diff mbox series

[v2,12/12] cachefiles: make on-demand read killable

Message ID 20240515084601.3240503-13-libaokun@huaweicloud.com (mailing list archive)
State New
Headers show
Series cachefiles: some bugfixes and cleanups for ondemand requests | expand

Commit Message

Baokun Li May 15, 2024, 8:46 a.m. UTC 
From: Baokun Li <libaokun1@huawei.com>

Replacing wait_for_completion() with wait_for_completion_killable() in
cachefiles_ondemand_send_req() allows us to kill processes that might
trigger a hunk_task if the daemon is abnormal.

But now only CACHEFILES_OP_READ is killable, because OP_CLOSE and OP_OPEN
is initiated from kworker context and the signal is prohibited in these
kworker.

Note that when the req in xas changes, i.e. xas_load(&xas) != req, it
means that a process will complete the current request soon, so wait
again for the request to be completed.

Suggested-by: Hou Tao <houtao1@huawei.com>
Signed-off-by: Baokun Li <libaokun1@huawei.com>
Reviewed-by: Jia Zhu <zhujia.zj@bytedance.com>
---
 fs/cachefiles/ondemand.c | 21 +++++++++++++++++++--
 1 file changed, 19 insertions(+), 2 deletions(-)
diff mbox series

Patch

diff --git a/fs/cachefiles/ondemand.c b/fs/cachefiles/ondemand.c
index a511d0a89109..bdc2d6dbadce 100644
--- a/fs/cachefiles/ondemand.c
+++ b/fs/cachefiles/ondemand.c
@@ -544,8 +544,25 @@  static int cachefiles_ondemand_send_req(struct cachefiles_object *object,
 		goto out;
 
 	wake_up_all(&cache->daemon_pollwq);
-	wait_for_completion(&req->done);
-	ret = req->error;
+wait:
+	ret = wait_for_completion_killable(&req->done);
+	if (!ret) {
+		ret = req->error;
+	} else {
+		xas_reset(&xas);
+		xas_lock(&xas);
+		if (xas_load(&xas) == req) {
+			xas_store(&xas, NULL);
+			ret = -EINTR;
+		}
+		xas_unlock(&xas);
+
+		/* Someone will complete it soon. */
+		if (ret != -EINTR) {
+			cpu_relax();
+			goto wait;
+		}
+	}
 	cachefiles_req_put(req);
 	return ret;
 out: