From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Yunsheng Lin <linyunsheng@huawei.com>,
"David S. Miller" <davem@davemloft.net>
Subject: [PATCH 4.19 13/37] net: sch_generic: aviod concurrent reset and enqueue op for lockless qdisc
Date: Fri, 25 Sep 2020 14:48:41 +0200 [thread overview]
Message-ID: <20200925124722.922202675@linuxfoundation.org> (raw)
In-Reply-To: <20200925124720.972208530@linuxfoundation.org>
From: Yunsheng Lin <linyunsheng@huawei.com>
[ Upstream commit 2fb541c862c987d02dfdf28f1545016deecfa0d5 ]
Currently there is concurrent reset and enqueue operation for the
same lockless qdisc when there is no lock to synchronize the
q->enqueue() in __dev_xmit_skb() with the qdisc reset operation in
qdisc_deactivate() called by dev_deactivate_queue(), which may cause
out-of-bounds access for priv->ring[] in hns3 driver if user has
requested a smaller queue num when __dev_xmit_skb() still enqueue a
skb with a larger queue_mapping after the corresponding qdisc is
reset, and call hns3_nic_net_xmit() with that skb later.
Reused the existing synchronize_net() in dev_deactivate_many() to
make sure skb with larger queue_mapping enqueued to old qdisc(which
is saved in dev_queue->qdisc_sleeping) will always be reset when
dev_reset_queue() is called.
Fixes: 6b3ba9146fe6 ("net: sched: allow qdiscs to handle locking")
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
net/sched/sch_generic.c | 49 ++++++++++++++++++++++++++++++++----------------
1 file changed, 33 insertions(+), 16 deletions(-)
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -1115,27 +1115,36 @@ static void dev_deactivate_queue(struct
struct netdev_queue *dev_queue,
void *_qdisc_default)
{
- struct Qdisc *qdisc_default = _qdisc_default;
- struct Qdisc *qdisc;
+ struct Qdisc *qdisc = rtnl_dereference(dev_queue->qdisc);
- qdisc = rtnl_dereference(dev_queue->qdisc);
if (qdisc) {
- bool nolock = qdisc->flags & TCQ_F_NOLOCK;
-
- if (nolock)
- spin_lock_bh(&qdisc->seqlock);
- spin_lock_bh(qdisc_lock(qdisc));
-
if (!(qdisc->flags & TCQ_F_BUILTIN))
set_bit(__QDISC_STATE_DEACTIVATED, &qdisc->state);
+ }
+}
- rcu_assign_pointer(dev_queue->qdisc, qdisc_default);
- qdisc_reset(qdisc);
+static void dev_reset_queue(struct net_device *dev,
+ struct netdev_queue *dev_queue,
+ void *_unused)
+{
+ struct Qdisc *qdisc;
+ bool nolock;
- spin_unlock_bh(qdisc_lock(qdisc));
- if (nolock)
- spin_unlock_bh(&qdisc->seqlock);
- }
+ qdisc = dev_queue->qdisc_sleeping;
+ if (!qdisc)
+ return;
+
+ nolock = qdisc->flags & TCQ_F_NOLOCK;
+
+ if (nolock)
+ spin_lock_bh(&qdisc->seqlock);
+ spin_lock_bh(qdisc_lock(qdisc));
+
+ qdisc_reset(qdisc);
+
+ spin_unlock_bh(qdisc_lock(qdisc));
+ if (nolock)
+ spin_unlock_bh(&qdisc->seqlock);
}
static bool some_qdisc_is_busy(struct net_device *dev)
@@ -1196,12 +1205,20 @@ void dev_deactivate_many(struct list_hea
dev_watchdog_down(dev);
}
- /* Wait for outstanding qdisc-less dev_queue_xmit calls.
+ /* Wait for outstanding qdisc-less dev_queue_xmit calls or
+ * outstanding qdisc enqueuing calls.
* This is avoided if all devices are in dismantle phase :
* Caller will call synchronize_net() for us
*/
synchronize_net();
+ list_for_each_entry(dev, head, close_list) {
+ netdev_for_each_tx_queue(dev, dev_reset_queue, NULL);
+
+ if (dev_ingress_queue(dev))
+ dev_reset_queue(dev, dev_ingress_queue(dev), NULL);
+ }
+
/* Wait for outstanding qdisc_run calls. */
list_for_each_entry(dev, head, close_list) {
while (some_qdisc_is_busy(dev))
next prev parent reply other threads:[~2020-09-25 12:55 UTC|newest]
Thread overview: 46+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-09-25 12:48 [PATCH 4.19 00/37] 4.19.148-rc1 review Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 01/37] af_key: pfkey_dump needs parameter validation Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 02/37] KVM: fix memory leak in kvm_io_bus_unregister_dev() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 03/37] kprobes: fix kill kprobe which has been marked as gone Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 04/37] mm/thp: fix __split_huge_pmd_locked() for migration PMD Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 05/37] cxgb4: Fix offset when clearing filter byte counters Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 06/37] geneve: add transport ports in route lookup for geneve Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 07/37] hdlc_ppp: add range checks in ppp_cp_parse_cr() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 08/37] ip: fix tos reflection in ack and reset packets Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 09/37] ipv6: avoid lockdep issue in fib6_del() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 10/37] net: DCB: Validate DCB_ATTR_DCB_BUFFER argument Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 11/37] net: dsa: rtl8366: Properly clear member config Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 12/37] net: ipv6: fix kconfig dependency warning for IPV6_SEG6_HMAC Greg Kroah-Hartman
2020-09-25 12:48 ` Greg Kroah-Hartman [this message]
2020-09-25 12:48 ` [PATCH 4.19 14/37] nfp: use correct define to return NONE fec Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 15/37] tipc: Fix memory leak in tipc_group_create_member() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 16/37] tipc: fix shutdown() of connection oriented socket Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 17/37] tipc: use skb_unshare() instead in tipc_buf_append() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 18/37] bnxt_en: return proper error codes in bnxt_show_temp Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 19/37] bnxt_en: Protect bnxt_set_eee() and bnxt_set_pauseparam() with mutex Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 20/37] net: phy: Avoid NPD upon phy_detach() when driver is unbound Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 21/37] net: qrtr: check skb_put_padto() return value Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 22/37] net: add __must_check to skb_put_padto() Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 23/37] ipv4: Update exception handling for multipath routes via same device Greg Kroah-Hartman
2020-09-25 16:51 ` Pavel Machek
2020-09-26 15:46 ` Greg Kroah-Hartman
2020-09-29 12:26 ` Pavel Machek
2020-09-25 12:48 ` [PATCH 4.19 24/37] MAINTAINERS: add CLANG/LLVM BUILD SUPPORT info Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 25/37] kbuild: add OBJSIZE variable for the size tool Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 26/37] Documentation/llvm: add documentation on building w/ Clang/LLVM Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 27/37] Documentation/llvm: fix the name of llvm-size Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 28/37] net: wan: wanxl: use allow to pass CROSS_COMPILE_M68k for rebuilding firmware Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 29/37] net: wan: wanxl: use $(M68KCC) instead of $(M68KAS) " Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 30/37] x86/boot: kbuild: allow readelf executable to be specified Greg Kroah-Hartman
2020-09-25 12:48 ` [PATCH 4.19 31/37] kbuild: remove AS variable Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 32/37] kbuild: replace AS=clang with LLVM_IAS=1 Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 33/37] kbuild: support LLVM=1 to switch the default tools to Clang/LLVM Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 34/37] mm: memcg: fix memcg reclaim soft lockup Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 35/37] tcp_bbr: refactor bbr_target_cwnd() for general inflight provisioning Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 36/37] tcp_bbr: adapt cwnd based on ack aggregation estimation Greg Kroah-Hartman
2020-09-25 12:49 ` [PATCH 4.19 37/37] serial: 8250: Avoid error message on reprobe Greg Kroah-Hartman
2020-09-25 17:39 ` [PATCH 4.19 00/37] 4.19.148-rc1 review Pavel Machek
2020-09-26 15:50 ` Greg Kroah-Hartman
2020-09-25 20:02 ` Shuah Khan
2020-09-26 12:35 ` Naresh Kamboju
2020-09-26 15:42 ` Guenter Roeck
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=20200925124722.922202675@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=davem@davemloft.net \
--cc=linux-kernel@vger.kernel.org \
--cc=linyunsheng@huawei.com \
--cc=stable@vger.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;
as well as URLs for NNTP newsgroup(s).