linux-kernel - Re: [PATCH v2] irq: generic-chip: provide irq_free_generic

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: <CAMpxmJW64GFxM86bjO_iyn8eFsCJT8rd=_t_9faYLurAiAoWjA@mail.gmail.com>
Date:   Thu, 9 Mar 2017 11:35:48 +0100
From:   Bartosz Golaszewski <bgolaszewski@...libre.com>
To:     Thomas Gleixner <tglx@...utronix.de>
Cc:     Marc Zyngier <marc.zyngier@....com>,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] irq: generic-chip: provide irq_free_generic_chip()

2017-03-09 11:17 GMT+01:00 Thomas Gleixner <tglx@...utronix.de>:
> On Wed, 8 Mar 2017, Bartosz Golaszewski wrote:
>
>> Some users of irq_alloc_generic_chip() are modules which can be
>> removed (e.g. gpio-ml-ioh) but have no means of freeing the allocated
>> generic chip. Provide a function for that.
>
> They have means, i.e. kfree(gc). If you want a wrapper for that then please
> make it a simple static inline w/o the bloat of an exported symbol.
>

Yes, I want a wrapper, since there are no guarantees as to what
irq_alloc_generic_chip() does internally.

> But, what you really want is a proper counterpart to
> irq_alloc_generic_chip() which undoes everything what
> irq_alloc_generic_chip() does.
>
> You have to call irq_remove_generic_chip() before calling kfree() or the
> wrapper anyway. And looking at the 3 users of irq_remove_generic_chip():
>
> Two of them (drivers/gpio/gpio-tb10x.c drivers/mfd/jz4740-adc.c) do
>
>     irq_remove_generic_chip(gc)
>     kfree(gc)
>
> The third one (arch/arm/mach-omap2/prm_common.c) leaks the generic chip
> memory....
>
> So the proper solution is to rename irq_remove_generic_chip() to
> irq_destroy_generic_chip() and do the kfree() at the end of that function.
>
> Hmm?
>

Sounds good!

Thanks,
Bartosz