netdev - Re: pegged softirq and NAPI race (?)

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: <09e7e3f8-dee8-71ea-7e57-4d0c92dcf13b@gmail.com>
Date:   Tue, 18 Sep 2018 14:25:38 -0700
From:   Florian Fainelli <f.fainelli@...il.com>
To:     Eric Dumazet <edumazet@...gle.com>, songliubraving@...com
Cc:     Alexei Starovoitov <ast@...com>, netdev <netdev@...r.kernel.org>,
        Jeff Kirsher <jeffrey.t.kirsher@...el.com>,
        Alexander Duyck <alexander.h.duyck@...el.com>,
        michael.chan@...adcom.com, kernel-team <Kernel-team@...com>
Subject: Re: pegged softirq and NAPI race (?)

On 09/18/2018 02:13 PM, Eric Dumazet wrote:
> On Tue, Sep 18, 2018 at 1:37 PM Song Liu <songliubraving@...com> wrote:
>>
> 
>> Looks like a patch like the following fixes the issue for ixgbe. But I
>> cannot explain it yet.
>>
>> Does this ring a bell?
> 
> I dunno, it looks like the NIC is  generating an interrupt while it should not,
> and constantly sets NAPI_STATE_MISSED.
> 
> Or maybe we need to properly check napi_complete_done() return value.
> 
> diff --git a/drivers/net/ethernet/intel/ixgb/ixgb_main.c
> b/drivers/net/ethernet/intel/ixgb/ixgb_main.c
> index d3e72d0f66ef428b08e4bd88508e05b734bc43a4..c4c565c982a98a5891603cedcdcb72dc1c401813
> 100644
> --- a/drivers/net/ethernet/intel/ixgb/ixgb_main.c
> +++ b/drivers/net/ethernet/intel/ixgb/ixgb_main.c
> @@ -1773,8 +1773,8 @@ ixgb_clean(struct napi_struct *napi, int budget)
>         ixgb_clean_rx_irq(adapter, &work_done, budget);
> 
>         /* If budget not fully consumed, exit the polling mode */
> -       if (work_done < budget) {
> -               napi_complete_done(napi, work_done);
> +       if (work_done < budget &&
> +           napi_complete_done(napi, work_done)) {
>                 if (!test_bit(__IXGB_DOWN, &adapter->flags))
>                         ixgb_irq_enable(adapter);
>         }

This would not be the only driver doing this unfortunately... should we
add a __must_check annotation to help catch those (mis)uses? Though that
could cause false positives for drivers using NAPI to clean their TX ring.
-- 
Florian