netdev - Re: [PATCH 4/7 v2] net: ethernet: i825xx: Fix platform_get

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: <791fdb7d-e0ec-eefb-c703-205c039eddc3@cogentembedded.com>
Date:   Tue, 5 Dec 2017 13:12:23 +0300
From:   Sergei Shtylyov <sergei.shtylyov@...entembedded.com>
To:     David Miller <davem@...emloft.net>, arvind.yadav.cs@...il.com
Cc:     wg@...ndegger.com, mkl@...gutronix.de, michal.simek@...inx.com,
        opendmb@...il.com, f.fainelli@...il.com,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
        netdev@...r.kernel.org
Subject: Re: [PATCH 4/7 v2] net: ethernet: i825xx: Fix platform_get_irq's
 error checking

On 12/4/2017 9:25 PM, David Miller wrote:

>> @@ -120,9 +120,10 @@ static int sni_82596_probe(struct platform_device *dev)
>>   	netdevice->dev_addr[5] = readb(eth_addr + 0x06);
>>   	iounmap(eth_addr);
>>   
>> -	if (!netdevice->irq) {
>> +	if (netdevice->irq <= 0) {
>>   		printk(KERN_ERR "%s: IRQ not found for i82596 at 0x%lx\n",
>>   			__FILE__, netdevice->base_addr);
>> +		retval = netdevice->irq ? netdevice->irq : -ENODEV;
>>   		goto probe_failed;
>>   	}
> 
> Ok, thinking about this some more...
> 
> It is impossible to use platform_get_irq() without every single call
> site having this funny:
> 
> 	ret = val ? val : -ENODEV;
> 
> sequence.
> 
> This is unnecessary duplication and it is also error prone, so I
> really think this logic belongs in platform_get_irq() itself.  It can
> convert '0' to -ENODEV and that way we need no special logic in the
> callers at all.

    Well, we can have:

	return r && r->start ? r->start : -ENXIO;

instead of:

	return r ? r->start : -ENXIO;

at the end of platform_get_irq(). But I don't really think it's worth doing -- 
request_irq() doesn't filter out IRQ0 anyway, IIRC...

MBR, Sergei