Convenient no-panic API for just waiting on notifications #1715
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
cbiffle
force-pushed
the
cbiffle/no-panic-notif-recv
branch
from
f81ffbb
to
a96d644
Compare
cbiffle
force-pushed
the
cbiffle/no-panic-notif-recv
branch
from
a96d644
to
eca0f2a
Compare
These comments by 2020-me didn't super make sense to 2024-me, so I have rewritten them to hopefully be a little more clear and reveal some context that was previously implicit.
cbiffle
force-pushed
the
cbiffle/no-panic-notif-recv
branch
from
eca0f2a
to
88b9f93
Compare
hawkw
approved these changes
Apr 3, 2024
Comment on lines
+273
to
+278
| // We're not using the extract_new_generation function here because | ||
| // that has a failure code path for cases where the code is not a | ||
| // dead code. In this case, sys_recv is defined as being _only_ | ||
| // capable of returning a dead code -- otherwise we have a serious | ||
| // kernel bug. So to avoid the introduction of a panic that can't | ||
| // trigger, we will do this manually: |
There was a problem hiding this comment.
This comment is great, thanks for adding it!
This is an alternate API for accessing the common "receive only kernel notifications" pattern. That specific use of the sys_recv syscall is defined as not being able to fail. Different users have handled that in different ways -- some panic, some ignore the result. The latter embeds the assumption of that behavior in many different places in the codebase, a pattern I'm trying to keep an eye on. This change adds a new sys_recv_notification userlib function that bundles up this common pattern. Because of its central placement in userlib, it seems like a more appropriate place to encode the assumption about the syscall not failing, and so I've done that, using unreachable_unchecked to explain it to the compiler. This removes a panic site from every notification-wait in programs that were previously using unwrap or unwrap_lite.
cbiffle
force-pushed
the
cbiffle/no-panic-notif-recv
branch
from
88b9f93
to
f734275
Compare
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.
This is an alternate API for accessing the common "receive only kernel
notifications" pattern. That specific use of the sys_recv syscall is
defined as not being able to fail. Different users have handled that in
different ways -- some panic, some ignore the result. The latter embeds
the assumption of that behavior in many different places in the
codebase, a pattern I'm trying to keep an eye on.
This change adds a new sys_recv_notification userlib function that
bundles up this common pattern. Because of its central placement in
userlib, it seems like a more appropriate place to encode the assumption
about the syscall not failing, and so I've done that, using
unreachable_unchecked to explain it to the compiler.
This removes a panic site from every notification-wait in programs that
were previously using unwrap or unwrap_lite.