Nested virtualization benchmark#9
Closed
ValentaTomas wants to merge 41 commits into
Closed
Conversation
Marks all unreleased changes as 1.12.0. Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
Adds v1.12 to the release status table and marks v1.10 as unsupported. Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
Intel Sapphire Rapids and ARM Graviton4 are now supported. Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
Update version number / CREDITS Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
`set -x` shows what commands ran, which helps us identify what failed in the release process. Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
I renamed the custom CPU template files uppercase, but I forgot to update their names in release.sh. Fixes: 2862941 ("chore: Use uppercase for custom CPU templates") Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
The current metric dimensions are so that we have 13k+ different sets of dimensions for each metric, which makes it impossible to plot in CloudWatch. This change adds a new set of dimensions with just the test name so that we reduce it to a few hundreds and are able to plot them (CW limit is 500). Signed-off-by: Riccardo Mancini <mancio@amazon.com> Signed-off-by: Patrick Roy <roypat@amazon.co.uk>
The `max_size` field is public, so no need for a getter. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
The size of queue set by the driver must be always less or equal to the queue size in FC. This is checked before device activation. This removes the need for `actual_size` function. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Currently block device has a guest notification logic inside it's request processing loop. This can create a situation when guest can continuously add more requests to the queue, making the whole request processing loop arbitrary long. This is an issue, since it block any other IO from being processed. The solution is to simply notify guest one time, after all current requests are processed. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk> Signed-off-by: Riccardo Mancini <mancio@amazon.com>
VIRTIO spec states: ``` After the device writes a descriptor index into the used ring: If the idx field in the used ring (which determined where that descriptor index was placed) was equal to used_event, the device MUST send a notification. ``` The current implementation does not follow this very precisely. It bumps used ring index when new descriptors are added to the used ring. But the check if the notification is needed is postponed to later processing stage. To be more VIRTIO spec compliant simply move the logic for updating the used ring index into the later processing stage as well, just before the check if the notification should be send. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Add a new changelog entry for the block device fairness fix. Signed-off-by: Riccardo Mancini <mancio@amazon.com>
Update version number / CHANGELOG / CREDITS Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Nested virtualization is now disabled and some cpu features/MSRs now should be marked as host only. These cpu features/MSRs are described the AMD reference manual at "SVM Revision and Feature Identification" section. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
Now this MSR is host only and not available in the guest. Check AMD reference manual at "SVM Revision and Feature Identification" section for more information. Signed-off-by: Egor Lazarchuk <yegorlz@amazon.co.uk>
The link we were using to download the Spectre/Meltdown checker script stopped working and caused our CI to fail consistently. Use this alternative link to download the script. (cherry picked from commit e5b4f26) Signed-off-by: Babis Chalios <bchalios@amazon.es>
Backport commit 8221464 from main. New Amazon Linux host kernels enable mitigation against VMScape that is IBPB before exit to userspace. However, our guest kernels still haven't had the patches yet. Note that Intel Ice Lake is not affected by VMScape as long as BHB clearing sequence is used to mitigate BHI. Signed-off-by: Takahiro Itazuri <itazur@amazon.com>
An upstream patch was backported to ubuntu 24.04 6.8.0-58 kernel that makes the nx hugepages recover thread a child of the firecracker process, thus increasing process count to 7. As we're not really interested in knowing how many threads we have in this test, let's remove the assertion altogether. Signed-off-by: Riccardo Mancini <mancio@amazon.com>
…-dev/firecracker into firecracker-v1.12-direct-mem
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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.
Changes
...
Reason
...
License Acceptance
By submitting this pull request, I confirm that my contribution is made under
the terms of the Apache 2.0 license. For more information on following Developer
Certificate of Origin and signing off your commits, please check
CONTRIBUTING.md.PR Checklist
tools/devtool checkbuild --allto verify that the PR passesbuild checks on all supported architectures.
tools/devtool checkstyleto verify that the PR passes theautomated style checks.
how they are solving the problem in a clear and encompassing way.
in the PR.
CHANGELOG.md.Runbook for Firecracker API changes.
integration tests.
TODO.rust-vmm.