[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <634db3fa-286d-1136-b2af-1bf45501275f@deltatee.com>
Date: Tue, 9 Apr 2019 17:32:13 -0600
From: Logan Gunthorpe <logang@...tatee.com>
To: Bjorn Helgaas <helgaas@...nel.org>,
Wesley Sheng <wesley.sheng@...rochip.com>
Cc: kurt.schwemmer@...rosemi.com, linux-pci@...r.kernel.org,
linux-kernel@...r.kernel.org, wesleyshenggit@...a.com,
kelvin.cao@...rochip.com
Subject: Re: [PATCH 1/2] switchtec: Fix false maximum supported PCIe function
number issue
On 2019-04-09 4:36 p.m., Bjorn Helgaas wrote:
> Hi Wesley,
>
> On Mon, Apr 08, 2019 at 10:34:47PM +0800, Wesley Sheng wrote:
>> The hardware supports up to 255 PFFs and the driver only supports 48, so
>> this patch updates the driver to support them all.
>> To be backward compatible, a new ioctl and corresponding data
>> structure are created, while keep the deprecated one.
>
> The above is either one paragraph that needs to be rewrapped, or two
> paragraphs that need a blank line between.
Wesley can address this.
> What's a PFF?
PFF is really a concept internal to the Switchtec device. It stands for
PCIe Function Framework. Essentially, there is a bank of registers for
every PCIe Function (aka endpoint) in the switch. When I originally
wrote the driver, I assumed incorrectly there would only ever be one PFF
per port and the maximum number of ports for Switchtec parts is 48. In
fact, the hardware supports up to 255 and there are typically two PFFs
per upstream port (one for the port itself and one for the management
endpoint).
Logan
Powered by blists - more mailing lists