Date: Thu, 1 Oct 2020 11:13:45 +0200 From: =?UTF-8?Q?Michal_Van=c4=8do?= <michal@microwave.sk> To: Hans Petter Selasky <hps@selasky.org>, freebsd-net@freebsd.org Subject: Re: mlx5 irq Message-ID: <3c64095f-8a45-0fb4-4835-7486bbd84663@microwave.sk> In-Reply-To: <c2fc9301-f085-189f-ca3a-42d1f97fd870@selasky.org> References: <0aa09fcc-dfcc-005e-8834-2a758ba6a03f@microwave.sk> <94978a05-94c6-cc55-229c-5a3c5352b29a@selasky.org> <c9f8bd7f-6d9d-bb6e-307c-a19c9730b564@microwave.sk> <c2fc9301-f085-189f-ca3a-42d1f97fd870@selasky.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On 01/10/2020 10:52, Hans Petter Selasky wrote: > On 2020-10-01 10:24, Michal Van=C4=8Do wrote: >> But why is the actual number of IRQ lines bigger than number of CPU >> cores? > > There are some dedicated IRQ's used for firmware management. > > Else the driver will use the number of online CPU's by default as the > number of rings, if the hardware supports it. Thanks for clarification. Is there any way to optimize this? In my case I have 2 CPU sockets with 8 cores each (SMT is disabled). NIC is connected via PCIe to the first CPU socket (numa domain 0). In this case, wouldn't it be better if all interrupts were firing only on cores of first socket?
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3c64095f-8a45-0fb4-4835-7486bbd84663>