From nobody Sat Oct 4 12:40:58 2025 Received: from m16.mail.163.com (m16.mail.163.com [117.135.210.3]) by smtp.subspace.kernel.org (Postfix) with ESMTP id 69F4038B; Sat, 16 Aug 2025 02:49:15 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=117.135.210.3 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755312558; cv=none; b=RB/Kiospmw//Yu2aadU1Elzc7ogHeeB8Qx4JYGPTj+GIXvDmvqsXpZ+EMPme5aI7A15oCqHsxJrSL5HBvedEdwzaK7jEZKwlpW7mdGsPFXvuDc2mhHPmwqr9JqzQk2XUoF/xeNAgHtP7go09MKsXUe9OYIK8RZUUOrZt8l7WW+Y= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1755312558; c=relaxed/simple; bh=hyFoRJfR+/+dfufmADJJETGSdWS8oikKKEtoGRU+TX4=; h=From:To:Cc:Subject:Date:Message-Id:MIME-Version; b=Jr2FVT6Z3nbcDDFOXfJYiObl00cMIPP+TPYEtY3D+z4VryJmxemESOwpwJPIE2TIV8w3Riz+pwK90Xspg8odBI2GLLT5NYL4ZZRIopyjrofTfiQr1ZQa4mYqBujKuV4W8ih9tyIMRWQVSHs0UiUUmRhTGKBy15yx85iTNt+EFzU= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=163.com; spf=pass smtp.mailfrom=163.com; dkim=pass (1024-bit key) header.d=163.com header.i=@163.com header.b=CgmbxXuR; arc=none smtp.client-ip=117.135.210.3 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=163.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=163.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=163.com header.i=@163.com header.b="CgmbxXuR" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=163.com; s=s110527; h=From:To:Subject:Date:Message-Id:MIME-Version; bh=3A DGn3+EY0cLRTbPVPUMRLpsDFaJI1jPaAcL8kcj18g=; b=CgmbxXuRFzH/7BCps2 APh3WrgqCi6swzU+OcrW1cZchvTNJnjmgwBQVFnPnof1V5k5qGkAsaoLczZAA/YP A0f96GWzTfe+8xFNFOcF0MZh0SIKTJmFmOcoacWdtefvtSltTfdAuZJ+o3zoKHPC AB2snqpXY1qDNqqV+jnnjAJ28= Received: from zhaoxin-MS-7E12.. (unknown []) by gzga-smtp-mtada-g0-3 (Coremail) with SMTP id _____wB3b+eA8Z9oAufhCA--.21748S2; Sat, 16 Aug 2025 10:48:33 +0800 (CST) From: Xin Zhao To: willemdebruijn.kernel@gmail.com, edumazet@google.com, ferenc@fejes.dev Cc: davem@davemloft.net, kuba@kernel.org, pabeni@redhat.com, horms@kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, Xin Zhao Subject: [PATCH net-next v3] net: af_packet: Use hrtimer to do the retire operation Date: Sat, 16 Aug 2025 10:48:31 +0800 Message-Id: <20250816024831.1451167-1-jackzxcui1989@163.com> X-Mailer: git-send-email 2.34.1 Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-CM-TRANSID: _____wB3b+eA8Z9oAufhCA--.21748S2 X-Coremail-Antispam: 1Uf129KBjvJXoW3WrWDWr1rAF1fGr48AFWUJwb_yoWxXry8pa y5WryxGw47Za1agw48Jrs7ZFyYgw1jyryUG393Xwsay3ZxtryrJa1j9r90gFWftFZFkw47 ArsYqF15Cw1ktrDanT9S1TB71UUUUU7qnTZGkaVYY2UrUUUUjbIjqfuFe4nvWSU5nxnvy2 9KBjDUYxBIdaVFxhVjvjDU0xZFpf9x0pMnQUUUUUUU= X-CM-SenderInfo: pmdfy650fxxiqzyzqiywtou0bp/1tbiMAGrCmif8MwNDgAAsY Content-Type: text/plain; charset="utf-8" In a system with high real-time requirements, the timeout mechanism of ordinary timers with jiffies granularity is insufficient to meet the demands for real-time performance. Meanwhile, the optimization of CPU usage with af_packet is quite significant. Use hrtimer instead of timer to help compensate for the shortcomings in real-time performance. In HZ=3D100 or HZ=3D250 system, the update of TP_STATUS_USER is not real-ti= me enough, with fluctuations reaching over 8ms (on a system with HZ=3D250). This is unacceptable in some high real-time systems that require timely processing of network packets. By replacing it with hrtimer, if a timeout of 2ms is set, the update of TP_STATUS_USER can be stabilized to within 3 ms. Signed-off-by: Xin Zhao --- Changes in v3: - return HRTIMER_NORESTART when pkc->delete_blk_timer is true as suggested by Willem de Bruijn; Changes in v2: - Drop the tov_in_msecs field of tpacket_kbdq_core added by the patch as suggested by Willem de Bruijn; - Link to v2: https://lore.kernel.org/all/20250815044141.1374446-1-jackzxcu= i1989@163.com/ Changes in v1: - Do not add another config for the current changes as suggested by Eric Dumazet; - Mention the beneficial cases 'HZ=3D100 or HZ=3D250' in the changelog as suggested by Eric Dumazet; - Add some performance details to the changelog as suggested by Ferenc Fejes; - Delete the 'pkc->tov_in_msecs =3D=3D 0' bounds check which is not necessa= ry as suggested by Willem de Bruijn; - Use hrtimer_set_expires instead of hrtimer_start_range_ns when retire tim= er needs update as suggested by Willem de Bruijn. Start the hrtimer in prb_setup_retire_b= lk_timer; - Just return HRTIMER_RESTART directly as all cases return the same value as suggested by Willem de Bruijn; - Link to v1: https://lore.kernel.org/all/20250813165201.1492779-1-jackzxcu= i1989@163.com/ - Link to v0: https://lore.kernel.org/all/20250806055210.1530081-1-jackzxcu= i1989@163.com/ --- --- net/packet/af_packet.c | 24 ++++++++++++++---------- net/packet/internal.h | 3 +-- 2 files changed, 15 insertions(+), 12 deletions(-) diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index a7017d7f0..763b0c968 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -203,7 +203,7 @@ static void prb_retire_current_block(struct tpacket_kbd= q_core *, static int prb_queue_frozen(struct tpacket_kbdq_core *); static void prb_open_block(struct tpacket_kbdq_core *, struct tpacket_block_desc *); -static void prb_retire_rx_blk_timer_expired(struct timer_list *); +static enum hrtimer_restart prb_retire_rx_blk_timer_expired(struct hrtimer= *); static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *); static void prb_fill_rxhash(struct tpacket_kbdq_core *, struct tpacket3_hd= r *); static void prb_clear_rxhash(struct tpacket_kbdq_core *, @@ -581,7 +581,7 @@ static __be16 vlan_get_protocol_dgram(const struct sk_b= uff *skb) =20 static void prb_del_retire_blk_timer(struct tpacket_kbdq_core *pkc) { - timer_delete_sync(&pkc->retire_blk_timer); + hrtimer_cancel(&pkc->retire_blk_timer); } =20 static void prb_shutdown_retire_blk_timer(struct packet_sock *po, @@ -603,9 +603,10 @@ static void prb_setup_retire_blk_timer(struct packet_s= ock *po) struct tpacket_kbdq_core *pkc; =20 pkc =3D GET_PBDQC_FROM_RB(&po->rx_ring); - timer_setup(&pkc->retire_blk_timer, prb_retire_rx_blk_timer_expired, - 0); - pkc->retire_blk_timer.expires =3D jiffies; + hrtimer_setup(&pkc->retire_blk_timer, prb_retire_rx_blk_timer_expired, + CLOCK_MONOTONIC, HRTIMER_MODE_REL_SOFT); + hrtimer_start(&pkc->retire_blk_timer, ms_to_ktime(pkc->retire_blk_tov), + HRTIMER_MODE_REL_SOFT); } =20 static int prb_calc_retire_blk_tmo(struct packet_sock *po, @@ -676,7 +677,6 @@ static void init_prb_bdqc(struct packet_sock *po, else p1->retire_blk_tov =3D prb_calc_retire_blk_tmo(po, req_u->req3.tp_block_size); - p1->tov_in_jiffies =3D msecs_to_jiffies(p1->retire_blk_tov); p1->blk_sizeof_priv =3D req_u->req3.tp_sizeof_priv; rwlock_init(&p1->blk_fill_in_prog_lock); =20 @@ -691,8 +691,8 @@ static void init_prb_bdqc(struct packet_sock *po, */ static void _prb_refresh_rx_retire_blk_timer(struct tpacket_kbdq_core *pkc) { - mod_timer(&pkc->retire_blk_timer, - jiffies + pkc->tov_in_jiffies); + hrtimer_set_expires(&pkc->retire_blk_timer, + ktime_add(ktime_get(), ms_to_ktime(pkc->retire_blk_tov))); pkc->last_kactive_blk_num =3D pkc->kactive_blk_num; } =20 @@ -719,8 +719,9 @@ static void _prb_refresh_rx_retire_blk_timer(struct tpa= cket_kbdq_core *pkc) * prb_calc_retire_blk_tmo() calculates the tmo. * */ -static void prb_retire_rx_blk_timer_expired(struct timer_list *t) +static enum hrtimer_restart prb_retire_rx_blk_timer_expired(struct hrtimer= *t) { + enum hrtimer_restart ret =3D HRTIMER_RESTART; struct packet_sock *po =3D timer_container_of(po, t, rx_ring.prb_bdqc.retire_blk_timer); struct tpacket_kbdq_core *pkc =3D GET_PBDQC_FROM_RB(&po->rx_ring); @@ -732,8 +733,10 @@ static void prb_retire_rx_blk_timer_expired(struct tim= er_list *t) frozen =3D prb_queue_frozen(pkc); pbd =3D GET_CURR_PBLOCK_DESC_FROM_CORE(pkc); =20 - if (unlikely(pkc->delete_blk_timer)) + if (unlikely(pkc->delete_blk_timer)) { + ret =3D HRTIMER_NORESTART; goto out; + } =20 /* We only need to plug the race when the block is partially filled. * tpacket_rcv: @@ -790,6 +793,7 @@ static void prb_retire_rx_blk_timer_expired(struct time= r_list *t) =20 out: spin_unlock(&po->sk.sk_receive_queue.lock); + return ret; } =20 static void prb_flush_block(struct tpacket_kbdq_core *pkc1, diff --git a/net/packet/internal.h b/net/packet/internal.h index 1e743d031..9812feb3d 100644 --- a/net/packet/internal.h +++ b/net/packet/internal.h @@ -47,10 +47,9 @@ struct tpacket_kbdq_core { =20 unsigned short retire_blk_tov; unsigned short version; - unsigned long tov_in_jiffies; =20 /* timer to retire an outstanding block */ - struct timer_list retire_blk_timer; + struct hrtimer retire_blk_timer; }; =20 struct pgv { --=20 2.34.1