This fixes a deadlock introduced in 1457b4fd4a.
We can't use the coreapi here because it will try to take the PinLock (RLock) again, so revert this small part of 1457b4fd4a.
This used cause a deadlock when concurrently running `ipfs dag import` concurrently with the GC.
The bug is that `ipfs dag import` takes an RLock with the PinLock.
*the cars are imported, leaving a wide window of time*
Then GC Takes a Lock on that same RWMutex while taking the GC Lock (it blocks because it waits for the RLock to be released).
Then the car imports are finished and `ipfs dag import` tries to aqcuire the PinLock (doing an RLock) again in `Api().Pin`.
However at this point the RWMutex is starved, the runtime put a fence in front of RLocks if a Lock has been waiting for too lock (else you could have an endless stream of RLock / RUnlock forever delaying a Lock to ever go through).
The issue is that `ipfs dag import`'s original RLock which is blocking everyone will be released once it returns, which only happens when `Api().Pin` completes.
So we have a deadlock (ABA kind ?), because `ipfs dag import` waits on the GC Lock, which waits on `ipfs dag import`.
Calling the Pinner directly does not acquire the PinLock again, and thus does not have this issue.
- don't bypass the CoreApi
- don't use a goroutine and return channel for `importWorker`, when what's happening is really just a synchronous call
- only `PinLock()` when we are going to pin
- use `cid.Set` instead of an explicit map
- fail the request early if any pinning fail, no need to try to pin more if the request failed already
* feat: add support for carv2 import
* update to multicodec 0.4.0
* add sharness test for carv2
Co-authored-by: Keenan Nemetz <keenan.nemetz@gmail.com>
* feat(cmds): add deprecated and experimental status
Added programmatic state annotation introduced in
https://github.com/ipfs/go-ipfs-cmds/pull/225
for already deprecated / experimental commands.
* chore: go-ipfs-cmds v0.7.0
Co-authored-by: Lucas Molas <schomatis@gmail.com>
Bubble ErrNotFound improvements.
The is a really nice UX improvement because it know tell you which ipld block is missing instead of a generic "not found".
* plumb through go-datastore context changes
* update go-libp2p to v0.16.0
* use LIBP2P_TCP_REUSEPORT instead of IPFS_REUSEPORT
* use relay config
* making deprecation notice match the go-ipfs-config key
* docs(config): circuit relay v2
* docs(config): fix links and headers
* feat(config): Internal.Libp2pForceReachability
This switches to config that supports setting and reading
Internal.Libp2pForceReachability OptionalString flag
* use configuration option for static relays
* chore: go-ipfs-config v0.18.0
https://github.com/ipfs/go-ipfs-config/releases/tag/v0.18.0
* feat: circuit v1 migration prompt when Swarm.EnableRelayHop is set (#8559)
* exit when Swarm.EnableRelayHop is set
* docs: Experimental.ShardingEnabled migration
This ensures existing users of global sharding experiment get notified
that the flag no longer works + that autosharding happens automatically.
For people who NEED to keep the old behavior (eg. have no time to
migrate today) there is a note about restoring it with
`UnixFSShardingSizeThreshold`.
* chore: add dag-jose code to the cid command output
* add support for setting automatic unixfs sharding threshold from the config
* test: have tests use low cutoff for sharding to mimic old behavior
* test: change error message to match the current error
* test: Add automatic sharding/unsharding tests (#8547)
* test: refactored naming in the sharding sharness tests to make more sense
* ci: set interop test executor to convenience image for Go1.16 + Node
* ci: use interop master
Co-authored-by: Marcin Rataj <lidel@lidel.org>
Co-authored-by: Marten Seemann <martenseemann@gmail.com>
Co-authored-by: Marcin Rataj <lidel@lidel.org>
Co-authored-by: Gus Eggert <gus@gus.dev>
Co-authored-by: Lucas Molas <schomatis@gmail.com>
* chore: replace go-merkledag walk with go-ipld-prime traversal for dag export
This is "safe" now because we can limit duplicate block loads like
go-merkledag does and won't get trapped taking a long time for complex
DAGs. We can do this while we're using an exhaustive selector (like
ExploreAll here) but will need an alternative strategy when we go for
arbitrary selectors.
* feat: report block count on `dag import`
* fix: clean-up dag import message format
* Only print stats when --stats flag is passed
This applies to both text and json output encoding.
- Stats data is now contained within a Stats datastructure
- Stats are printed after root so that first line of output is the same as previously, even when stats are output using --stats
* fix sharness test
* Add PayloadBytesCount to stats
* Attempt to stabilize flaky tests
* Rename PayloadBytesCount to BlockBytesCount
* Correctly calculate size or imported dag
* Use RawSize of original block for import bytes calc
* test: dag import without --stats
basic regression tests for the default output (text and json)
Co-authored-by: gammazero <gammazero@users.noreply.github.com>
Co-authored-by: Marcin Rataj <lidel@lidel.org>
* feat: switch to using go-ipld-prime for codecs, path resolution, and the `dag put/get` commands
* fix: `dag put/get` not roundtripping due to an extra new line being added (https://github.com/ipfs/go-ipfs/issues/3503)
More detailed information is in the CHANGELOG.md file. Very high level:
* IPLD codecs (and their plugins) must use go-ipld-prime
* Added support for the dag-json codec
* `dag get/put` use IPLD codec names from the multicodec table
* `dag get` defaults to dag-json output instead of json, but may output with other codecs
* Data model pathing can be achieved using the /ipld prefix. For example, you can use `/ipld/QmFoo/Links/0/Hash` to traverse through a DagPB node
* With `dag get/put` the DagPB field names have been changed to match the ones in the protobuf listed in the specification
Co-authored-by: hannahhoward <hannah@hannahhoward.net>
Co-authored-by: Daniel Martí <mvdan@mvdan.cc>
Co-authored-by: acruikshank <acruikshank@example.com>
Co-authored-by: Steven Allen <steven@stebalien.com>
Co-authored-by: Will Scott <will.scott@protocol.ai>
Co-authored-by: Will Scott <will@cypherpunk.email>
Co-authored-by: Rod Vagg <rod@vagg.org>
Co-authored-by: Adin Schmahmann <adin.schmahmann@gmail.com>
Co-authored-by: Eric Myhre <hash@exultant.us>
Unfortunately, API docs are genereted from go-ipfs sources
via https://github.com/ipfs/http-api-docs
so it all gets lost when API docs for new version are re-generated.
This re-applies manual fixes from:
https://docs.ipfs.io/reference/http/api/
so the next time http-api-docs are re-generated, those changes stay.
This still works over "loosely defined" .car files
Please refer to the sharness tests for extra info
We can tighten this up if the sentiment is "Postel was wrong"
