linux-kernel - Re: [PATCH] clk: Fix memory leak in clk

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 22 Oct 2019 11:51:49 -0700
From:   Stephen Boyd <sboyd@...nel.org>
To:     Kishon Vijay Abraham I <kishon@...com>,
        Michael Turquette <mturquette@...libre.com>
Cc:     linux-clk@...r.kernel.org, linux-kernel@...r.kernel.org,
        Kishon Vijay Abraham I <kishon@...com>,
        Tomi Valkeinen <tomi.valkeinen@...com>,
        Tero Kristo <t-kristo@...com>
Subject: Re: [PATCH] clk: Fix memory leak in clk_unregister()

Quoting Kishon Vijay Abraham I (2019-10-22 00:11:53)
> Memory allocated in alloc_clk() for 'struct clk' and
> 'const char *con_id' while invoking clk_register() is never freed
> in clk_unregister(), resulting in kmemleak showing the following
> backtrace.
> 
>   backtrace:
>     [<00000000546f5dd0>] kmem_cache_alloc+0x18c/0x270
>     [<0000000073a32862>] alloc_clk+0x30/0x70
>     [<0000000082942480>] __clk_register+0xc8/0x760
>     [<000000005c859fca>] devm_clk_register+0x54/0xb0
>     [<00000000868834a8>] 0xffff800008c60950
>     [<00000000d5a80534>] platform_drv_probe+0x50/0xa0
>     [<000000001b3889fc>] really_probe+0x108/0x348
>     [<00000000953fa60a>] driver_probe_device+0x58/0x100
>     [<0000000008acc17c>] device_driver_attach+0x6c/0x90
>     [<0000000022813df3>] __driver_attach+0x84/0xc8
>     [<00000000448d5443>] bus_for_each_dev+0x74/0xc8
>     [<00000000294aa93f>] driver_attach+0x20/0x28
>     [<00000000e5e52626>] bus_add_driver+0x148/0x1f0
>     [<000000001de21efc>] driver_register+0x60/0x110
>     [<00000000af07c068>] __platform_driver_register+0x40/0x48
>     [<0000000060fa80ee>] 0xffff800008c66020
> 
> Fix it here.

Do you have some Fixes tag for this? Looks OK to me, but I wonder if we
should also assign hw->clk to NULL after unregister frees it. There are
probably other problems around unregistration and reference counting
that we haven't found so far so I'm worried this doesn't solve all the
problems by just freeing the clk pointer.

For example, anything referencing the clk pointer freed here will now
start trying to dereference freed memory. For most cases we've replaced
struct clk with struct clk_core in the clk framework so I hope this
pointer isn't used very much at all. A quick grep shows that it is
returned from clk_get_parent() and __clk_lookup(). We really need to
kill off __clk_lookup() and replace it with something else, and
clk_get_parent() needs to generate a different clk and have callers call
clk_put() on the returned pointer. Long story short I think this is OK!