Re: [RFC 1/1] s390/pci: expose a PCI device's UID as its index

From: Viktor Mihajlovski <mihajlov@linux.ibm.com>
To: "Niklas Schnelle" <schnelle@linux.ibm.com>,
	"Krzysztof Wilczyński" <kw@linux.com>
Cc: Bjorn Helgaas <helgaas@kernel.org>,
	Narendra K <narendra_k@dell.com>,
	Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	Stefan Raspl <raspl@linux.ibm.com>,
	Peter Oberparleiter <oberpar@linux.ibm.com>,
	linux-pci@vger.kernel.org, linux-kernel@vger.kernel.org,
	linux-s390@vger.kernel.org
Subject: Re: [RFC 1/1] s390/pci: expose a PCI device's UID as its index
Date: Mon, 8 Mar 2021 09:42:37 +0100	[thread overview]
Message-ID: <e2f45e3e-210b-0b98-c8bd-bbe6609b85f1@linux.ibm.com> (raw)
In-Reply-To: <c714ca55-7189-e196-7b8d-f02da555b399@linux.ibm.com>


On 3/8/21 8:02 AM, Niklas Schnelle wrote:
> 
> 
> On 3/7/21 9:46 PM, Krzysztof Wilczyński wrote:
>> Hi Niklas,
>>
>> [...]
>>> +static ssize_t index_show(struct device *dev,
>>> +			  struct device_attribute *attr, char *buf)
>>> +{
>>> +	struct zpci_dev *zdev = to_zpci(to_pci_dev(dev));
>>> +	u32 index = ~0;
>>> +
>>> +	if (zpci_unique_uid)
>>> +		index = zdev->uid;
>>> +
>>> +	return sprintf(buf, "%u\n", index);
>> [...]
>>
>> Would it be possible to use the new sysfs_emit() rather than sprintf()
>> even though the zpci_attr macro and still use mio_enabled_show() still
>> would use sprintf().  What do you think?
>>
>> See https://www.kernel.org/doc/html/latest/filesystems/sysfs.html for
>> the changes in the internal API.
>>
>> Krzysztof
>>
> 
> Of course that makes sense and thanks for pointing me to this API!
> @Viktor, may I carry your R-b over?
> 
Sure, please go ahead.
> I'll also update the other attributes in a clean up patch.
> 
> Thanks,
> Niklas
> 

-- 
Kind Regards,
    Viktor