Re: [net-next 10/10] net/mlx5e: Add support for PCI relaxed ordering

From: Ding Tianhong <dingtianhong@huawei.com>
To: "Raj, Ashok" <ashok.raj@intel.com>, Bjorn Helgaas <helgaas@kernel.org>
Cc: Aya Levin <ayal@mellanox.com>, Jakub Kicinski <kuba@kernel.org>,
	"Saeed Mahameed" <saeedm@mellanox.com>,
	"mkubecek@suse.cz" <mkubecek@suse.cz>,
	"davem@davemloft.net" <davem@davemloft.net>,
	"netdev@vger.kernel.org" <netdev@vger.kernel.org>,
	Tariq Toukan <tariqt@mellanox.com>, <linux-pci@vger.kernel.org>,
	Alexander Duyck <alexander.h.duyck@linux.intel.com>,
	Casey Leedom <leedom@chelsio.com>
Subject: Re: [net-next 10/10] net/mlx5e: Add support for PCI relaxed ordering
Date: Tue, 30 Jun 2020 15:32:12 +0800	[thread overview]
Message-ID: <edad6af6-c7b9-c6ae-9002-71a92bcd81ee@huawei.com> (raw)
In-Reply-To: <20200629195759.GA255688@otc-nc-03>

在 2020/6/30 3:57, Raj, Ashok 写道:
> Hi Bjorn
> 
> 
> On Mon, Jun 29, 2020 at 02:33:16PM -0500, Bjorn Helgaas wrote:
>> [+cc Ashok, Ding, Casey]
>>
>> On Mon, Jun 29, 2020 at 12:32:44PM +0300, Aya Levin wrote:
>>> I wanted to turn on RO on the ETH driver based on
>>> pcie_relaxed_ordering_enabled().
>>> From my experiments I see that pcie_relaxed_ordering_enabled() return true
>>> on Intel(R) Xeon(R) CPU E5-2650 v3 @ 2.30GHz. This CPU is from Haswell
>>> series which is known to have bug in RO implementation. In this case, I
>>> expected pcie_relaxed_ordering_enabled() to return false, shouldn't it?
>>
>> Is there an erratum for this?  How do we know this device has a bug
>> in relaxed ordering?
> 
> https://software.intel.com/content/www/us/en/develop/download/intel-64-and-ia-32-architectures-optimization-reference-manual.html
> 
> For some reason they weren't documented in the errata, but under
> Optimization manual :-)
> 
> Table 3-7. Intel Processor CPU RP Device IDs for Processors Optimizing PCIe
> Performance
> Processor CPU RP Device IDs
> Intel Xeon processors based on Broadwell microarchitecture 6F01H-6F0EH
> Intel Xeon processors based on Haswell microarchitecture 2F01H-2F0EH
> 
> These are the two that were listed in the manual. drivers/pci/quirks.c also
> has an eloborate list of root ports where relaxed_ordering is disabled. Did
> you check if its not already covered here?
> 
> Send lspci if its not already covered by this table.
> 

Looks like the chip series is not in the errta list, but it is really difficult to distinguish and test.

> 
>>
>>> In addition, we are worried about future bugs in new CPUs which may result
>>> in performance degradation while using RO, as long as the function
>>> pcie_relaxed_ordering_enabled() will return true for these CPUs. 
>>
>> I'm worried about this too.  I do not want to add a Device ID to the
>> quirk_relaxedordering_disable() list for every new Intel CPU.  That's
>> a huge hassle and creates a real problem for old kernels running on
>> those new CPUs, because things might work "most of the time" but not
>> always.
> 
> I'll check when this is fixed, i was told newer ones should work properly.
> But I'll confirm.
> 

Maybe prevent the Relax Ordering for all Intel CPUs is a better soluton, looks like
it will not break anything.

Ding
> 
> .
> 