linux-kernel - Re: devm

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] [day] [month] [year] [list]

Message-ID: <alpine.DEB.2.00.1112270710300.7818@pl-59034.rocq.inria.fr>
Date:	Tue, 27 Dec 2011 07:10:42 +0100 (CET)
From:	Julia Lawall <julia.lawall@...6.fr>
To:	Adam Jiang <jiang.adam@...il.com>
cc:	kernel-janitors@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: devm_kfree

On Tue, 27 Dec 2011, Adam Jiang wrote:

> 2011/12/26 Julia Lawall <julia.lawall@...6.fr>:
>> Is it reasonable to use devm_kfree in a probe function that fails or a
>> remove function that succeeds?
>
> I cannot catch your question. If the memory is allocated by
> devm_kzalloc(), it is not necessary to release with devm_kfree() at
> all. It would be release once the driver was detached or destroyed.

Perhaps it is undesirable to keep it around until then, as in my last
example below where the probe function seems to want to go with a
different solution in some cases?

I was also worried that one call could make some information required by
another one unavailable.  For example in the following case:

drivers/dma/mpc512x_dma.c

         devm_free_irq(dev, mdma->irq, mdma);
         irq_dispose_mapping(mdma->irq);

Is it OK to call irq_dispose_mapping and then let devm_free_irq be called
implicitly afterwards?

julia



>
>>
>> Examples from drivers/watchdog/shwdt.c:
>>
>> sh_wdt_probe:
>>
>>        ...
>>  out_unreg:
>>        unregister_reboot_notifier(&sh_wdt_notifier);
>>  out_unmap:
>> -       devm_iounmap(&pdev->dev, wdt->base);
>>  out_err:
>> -       devm_kfree(&pdev->dev, wdt);
>
> It is necessary?
>
>>  out_release:
>> -       devm_release_mem_region(&pdev->dev, res->start, resource_size(res));
>
> It is necessary?
>
>>
>>        return rc;
>>  }
>>
>>
>> sh_wdt_remove:
>>
>>        ...
>>        unregister_reboot_notifier(&sh_wdt_notifier);
>>        devm_release_mem_region(&pdev->dev, res->start, resource_size(res));
>>        devm_iounmap(&pdev->dev, wdt->base);
>>        devm_kfree(&pdev->dev, wdt);
>
> These lines should be deleted if the driver has detached.
>
>>
>>        return 0;
>> }
>>
>>
>> On the other hand, perhaps devm_kfree is needed in probe function that
>> succeeds, eg in drivers/input/keyboard/samsung-keypad.c:
>>
>>        if (pdev->dev.of_node) {
>>                devm_kfree(&pdev->dev, (void *)pdata->keymap_data->keymap);
>>                devm_kfree(&pdev->dev, (void *)pdata->keymap_data);
>>                devm_kfree(&pdev->dev, (void *)pdata);
>>        }
>>        return 0;
>>
>> julia
>> --
>> 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/
> --
> To unsubscribe from this list: send the line "unsubscribe kernel-janitors" in
> the body of a message to majordomo@...r.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>