netdev - Re: [PATCH 3/3] can: flexcan: handle S32G2/S32G3 separate interrupt lines

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: <7a91c06f-6ea3-4262-82a3-9a1daf481f82@wanadoo.fr>
Date: Tue, 19 Nov 2024 20:36:30 +0900
From: Vincent Mailhol <mailhol.vincent@...adoo.fr>
To: Ciprian Marian Costea <ciprianmarian.costea@....nxp.com>,
 Marc Kleine-Budde <mkl@...gutronix.de>, Andrew Lunn <andrew+netdev@...n.ch>,
 "David S . Miller" <davem@...emloft.net>, Eric Dumazet
 <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley <conor+dt@...nel.org>
Cc: linux-can@...r.kernel.org, netdev@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
 imx@...ts.linux.dev, NXP Linux Team <s32@....com>,
 Christophe Lizzi <clizzi@...hat.com>, Alberto Ruiz <aruizrui@...hat.com>,
 Enric Balletbo <eballetb@...hat.com>
Subject: Re: [PATCH 3/3] can: flexcan: handle S32G2/S32G3 separate interrupt
 lines

On 19/11/2024 at 20:26, Vincent Mailhol wrote:
> On 19/11/2024 at 19:01, Ciprian Marian Costea wrote:
>> On 11/19/2024 11:26 AM, Vincent Mailhol wrote:
>>> On 19/11/2024 at 17:10, Ciprian Costea wrote:
> 
> (...)
> 
>>>>   +    if (priv->devtype_data.quirks & FLEXCAN_QUIRK_SECONDARY_MB_IRQ) {
>>>> +        err = request_irq(priv->irq_secondary_mb,
>>>> +                  flexcan_irq, IRQF_SHARED, dev->name, dev);
>>>> +        if (err)
>>>> +            goto out_free_irq_err;
>>>> +    }
>>>
>>> Is the logic here correct?
>>>
>>>    request_irq(priv->irq_err, flexcan_irq, IRQF_SHARED, dev->name, dev);
>>>
>>> is called only if the device has the FLEXCAN_QUIRK_NR_IRQ_3 quirk.
>>>
>>> So, if the device has the FLEXCAN_QUIRK_SECONDARY_MB_IRQ but not the
>>> FLEXCAN_QUIRK_NR_IRQ_3, you may end up trying to free an irq which was
>>> not initialized.
>>>
>>> Did you confirm if it is safe to call free_irq() on an uninitialized irq?
>>>
>>> (and I can see that currently there is no such device with
>>> FLEXCAN_QUIRK_SECONDARY_MB_IRQ but without FLEXCAN_QUIRK_NR_IRQ_3, but
>>> who knows if such device will be introduced in the future?)
>>>
>>
>> Hello Vincent,
>>
>> Thanks for your review. Indeed this seems to be an incorrect logic since
>> I do not want to create any dependency between 'FLEXCAN_QUIRK_NR_IRQ_3'
>> and 'FLEXCAN_QUIRK_SECONDARY_MB_IRQ'.
>>
>> I will change the impacted section to:
>>     if (err) {
>>         if (priv->devtype_data.quirks & FLEXCAN_QUIRK_NR_IRQ_3)
>>             goto out_free_irq_err;
>>         else
>>             goto out_free_irq;
>>     }
> 
> This is better. Alternatively, you could move the check into the label:
> 
>   out_free_irq_err:
>   	if (priv->devtype_data.quirks & FLEXCAN_QUIRK_NR_IRQ_3)
>   		free_irq(priv->irq_err, dev);
> 
> But this is not a strong preference, I let you pick the one which you
> prefer.

On second thought, it is a strong preference. If you keep the

	if (priv->devtype_data.quirks & FLEXCAN_QUIRK_NR_IRQ_3)
		goto out_free_irq_err;
	else
		goto out_free_irq;

then what if more code with a clean-up label is added to flexcan_open()?
I am thinking of this:

  out_free_foo:
  	free(foo);
  out_free_irq_err:
  	free_irq(priv-irq_err, dev);
  out_free_irq_boff:
  	free_irq(priv->irq_boff, dev);

Jumping to out_free_foo would now be incorrect because the
out_free_irq_err label would also be visited.

>>>>       flexcan_chip_interrupts_enable(dev);
>>>>         netif_start_queue(dev);
>>>>         return 0;
>>>>   + out_free_irq_err:
>>>> +    free_irq(priv->irq_err, dev);
>>>>    out_free_irq_boff:
>>>>       free_irq(priv->irq_boff, dev);
>>>>    out_free_irq:
> 
> (...)

Yours sincerely,
Vincent Mailhol