netdev - Re: [PATCH] net: hinic: avoid kernel hung in hinic_get

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANn89iLpW4zFf2ABADbMNERPFr=OrAXEMm6ZgCxYA5VpcDpYTw@mail.gmail.com>
Date:   Thu, 30 Jun 2022 12:23:26 +0200
From:   Eric Dumazet <edumazet@...gle.com>
To:     Qiao Ma <mqaio@...ux.alibaba.com>
Cc:     David Miller <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, gustavoars@...nel.org,
        cai.huoqing@...ux.dev, Aviad Krawczyk <aviad.krawczyk@...wei.com>,
        zhaochen6@...wei.com, netdev <netdev@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] net: hinic: avoid kernel hung in hinic_get_stats64()

On Wed, Jun 29, 2022 at 9:28 AM Qiao Ma <mqaio@...ux.alibaba.com> wrote:
>
> When using hinic device as a bond slave device, and reading device stats of
> master bond device, the kernel may hung.
>
> The kernel panic calltrace as follows:
> Kernel panic - not syncing: softlockup: hung tasks
> Call trace:
>   native_queued_spin_lock_slowpath+0x1ec/0x31c
>   dev_get_stats+0x60/0xcc
>   dev_seq_printf_stats+0x40/0x120
>   dev_seq_show+0x1c/0x40
>   seq_read_iter+0x3c8/0x4dc
>   seq_read+0xe0/0x130
>   proc_reg_read+0xa8/0xe0
>   vfs_read+0xb0/0x1d4
>   ksys_read+0x70/0xfc
>   __arm64_sys_read+0x20/0x30
>   el0_svc_common+0x88/0x234
>   do_el0_svc+0x2c/0x90
>   el0_svc+0x1c/0x30
>   el0_sync_handler+0xa8/0xb0
>   el0_sync+0x148/0x180
>
> And the calltrace of task that actually caused kernel hungs as follows:
>   __switch_to+124
>   __schedule+548
>   schedule+72
>   schedule_timeout+348
>   __down_common+188
>   __down+24
>   down+104
>   hinic_get_stats64+44 [hinic]
>   dev_get_stats+92
>   bond_get_stats+172 [bonding]
>   dev_get_stats+92
>   dev_seq_printf_stats+60
>   dev_seq_show+24
>   seq_read_iter+964
>   seq_read+220
>   proc_reg_read+164
>   vfs_read+172
>   ksys_read+108
>   __arm64_sys_read+28
>   el0_svc_common+132
>   do_el0_svc+40
>   el0_svc+24
>   el0_sync_handler+164
>   el0_sync+324
>
> When getting device stats from bond, kernel will call bond_get_stats().
> It first holds the spinlock bond->stats_lock, and then call
> hinic_get_stats64() to collect hinic device's stats.
> However, hinic_get_stats64() calls `down(&nic_dev->mgmt_lock)` to
> protect its critical section, which may schedule current task out.
> And if system is under high pressure, the task cannot be woken up
> immediately, which eventually triggers kernel hung panic.
>
> Fixes: edd384f682cc ("net-next/hinic: Add ethtool and stats")
> Signed-off-by: Qiao Ma <mqaio@...ux.alibaba.com>
> ---
>  drivers/net/ethernet/huawei/hinic/hinic_dev.h  | 1 +
>  drivers/net/ethernet/huawei/hinic/hinic_main.c | 7 +++----
>  2 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/net/ethernet/huawei/hinic/hinic_dev.h b/drivers/net/ethernet/huawei/hinic/hinic_dev.h
> index fb3e89141a0d..1fb343d03fd5 100644
> --- a/drivers/net/ethernet/huawei/hinic/hinic_dev.h
> +++ b/drivers/net/ethernet/huawei/hinic/hinic_dev.h
> @@ -97,6 +97,7 @@ struct hinic_dev {
>
>         struct hinic_txq_stats          tx_stats;
>         struct hinic_rxq_stats          rx_stats;
> +       spinlock_t                      stats_lock;
>
>         u8                              rss_tmpl_idx;
>         u8                              rss_hash_engine;
> diff --git a/drivers/net/ethernet/huawei/hinic/hinic_main.c b/drivers/net/ethernet/huawei/hinic/hinic_main.c
> index 56a89793f47d..32a3d700ad26 100644
> --- a/drivers/net/ethernet/huawei/hinic/hinic_main.c
> +++ b/drivers/net/ethernet/huawei/hinic/hinic_main.c
> @@ -125,11 +125,13 @@ static void update_nic_stats(struct hinic_dev *nic_dev)
>  {
>         int i, num_qps = hinic_hwdev_num_qps(nic_dev->hwdev);
>
> +       spin_lock(&nic_dev->stats_lock);
>         for (i = 0; i < num_qps; i++)
>                 update_rx_stats(nic_dev, &nic_dev->rxqs[i]);
>
>         for (i = 0; i < num_qps; i++)
>                 update_tx_stats(nic_dev, &nic_dev->txqs[i]);
> +       spin_unlock(&nic_dev->stats_lock);
>  }
>
>  /**
> @@ -859,13 +861,9 @@ static void hinic_get_stats64(struct net_device *netdev,
>         nic_rx_stats = &nic_dev->rx_stats;
>         nic_tx_stats = &nic_dev->tx_stats;
>
> -       down(&nic_dev->mgmt_lock);
> -
>         if (nic_dev->flags & HINIC_INTF_UP))
>                 update_nic_stats(nic_dev);
>
> -       up(&nic_dev->mgmt_lock);
> -

Note: The following is racy, because multiple threads can call
hinic_get_stats64() at the same time.
It needs a loop, see include/linux/u64_stats_sync.h for detail.

>         stats->rx_bytes   = nic_rx_stats->bytes;
>         stats->rx_packets = nic_rx_stats->pkts;
>         stats->rx_errors  = nic_rx_stats->errors;

Same remark for nic_tx_stats->{bytes|pkts|errors}

> @@ -1239,6 +1237,7 @@ static int nic_dev_init(struct pci_dev *pdev)
>
>         u64_stats_init(&tx_stats->syncp);
>         u64_stats_init(&rx_stats->syncp);
> +       spin_lock_init(&nic_dev->stats_lock);
>
>         nic_dev->vlan_bitmap = devm_bitmap_zalloc(&pdev->dev, VLAN_N_VID,
>                                                   GFP_KERNEL);
> --
> 1.8.3.1
>