On 7/29/24 23:38, Olivier Langlois wrote:
this info may be useful when attempting to debug a problem
involving a ring using the feature.
While on the topic of busy polling, there is a function
io_napi_adjust_timeout(), it ensures that we don't busy poll for longer
than the passed wait timeout.
Do you use it? I have some doubts in regards to its usefulness, and
would prefer to try get rid of it if there are no users since it's a
hustle.
CqOverflowList:
NAPI: enabled
napi_busy_poll_to: 1
napi_prefer_busy_poll: true
Signed-off-by: Olivier Langlois <olivier@xxxxxxxxxxxxxx>
---
io_uring/fdinfo.c | 13 ++++++++++++-
1 file changed, 12 insertions(+), 1 deletion(-)
diff --git a/io_uring/fdinfo.c b/io_uring/fdinfo.c
index b1e0e0d85349..3ba42e136a40 100644
--- a/io_uring/fdinfo.c
+++ b/io_uring/fdinfo.c
@@ -221,7 +221,18 @@ __cold void io_uring_show_fdinfo(struct seq_file *m, struct file *file)
cqe->user_data, cqe->res, cqe->flags);
}
-
+#ifdef CONFIG_NET_RX_BUSY_POLL
+ if (ctx->napi_enabled) {
+ seq_puts(m, "NAPI:\tenabled\n");
+ seq_printf(m, "napi_busy_poll_dt:\t%llu\n", ctx->napi_busy_poll_dt);
+ if (ctx->napi_prefer_busy_poll)
+ seq_puts(m, "napi_prefer_busy_poll:\ttrue\n");
+ else
+ seq_puts(m, "napi_prefer_busy_poll:\tfalse\n");
+ } else {
+ seq_puts(m, "NAPI:\tdisabled\n");
+ }
+#endif
spin_unlock(&ctx->completion_lock);
}
#endif
--
Pavel Begunkov
