Received-SPF: pass (google.com: domain of linux-kernel+bounces-150596-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Date: Thu, 18 Apr 2024 14:50:59 -0300
From: Jason Gunthorpe <jgg@ziepe.ca>
To: Shradha Gupta <shradhagupta@linux.microsoft.com>
Cc: Andrew Lunn <andrew@lunn.ch>, Zhu Yanjun <yanjun.zhu@linux.dev>,
	linux-kernel@vger.kernel.org, linux-hyperv@vger.kernel.org,
	linux-rdma@vger.kernel.org, netdev@vger.kernel.org,
	Eric Dumazet <edumazet@google.com>,
	Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>,
	Ajay Sharma <sharmaajay@microsoft.com>,
	Leon Romanovsky <leon@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
	"K. Y. Srinivasan" <kys@microsoft.com>,
	Haiyang Zhang <haiyangz@microsoft.com>,
	Wei Liu <wei.liu@kernel.org>, Dexuan Cui <decui@microsoft.com>,
	Long Li <longli@microsoft.com>,
	Michael Kelley <mikelley@microsoft.com>,
	Shradha Gupta <shradhagupta@microsoft.com>,
	Yury Norov <yury.norov@gmail.com>,
	Konstantin Taranov <kotaranov@microsoft.com>,
	Souradeep Chakrabarti <schakrabarti@linux.microsoft.com>
Subject: Re: [PATCH net-next] net: mana: Add new device attributes for mana
Message-ID: <20240418175059.GZ223006@ziepe.ca>
References: <1713174589-29243-1-git-send-email-shradhagupta@linux.microsoft.com>
 <20240415161305.GO223006@ziepe.ca>
 <56b0a8c1-50f6-41a9-9ea5-ed45ada58892@linux.dev>
 <b34bfb11-98a3-4418-b482-14f2e50745d3@lunn.ch>
 <20240418060108.GB13182@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240418060108.GB13182@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net>

On Wed, Apr 17, 2024 at 11:01:08PM -0700, Shradha Gupta wrote:

> > > > > +static ssize_t mana_attr_show(struct device *dev,
> > > > > +			      struct device_attribute *attr, char *buf)
> > > > > +{
> > > > > +	struct pci_dev *pdev = to_pci_dev(dev);
> > > > > +	struct gdma_context *gc = pci_get_drvdata(pdev);
> > > > > +	struct mana_context *ac = gc->mana.driver_data;
> > > > > +
> > > > > +	if (strcmp(attr->attr.name, "mport") == 0)
> > > > > +		return snprintf(buf, PAGE_SIZE, "%d\n", ac->num_ports);
> > > > > +	else if (strcmp(attr->attr.name, "adapter_mtu") == 0)
> > > > > +		return snprintf(buf, PAGE_SIZE, "%d\n", gc->adapter_mtu);
> > > > > +	else if (strcmp(attr->attr.name, "msix") == 0)
> > > > > +		return snprintf(buf, PAGE_SIZE, "%d\n", gc->max_num_msix);
> > > > > +	else
> > > > > +		return -EINVAL;
> > > > > +
> > > > 
> > > > That is not how sysfs should be implemented at all, please find a
> > > > good example to copy from. Every attribute should use its own function
> > > > with the macros to link it into an attributes group and sysfs_emit
> > > > should be used for printing
> > > 
> > > Not sure if this file drivers/infiniband/hw/usnic/usnic_ib_sysfs.c is a good
> > > example or not.
> > 
> > The first question should be, what are these values used for? You can
> > then decide on debugfs or sysfs. debugfs is easier to use, and you
> > avoid any ABI, which will make long term support easier.
> 
> Hi Andrew,
> We want to eventually use these attributes to make the device settings configurable
> and also improve debuggability for MANA devices. I feel having these attributes 
> in sysfs would make more sense as we plan to extend the attribute list and also make
> them settable.

From an RDMA perspective this is all available from other APIs already
at least and I wouldn't want to see new sysfs unless there is a netdev
justification.

Jason