-
Notifications
You must be signed in to change notification settings - Fork 338
DAOS-18017 build: update mercury to 2.4.1 #16943
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
soumagne
wants to merge
2
commits into
master
Choose a base branch
from
soumagne/mercury_241rc4
base: master
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
+1,609
−538
Open
Changes from all commits
Commits
Show all changes
2 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,71 @@ | ||
| From c0d62973299b4ba98a9aaf9f82ed000a11c0f7b0 Mon Sep 17 00:00:00 2001 | ||
| From: Jerome Soumagne <[email protected]> | ||
| Date: Mon, 24 Nov 2025 16:44:14 -0600 | ||
| Subject: [PATCH] NA OFI/UCX: print version infos | ||
|
|
||
| NA OFI: fix log warning | ||
| --- | ||
| src/na/na_ofi.c | 12 +++++++----- | ||
| src/na/na_ucx.c | 10 ++++++++++ | ||
| 2 files changed, 17 insertions(+), 5 deletions(-) | ||
|
|
||
| diff --git a/src/na/na_ofi.c b/src/na/na_ofi.c | ||
| index c7c3e0b3..682efe65 100644 | ||
| --- a/src/na/na_ofi.c | ||
| +++ b/src/na/na_ofi.c | ||
| @@ -8048,8 +8048,10 @@ na_ofi_check_protocol(const char *protocol_name) | ||
| uint32_t runtime_version = fi_version(); | ||
| na_return_t na_ret; | ||
|
|
||
| - NA_LOG_SUBSYS_DEBUG(cls, "Querying info on libfabric v%d.%d", | ||
| - FI_MAJOR(runtime_version), FI_MINOR(runtime_version)); | ||
| + NA_LOG_SUBSYS_INFO(cls, | ||
| + "Querying info on libfabric (runtime v%d.%d, API v%d.%d)", | ||
| + FI_MAJOR(runtime_version), FI_MINOR(runtime_version), | ||
| + FI_MAJOR(FI_COMPILE_VERSION), FI_MINOR(FI_COMPILE_VERSION)); | ||
| NA_CHECK_SUBSYS_ERROR(cls, FI_VERSION_LT(runtime_version, NA_OFI_VERSION), | ||
| out, accept, false, | ||
| "runtime libfabric version (v%d.%d) is lower than required version " | ||
| @@ -9105,7 +9107,7 @@ na_ofi_mem_handle_create(na_class_t NA_UNUSED *na_class, void *buf, | ||
|
|
||
| NA_LOG_SUBSYS_DEBUG(mem, | ||
| "Created mem handle %p (iov_base=%p, iov_len=%zu, iovcnt=1, " | ||
| - "flags=0x%lx, len=%zu)", | ||
| + "flags=0x%" PRIx8 ", len=%" PRIu64 ")", | ||
| (void *) na_ofi_mem_handle, na_ofi_mem_handle->desc.iov.s[0].iov_base, | ||
| na_ofi_mem_handle->desc.iov.s[0].iov_len, | ||
| na_ofi_mem_handle->desc.info.flags, na_ofi_mem_handle->desc.info.len); | ||
| @@ -9444,8 +9446,8 @@ na_ofi_mem_handle_deserialize(na_class_t NA_UNUSED *na_class, | ||
| na_ofi_mem_handle->desc.info.iovcnt); | ||
|
|
||
| NA_LOG_SUBSYS_DEBUG(mem, | ||
| - "Deserialized mem handle %p (iov_base=%p, iov_len=%zu, iovcnt=%zu, " | ||
| - "flags=0x%lx, len=%zu)", | ||
| + "Deserialized mem handle %p (iov_base=%p, iov_len=%zu, iovcnt=%" PRIu64 | ||
| + ", flags=0x%" PRIx8 ", len=%" PRIu64 ")", | ||
| (void *) na_ofi_mem_handle, na_ofi_mem_handle->desc.iov.s[0].iov_base, | ||
| na_ofi_mem_handle->desc.iov.s[0].iov_len, | ||
| na_ofi_mem_handle->desc.info.iovcnt, na_ofi_mem_handle->desc.info.flags, | ||
| diff --git a/src/na/na_ucx.c b/src/na/na_ucx.c | ||
| index 96501b27..7c0ac8d4 100644 | ||
| --- a/src/na/na_ucx.c | ||
| +++ b/src/na/na_ucx.c | ||
| @@ -3433,8 +3433,18 @@ na_ucx_check_protocol(const char *protocol_name) | ||
| .field_mask = UCP_PARAM_FIELD_FEATURES, .features = NA_UCX_FEATURES}; | ||
| ucp_context_h context = NULL; | ||
| ucs_status_t status; | ||
| + unsigned int runtime_major_version, runtime_minor_version, | ||
| + runtime_patch_version; | ||
| bool accept = false; | ||
|
|
||
| + ucp_get_version( | ||
| + &runtime_major_version, &runtime_minor_version, &runtime_patch_version); | ||
| + | ||
| + NA_LOG_SUBSYS_INFO(cls, | ||
| + "Querying info on UCX (runtime v%u.%u.%u, API v%u.%u)", | ||
| + runtime_major_version, runtime_minor_version, runtime_patch_version, | ||
| + UCP_API_MAJOR, UCP_API_MINOR); | ||
| + | ||
| status = ucp_config_read(NULL, NULL, &config); | ||
| NA_CHECK_SUBSYS_ERROR_NORET(cls, status != UCS_OK, done, | ||
| "ucp_config_read() failed (%s)", ucs_status_string(status)); |
This file was deleted.
Oops, something went wrong.
This file was deleted.
Oops, something went wrong.
Oops, something went wrong.
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@phender this is the best and easiest solution I think for now (using the PR daos-stack/pipeline-lib#480), this way there is only minimal changes required to the Jenkinsfiles.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Reminder to ourselves that we need to land the pipeline-lib change first
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
to be frank, I'm not completely sure of the order that should go in, if I look at the comment it says:
so in theory I think we want to land that PR first so that it uses the branch when it lands, then land the pipeline-lib PR without deleting the branch, and finally merge a new PR that comments out that line again so it uses the default branch. That's how I see it but maybe I'm missing something?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Usually we land the pipeline-lib PR (and delete the branch). Then update this daos PR back to the default / uncommented pipeline-lib. Then land this PR if testing is okay
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
well imo if you do that then you are going to break all PRs / CIs until this PR has landed :)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ok, I misunderstood what this PR was doing. I did some testing locally and for example trying to install this new mercury with 2.6.4 works as expected
So, PRs that do NOT have this change will be okay because their version of
daosstill depends onmercuryand onlibfabric.Then, when pipeline-lib says to install
mercury-libfabricit doesn't actually matter for those PRs becausemercury-libfabricreally just installsmercuryandlibfabric.There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So I see two options. Either what you mentioned about leaving the pipeline-lib reference in the landing of this PR, OR something like this:
daos -> libfabricdependency for now for compatibilitydaos -> libfabricdependencyIMO I prefer this approach so we don't have to leave the pipeline-lib reference in place
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
you are allowed to install mercury without mercury-libfabric or mercury-ucx but all tests will fail because the providers / plugins won't be found. What this PR does is separate the libfabric mercury plugin that allows mercury to interface with libfabric. Without it, mercury cannot use libfabric.
No, this does not help in any way and no with (1) and (2) it won't work because you do need mercury-libfabric. This is not a simple alias of some sort, to expand on what I just mentioned above, it installs a separate library that is similar to mercury-ucx (you can look at that as an example). This library is then dynamically loaded by mercury.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Then I don't know how to land any of this without breaking CI
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Well that's what I've just been trying to explain :) I'm glad we're reaching the same conclusion. The real issue is that rpms are globally visible once merged, so because of that, any change of that sort will face the same issues.