On Wed, Jul 16, 2025 at 10:02 AM Mario Limonciello <superm1@xxxxxxxxxx> wrote:
>
> On 7/16/25 11:32 AM, Matthew Wood wrote:
> > Add a single sysfs read-only interface for reading PCIe device serial
> > numbers from userspace in a programmatic way. This device attribute
> > uses the same hexadecimal 1-byte dashed formatting as lspci serial number
> > capability output. If a device doesn't support the serial number
> > capability, the device_serial_number sysfs attribute will not be visible.
> >
> > Signed-off-by: Matthew Wood <thepacketgeek@xxxxxxxxx>
> > ---
> > Documentation/ABI/testing/sysfs-bus-pci | 7 +++++++
> > drivers/pci/pci-sysfs.c | 27 ++++++++++++++++++++++---
> > 2 files changed, 31 insertions(+), 3 deletions(-)
> >
> > diff --git a/Documentation/ABI/testing/sysfs-bus-pci b/Documentation/ABI/testing/sysfs-bus-pci
> > index 69f952fffec7..f7e84b3a4204 100644
> > --- a/Documentation/ABI/testing/sysfs-bus-pci
> > +++ b/Documentation/ABI/testing/sysfs-bus-pci
> > @@ -612,3 +612,10 @@ Description:
> >
> > # ls doe_features
> > 0001:01 0001:02 doe_discovery
> > +
> > +What: /sys/bus/pci/devices/.../device_serial_number
> > +Date: July 2025
> > +Contact: Matthew Wood <thepacketgeek@xxxxxxxxx>
> > +Description:
> > + This is visible only for PCIe devices that support the serial
> > + number extended capability. The file is read only.
> > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
> > index 268c69daa4d5..b7b52dea6e31 100644
> > --- a/drivers/pci/pci-sysfs.c
> > +++ b/drivers/pci/pci-sysfs.c
> > @@ -239,6 +239,22 @@ static ssize_t current_link_width_show(struct device *dev,
> > }
> > static DEVICE_ATTR_RO(current_link_width);
> >
> > +static ssize_t device_serial_number_show(struct device *dev,
> > + struct device_attribute *attr, char *buf)
> > +{
> > + struct pci_dev *pci_dev = to_pci_dev(dev);
> > + u64 dsn;
> > +
> > + dsn = pci_get_dsn(pci_dev);
> > + if (!dsn)
> > + return -EIO;
> > +
> > + return sysfs_emit(buf, "%02llx-%02llx-%02llx-%02llx-%02llx-%02llx-%02llx-%02llx\n",
> > + dsn >> 56, (dsn >> 48) & 0xff, (dsn >> 40) & 0xff, (dsn >> 32) & 0xff,
> > + (dsn >> 24) & 0xff, (dsn >> 16) & 0xff, (dsn >> 8) & 0xff, dsn & 0xff);
> > +}
> > +static DEVICE_ATTR_RO(device_serial_number);
>
> The serial number /could/ be considered sensitive information. I think
> it's better to use DEVICE_ATTR_ADMIN_RO.
That makes sense, thanks for the suggestion.
>
> Also, as this is a "device" attribute is it really necessary to encode
> the extra word and "number"?
Another good point. I will remove the device_ prefix however I think
"serial_number" is helpful to keep as "serial" alone is too easy to
conflate with any of the other usages of the serial word.
>
> > +
> > static ssize_t secondary_bus_number_show(struct device *dev,
> > struct device_attribute *attr,
> > char *buf)
> > @@ -660,6 +676,7 @@ static struct attribute *pcie_dev_attrs[] = {
> > &dev_attr_current_link_width.attr,
> > &dev_attr_max_link_width.attr,
> > &dev_attr_max_link_speed.attr,
> > + &dev_attr_device_serial_number.attr,
> > NULL,
> > };
> >
> > @@ -1749,10 +1766,14 @@ static umode_t pcie_dev_attrs_are_visible(struct kobject *kobj,
> > struct device *dev = kobj_to_dev(kobj);
> > struct pci_dev *pdev = to_pci_dev(dev);
> >
> > - if (pci_is_pcie(pdev))
> > - return a->mode;
> > + if (!pci_is_pcie(pdev))
> > + return 0;
> > +
> > + if (a == &dev_attr_device_serial_number.attr && !pci_get_dsn(pdev))
> > + return 0;
> > +
> > + return a->mode;
> >
> > - return 0;
> > }
> >
> > static const struct attribute_group pci_dev_group = {
>
