Re: [PATCH v2] clk: tegra: tegra124-emc: fix krealloc() memory leak

Thierry Reding <[email protected]> Thu, 30 Jul 2026 17:55:01 +0200

Newsgroups	org.kernel.vger.linux-clk,dev.linux.lists.llvm,org.kernel.vger.linux-kernel,org.kernel.vger.linux-tegra
Message-ID	<amtyoOXfiPic0f5H@orome>

On Sun, May 31, 2026 at 09:52:52PM +0200, Alexander A. Klimov wrote:
> Don't just overwrite the original pointer passed to krealloc()
> with its return value without checking latter:
> 
>     MEM = krealloc(MEM, SZ, GFP);
> 
> If krealloc() returns NULL, that erases the pointer
> to the still allocated memory, hence leaks this memory.
> Instead, use a temporary variable, check it's not NULL
> and only then assign it to the original pointer:
> 
>     TMP = krealloc(MEM, SZ, GFP);
>     if (!TMP) return;
>     MEM = TMP;
> 
> Fixes: 888ca40e2843 ("clk: tegra: emc: Support multiple RAM codes")
> Signed-off-by: Alexander A. Klimov <[email protected]>
> ---
>  v2: Separate variable declaration/init
>  v2: While on it, enhance variable name
>  v2: While on it, explicit variable type
>  v2: While on it, re-order variables (reverse Xmas tree)
> 
>  [✓] scripts/checkpatch.pl --strict
>  [✓] allmodconfig compiled (i686, LLVM)
>  [✓] localyesconfig booted (IBM T43)

The latter two are a bit pointless because they are not going to hit
this because the driver depends on ARCH_TEGRA which doesn't exist on
i686 allmodconfig and booting this on a T43 isn't going to hit this
either.

>  Note to myself: use --in-reply-to=ahilgKKwkttOd9H6@orome

It's not generally useful to send as replies to earlier versions. Just
regular sending is good enough. We have tools to track versions and
such.

> 
>  drivers/clk/tegra/clk-tegra124-emc.c | 6 ++++--
>  1 file changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/clk/tegra/clk-tegra124-emc.c b/drivers/clk/tegra/clk-tegra124-emc.c
> index f3b2c96fdcfc..81e4c02a807c 100644
> --- a/drivers/clk/tegra/clk-tegra124-emc.c
> +++ b/drivers/clk/tegra/clk-tegra124-emc.c
> @@ -445,15 +445,17 @@ static int load_timings_from_dt(struct tegra_clk_emc *tegra,
>  {
>  	struct emc_timing *timings_ptr;
>  	int child_count = of_get_child_count(node);
> +	struct emc_timing *timings;

I guess we could've reused timings_ptr. Or remove timings_ptr and use
the new one, but I guess this is fine, too:

Reviewed-by: Thierry Reding <[email protected]>

>  	int i = 0, err;
>  	size_t size;
>  
>  	size = (tegra->num_timings + child_count) * sizeof(struct emc_timing);
>  
> -	tegra->timings = krealloc(tegra->timings, size, GFP_KERNEL);
> -	if (!tegra->timings)
> +	timings = krealloc(tegra->timings, size, GFP_KERNEL);
> +	if (!timings)
>  		return -ENOMEM;
>  
> +	tegra->timings = timings;
>  	timings_ptr = tegra->timings + tegra->num_timings;
>  	tegra->num_timings += child_count;

signature.asc (application/pgp-signature, 833 B)