Shard pure-serve across cores

A burst now looks at the store once (pure_store_pin) instead of once per
request; another shard's commit ends before its answer is sent, so a
request made after that answer still reads it. With this alone one thread
serves 125k small keep-alive reads per second, as many as axum on three
cores, with a shorter p99.

--threads N runs a shard per thread: its own SO_REUSEPORT listener, epoll
set, runtime domain, app and store handle. Writer sessions of the tree
store now take turns inside a process with a mutex (the file lock belongs
to the process and does not keep threads apart), and revisions settle
concurrent edits of a key. A module with states runs on one thread. Three
shards serve 151k reads per second; they still commit one by one, so
durable writes and 1 MiB reads are better on one thread for now, which
stays the default. Note 075 records the numbers.

1 files changed, 4 insertions, 0 deletions

diff --git a/prototypes/connections/state.h b/prototypes/connections/state.h
index 25c0f61..60d333c 100644
--- a/prototypes/connections/state.h
+++ b/prototypes/connections/state.h
@@ -30,6 +30,10 @@ PureResult pure_store_open(const char *path, PureStore **out);
 void pure_store_close(PureStore *);
 PureResult pure_store_draft(PureStore *, const char *key, PureDraft **out);
 PureResult pure_store_state(PureStore *, const char *name, PureDraft **out);
+/* Pinned, drafts open at the version found when pinning, without checking
+ * the file each time (a host pins for one burst of requests); commits made
+ * through the handle still move it on. */
+PureResult pure_store_pin(PureStore *, bool pinned);
 /* Memory only. Fails without change on a bad range or when memory runs out. */
 PureResult pure_draft_splice(PureDraft *, size_t offset, size_t remove, PureSpan insert);
 size_t pure_draft_size(const PureDraft *);