netdev - Re: [PATCH 2/2] of: net: fix of_get_mac_addr

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1b7e58ba2ec798ddda77a9a3ab72338c@walle.cc>
Date:   Tue, 06 Apr 2021 10:59:52 +0200
From:   Michael Walle <michael@...le.cc>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH 2/2] of: net: fix of_get_mac_addr_nvmem() for PCI and DSA
 nodes

Am 2021-04-06 00:13, schrieb Andrew Lunn:
> On Mon, Apr 05, 2021 at 11:46:04PM +0200, Michael Walle wrote:
>> Hi Andrew,
>> 
>> Am 2021-04-05 23:34, schrieb Andrew Lunn:
>> > > -static int of_get_mac_addr_nvmem(struct device_node *np, u8 addr)
>> > > +static int of_get_mac_addr_nvmem(struct device_node *np, u8 *addr)
>> > >  {
>> > >  	struct platform_device *pdev = of_find_device_by_node(np);
>> > > +	struct nvmem_cell *cell;
>> > > +	const void *mac;
>> > > +	size_t len;
>> > >  	int ret;
>> > >
>> > > -	if (!pdev)
>> > > -		return -ENODEV;
>> > > +	/* Try lookup by device first, there might be a nvmem_cell_lookup
>> > > +	 * associated with a given device.
>> > > +	 */
>> > > +	if (pdev) {
>> > > +		ret = nvmem_get_mac_address(&pdev->dev, addr);
>> > > +		put_device(&pdev->dev);
>> > > +		return ret;
>> > > +	}
>> >
>> > Can you think of any odd corner case where nvmem_get_mac_address()
>> > would fail, but of_nvmem_cell_get(np, "mac-address") would work?
>> 
>> You mean, it might make sense to just return here when
>> nvmem_get_mac_address() will succeed and fall back to the
>> of_nvmem_cell_get() in case of an error?
> 
> I've not read the documentation for nvmem_get_mac_address(). I was
> thinking we might want to return real errors, and -EPROBE_DEFER.

I can't follow, nvmem_get_mac_address() should already return those.

> But maybe with -ENODEV we should try of_nvmem_cell_get()?

And if this happens - that is nvmem_get_mac_address(&pdev->dev) returns
-ENODEV - then of_nvmem_cell_get(np) will also return -ENODEV.

Because pdev->dev.of_node == np and nvmem_get_mac_address(&pdev->dev)
tries of_nvmem_cell_get(pdev->dev.of_node) first.

> But i'm not sure if there are any real use cases? The only thing i can
> think of is if np points to something deeper inside the device tree
> than what pdev does?

But then pdev will be NULL and nvmem_get_mac_address() won't be called
at all, no?

-michael