Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <7bb6e2c9-835c-c8d9-f8a5-baa3d3b03b12@gmail.com>
Date:   Thu, 11 May 2023 22:17:18 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.11.0
Content-Language: en-US
To:     Lukas Wunner <lukas@wunner.de>
Cc:     =?UTF-8?Q?Ilpo_J=c3=a4rvinen?= <ilpo.jarvinen@linux.intel.com>,
        linux-pci@vger.kernel.org, Bjorn Helgaas <helgaas@kernel.org>,
        Rob Herring <robh@kernel.org>,
        Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>,
        Krzysztof Wilczy??ski <kw@linux.com>, nic_swsd@realtek.com,
        "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <20230511131441.45704-1-ilpo.jarvinen@linux.intel.com>
 <20230511131441.45704-15-ilpo.jarvinen@linux.intel.com>
 <98b3b70a-86c0-78c0-b734-0764bb5a21fc@gmail.com>
 <20230511200244.GA31598@wunner.de>
From:   Heiner Kallweit <hkallweit1@gmail.com>
Subject: Re: [PATCH 14/17] r8169: Use pcie_lnkctl_clear_and_set() for changing
 LNKCTL
In-Reply-To: <20230511200244.GA31598@wunner.de>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Precedence: bulk

On 11.05.2023 22:02, Lukas Wunner wrote:
> On Thu, May 11, 2023 at 09:49:52PM +0200, Heiner Kallweit wrote:
>> On 11.05.2023 15:14, Ilpo Järvinen wrote:
>>> Don't assume that only the driver would be accessing LNKCTL. ASPM
>>> policy changes can trigger write to LNKCTL outside of driver's control.
>>>
>>> Use pcie_lnkctl_clear_and_set() which does proper locking to avoid
>>> losing concurrent updates to the register value.
>>
>> Wouldn't it be more appropriate to add proper locking to the
>> underlying pcie_capability_clear_and_set_word()?
> 
> PCI config space accessors such as this one are also used in hot paths
> (e.g. interrupt handlers).  They should be kept lean (and lockless)

I *think* in case the system uses threaded interrupts you may need locking
also in interrupt handlers.

> by default.  We only need locking for specific PCIe Extended Capabilities
> which are concurrently accessed by PCI core code and drivers.
> 
> Thanks,
> 
> Lukas