Re: [PATCH] platform/x86/intel/pmc: validate LPM mode without stale index
From: David Box
Date: Fri Jul 10 2026 - 12:15:03 EST
On Tue, Jun 30, 2026 at 12:51:25PM +0200, Yousef Alhouseen wrote:
> pmc_lpm_mode_write() compares the requested mode with the loop variable
> after pmc_for_each_mode(). If no low-power modes are enabled,
> the loop never assigns that variable and the comparison reads an
> uninitialized value.
>
> Track a successful match explicitly, including the empty-mode case.
>
> Signed-off-by: Yousef Alhouseen <alhouseenyousef@xxxxxxxxx>
> ---
> drivers/platform/x86/intel/pmc/core.c | 14 ++++++++++----
> 1 file changed, 10 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/platform/x86/intel/pmc/core.c b/drivers/platform/x86/intel/pmc/core.c
> index 9f77c0716e59..9833df172d4e 100644
> --- a/drivers/platform/x86/intel/pmc/core.c
> +++ b/drivers/platform/x86/intel/pmc/core.c
> @@ -1147,6 +1147,7 @@ static ssize_t pmc_core_lpm_latch_mode_write(struct file *file,
> struct pmc_dev *pmcdev = s->private;
> struct pmc *pmc = pmcdev->pmcs[PMC_IDX_MAIN];
> bool clear = false, c10 = false;
> + bool mode_enabled = false;
> unsigned char buf[8];
> int mode;
> u32 reg;
> @@ -1167,11 +1168,16 @@ static ssize_t pmc_core_lpm_latch_mode_write(struct file *file,
> mode = sysfs_match_string(pmc_lpm_modes, buf);
>
Since you're now checking mode right after assignment remove this empty line to
tie it to the test.
> /* Check string matches enabled mode */
> - pmc_for_each_mode(m, pmc)
> - if (mode == m)
> - break;
> + if (mode >= 0) {
> + pmc_for_each_mode(m, pmc) {
> + if (mode == m) {
> + mode_enabled = true;
> + break;
> + }
> + }
> + }
>
> - if (mode != m || mode < 0) {
> + if (!mode_enabled) {
> if (sysfs_streq(buf, "clear"))
> clear = true;
> else if (sysfs_streq(buf, "c10"))
> --
> 2.54.0
>
With the above change,
Reviewed-by: David E. Box <david.e.box@xxxxxxxxxxxxxxx>
Thanks