Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8;
        delsp=yes;
        format=flowed
Mime-Version: 1.0 (Mac OS X Mail 12.4 \(3445.104.8\))
Subject: Re: [PATCH] nvme-pci: Use non-operational power state instead of D3
 on Suspend-to-Idle
From:   Kai-Heng Feng <kai.heng.feng@canonical.com>
In-Reply-To: <20190508191624.GA8365@localhost.localdomain>
Date:   Thu, 9 May 2019 03:30:18 +0800
Cc:     Keith Busch <keith.busch@intel.com>, axboe@fb.com, hch@lst.de,
        sagi@grimberg.me, linux-nvme@lists.infradead.org,
        linux-kernel@vger.kernel.org,
        Mario Limonciello <mario.limonciello@dell.com>
Content-Transfer-Encoding: 8bit
Message-Id: <3CDA9F13-B17C-456F-8CE1-3A63C6E0DC8F@canonical.com>
References: <20190508185955.11406-1-kai.heng.feng@canonical.com>
 <20190508191624.GA8365@localhost.localdomain>
To:     Keith Busch <kbusch@kernel.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

at 03:16, Keith Busch <kbusch@kernel.org> wrote:

> On Thu, May 09, 2019 at 02:59:55AM +0800, Kai-Heng Feng wrote:
>> +static int nvme_do_resume_from_idle(struct pci_dev *pdev)
>> +{
>> +	struct nvme_dev *ndev = pci_get_drvdata(pdev);
>> +	int result;
>> +
>> +	pdev->dev_flags &= ~PCI_DEV_FLAGS_NO_D3;
>> +	ndev->ctrl.suspend_to_idle = false;
>> +
>> +	result = pci_alloc_irq_vectors(pdev, 1, 1, PCI_IRQ_ALL_TYPES);
>> +	if (result < 0)
>> +		goto out;
>> +
>> +	result = nvme_pci_configure_admin_queue(ndev);
>> +	if (result)
>> +		goto out;
>> +
>> +	result = nvme_alloc_admin_tags(ndev);
>> +	if (result)
>> +		goto out;
>> +
>> +	ndev->ctrl.max_hw_sectors = NVME_MAX_KB_SZ << 1;
>> +	ndev->ctrl.max_segments = NVME_MAX_SEGS;
>> +	mutex_unlock(&ndev->shutdown_lock);
>
> This lock was never locked.

Yea, will fix this.

>
> But I think these special suspend/resume routines are too similar to the
> existing ones, they should just incorporate this feature if we need to
> do this.

That’s my original direction, but this requires a new bool to  
nvme_dev_disable(), so it becomes
static void nvme_dev_disable(struct nvme_dev *dev, bool shutdown, bool  
suspend_to_idle)

And it starts to get messy.

Kai-Heng