The refactoring of ice_dpll_rclk_state_on_pin_get() to use
ice_dpll_pin_get_parent_idx() omitted the base_rclk_idx adjustment that was
correctly added in the ice_dpll_rclk_state_on_pin_set() path. This breaks
E810 devices where base_rclk_idx is non-zero, causing the wrong hardware
index to be used for pin state lookup and incorrect recovered clock state
to be reported via the DPLL subsystem. E825C is unaffected as its
base_rclk_idx is 0.
While at it, add bounds check against ICE_DPLL_RCLK_NUM_MAX on hw_idx after
the base_rclk_idx subtraction in both ice_dpll_rclk_state_on_pin_{get,set}()
to prevent out-of-bounds access on the pin state array.
Fixes: ad1df4f2d591 ("ice: dpll: Support E825-C SyncE and dynamic pin discovery")
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20260506-jk-iwl-net-2026-05-04-v2-7-a5ea4dc837a9@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
if (hw_idx < 0)
goto unlock;
hw_idx -= pf->dplls.base_rclk_idx;
+ if (hw_idx >= ICE_DPLL_RCLK_NUM_MAX)
+ goto unlock;
if ((enable && p->state[hw_idx] == DPLL_PIN_STATE_CONNECTED) ||
(!enable && p->state[hw_idx] == DPLL_PIN_STATE_DISCONNECTED)) {
hw_idx = ice_dpll_pin_get_parent_idx(p, parent_pin);
if (hw_idx < 0)
goto unlock;
+ hw_idx -= pf->dplls.base_rclk_idx;
+ if (hw_idx >= ICE_DPLL_RCLK_NUM_MAX)
+ goto unlock;
ret = ice_dpll_pin_state_update(pf, p, ICE_DPLL_PIN_TYPE_RCLK_INPUT,
extack);