-------------------------------------------------------------------
Thu Jun 18 17:15:55 UTC 2020 - Michał Rostecki <mrostecki@suse.com>

- Add patch which fixes the ipv6 sysctl error on systems without
  IPv6 enabled (bsc#1173039)
  * 0006-only-enable-ipv6-if-ipv6-is-enabled.patch

-------------------------------------------------------------------
Thu May 21 21:57:25 UTC 2020 - Swaminathan Vasudevan <svasudevan@suse.com>

- Adds patch 0005-fix-issue-where-cilium-agent-fails-to-start-on-nodes-without-a-default-gateway.patch (bsc#1171592)

-------------------------------------------------------------------
Mon May  4 05:48:30 UTC 2020 - Swaminathan Vasudevan <svasudevan@suse.com>

- Adds a couple of patches that fixes bpf load error (bsc#1151876)
  * 0001-rename-PolicyMapMaxEntries-to-PolicyMapEntries-and-define-policy-map-size-limits-as-consts.patch(combined)
  * 0002-allow-to-configure-bpf-nat-global-max-using-helm.patch
  * 0003-reduce-default-number-for-TCP-CT-and-NAT-table-max-entries.patch
  * 0004-add-option-to-dynamically-size-BPF-maps-based-on-system-memory.patch

-------------------------------------------------------------------
Fri Mar  6 10:50:09 UTC 2020 - Michał Rostecki <mrostecki@opensuse.org>

- Add bpftool as a runtime dependency.

-------------------------------------------------------------------
Tue Mar  3 17:22:12 UTC 2020 - Michał Rostecki <mrostecki@suse.com>

- Update to version 1.6.6:
  * Highlights
    * KVStore free operation
    * 100% Kube-proxy replacement
    * Socket-based load-balancing
    * Policy scalability improvements
    * Generic CNI chaining
    * Native AWS ENI mode
  * Bugfixes
    * Fix to allocate a global identity for an empty container
      label-set.
    * Enable IP forwarding on daemon start
    * eni: Fix releases of excess IPs
    * cni: Fix IP leak when CNI ADD times out
    * cni: Fix noisy warning "Unknown CNI chaining configuration"
    * Fix cilium installation in GCloud beta "rapid" channel
    * garbage collect stale distributed locks
    * fqdn: Support setting tofqdns-min-ttl to 0
  * Misc Changes
    * Add missing words to spelling_wordlist
    * Fix GC Locks bugs
    * nodeinit/templates: fix indentation of sys-fs-bpf
    * v1.6: install: Update the chart versions

-------------------------------------------------------------------
Fri Feb 28 11:01:57 UTC 2020 - Michał Rostecki <mrostecki@suse.com>

- Update to version 1.5.12:
  * Bugfixes
    * Fix to allocate a global identity for an empty container
      label-set.
    * Enable IP forwarding on daemon start
    * cni: Fix IP leak when CNI ADD times out
    * garbage collect stale distributed locks
  * Misc Changes
    * Fix GC Locks bugs
- Use %requires_eq for cilium-proxy.

-------------------------------------------------------------------
Mon Jan 13 10:59:15 UTC 2020 - Michał Rostecki <mrostecki@suse.com>

- Update to version 1.5.11+git.20200110:
  * Important Bug Fixes
    * Compatibility with Envoy 1.12.2, which includes important
      security fixes for CVE-2019-18801, CVE-1019-18802,
      CVE-1019-18838. For more information, see Envoy 1.12.2
      Release Notes.
    * Allow for future endpoint regenerations to try to re-attempt
      to create the template for a given hash of endpoint
      configuration if a previous endpoint build with the same hash
      of endpoint configuration failed
    * Fix regression in datapath where traffic coming from the
      outside world has its packets fragmented since the MTU of
      the cilium_host interface is lower than the MTU of the
      external facing device
    * Fix setting of route MTU depending on whether prefix in route
      is local or remote
    * Fix breakage of "externaltrafficpolicy: local" by limiting
      host->service IP SNAT to local traffic
    * Fix transient rules to use allocation CIDR, since they will
      be dropped because the rules were matching on lxc+
    * Atomically replace filters when possible to avoid momentary
      drops of packets upon daemon restart
    * Fix proxy port leak on endpoint delete
    * Fix deadlock when endpoint EventQueue is full when endpoint
      is being deleted
    * Fix deadlock when service events channel is full, which
      caused no service-related events from
    * Fix service not getting removed from service cache
    * Remove no longer existing services from the BPF loadbalancer
      maps
    * Fix error where warning_error metric was not being reported
    * Do not log error when endpoints are trying to transition
      from waiting-to-regenerate to waiting-to-regenerate state,
      since this action just means that no action needs to be
      performed
    * Change endpoint state machine to ensure that endpoint
      restore goroutines are the only goroutines which can
      regenerate endpoints
    * Kubernetes to be plumbed by Cilium
    * Return error if Enqueue fails for EventQueue
    * Protect against panic resulting from enqueueing same Event
      twice
    * Fix delayed node delete event to avoid deleting routes to
      nodes that are still in the cluster
  * Bug fixes
    * Fix segfault in node equality check
    * Re-add node to kv-store once it restarts after being down for
      more than 15 minutes
    * Delete service ports from datapath if they are removed with a
      k8s update
    * Re-fetch CEP from kube-apiserver in case of update conflict
    * Update toServices policy when service endpoints are modified
    * Do not remove [kube|core]-dns pods if unmanaged-pod-watcher-
      interval == 0
    * Fix vishvananda/netlink library's VethPeerIndex() stack
      corruption with 4.20+ kernels.
  * Datapath Fixes
    * install transient rules during agent restart to avoid
      momentary packet drops during agent restart
    * mtu: Use Route MTU for MTU source for cilium_host
    * bpf: Do not fail if route contains gw equal to dst
    * iptables: mount xtables.lock to allow for concurrent
      iptables access to avoid Cilium crashing upon bootstrap when
      other processes are accessing iptables
    * iptables: ensure 0xd00 and 0xe00 marks dont cause collisions
      with kube-proxy
    * bpf: add skb_pull_data to bpf_network to avoid revalidate
      error
    * bpf: refactor duplicate code into revalidate_data_first()
      and utilize it
    * bpf: fix verifier error due to repulling of skb->data/end
    * remove old probe content before restoring assets to avoid
      compilation failures upon downgrade
  * Encryption Fixes
    * Always use CiliumInternal IP instead of Router IP for ipsec
  * Proxy
    * Perform dnsproxy Close() in the returned finalizeFunc
    * Create redirects before bpf map updates.
    * Do not error out if reading of open ports fails
    * Fix proxylib injection
    * Use LPM ipcache instead of xDS when available
    * Reduce logging output
    * Add SO_MARK option to listener config
  * Enhancements
    * Add way to use usr prandom as slave selection in lb
    * Make all ct timeouts configurable
    * Allow "_" in DNS names to support service discovery schemes
      to ToFQDNs policy
  * Dependencies
    * Bump K8s dependency to v1.15.5
  * Documentation
    * Clarify usage of bpf fs mount
    * Fix clustermesh secrets namespace
    * Update direct routing limitations for policy
    * Fix getting started guide command which creates secrets
  * Misc
    * Add github actions to cilium
    * Update golang to 1.12.14
    * bugtool: add cilium node list output

-------------------------------------------------------------------
Thu Oct 17 15:47:04 UTC 2019 - Richard Brown <rbrown@suse.com>

- Remove obsolete Groups tag (fate#326485)

-------------------------------------------------------------------
Mon Jul 29 11:38:56 UTC 2019 - mrostecki@opensuse.org

- Update to version 1.5.5:
  * lbmap: Get rid of bpfService cache lock
  * retry vm provisioning, increase timeout
  * daemon: Remove svc-v2 maps when restore is disabled
  * daemon: Do not remove revNAT if removing svc fails
  * pkg/k8s: add conversion for DeleteFinalStateUnknown objects
  * cli: fix panic in cilium bpf sha get command
  * Retry provisioning vagrant vms in CI
  * pkg/k8s: hold mutex while adding events to the queue
  * Change nightly CI job label from fixed to baremetal
  * test: set 1.15 by default in CI Vagrantfile
  * daemon: Change loglevel of "ipcache entry owned by kvstore or agent"
  * pkg/kvstore: add etcd lease information into cilium status
  * pkg/k8s: do not parse empty annotations
  * maps/lbmap: protect service cache refcount with concurrent access
  * operator: add warning message if status returns an error
  * pkg/kvstore: fix nil pointer in error while doing a transaction in etcd
  * examples/kubernetes: bump cilium to v1.5.4
  * bpf: Remove unneeded debug instructions to stay below instruction limit
  * bpf: Prohibit encapsulation traffic from pod when running in encapsulation mode
  * pkg/endpointmanager: protecting endpoints against concurrent access
  * test: set k8s 1.15 as default k8s version
  * CI: Clean VMs and reclaim disk in nightly test
  * allocator: fix race condition when allocating local identities upon bootstrap
  * identity: Initialize well-known identities before the policy repository.
  * cilium: docker.go ineffectual assignment
  * Disable automatic direct node routes test
  * kubernetes-upstream: add seperate stage to run tests
  * docs: update documentation with k8s 1.15 support
  * test: run k8s 1.15.0 by default in all PRs
  * test: test against 1.15.0
  * vendor: update k8s to v1.15.0
  * bpf: Set random MAC addrs for cilium interfaces
  * endpoint: Set random MAC addrs for veth when creating it
  * vendor: Update vishvananda/netlink
  * mac: Add function to generate a random MAC addr
  * test: remove unused function
  * test: introduce `ExecShort` function
  * docs: Clarify about legacy services enabled by default
  * pkg/metrics: re-register newStatusCollector function
  * CI: Clean workspace when all stages complete
  * CI: Clean VMs and reclaim disk after jobs complete
  * CI: Report last seen error in CiliumPreFlightCheck
  * fqdn: correctly populate Source IP and Port in `notifyOnDNSMsg`
  * test: do not overwrite context in `GetPodNamesContext`
  * test: change `GetPodNames` to have a timeout
  * test: make sure that `GetPodNames` times out after 30 seconds
  * CI: Ensure k8s execs cancel contexts
  * test: Fix NodeCleanMetadata by using --overwrite
  * test: add timeout to `waitToDeleteCilium` helper function
  * .travis: update travis golang to 1.12.5
  * Don't set debug to true in monitor test
  * pkg/lock: fix RUnlockIgnoreTime
  * daemon: fix endpoint restore when endpoints are not available
  * Preload vagrant boxes in k8s upstream jenkinsfile
  * pkg/health: Fix IPv6 URL format in HTTP probe
  * test: use context with timeout to ensure that Cilium log gathering takes <= 5 minutes
  * k8s: Introduce test for multiple From/To selectors
  * k8s: Fix policies with multiple From/To selectors
  * test: create session and run commands asynchronously
  * test: bump to k8s 1.14.3
  * test: error out if no-spec policies is allowed in k8s >= 1.15
  * test/provision: upgrade k8s 1.15 to 1.15.0-beta.2
  * test: have timeout for `Exec`
  * pkg/kvstore: introduced a dedicated session for locks
  * pkg/kvstore: implement new *IfLocked methods for etcd
  * kvstore/allocator: make the allocator aware of kvstore lock holding
  * pkg/kvstore: add Comparator() to KVLocker
  * pkg/kvstore: add new *IfLocked methods to perform txns
  * test: bump k8s 1.13 to 1.13.7
  * test: Enable IPv6 forwarding in test VMs
  * docs: Remove architecture target links
  * test: add serial ports to CI VMs
  * *.Jenkinsfile: remove leftover failFast
  * endpoint: make sure `updateRegenerationStatistics` is called within anonymous function
  * Prepare for v1.5.3
  * test: do not spawn goroutines to wait for canceled context in `RunCommandContext`
  * node/store: Do not delete node key in kvstore on node registration failure
  * kvstore/store: Do not remove local key on sync failure
  * node: Delay handling of node delete events received via kvstore
  * test/provision: bump k8s 1.12 to 1.12.9
  * pkg/kvstore: do not always UpdateIfDifferent with and without lease
  * Don't overwrite minRequired in WaitforNPods
  * daemon: Don't log endpoint restore if IP alloc fails
  * daemon: Refactor individual endpoint restore
  * test: provide context which will be cancled to `CiliumExecContext`
  * Jenkinsfile: backport all Jenkinsfile from master
  * doc: Document regressions in 1.5.0 and 1.5.1
  * Prepare for release v1.5.2
  * test: Disable unstable K8sDatapathConfig Encapsulation Check connectivity with transparent encryption and VXLAN encapsulation
  * Add kvstore quorum check to Cilium precheck
  * pkg/kvstore: acquire a random initlock
  * kvstore: Wait for kvstore to reach quorum
  * ipcache: Fix automatic recovery of deleted ipcache entries
  * tests, k8s: add monitor dump helper for debugging
  * bugtool: add raw dumps of all lb and lb-related maps
  * pkg/labels: ignore all labels that match the regex "annotation.*"
  * docs: Add note about keeping enable-legacy-services
  * docs: Add note about running preflight-with-rm-svc-v2.yaml
  * examples: Add preflight DaemonSet for svc-v2 removal
  * operator: Fix health check API
  * doc: Add EKS node-init DaemonSet to mount BPF filesystem
  * pkg/kvstore: perform update if value or lease are different
  * kvstore/allocator: do not immediately delete master keys if unused
  * pkg/kvstore: store Modified Revision number KeyValuePairs map
  * kvstore/allocator: do not re-allocate localKeys
  * kvstore/allocator: move invalidKey to cache.go
  * kvstore/allocator: add lookupKey method
  * allocator: Provide additional info message on key allocation and deletion
  * allocator: Fix garbage collector to compare prefix
  * allocator: Make GetNoCache() deterministic
  * kvstore/allocator: protect concurrent access of slave keys
  * kvstore/allocator: release ID from idpool on error
  * kvstore/allocator: do not re-get slave key on allocation
  * pkg/kvstore: Run GetPrefix with limit of 1
  * allocator: Verify locally allocated key
  * envoy: Prevent resending NACKed resources also when there are no ACK observers.
  * endpoint: Guard against deleted endpoints in regenerate
  * service: Reduce backend ID allocation space
  * cilium: fix up source address selection for cluster ip
  * CI: Log at INFO and above for all unit tests
  * bpf: Fix dump parsers of encrypt and sockmap maps
  * pkg/maps: use pointer in receivers for GetKeyPtr and GetValuePtr
  * test: fix incorrect deletion statement for policy
  * proxylib: Fix egress enforcement
  * Recover from ginkgo fail in WithTimeout helper
  * docs: move well known identities to the concepts section
  * docs: update well-known-identities documentation
  * add support for k8s 1.14.2
  * test: add v1.15.0-beta.0 to the CI
  * cni: Fix incorrect logging in failure case
  * daemon: Make policymap size configurable
  * Add jenkins stage for loading vagrant boxes
  * bpf: Remove several debug messages
  * Revert "pkg/bpf: add DeepCopyMapKey and DeepCopyMapValue"
  * Revert "pkg/{bpf,datapath,maps}: use same MapKey and MapValue in map iterations"
  * Revert "pkg/bpf: add newer LookupElement, GetNextKey and UpdateElement functions"
  * Revert "pkg/bpf: use own binary which does not require to create buffers"
  * Revert "maps/ctmap: add ctmap benchmark"
  * bpf: force recreation of regular ct entry upon service collision
  * pkg/endpoint: fix assignment in nil map on restore
  * pkg/ipcache: initialize globalmap at import time
  * test/provision: bump k8s testing to v1.13.6
  * bpf: do propagate backend, and rev nat to new entry
  * datapath: Redo backend selection if stale CT_SERVICE entry is found
  * daemon/Makefile: rm -f on make clean for links
  * CI: Consolidate Vagrant box information into 1 file
  * cilium: encode table attribute in Route delete
  * daemon: Remove stale maps only after restoring all endpoints
  * envoy: Do not use deprecated configuration options.
  * cilium: IsLocal() needs to compare both Name and Cluster
  * daemon: Do not restore service if adding to cache fails
  * daemon: Improve logging of service restoration
  * doc: Adjust documentation with new dynamic gc interval
  * ctmap: Introduce variable conntrack gc interval
  * pkg/envoy: use proto.Equal instead comparing strings
  * test: replace guestbook test docker image
  * docs: give better troubleshooting for conntrack-gc-interval
  * operator: fix concurrent access of variable in cnp garbage collection
  * Bump vagrant box version for tests to 151
  * cni: Fix unexpected end of JSON input on errors
  * docs: add missing cilium-operator-sa.yaml for k8s 1.14 upgrade guide
  * maps: Remove disabled svc v2 maps
  * fqdn: DNSProxy does not fold similar DNS requests
  * docs: fix architecture images' URL
  * CI: Consolidate WaitforNPods and WaitForPodsRunning
  * CI: WaitForNPods uses count of pods
  * Dockerfile: update golang to 1.12.5
  * metrics: add map_ops_total by default
  * Bump vagrant box versions for tests
  * Jenkins separate directories for parallel builds

-------------------------------------------------------------------
Fri Jun  7 13:36:27 UTC 2019 - Michal Rostecki <mrostecki@opensuse.org>

- Switch container image URI from devel:kubic:containers to
  openSUSE:Containers:Tumbleweed.

-------------------------------------------------------------------
Mon Jun  7 13:34:10 CEST 2019 - ndas@suse.de

- Update to version 1.5.3:
  * pkg/kvstore: do not always UpdateIfDifferent with and without lease
  * daemon: Refactor individual endpoint restore
  * daemon: Don't log endpoint restore if IP alloc fails
  * Don't overwrite minRequired in WaitforNPods
  * node: Delay handling of node delete events received via kvstore
  * kvstore/store: Do not remove local key on sync failure
  * node/store: Do not delete node key in kvstore on node registration failure
  * Jenkinsfile: backport all Jenkinsfile from master
  * test/provision: bump k8s 1.12 to 1.12.9
  * test: do not spawn goroutines to wait for canceled context in `RunCommandContext`
  * test: provide context which will be cancled to `CiliumExecContext`

-------------------------------------------------------------------
Mon Jun  3 13:34:10 CEST 2019 - ndas@suse.de

- Add cniVersion in cilium cni config

-------------------------------------------------------------------
Fri May 10 10:20:32 UTC 2019 - Michal Rostecki <mrostecki@opensuse.org>

- Update to version 1.5.1:
  * Important Bugfixes:
    * Fix bug where Cilium would refuse to start if ipv6 netfilter
      modules are unavailable.
    * Warn when iptables modules are not available.
    * Use all labels to restore endpoint identity to correctly
      filter labels upon restart.
    * Fix cases where multiple bindings are provided to CLI flags.
  * New Functionality / Enhancements:
    * Add node-init script to automatically restart pods managed by
      kubenet on GKE
    * Add functionality to enable or disable metrics for specific
      subsystems
    * bpf syscall metrics are disabled by default for performance
    * Update node, node/status to allow for patch operations in
      Cilium RBAC
    * Patch, instead of update, node annotations for better
      performance
    * Annotate node status with NetworkUnavailable as false
    * Performance increase by not allocating any memory when
      iterating over BPF maps
    * CLI now prints tunnel endpoint for RemoteEndpointInfo
    * Try to register node forever in nodediscovery
    * Remove unused buildqueue package
  * Minor Bug Fixes:
    * endpoint: do not serialize JSON for EventQueue field
    * Avoid unlocked access of endpoint security identity when
      calculating what rules select an endpoint
    * Only dump bpf lb list if map exists
    * Fix bug where endpoint state metrics get stuck with nonzero
      endpoints in restoring state
    * Do not init config when running with --cmdref parameter
    * Improve separation between cilium-agent and cilium CLI
    * Add cilium namespace to fqdn_gc_deletions_total metric
    * Force preallocation for SNAT maps of LRU type
    * Set BPF_F_NO_PREALLOC before comparing maps
  * Operator:
    * Improve cilium-operator bootstrap sequence (Start health API
      earlier, add more logging to see where the operator blocks
      on startup)
    * Add ca-certificates to operator
  * Documentation:
    * Add upgrade guide from >=1.4.0 to 1.5
    * Mention enable-legacy-services flag in upgrade docs
    * Add k8s 1.14 to supported versions for testing
    * Improve configmap documentation
    * Document how to get started with MicroK8s, and provide example
      YAMLs
    * Fix typo in encryption algorithm: GMC -> GCM
    * Fix up Ubuntu apt-get install command
    * Minor fixes to AWS EKS and AWS Metadata filtering GSGs
  * CI:
    * Wait for endpoints to be ready after containers are created,
      deleted
    * Ensure that `go fmt` check always runs correctly in CI 
    * Increase test suite timeouts to allow for cases where tests
      take longer
    * Do not set enable-legacy-services in v1.4 ConfigMap
    * Update k8s testing versions to v1.11.10 and v1.12.8
    * Make function provided to WithTimeout run asynchronously to
      avoid test suites getting stuck
- Add cilium-k8s-yaml package with Kubernetes yaml file to run
  Cilium containers.

-------------------------------------------------------------------
Fri May 10 12:02:55 CEST 2019 - ndas@suse.de

- Add missing gzip package, cilium does zgrep of /proc/config.gz

-------------------------------------------------------------------
Mon May 06 13:53:28 UTC 2019 - Michal Rostecki <mrostecki@opensuse.org>

- Update to version 1.5.0:
  * BPF programs templating which alows to inject information into
    ELF files instead of compiling separate programs with separate
    data for each endpoint.
  * BPF-based masquerading support - a native BPF-based SNAT
    engine.
  * Optimizations for policy engine and load balancer.
- Remove patches which are accepted upstream:
  * cilium-allow-to-add-extra-go-build-flags.patch
  * cilium-allow-to-specify-cni-install-dirs.patch

-------------------------------------------------------------------
Tue Apr 16 12:53:38 UTC 2019 - Michal Rostecki <mrostecki@opensuse.org>

- Add cilium-operator package which provides the Kubernetes
  operator that does garbage collector work for Cilium.
- Do not require cilium and docker in cilium-init package.

-------------------------------------------------------------------
Fri Apr 12 10:51:14 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Add cilium-init package, which provides the script for Cilium
  init container.

-------------------------------------------------------------------
Fri Mar 29 15:59:38 UTC 2019 - mrostecki@opensuse.org

- Update to version 1.4.2:
  * Prepare for v1.4.2 release
  * cilium: ipsec, zero cb[0] to avoid incorrectly encrypting
  * contrib: Update backporting README
  * contrib: Fix cherry-pick to avoid omitting parts of patch
  * cilium: push decryption up so we can decrypt even if not endpoint
  * cilium: populate wildcard src->dst policy for ipsec
  * daemon: Remove old health EP state dirs in restore
  * api: Return 500 when API handlers panic.
  * ipcache: Protect from delete events for alive IP but mismatching key
  * store: Protect from deletion of local key via kvstore event
  * test: Wait for cilium to start in runtime provision
  * contrib: fix extraction of cilium-docker binary
  * contrib: Update rebase-bindata to use fix-sha.sh
  * contrib: Add new script to auto-fix bpf.sha
  * cherry-pick: Print sha when applying patch.
  * check-stable: Sort PRs by merge date
  * workloads: Don't spin up receive queue in periodic watcher
  * workloads: Change watcher interval from 30 seconds to 5 minutes
  * workloads: Synchroneous handling of container events
  * endpoints: Add optional callback to WaitForPolicyRevision
  * daemon: Track policy implementation delay by source
  * agent: Wait to regenerate restore endpoints until ipcache has been populated
  * ipcache: Provide WaitForInitialSync() to wait for kvstore sync
  * pkg/kvstore: add 15 min TTL for the first session lease
  * policy: Add missing import error metric calls
  * endpoint: Fix ENABLE_NAT46 endpoint config validation
  * endpoint: Fix and quieten endpoint revert logs
  * test: Get rid of JoinEP flakes
  * ctmap: Print source addresses in ctmap cli
  * cilium: fix bailing out on auto-complete when v4/v6 ranges are specified
  * test: Test upgrade from v1.3 to master
  * doc: Fix --tofqdns-pre-cache reference
  * doc: Fix delete pod commend in clustermesh guide
  * bpf: Enable pipefail option in init.sh
  * cilium: bpftool included DS reports error on bpf_sockops load
  * cilium: sockmap remove socket.h dependency
  * cilium: sockmap, convert BPF_ANY to BPF_NOEXIST
  * 1: fix when have black hole route container pod CIDR can cause postIpAMFailure range is full
  * pkg/kvstore: do not use default instance to create new instance module
  * bpf: Do not account tx for CT_SERVICE
  * cilium.io/v2: set DerivativePolicies json to derivativePolicies
  * fqdn-poller: Ensure monitor events contain all data
  * ctmap: Fix order of CtKey{4,6} struct fields
  * release: fix uploadrev script to work with changes made after 1.3
  * datapath: Fix nil dereference in logging statement
  * Prepare 1.4.1 release
  * k8s/utils: wrap kubernetes controller with ControllerSyncer
  * k8s/utils: make the ControllerSynced fields public
  * allocator: Wait until kvstore is connected before allocating global identities
  * policy: Fix ipcache synchronization on startup
  * cilium: ipsec, fix kube-proxy compatability
  * cilium: ipsec, remove bogus mark set
  * cilium: ipsec, zero CB_SRC_IDENTITY to ensure we don't incorrectly encrypt
  * cilium: k8s watcher, push internal Cilium IPs through annotations
  * policy: Add unit tests for ResolvePolicy() for L7 + ingress wildcards
  * identity/cache: Allow using GetIdentityCache() without initializing allocator
  * Change endpoint policy status map to regular map
  * Minor disambiguation to 1.4 release/upgrade doc
  * examples: Fix docker-compose mount points
  * docs: Add note about triggering builds with net-next
  * FQDN: Set always a empty ToCIDRSet in case of no entries in cache.
  * docs: re write k8s setup for ipsec
  * datapath/linux: log errors for ipsec setup
  * linux/ipsec: decode ipsec keys from hex
  * cilium preflight command for FQDN poller upgrade
  * docs: Add FQDN Poller upgrade impact & instructions
  * docs: Small changes to toFQDN and DNS sections
  * docs: Move "Obtaining DNS Data" to L7 section
  * cilium preflight container prepares tofqdn-pre-cache
  * pkg/identity: add well known identity for cilium-etcd-operator
  * pkg/kvstore: wait until etcd configuration files are available
  * policy/api: generate missing deepcopy code
  * vendor: fix Gopkg.lock
  * datapath: Clean up stale ipvlan maps
  * cilium, bpf: only account tx for egress direction
  * examples: Update docker-compose examples
  * lookup rule for the given IP family
  * cilium-operator.Dockerfile: set `klog` logging values from cilium-operator
  * datapath: Clean up config map on startup
  * datapath: Fix map cleanup for CT maps
  * Update k8s-install-gke.rst
  * cilium-docker-plugin: set default CMD to /usr/bin/cilium-docker
  * api/v1: remove requirements of labels in endpoints API
  * apis/cilium.io: do not regenerate deepcopy for unnecessary structs

-------------------------------------------------------------------
Mon Mar 11 14:31:04 UTC 2019 - ndas@suse.de

- Move cilium-docker files to cilium-cni

-------------------------------------------------------------------
Mon Mar  4 14:43:27 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Add gcc as a runtime dependency. BPF programs need to have libgcc
  and libgcc_s linked in.
  https://github.com/cilium/cilium/issues/7273

-------------------------------------------------------------------
Mon Mar  4 10:38:19 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Provide an explanation why glibc-devel-32bit is needed.
- Ship cilium-cni and cilium-docker in separate packages.

-------------------------------------------------------------------
Fri Mar  1 15:23:36 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Add missing runtime dependencies which are needed to execute
  scripts shipped with Cilium and to compile BPF programs.

-------------------------------------------------------------------
Wed Feb 27 15:52:38 UTC 2019 - ndas@suse.de

- Fix license. BPF code templates are licensed under GPLv2 while
  the rest is under Apache License, v2
  (see https://github.com/cilium/cilium#license)

  Cilium (the component licensed on Apache 2.0, written in Go) does
  two things with BPF program sources (licensed on GPL-2.0):

  * it executes llvm/clang to compile BPF program sources to object
    files
  * it executes tc (a utility which is a part of iproute2) to load
    object files into the kernel

  So, Cilium as a Go program only does execv calls on external
  utilities (llvm and iproute2) to perform some actions on BPF
  program sources and objects.

-------------------------------------------------------------------
Mon Feb 25 09:56:48 CET 2019 - ndas@suse.de

- Add missing GPL2 License for eBPF source codes

-------------------------------------------------------------------
Wed Feb 13 10:09:55 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Update to version 1.4.0:
  * doc: Fix key generation for encryption
  * doc: Add validation and troubleshooting section to encryption
    GSG
  * datapath: Report IPsec route installation errors
  * datapath: Fix IPsec with IPv4 or IPv6 disabled
  * docs: Add ipvlan-based datapath limitations and requirements
  * doc, configmap: add missing entries
  * examples/kubernetes: Add tofqdns-enable-poller option
  * doc: Minor update to encryption guide
  * cilium: transparent encryption with ipsec getting started docs
  * Note about apiserver outside of cluster
- Add upstream patch which allows to set additional `go build`
  flags
  * cilium-allow-to-add-extra-go-build-flags.patch
- Add upstream patch which allows to specify installation
  directories for CNI files
  * cilium-allow-to-specify-cni-install-dirs.patch
- Make use of golang-packaging macros.
- Add rc* symlinks.

-------------------------------------------------------------------
Thu Feb  7 12:46:51 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Run code checkers/linters only on openSUSE Tumbleweed.

-------------------------------------------------------------------
Wed Feb  6 14:30:47 UTC 2019 - Michał Rostecki <mrostecki@opensuse.org>

- Add devel package which contains a header and .so file.
- Improve descriptions of all packages.
- Set BINDIR, DESTDIR and LIBDIR variables properly instead of
  manual installation of files in those destinations.
- Install bash completion script.
- Execute ldconfig in post and postun phases of the lib package.
- Fix Source attribute.

-------------------------------------------------------------------
Tue Feb  5 17:44:40 CET 2019 - ndas@suse.de

- Updated to 1.4-rc7
   *pkg/datapath/ipcache: stop leaking FD
   *pkg/fqdn: make any operation in the sourceRuleCopy
   *daemon: change policyAdd message type from Info to Debug for dns policies
   *pkg/endpoint: do not leak go routines if endpoint is disconnected
   *pkg/endpoint: ignore negative time durations in metrics
   *Endpoint: set a new context per endpoint regeneration
   *endpoint: revert endpoint BPF config map update if regenerateBPF fails
   *bpf: pin endpoint configuration map
   *endpoint: Unlock endpoint to prevent deadlocks.
   *daemon: Allow releasing builder while waiting for proxy ACKs
   *endpoint: Make regenaration timeout greater than ExecTimeout
   *endpoint: Eliminate ExecTimeout, ctx.
   *daemon: Use sync.Once, rewamp comments.
   *bpf: Fix node-port access to l7 proxy
   *bpf: Templatize endpoint configuration
   *maps: Add BPFConfigMap for endpoint configuration
   *endpoint: Support dynamic BPF configuration
   *bpf: Relax verifier in IPv6 drop case
   *bpf: Fix tcp flag access
   *bpf: Don't reset TCP timer on final ACK
   *cilium: spelling: sha is an acronym replace with SHA
   *bpf: Provide more specific drop reasons
   *proxylib: Update proxylib.h with go 1.11
   *agent: Fix invalid printf style invocations
   *gitignore: Ingore cilium-ring-dump binary
   *lbmap: Retrieve service ID when dumping BPF map
   *service: Restore service IDs before connecting to Kubernetes apiserver
   *service: Restore bpfservie cache on startup
   *lbmap: Add unit test for getBackends()
   *idpool: Factor out IDPool from allocator into package for reuse
   *idpool: Fix leaseAvailableID() and slice out of bounds
   *node: Don't insert own node into tunnel map
   *bpf: Avoid routing loops for former local endpoint IPs
   *test: Use cilium-etcd-operator
   *clustermesh: Fix race when shutting down clustermesh
   *clustermesh: Wait for controllers to be shutdown when closing
   *cni: Synchroneous pod label retrieval on CNI add
   *identity: Block createEndpoint() while identity is being resolved
   *bpf: Remove source MAC address validation
   *bpf: Remove destination MAC address verification
   *agent: Ignore IPV4_GATEWAY=0x0 when restoring
   - details changelogs are in https://github.com/cilium/cilium/projects/11
- disable bash completion
- added a new package libcilium1
- build with go1.10(need fix for cgo alignchecker issue)


-------------------------------------------------------------------
Tue Sep  4 15:58:32 CEST 2018 - ndas@suse.de

- change  00-cilium-cni.conf -> 10-cilium-cni.conf to keep sync with salt

-------------------------------------------------------------------
Mon Sep  3 14:06:13 CEST 2018 - ndas@suse.de

- Use proper bash-completion dir
- Updated to 1.2.1
  *docker, bpf: add iproute2 version which works around missing af_alg
  *docker, bpf: add bpftool for debugging and introspection
  *test/k8sT: use specific commit for cilium/star-wars-demo YAMLs
  *pkg/k8s: properly handle empty NamespaceSelector
  *lxcmap: Improve error messages in DeleteElement()
  *lxcmap: Fix always returning an error on delete
  *ctmap: Mark IPv6 CT GC as completed on success
  *endpoint: Fix endpoint regeneration failure metric
  *Block locked code in TriggerPolicyUpdates
  *Ignore non-existing link error in cni del
  *fqdn: Strip toCIDRSet rules to be more resilient
  *fqdn: Use UUIDs to manage rules
  *fqdn: Inject IPs on initial rule insert
  *xds: Ignore completion timeouts on resource upsert and delete
  *endpoint: Log when BPF regeneration times out not because of Envoy
  *endpoint: In BPF regeneration, create/remove listeners early
  *doc: Restructure and simplify upgrade guide
  *doc: Restructure installation guides
  *doc: AWS EKS installation guide
  *identity: Wait for initial set of security identities before restoring endpoints

-------------------------------------------------------------------
Wed Aug  8 12:06:50 CEST 2018 - ndas@suse.de

- Updated to 1.2.0-rc1
  * Inter cluster service routing
  * BPF based flow aggregation
  * BGP with kube-router 
  more at https://github.com/cilium/cilium/releases/tag/v1.2.0-rc1
- Add cilium group

-------------------------------------------------------------------
Mon Jun  4 16:04:59 UTC 2018 - dcassany@suse.com

- Refactor %license usage to simpler form 

-------------------------------------------------------------------
Mon Jun  4 09:50:42 UTC 2018 - dcassany@suse.com

- Make use of %license macro

-------------------------------------------------------------------
Wed Apr 25 10:54:45 CEST 2018 - ndas@suse.de

- Updated to v1.0.0
   Bugfixes Changes:
    *etcd: Clear the etcd status error when connectivity is OK (3824, @rlenglet)
    *ipcache: Fix ipcache deletion of old identities on update (3865, @rlenglet)
    *bpf: Fix tracing message for egress policy (3806, @joestringer)
    [- envoy-optional.patch]
- use url for source

-------------------------------------------------------------------
Wed Apr 18 13:49:11 CEST 2018 - ndas@suse.de

- skip doc, less depedency
- remove libelf1, zypper/rpm should auto resolve
- define _fillupdir if not so

-------------------------------------------------------------------
Mon Apr 16 18:14:11 CEST 2018 - ndas@suse.de

- clean up spec file
- use %fillup_only for cilium sysconfig
- move cilium-cni to %{_libexecdir}/cni like all other cni-plugins

-------------------------------------------------------------------
Mon Apr 16 14:20:01 UTC 2018 - jengelh@inai.de

- Combine %service_* macro calls to reduce generated code.
- Trim filler wording from description.
- Use modern tar invocation syntax.

-------------------------------------------------------------------
Mon Apr  9 11:42:11 UTC 2018 - mrostecki@suse.com

- Updated to v1.0.0-rc10
* API preparation for 1.0
    Changed the base prefix of the API from /v1beta to /v1 tada. The API will become stable with the 1.0 release. This makes client binaries with version < 1.0.0-rc10.

* Bugfixes Changes
    policymap: Avoid using golang arrays in entry (#3506, @joestringer)
    etcd: Run etcd version check in the background (#3499, @tgraf)
    Test: Fix bugtool on kubernetes 1.7 (#3487, @eloycoto)
    Fix L4-only policy egress to world and CIDR-only egress to world (#3486, @joestringer)
    proxy: Use the same proxy map size as in BPF (#3485, @rlenglet)
    bpf: Do not route packets from egress proxy back into cilium_host (#3473, @tgraf)
    Continue to show timestamps in error cases in CiliumNetworkPolicy NodeStatus. (#3461, @aanm)
    policy: Add missing EntitySlice autogen code (#3458, @raybejjani)
    Fix l3-dependent L4/L7 rules applying to CIDR egress traffic (#3434, @joestringer)

Other Changes
    bugtool: add ip rule and cilium-health status commands (#3500, @ianvernon)
    Policy: Kafka multi-topic request support (#3445, @manalibhutiyani)
- build cilium without envoy
  [+envoy-optional.patch]

-------------------------------------------------------------------
Fri Jan 19 14:02:21 CET 2018 - ndas@suse.de

- Updated to v1.0.0-rc2
*Major Changes

    Tech preview of Envoy as Cilium HTTP proxy, adding HTTP2 and gRPC support. (#1580, @jrajahalme)
    Introduce "cilium-health", a new tool for investigating cluster connectivity issues. (#2052, @joestringer)
    cilium-agent collects and serves prometheus metrics (#2127, @raybejjani)
    bugtool and debuginfo (#2044, @scanf)
    Add nightly test infrastructure (#2212, @ianvernon)
    Separate ingress and egress default deny modes with better control (#2156, @manalibhutiyani)
    k8s: add support for IPBlock and Egress Rules with IPBlock (#2096, @ianvernon)
    Kafka: Support access logging for Kafka requests/responses (#1870, @manalibhutiyani)
    Added cilium endpoint log command that returns the endpoint's status log (#2060, @raybejjani)
        Change endpoint status log in cilium endpoint get to show only the most recent log
    Routes connecting the host to the Cilium IP space is now implemented as
    individual route for each node in the cluster. This allows to assign IPs
    which are part of the cluster CIDR to endpoints outside of the cluster
    as long as the IPs are never used as node CIDRs. (#1888, @tgraf)
    Standardized structured logging (#1801, #1828, #1836, #1826, #1833, #1834, #1827, #1829, #1832, #1835, @raybejjani)

*Bugfixes Changes

    Fix L4Filter JSON marshalling (#1871, @joestringer)
    Fix swapped src dst IPs on Conntrack related messages on the monitor's output (#2228, @aanm)
    Fix output of cilium endpoint list for endpoints using multiple labels. (#2225, @aanm)
    bpf: fix verifier error in dameon debug mode with newer LLVM versions (#2181, @borkmann)
    pkg/kvstore: fixed race in internal mutex map (#2179, @aanm)
    Proxy ingress policy fix for LLVM 4.0 and greater. Resolves return code 500 'Internal Error' seen with some policies and traffic patterns. (#2162, @jrfastab)
    Printing patch clang and kernel patch versions when starting cilium. (#2137, @aanm)
    Clean up Connection Tracking entries when a new policy no longer allows it. #1667, #1823 (#2136, @aanm)
    k8s: fix data race in d.loadBalancer.K8sEndpoints (#2129, @aanm)
    Add internal queue for k8s watcher updates #1966 (#2123, @aanm)
    k8s: fix missing deep copy when updating status (#2115, @aanm)
    Accept traffic to Cilium in FORWARD chain (#2112, @tgraf)
        Also clear the masquerade bit in the FORWARD chain to skip the masquerade rule installed by kube-proxy
    Fix SNAT issue in combination with kube-proxy, when masquerade rule installed by kube-proxy takes precedence over rule installed by Cilium. (#2108, @tgraf)
    Fixed infinite loop when importing CNP to kubernetes with an empty kafka version (#2090, @aanm)
    Mark cilium pod as CriticalPod in the DaemonSet (#2024, @manalibhutiyani)
    proxy: Provide identities { host | world | cluster } in SourceEndpoint (#2022, @manalibhutiyani)
    In kubernetes mode, fixed bug that was allowing cilium to start up even if the kubernetes api-server was not reachable #1973 (#2014, @aanm)
    Support policy with EndpointSelector missing (#1987, @raybejjani)
    Implemented deep copy functionality when receiving events from kubernetes watcher #1885 (#1986, @aanm)
    pkg/labels: Filter out pod-template-generation label (#1979, @michi-covalent)
    bpf: Double timeout on building BPF programs (#1949, @raybejjani)
    policy: add PolicyTrace msg to AllowsRLocked() when L4 policies not evaluated (#1939, @gnahckire)
    Handle Kafka responses correctly (#1924, @manalibhutiyani)
    bpf: Avoid excessive proxymap updates (#2210, @joestringer)
    cilium-agent correctly restarts listening for CiliumNetworkPolicy changes when it sees decoding errors (#1899, @raybejjani)

-------------------------------------------------------------------
Wed Nov  8 12:46:02 CET 2017 - ndas@suse.de

- Initial version 0.12
