linux-kernel - Re: [PATCH 6/7] drivers: net: cpsw: fix error return code

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: <567EF457.8020602@cogentembedded.com>
Date:	Sat, 26 Dec 2015 23:11:03 +0300
From:	Sergei Shtylyov <sergei.shtylyov@...entembedded.com>
To:	Julia Lawall <julia.lawall@...6.fr>
Cc:	netdev@...r.kernel.org, kernel-janitors@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 6/7] drivers: net: cpsw: fix error return code

On 12/26/2015 11:07 PM, Julia Lawall wrote:

>>>>>> diff --git a/drivers/net/ethernet/ti/cpsw.c
>>>>>> b/drivers/net/ethernet/ti/cpsw.c
>>>>>> index 3409e80..6a76992 100644
>>>>>> --- a/drivers/net/ethernet/ti/cpsw.c
>>>>>> +++ b/drivers/net/ethernet/ti/cpsw.c
>>>>>> @@ -2448,8 +2448,10 @@ static int cpsw_probe(struct platform_device
>>>>>> *pdev)
>>>>>>
>>>>>>         /* RX IRQ */
>>>>>>         irq = platform_get_irq(pdev, 1);
>>>>>> -    if (irq < 0)
>>>>>> +    if (irq < 0) {
>>>>>> +        ret = -ENOENT;
>>>>>
>>>>>      Why not just propagate an error returned by that function?
>>>>
>>>> OK, I did what was done a few lines before in the same function:
>>>>
>>>>           ndev->irq = platform_get_irq(pdev, 1);
>>>>           if (ndev->irq < 0) {
>>>>          dev_err(priv->dev, "error getting irq resource\n");
>>>>                   ret = -ENOENT;
>>>>                   goto clean_ale_ret;
>>>>           }
>>>>
>>>> Maybe they should all be changed?
>>>
>>>      Yeah, I'd vote for it. I'm seeing no sense in overriding an actual
>>> error.
>>
>>     Hm, I decided to check drivers/base/dd.c and I think I maybe know the
>> reason now: -ENXIO, usually returned by platform_get_irq(), is silently
>> "swallowed" by really_probe(); to be precise, -ENODEV and -ENXIO are only
>> reported with pr_debug(), while -ENOENT causes printk(KERN_WARNING, ...)...

> Sorry, I'm confused...  What should it be?  v1 or v2?  Here are the counts
> of the different constants returned on failure of platform_get_irq:

    I was somewhat confused myself but then I remembered about the deferred 
probing -- overriding error code basically just disables it in this case.

> ENODEV: 84
> ENXIO: 67

   Those 2 totally make no sense. :-)

> EINVAL: 61
> ENOENT: 29
> EBUSY: 11

    Hm...

> EIO: 2
> EPROBE_DEFER: 1

    Hm, and that last one is unconditional?

> julia

MBR, Sergei

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/