chore: update to vm-memory 0.13.1 #4175
Merged
Conversation
This file contains 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
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #4175 +/- ##
==========================================
- Coverage 83.00% 82.97% -0.04%
==========================================
Files 222 221 -1
Lines 28642 28505 -137
==========================================
- Hits 23775 23652 -123
+ Misses 4867 4853 -14
Flags with carried forward coverage won't be shown. Click here to find out more.
☔ View full report in Codecov by Sentry. |
roypat
force-pushed
the
update-vm-memory-0-13-1
branch
22 times, most recently
from
9a521d3 to
3a35f20
Compare
roypat
added
the
Status: Awaiting review
pb8o
reviewed
Oct 20, 2023
pb8o
reviewed
Oct 20, 2023
roypat
force-pushed
the
update-vm-memory-0-13-1
branch
from
6f37e97 to
f58cd48
Compare
bchalios
requested changes
Oct 20, 2023
roypat
force-pushed
the
update-vm-memory-0-13-1
branch
2 times, most recently
from
6eeab86 to
d01913c
Compare
As always, the linux-loader update has to piggy-back off of the
vm-memory update. This allows us to remove the {Read,Write}Volatile
traits, instead using the ones from upstream.
Also eliminates some more usages of now-deprecated vm-memory functions
that I missed when I first wrote the volatile-io patch for firecracker.
Signed-off-by: Patrick Roy <[email protected]>
Make all Queue functions generic over their GuestMemory parameter. This allows kani to mock out rust-vmm's "multi-region support" guest memory implementation with a simpler one that only supports a single region, allowing significant harness speedups. Signed-off-by: Patrick Roy <[email protected]>
GuestMemoryMmap supports arbitrary guest memory regions, and thus needs logic that deals with memory accesses spanning multiple regions. Our kani harnesses only use a single region, and the logic for supporting multiple regions causes kani to severely degrade in performance. Signed-off-by: Patrick Roy <[email protected]>
Calls to set_avail_event tend to cause memory to grow unboundedly during verification. The function writes to the `avail_event` of the virtio queue, which is not read from by the device. It is only intended to be used by guest. Therefore, it does not affect any device functionality (e.g. its only call site, try_enable_notification, will behave independently of what value was written here). Thus we can stub it out with a no-op. Note that we have a separate harness for set_avail_event, to ensure the function itself is sound. Signed-off-by: Patrick Roy <[email protected]>
m6a.metal has significantly more memory, and is also slightly faster. Signed-off-by: Patrick Roy <[email protected]>
To get the runtime of the harnesses back below 15 minutes. This reduction does not loose any generality, as the only real distinction of cases here is "0, 1 or 2 buffers". Anything above that is just nice-to-have. Signed-off-by: Patrick Roy <[email protected]>
The --enable-stubbing option is deprecated in favor of -Z stubbing. Signed-off-by: Patrick Roy <[email protected]>
Using dirty bitmap tracking will cause verification performance to drastically decrease, as it involves non-trivial arithmetic on arbitrary values. Signed-off-by: Patrick Roy <[email protected]>
Kani, running using nightly rustc, was warning about this comment. Fix it to resolve the warning during verification, to get slightly cleaner logs. Signed-off-by: Patrick Roy <[email protected]>
roypat
force-pushed
the
update-vm-memory-0-13-1
branch
from
0aad9f6 to
864039b
Compare
bchalios
approved these changes
Oct 20, 2023
pb8o
approved these changes
Oct 20, 2023
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
As always, the linux-loader update has to biggy-back off of the vm-memory update. This allows us to remove the {Read,Write}Volatile traits, instead using the ones from upstream.
Also eliminates some more usages of now-deprecated vm-memory functions that I missed when I first wrote the volatile-io patch for firecracker.
Lastly, this PR also includes a lot of kani performance tweaking that became necessary with the new vm-memory update for some 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
CHANGELOG.md.TODOs link to an issue.rust-vmm.