigger
a panic stems from the fact that if the item's path is zero, it
means something unexpected occurred in the validation logic and
we should not proceed. The user who originally reported this bug
likely hit a config issue that is no longer possible after the
fix for FAB-5606.

Change-Id: I4052174c04e0810714823c3ca8a533ab2f261804
Signed-off-by: Will Lahti <wtlahti@us.ibm.com>

* changes:
  [FAB-6141] Add more flexible channel create helper
  [FAB-6140] Remove the provisional config encoder

This change set makes the private data push logic to disseminate
to a set of peers that doesn't take into account their belonging
to any organization.

Change-Id: Ic1969c50861c5b623042f37a8974b3d93a9ee20b
Signed-off-by: yacovm <yacovm@il.ibm.com>

This change set makes the pull mechanism of private data to try
select endorsers of a certain private RWSet over others.

Change-Id: If5f2e0a0031a208fee06c749216f708b81342847
Signed-off-by: yacovm <yacovm@il.ibm.com>

This change set makes the coordinator extract the endorsers
of the transaction in order to pass that to the pull mechanism
which would prioritize pulling from endorsers.

Change-Id: I35b317c73435dc625d14699f3fa02faea431eab0
Signed-off-by: yacovm <yacovm@il.ibm.com>

The MSP support for gossip (MessageCryptoService) takes into account
The entire PEM of an x509 based identity, which is too lenient.

Instead of enforcing explicit sanitation of input for GetPKIidOfCert(),
We will just make identities that have the wrong PEM type - invalid.

This is much more efficient because the GetPKIidOfCert is called many times
in the code, and sanitizing its input is therefore costly,
while if we just make identities that have the wrong type - invalid,
it would implicitly force sanitation of identities.
Another argument for doing so, would be that the Validate() is costly
anyway and is rarely performed in the peer's life cycle, therefore -
adding an additional slight overhead to it doesn't affect much the cost
of validating an identity.

The BCCSP-MSP Serialize() method already puts the right PEM type as of
an earlier commit that is targeted for v1.1 and isn't in v1.0.x

This change set makes the MessageCryptoService consider x509 based
identities valid only if they have a PEM type of either CERTIFICATE
or a blank one.

Change-Id: Ic332f592cf2ebfe364f6c99ae762b2789d9c2b51
Signed-off-by: yacovm <yacovm@il.ibm.com>

One of the most common questions when attempting to add an organization
to a channel is how to get the JSON definition for that organization.
The only real way to do this at present is to generate a genesis block,
decode the block, then go look for the org definition.

This CR adds a '-printOrg' flag to configtxgen which will print the
organization's definition as JSON to make this process easier.

Change-Id: Id12098f36feae8ca4751b82afaf2914166077e50
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

Now that almost all references the config templating code is gone, it
should be removed entirely from the codebase.

This final CR in the channel config generation cleanup series shows a
net decrease of over 400 lines, with an even larger shift of lines of
code from the orderer/peer production binaries and into the configtxgen
binary.

Change-Id: I614a10113c0f90b822da5103319a91f2f28b66b5
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

Presently, MakeChainCreationTransaction is a function of the
channelconfig package.  Although it does seem related to the
channelconfig and references many of the constants defined there, it is
actually unneeded during the runtime for either the orderer or peer
binaries, and is only used by CLI and tests.

Therefore, it is a better fit in the configtxgen/encoder package.

This CR also renames this method to MakeChannelCreationTransaction and
switches it to use the newer underlying config group generation
introduced earlier in this series.

Change-Id: Iece267073dd612c02cc8d1b2d7feea3c02731d8f
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

The original channel creation tx generation code assumes that no
modifications have been made to the organization definitions in the
ordering system channel.  For longer running systems this assumption may
not hold.

Consequently, a new function which accepts the ordering system channel's
config is required so that it can set the correct versions for the
organization's in the channel creation transaction.  This CR adds this
new function.

Change-Id: Iebeb247bbea5c688a14c6133f793373da7357c9b
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

With the introduction of FAB-6139 a new more readable and flexible
config encoder is available.  This means all existing usage of the old
deprecated provisional encoder should be removed.  This CR does that.

Change-Id: If64e0e7dc920c709cc39a111547c61c3ae4dc3f0
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

Change-Id: Ib88f2a69db9946dba9acf0614effe8b91b077bfb
Signed-off-by: Divyank Katira <Divyank.Katira@securekey.com>

This CR ensures that filtered transactions within filtered
block contain a nil CcEvent if no chaincode event
is present. It previously included an empty ChaincodeEvent
message.

Change-Id: If872b9e2013a5c43a6eb8cc11d24fa771a0b483e
Signed-off-by: Will Lahti <wtlahti@us.ibm.com>

Add links to the API Reference website to the main fabric docs

Change-Id: If3205418334b464a4140b0350f5e5fcafdaffec2
Signed-off-by: Jim Zhang <jzhang@us.ibm.com>

The ACL work preliminarily put all of the policy references and polici
at the root level of the config tree.  Because the peer resources tree
is now intended to support more than ACLs, keeping these at the root
level does not make sense.  Instead, these need to be moved down into
the tree.

This CR moves the old resource policy references from

 /Resources

to

 /Resources/APIs

It also creates

 /Resources/PeerPolicies

as a global location to declare policies which may be referenced both
for API ACL purposes, as well as later by other parts of the resources
tree.

Change-Id: I5c8eeba0472f480ac88ddbf02a0a1a3d90092463
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>

There used to be a config/channel and a config/resources.  This
structure was modified and config/channel moved to channelconfig.  This
CR simply cleans up the structure for symmetry's sake.

Change-Id: I2fde9dc0eba1dd0f401397673ebfad823264beb8
Signed-off-by: Jason Yellick <jyellick@us.ibm.com>