[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <bcb1b9e9-150d-38d1-d61e-2e0bf440d310@gmail.com>
Date: Thu, 18 Oct 2018 18:51:56 -0700
From: Eric Dumazet <eric.dumazet@...il.com>
To: Francois Romieu <romieu@...zoreil.com>
Cc: Heiner Kallweit <hkallweit1@...il.com>,
David Miller <davem@...emloft.net>,
Realtek linux nic maintainers <nic_swsd@...ltek.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: [PATCH v2 net] r8169: fix NAPI handling under high load
On 10/18/2018 03:59 PM, Francois Romieu wrote:
> Eric Dumazet <eric.dumazet@...il.com> :
> [...]
>> One has to wonder why rtl8169_poll(), which might be called in a loop under DOS,
>> has to call rtl_ack_events() ?
>
> So as to cover a wider temporal range before any event can trigger an
> extra irq. I was more worried about irq cost than about IO cost (and
> I still am).
>
Normally the IRQ would not be enabled under DOS.
Only when a poll would receive less packets than the budget
the following would normally allow the device to send another IRQ
if (work_done < budget) {
napi_complete_done(napi, work_done);
rtl_irq_enable(tp, enable_mask);
mmiowb();
}
Powered by blists - more mailing lists