public inbox for netdev@vger.kernel.org
 help / color / mirror / Atom feed
From: David Howells <dhowells@redhat.com>
To: netdev@vger.kernel.org
Cc: David Howells <dhowells@redhat.com>,
	Marc Dionne <marc.dionne@auristor.com>,
	Jakub Kicinski <kuba@kernel.org>,
	"David S. Miller" <davem@davemloft.net>,
	Eric Dumazet <edumazet@google.com>,
	Paolo Abeni <pabeni@redhat.com>,
	linux-afs@lists.infradead.org, linux-kernel@vger.kernel.org,
	Marc Dionne <marc.c.dionne@gmail.com>,
	Jeffrey Altman <jaltman@auristor.com>,
	Simon Horman <horms@kernel.org>,
	stable@kernel.org
Subject: [PATCH net v5 09/21] rxrpc: Fix to request an ack if window is limited
Date: Wed,  8 Apr 2026 13:12:37 +0100	[thread overview]
Message-ID: <20260408121252.2249051-10-dhowells@redhat.com> (raw)
In-Reply-To: <20260408121252.2249051-1-dhowells@redhat.com>

From: Marc Dionne <marc.c.dionne@gmail.com>

Peers may only send immediate acks for every 2 UDP packets received.
When sending a jumbogram, it is important to check that there is
sufficient window space to send another same sized jumbogram following
the current one, and request an ack if there isn't.  Failure to do so may
cause the call to stall waiting for an ack until the resend timer fires.

Where jumbograms are in use this causes a very significant drop in
performance.

Fixes: fe24a5494390 ("rxrpc: Send jumbo DATA packets")
Signed-off-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Jeffrey Altman <jaltman@auristor.com>
cc: Eric Dumazet <edumazet@google.com>
cc: "David S. Miller" <davem@davemloft.net>
cc: Jakub Kicinski <kuba@kernel.org>
cc: Paolo Abeni <pabeni@redhat.com>
cc: Simon Horman <horms@kernel.org>
cc: linux-afs@lists.infradead.org
cc: netdev@vger.kernel.org
cc: stable@kernel.org
---
 include/trace/events/rxrpc.h | 1 +
 net/rxrpc/ar-internal.h      | 2 +-
 net/rxrpc/output.c           | 2 ++
 net/rxrpc/proc.c             | 5 +++--
 4 files changed, 7 insertions(+), 3 deletions(-)

diff --git a/include/trace/events/rxrpc.h b/include/trace/events/rxrpc.h
index f7f559204b87..578b8038b211 100644
--- a/include/trace/events/rxrpc.h
+++ b/include/trace/events/rxrpc.h
@@ -521,6 +521,7 @@
 #define rxrpc_req_ack_traces \
 	EM(rxrpc_reqack_ack_lost,		"ACK-LOST  ")	\
 	EM(rxrpc_reqack_app_stall,		"APP-STALL ")	\
+	EM(rxrpc_reqack_jumbo_win,		"JUMBO-WIN ")	\
 	EM(rxrpc_reqack_more_rtt,		"MORE-RTT  ")	\
 	EM(rxrpc_reqack_no_srv_last,		"NO-SRVLAST")	\
 	EM(rxrpc_reqack_old_rtt,		"OLD-RTT   ")	\
diff --git a/net/rxrpc/ar-internal.h b/net/rxrpc/ar-internal.h
index 36d6ca0d1089..96ecb83c9071 100644
--- a/net/rxrpc/ar-internal.h
+++ b/net/rxrpc/ar-internal.h
@@ -117,7 +117,7 @@ struct rxrpc_net {
 	atomic_t		stat_tx_jumbo[10];
 	atomic_t		stat_rx_jumbo[10];
 
-	atomic_t		stat_why_req_ack[8];
+	atomic_t		stat_why_req_ack[9];
 
 	atomic_t		stat_io_loop;
 };
diff --git a/net/rxrpc/output.c b/net/rxrpc/output.c
index d70db367e358..870e59bf06af 100644
--- a/net/rxrpc/output.c
+++ b/net/rxrpc/output.c
@@ -479,6 +479,8 @@ static size_t rxrpc_prepare_data_subpacket(struct rxrpc_call *call,
 		why = rxrpc_reqack_old_rtt;
 	else if (!last && !after(READ_ONCE(call->send_top), txb->seq))
 		why = rxrpc_reqack_app_stall;
+	else if (call->tx_winsize <= (2 * req->n) || call->cong_cwnd <= (2 * req->n))
+		why = rxrpc_reqack_jumbo_win;
 	else
 		goto dont_set_request_ack;
 
diff --git a/net/rxrpc/proc.c b/net/rxrpc/proc.c
index 59292f7f9205..7755fca5beb8 100644
--- a/net/rxrpc/proc.c
+++ b/net/rxrpc/proc.c
@@ -518,11 +518,12 @@ int rxrpc_stats_show(struct seq_file *seq, void *v)
 		   atomic_read(&rxnet->stat_rx_acks[RXRPC_ACK_IDLE]),
 		   atomic_read(&rxnet->stat_rx_acks[0]));
 	seq_printf(seq,
-		   "Why-Req-A: acklost=%u mrtt=%u ortt=%u stall=%u\n",
+		   "Why-Req-A: acklost=%u mrtt=%u ortt=%u stall=%u jwin=%u\n",
 		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_ack_lost]),
 		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_more_rtt]),
 		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_old_rtt]),
-		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_app_stall]));
+		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_app_stall]),
+		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_jumbo_win]));
 	seq_printf(seq,
 		   "Why-Req-A: nolast=%u retx=%u slows=%u smtxw=%u\n",
 		   atomic_read(&rxnet->stat_why_req_ack[rxrpc_reqack_no_srv_last]),


  parent reply	other threads:[~2026-04-08 12:14 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-04-08 12:12 [PATCH net v5 00/21] rxrpc: Miscellaneous fixes David Howells
2026-04-08 12:12 ` [PATCH net v5 01/21] rxrpc: Fix key quota calculation for multitoken keys David Howells
2026-04-08 12:12 ` [PATCH net v5 02/21] rxrpc: Fix key parsing memleak David Howells
2026-04-08 12:12 ` [PATCH net v5 03/21] rxrpc: Fix anonymous key handling David Howells
2026-04-08 12:12 ` [PATCH net v5 04/21] rxrpc: Fix call removal to use RCU safe deletion David Howells
2026-04-08 12:12 ` [PATCH net v5 05/21] rxrpc: Fix RxGK token loading to check bounds David Howells
2026-04-08 12:12 ` [PATCH net v5 06/21] rxrpc: Fix use of wrong skb when comparing queued RESP challenge serial David Howells
2026-04-08 12:12 ` [PATCH net v5 07/21] rxrpc: Fix rack timer warning to report unexpected mode David Howells
2026-04-08 12:12 ` [PATCH net v5 08/21] rxrpc: Fix key reference count leak from call->key David Howells
2026-04-08 12:12 ` David Howells [this message]
2026-04-08 12:12 ` [PATCH net v5 10/21] rxrpc: Only put the call ref if one was acquired David Howells
2026-04-08 12:12 ` [PATCH net v5 11/21] rxrpc: reject undecryptable rxkad response tickets David Howells
2026-04-08 12:12 ` [PATCH net v5 12/21] rxrpc: fix RESPONSE authenticator parser OOB read David Howells
2026-04-08 12:12 ` [PATCH net v5 13/21] rxrpc: fix oversized RESPONSE authenticator length check David Howells
2026-04-08 12:12 ` [PATCH net v5 14/21] rxrpc: fix reference count leak in rxrpc_server_keyring() David Howells
2026-04-08 12:12 ` [PATCH net v5 15/21] rxrpc: Fix key/keyring checks in setsockopt(RXRPC_SECURITY_KEY/KEYRING) David Howells
2026-04-08 12:12 ` [PATCH net v5 16/21] rxrpc: Fix missing error checks for rxkad encryption/decryption failure David Howells
2026-04-08 12:12 ` [PATCH net v5 17/21] rxrpc: Fix integer overflow in rxgk_verify_response() David Howells
2026-04-08 12:12 ` [PATCH net v5 18/21] rxrpc: Fix leak of rxgk context " David Howells
2026-04-08 12:12 ` [PATCH net v5 19/21] rxrpc: Fix buffer overread in rxgk_do_verify_authenticator() David Howells
2026-04-08 12:12 ` [PATCH net v5 20/21] rxrpc: only handle RESPONSE during service challenge David Howells
2026-04-08 12:12 ` [PATCH net v5 21/21] rxrpc: proc: size address buffers for %pISpc output David Howells

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20260408121252.2249051-10-dhowells@redhat.com \
    --to=dhowells@redhat.com \
    --cc=davem@davemloft.net \
    --cc=edumazet@google.com \
    --cc=horms@kernel.org \
    --cc=jaltman@auristor.com \
    --cc=kuba@kernel.org \
    --cc=linux-afs@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=marc.c.dionne@gmail.com \
    --cc=marc.dionne@auristor.com \
    --cc=netdev@vger.kernel.org \
    --cc=pabeni@redhat.com \
    --cc=stable@kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox