[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20180531045434.GS39853@bhelgaas-glaptop.roam.corp.google.com>
Date: Wed, 30 May 2018 23:54:34 -0500
From: Bjorn Helgaas <helgaas@...nel.org>
To: Lukas Wunner <lukas@...ner.de>
Cc: linux-pci@...r.kernel.org, Oza Pawandeep <poza@...eaurora.org>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 1/2] PCI/AER: Decode Error Source Requester ID
On Wed, May 30, 2018 at 08:32:41PM +0200, Lukas Wunner wrote:
> On Wed, May 30, 2018 at 12:54:15PM -0500, Bjorn Helgaas wrote:
> > void aer_print_port_info(struct pci_dev *dev, struct aer_err_info *info)
> > {
> > - pci_info(dev, "AER: %s%s error received: id=%04x\n",
> > + u8 bus = info->id >> 8;
> > + u8 devfn = info->id & 0xff;
> > +
> > + pci_info(dev, "AER: %s%s error received: %04x:%02x:%02x.%d\n",
> > info->multi_error_valid ? "Multiple " : "",
> > - aer_error_severity_string[info->severity], info->id);
> > + aer_error_severity_string[info->severity],
> > + pci_domain_nr(dev->bus), bus, devfn >> 3, devfn & 0x7);
>
> I think PCI_SLOT(devfn), PCI_FUNC(devfn) is a bit more readable.
I used those originally, but of course those definitions predate PCIe
so they aren't clearly related to a Requester ID.
I searched the PCIe spec for the specifics of the Requester ID
composition. It was surprisingly hard to find a clear statement. The
best I found was PCIe r4.0, sec 6.13, which says
Routing IDs, Requester IDs, and Completer IDs are 16-bit
identifiers traditionally composed of three fields: an 8-bit Bus
Number, a 5-bit Device Number, and a 3-bit Function Number.
Even that isn't specific about where the fields are,
But it's probably not worth obsessing over this and PCI_SLOT() and
PCI_FUNC() are definitely more readable, so I changed them.
Powered by blists - more mailing lists