From: Ulf Hansson ulf.hansson@linaro.org
[ Upstream commit 99012014c902cd9ad85fd288d8a107f33a69855e ]
If genpd_alloc_data() allocates data for the default power-states for the genpd, let's make sure to also reset the pointer in the error path. This makes sure a genpd provider driver doesn't end up trying to free the data again, but using an invalid pointer.
Signed-off-by: Ulf Hansson ulf.hansson@linaro.org Reviewed-by: Dhruva Gole d-gole@ti.com Link: https://lore.kernel.org/r/20250402120613.1116711-1-ulf.hansson@linaro.org Signed-off-by: Sasha Levin sashal@kernel.org ---
Based on the detailed analysis of the code and the commit changes, I can now provide a comprehensive answer. **YES** This commit should be backported to stable kernel trees. Here's my extensive analysis: ## Bug Analysis The commit fixes a memory management bug in the power domain subsystem's error handling path. In `genpd_alloc_data()`, when default power states are allocated via `genpd_set_default_power_state()`, the function sets: - `genpd->states = state` (pointing to allocated memory) - `genpd->free_states = genpd_free_default_power_state` However, in the error path (`put:` label at line 2197-2198), the code frees the allocated states memory with `kfree(genpd->states)` but **does not reset the pointer to NULL**. This leaves a dangling pointer in `genpd->states`. ## The Problem If `genpd_alloc_data()` fails and returns an error, but the genpd provider driver later tries to clean up and calls `genpd_free_data()` (which happens in `genpd_remove()` at line 2214), the following sequence occurs: 1. `genpd_free_data()` checks `if (genpd->free_states)` - this is still set to `genpd_free_default_power_state` 2. It calls `genpd->free_states(genpd->states, genpd->state_count)` 3. This results in `kfree(genpd->states)` being called on already-freed memory (double- free) This is a classic use-after-free/double-free vulnerability that can lead to memory corruption, crashes, or potential security issues. ## The Fix The fix is simple but crucial: after `kfree(genpd->states)` in the error path, it adds `genpd->states = NULL`. This ensures that: 1. The dangling pointer is eliminated 2. Future calls to `genpd_free_data()` will call `kfree(NULL)`, which is safe and does nothing 3. No double-free occurs ## Backport Criteria Assessment **1. Does it fix a bug that affects users?** YES - This can cause kernel crashes and memory corruption when power domain initialization fails. **2. Is the fix small and contained?** YES - It's a single line addition (`genpd->states = NULL`). **3. Does it have clear side effects?** NO - Setting a pointer to NULL after freeing it is a standard safe practice with no negative side effects. **4. Major architectural changes?** NO - This is a simple bug fix in error handling. **5. Critical kernel subsystem?** YES - Power domain management is critical for power management across many devices. **6. Explicit stable mention?** NO - But this type of memory safety fix is exactly what stable trees are for. **7. Follows stable rules?** YES - Important bugfix with minimal risk. ## Risk Assessment **Risk: MINIMAL** - The change only affects the error path and makes the code safer by preventing double-free. There's virtually no risk of regression since it only changes behavior when an error has already occurred. ## Similar Commit Patterns This follows the same pattern as **Similar Commit #1 and #4** which were both marked as **YES** for backporting - they fix memory management bugs in the same subsystem with minimal, contained changes that improve robustness without architectural changes. The fix prevents a potential security vulnerability (memory corruption) and improves system stability, making it an excellent candidate for stable tree backporting.
drivers/pmdomain/core.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/pmdomain/core.c b/drivers/pmdomain/core.c index 949445e929731..951071e1093ac 100644 --- a/drivers/pmdomain/core.c +++ b/drivers/pmdomain/core.c @@ -2194,8 +2194,10 @@ static int genpd_alloc_data(struct generic_pm_domain *genpd) return 0; put: put_device(&genpd->dev); - if (genpd->free_states == genpd_free_default_power_state) + if (genpd->free_states == genpd_free_default_power_state) { kfree(genpd->states); + genpd->states = NULL; + } free: if (genpd_is_cpu_domain(genpd)) free_cpumask_var(genpd->cpus);