Skip to content

Conversation

@xemul
Copy link
Contributor

@xemul xemul commented Jan 21, 2026

The bitmap is used to allocate ID for newly created scheduling group. This PR replaces it with the logic applied in supergroup ID allocation -- the ID is "allocated" by looking up free slot in sched group pointers on shard-0, then the existing group initialization proceeds.

The main intent is to keep the sched group counting control in one place, specifically after applying per-supergroup maintenance (see #3212).

xemul added 4 commits January 21, 2026 14:36
To facilitate further patching. Also, to be symmetrical with
create_scheduling_supergroup().

Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
The helper initializes specific_vals for newly created sched group. The
very same code will need to be added to another function in the next
patch, so prepare in advance.

Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
This patch adds and calls init_scheduling_group() overload that also
_creates_ the scheduling group. For that it needs to allocate an id for
it, and this allocation happens by scanning the task_queue-s finding for
free slot in it. Other shards then just use the provided id.

The goal here is to remove global bitmap with sched group ids and to
make it symmetrical with how scheduling supergroup ids are allocated.

Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
It's now unused and can be safely removed.

Signed-off-by: Pavel Emelyanov <xemul@scylladb.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant