From: Kuniyuki Iwashima <kuniyu@xxxxxxxxxx>
When commit 390390240145 ("nfsd: don't allow nfsd threads to be
signalled.") is backported to 5.10, it was adjusted considering commit
3feac2b55293 ("sunrpc: exclude from freezer when waiting for requests:").
However, 3feac2b55293 is based on commit f6e70aab9dfe ("SUNRPC: refresh
rq_pages using a bulk page allocator"), which converted page-by-page
allocation to a batch allocation, so schedule_timeout() is placed
un-nested.
As a result, the backported commit 7229200f6866 ("nfsd: don't allow nfsd
threads to be signalled.") placed freezable_schedule_timeout() in the wrong
place.
Now, freezable_schedule_timeout() is called after every successful page
allocation, and we see 30%+ performance regression on 5.10.220 in our
test suite.
Let's move it to the correct place so that freezable_schedule_timeout()
is called only when page allocation fails.
Fixes: 7229200f6866 ("nfsd: don't allow nfsd threads to be signalled.")
Reported-by: Hughdan Liu <hughliu@xxxxxxxxxx>
Signed-off-by: Kuniyuki Iwashima <kuniyu@xxxxxxxxxx>
Reviewed-by: Chuck Lever <chuck.lever@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
net/sunrpc/svc_xprt.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/net/sunrpc/svc_xprt.c
+++ b/net/sunrpc/svc_xprt.c
@@ -679,8 +679,8 @@ static int svc_alloc_arg(struct svc_rqst
set_current_state(TASK_RUNNING);
return -EINTR;
}
+ freezable_schedule_timeout(msecs_to_jiffies(500));
}
- freezable_schedule_timeout(msecs_to_jiffies(500));
rqstp->rq_pages[i] = p;
}
rqstp->rq_page_end = &rqstp->rq_pages[i];
Patches currently in stable-queue which might be from kuniyu@xxxxxxxxxx are
queue-5.10/nfsd-don-t-call-freezable_schedule_timeout-after-each-successful-page-allocation-in-svc_alloc_arg.patch
queue-5.10/kcm-serialise-kcm_sendmsg-for-the-same-socket.patch
