Re: [PATCH] pci/pci-sysfs: Set pci interface in uppercase

From: Ricardo Ribalda Delgado
Date: Wed Aug 27 2014 - 16:57:22 EST


No worries,

I have to mark for stable it or Bjorn? It it is me, how :) ?


Regards!


ps: For other people reading this thread, the kmod/modprobe is in
http://anonscm.debian.org/cgit/users/md/kmod.git/tree/libkmod/libkmod-index.c
and handles all the modalias as strings without differing the type.

On Wed, Aug 27, 2014 at 10:51 PM, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote:
> On Wed, Aug 27, 2014 at 01:23:01PM -0700, Greg KH wrote:
>> On Wed, Aug 27, 2014 at 02:57:57PM +0200, Ricardo Ribalda Delgado wrote:
>> > There is a missmatch between the way file2alias generates the modalias
>> > and the way the pci driver generates it.
>> >
>> > Some implementations of modprobe will fail to load the driver for a pci
>> > device automatically when the pci interface is defined on the driver. As
>> > one will be in uppercase and the other in lowercase.
>> >
>> > Fortunatelly not many drivers define this.
>> >
>> > Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@xxxxxxxxx>
>> > ---
>> > drivers/pci/pci-sysfs.c | 2 +-
>> > 1 file changed, 1 insertion(+), 1 deletion(-)
>> >
>> > diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
>> > index 9ff0a90..76ef791 100644
>> > --- a/drivers/pci/pci-sysfs.c
>> > +++ b/drivers/pci/pci-sysfs.c
>> > @@ -177,7 +177,7 @@ static ssize_t modalias_show(struct device *dev, struct device_attribute *attr,
>> > {
>> > struct pci_dev *pci_dev = to_pci_dev(dev);
>> >
>> > - return sprintf(buf, "pci:v%08Xd%08Xsv%08Xsd%08Xbc%02Xsc%02Xi%02x\n",
>> > + return sprintf(buf, "pci:v%08Xd%08Xsv%08Xsd%08Xbc%02Xsc%02Xi%02X\n",
>> > pci_dev->vendor, pci_dev->device,
>> > pci_dev->subsystem_vendor, pci_dev->subsystem_device,
>> > (u8)(pci_dev->class >> 16), (u8)(pci_dev->class >> 8),
>>
>> As said in the other thread about this issue, no, this code has been
>> here for over 9 years just fine. Please fix your userspace code that is
>> trying to compare hex values as a string and not a numeric value, that
>> is the stuff that is wrong, not the kernel.
>
> Oh wait, I see what you are worried about now, the mis-match for just
> the upper bits of the class value.
>
> Yeah, that's a bug, sorry about that, a 9+ year old one, nice catch :)
>
> Bjorn, feel free to apply this, sorry for the earlier objection. Also
> please mark it for stable tree inclusion so this gets backported
> properly.
>
> thanks,
>
> greg k-h



--
Ricardo Ribalda
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/