[dpdk] Update to version v22.03

[ljishen]

Describe the pull request

• What does your PR fix?

  Update dpdk to version v22.03

• Which triplets are supported/not supported? Have you updated the CI baseline?

  I've removed the x64-linux=fail from the file to mark that this port is supported by that Linux triplet.

• Does your PR follow the maintainer guide?

  Yes

• If you have added/updated a port: Have you run ./vcpkg x-add-version --all and committed the result?

  Yes

[dg0yt]

I think both the old and the new attempt to provide a CMake config are full of problems. In addition, it is not from upstream, so it is should be in the unofficial namespace.

[Cheney-W]

@ljishen Could you please handle the failure in CI test? Failure log

[ljishen]

@ljishen Could you please handle the failure in CI test? Failure log

The error is expected currently without the #25372. I've figured out how to separate static from dynamic build in dpdk according to @dg0yt's requirement. There is one more issue I need to fix before I can update this PR. Sorry about that.

[ljishen]

@Cheney-W Is there a way to view the CI error log? Because I don't see failures in my local tests. I would like to see what's inside of /mnt/vcpkg-ci/buildtrees/dpdk/config-x64-linux-dbg-out.log reported by the CI.

[Cheney-W]

Could you please open this: https://dev.azure.com/vcpkg/public/_build/results?buildId=74149&view=artifacts&pathAsName=false&type=publishedArtifacts ?
All failure logs are stored there.

[ljishen]

Could you please open this: https://dev.azure.com/vcpkg/public/_build/results?buildId=74149&view=artifacts&pathAsName=false&type=publishedArtifacts ? All failure logs are stored there.

Thanks so much, @Cheney-W. This PR is ready to be reviewed.

[Cheney-W]

Please remove the record about dpdk from ci.baseline.txt, because you have added "supports": "linux | freebsd" into the vcpkg.json file. Thank you!

[ljishen]

@Cheney-W I've removed them. Please take a look. Thanks!

[ljishen]

@Cheney-W Is this PR ready to be merged? Thanks.

[Cheney-W]

@ljishen Could you please handle the suggestions of dg0yt?
Also, I don't think you should let a PR that hasn't been merged block your PR.

[ljishen]

@ljishen Could you please handle the suggestions of dg0yt? Also, I don't think you should let a PR that hasn't been merged block your PR.

Great. Thanks for your support! I'll fix the code.

[github-actions]

This is a new experimental fast check for PR issues. Please let us know if this bot is helpful!

PRs must add only one version and must not modify any published versions

When making any changes to a library, the version or port-version in vcpkg.json or CONTROL must be modified.

error: checked-in files for dpdk have changed but the version was not updated
version: 22.03
old SHA: 4b1f50b8abd4b51bc126eaf5f3686d16caf6b0ff
new SHA: 17a389722a444ba5834d74a86717b6ee963be519
Did you remember to update the version or port version?
Use --overwrite-version to bypass this check
***No files were updated***

[ljishen]

@dan-shaw @dg0yt @Cheney-W Could you help to merge this one, please?

[autoantwort]

This package contains the following absolute path in include/rte_build_config.h:

#define RTE_EAL_PMD_PATH "/home/leander/vcpkg/packages/dpdk_x64-linux/lib/dpdk/pmds-22.1"

and it seems that this path is used in the code. The path is not existing when the package is restored from a binary cache. How should we handle this @ljishen?

[ljishen]

@autoantwort With the linux triplet (e.g., x64-linux), this port will only build static libraries. The absolute path you asked contains dpdk's shared libraries and should be empty based on the triplet's requirement. Can you elaborate on the problem you have with that macro definition in the code? Did you have a customized triplet to use dpdk as a shared library?

[dg0yt]

Since 10 hours, there is now an x64-linux-dynamic community triplet.
And in general, there must be no absolute paths in vcpkg packages. They must be relocatable.

[ljishen]

OK. I'll update the dpdk port to fix the absolute part problem.

[autoantwort]

@ljishen Is there already a PR to fix the absolute path problem? :)