Re: [PATCH v2 05/11] pci: use helper for safer setting of driver_override

[Bjorn Helgaas <helgaas@xxxxxxxxxx>]

On Fri, Feb 25, 2022 at 10:36:20AM +0100, Krzysztof Kozlowski wrote:
> On 25/02/2022 00:52, Bjorn Helgaas wrote:
> > On Thu, Feb 24, 2022 at 08:49:15AM +0100, Krzysztof Kozlowski wrote:
> >> On 23/02/2022 22:51, Bjorn Helgaas wrote:
> >>> In subject, to match drivers/pci/ convention, do something like:
> >>>
> >>>   PCI: Use driver_set_override() instead of open-coding
> >>>
> >>> On Wed, Feb 23, 2022 at 08:13:04PM +0100, Krzysztof Kozlowski wrote:
> >>>> Use a helper for seting driver_override to reduce amount of duplicated
> >>>> code.
> >>>> @@ -567,31 +567,15 @@ static ssize_t driver_override_store(struct device *dev,
> >>>>  				     const char *buf, size_t count)
> >>>>  {
> >>>>  	struct pci_dev *pdev = to_pci_dev(dev);
> >>>> -	char *driver_override, *old, *cp;
> >>>> +	int ret;
> >>>>  
> >>>>  	/* We need to keep extra room for a newline */
> >>>>  	if (count >= (PAGE_SIZE - 1))
> >>>>  		return -EINVAL;
> >>>
> >>> This check makes no sense in the new function.  Michael alluded to
> >>> this as well.
> >>
> >> I am not sure if I got your comment properly. You mean here:
> >> 1. Move this check to driver_set_override()?
> >> 2. Remove the check entirely?
> > 
> > I was mistaken about the purpose of the comment and the check.  I
> > thought it had to do with *this* function, and this function doesn't
> > add a newline, and there's no obvious connection with PAGE_SIZE.
> > 
> > But looking closer, I think the "extra room for a newline" is really
> > to make sure that *driver_override_show()* can add a newline and have
> > it still fit within the PAGE_SIZE sysfs limit.
> > 
> > Most driver_override_*() functions have the same comment, so maybe
> > this was obvious to everybody except me :)  I do see that spi.c adds
> > "when displaying value" at the end, which helps a lot.
> > 
> > Sorry for the wild goose chase.
> 
> I think I will move this check anyway to driver_set_override() helper,
> because there is no particular benefit to have duplicated all over. The
> helper will receive "count" argument so can perform all checks.

Thanks, I think that would be good!

Bjorn