Received-SPF: pass (google.com: domain of linux-kernel+bounces-176806-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
From: Thomas Gleixner <tglx@linutronix.de>
To: Keith Busch <kbusch@kernel.org>, Ming Lei <ming.lei@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>, Keith Busch <kbusch@meta.com>,
 linux-nvme@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/2] nvme-pci: allow unmanaged interrupts
In-Reply-To: <Zj6-1sXvUNZWO1pB@kbusch-mbp.dhcp.thefacebook.com>
References: <20240510141459.3207725-1-kbusch@meta.com>
 <20240510141459.3207725-2-kbusch@meta.com> <20240510151047.GA10486@lst.de>
 <Zj5JMqWRY187PqnD@kbusch-mbp.dhcp.thefacebook.com>
 <Zj6yvTxIpUnOXl7R@fedora>
 <Zj6-1sXvUNZWO1pB@kbusch-mbp.dhcp.thefacebook.com>
Date: Sun, 12 May 2024 08:35:55 +0200
Message-ID: <87r0e7mt9w.ffs@tglx>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain

On Fri, May 10 2024 at 18:41, Keith Busch wrote:
> On Sat, May 11, 2024 at 07:50:21AM +0800, Ming Lei wrote:
>> Can you explain a bit why it is a no-op? If only isolated CPUs are
>> spread on one queue, there will be no IO originated from these isolated
>> CPUs, that is exactly what the isolation needs.
>
> The "isolcpus=managed_irq," option doesn't limit the dispatching CPUs.
> It only limits where the managed irq will assign the effective_cpus as a
> best effort.
>
> Example, I boot with a system with 4 threads, one nvme device, and
> kernel parameter:
>
>   isolcpus=managed_irq,2-3
>
> Run this:
>
>   for i in $(seq 0 3); do taskset -c $i dd if=/dev/nvme0n1 of=/dev/null bs=4k count=1000 iflag=direct; done
>
> Check /proc/interrupts | grep nvme0:
>
>            CPU0       CPU1       CPU2       CPU3
> ..
>  26:       1000          0          0          0  PCI-MSIX-0000:00:05.0   1-edge      nvme0q1
>  27:          0       1004          0          0  PCI-MSIX-0000:00:05.0   2-edge      nvme0q2
>  28:          0          0       1000          0  PCI-MSIX-0000:00:05.0   3-edge      nvme0q3
>  29:          0          0          0       1043  PCI-MSIX-0000:00:05.0   4-edge      nvme0q4
>
> The isolcpus did nothing becuase the each vector's mask had just one
> cpu; there was no where else that the managed irq could send it. The
> documentation seems to indicate that was by design as a "best effort".

That's expected as you pin the I/O operation on the isolated CPUs which
in turn makes them use the per CPU queue.

The isolated CPUs are only excluded for device management interrupts,
but not for the affinity spread of the queues.

Thanks,

        tglx