[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <aH7jM_q3y85o2Daf@ketchup>
Date: Tue, 22 Jul 2025 01:02:43 +0000
From: Haylen Chu <heylenay@....org>
To: Akhilesh Patil <akhilesh@...iitb.ac.in>, mturquette@...libre.com,
sboyd@...nel.org, dlan@...too.org, elder@...cstar.com,
inochiama@...look.com, robh@...nel.org, krzk+dt@...nel.org,
conor+dt@...nel.org, heylenay@...look.com, paul.walmsley@...ive.com,
palmer@...belt.com, aou@...s.berkeley.edu, alex@...ti.fr
Cc: linux-clk@...r.kernel.org, linux-riscv@...ts.infradead.org,
spacemit@...ts.linux.dev, linux-kernel@...r.kernel.org,
unicornxdotw@...mail.com, jszhang@...nel.org,
zhangmeng.kevin@...ux.spacemit.com, akhileshpatilvnit@...il.com,
skhan@...uxfoundation.org
Subject: Re: [PATCH] clk: spacemit: fix error handling in
ccu_pll_set_rate/_round_rate
On Tue, Jul 22, 2025 at 12:29:23AM +0530, Akhilesh Patil wrote:
> Initialize best_entry pointer with NULL in ccu_pll_lookup_best_rate()
> to avoid returning garbage value when function fails to assign it
> a valid rate entry.
This doesn't sound very reasonable to me. Looking through
ccu_pll_lookup_best_rate(),
static const struct ccu_pll_rate_tbl *
ccu_pll_lookup_best_rate(struct ccu_pll *pll, unsigned long rate)
{
struct ccu_pll_config *config = &pll->config;
const struct ccu_pll_rate_tbl *best_entry;
unsigned long best_delta = ULONG_MAX;
int i;
for (i = 0; i < config->tbl_num; i++) {
const struct ccu_pll_rate_tbl *entry = &config->rate_tbl[i];
unsigned long delta = abs_diff(entry->rate, rate);
if (delta < best_delta) {
best_delta = delta;
best_entry = entry;
}
}
return best_entry;
}
best_entry is assigned as long as there's one entry fits the delta
better. Since best_delta is set to ULONG_MAX, any entry with non-zero
rates fits the required rate "better" at start of the loop. As long as
we have at least one non-zero entry defined for the PLL, best_entry is
always initialized and ccu_pll_lookup_best_rate() cannot return an
invalid pointer. And all existing PLLs do define at least one entry.
> Avoid passing invalid rate entry reference to
> ccu_pll_update_param by adding appropriate error handling in
> ccu_pll_set_rate and ccu_pll_round_rate.
> Address the effects of uninitialized pointer as reported
> by smatch and coverity static code analysis tools.
>
> Addresses-Coverity-ID: 1649164
> Reported-by: kernel test robot <lkp@...el.com>
> Closes: https://lore.kernel.org/r/202505111057.ejK2J56K-lkp@intel.com/
Thus this looks like a false-positive to me.
> Signed-off-by: Akhilesh Patil <akhilesh@...iitb.ac.in>
Regards,
Haylen Chu
> ---
> drivers/clk/spacemit/ccu_pll.c | 11 +++++++++--
> 1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/clk/spacemit/ccu_pll.c b/drivers/clk/spacemit/ccu_pll.c
> index 4427dcfbbb97..3fc6a30f98b7 100644
> --- a/drivers/clk/spacemit/ccu_pll.c
> +++ b/drivers/clk/spacemit/ccu_pll.c
> @@ -21,7 +21,7 @@ static const struct ccu_pll_rate_tbl *ccu_pll_lookup_best_rate(struct ccu_pll *p
> unsigned long rate)
> {
> struct ccu_pll_config *config = &pll->config;
> - const struct ccu_pll_rate_tbl *best_entry;
> + const struct ccu_pll_rate_tbl *best_entry = NULL;
> unsigned long best_delta = ULONG_MAX;
> int i;
>
> @@ -107,6 +107,10 @@ static int ccu_pll_set_rate(struct clk_hw *hw, unsigned long rate,
> const struct ccu_pll_rate_tbl *entry;
>
> entry = ccu_pll_lookup_best_rate(pll, rate);
> +
> + if (!entry)
> + return -EINVAL;
> +
> ccu_pll_update_param(pll, entry);
>
> return 0;
> @@ -129,8 +133,11 @@ static long ccu_pll_round_rate(struct clk_hw *hw, unsigned long rate,
> unsigned long *prate)
> {
> struct ccu_pll *pll = hw_to_ccu_pll(hw);
> + const struct ccu_pll_rate_tbl *entry;
> +
> + entry = ccu_pll_lookup_best_rate(pll, rate);
>
> - return ccu_pll_lookup_best_rate(pll, rate)->rate;
> + return entry ? entry->rate : 0;
> }
>
> static int ccu_pll_init(struct clk_hw *hw)
> --
> 2.34.1
>
Powered by blists - more mailing lists