[PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Runyu Xiao]

From: Maarten Lankhorst <dev@lankhorst.se>

The LPE audio interrupt comes from the i915 interrupt handler. It
should be in irq disabled context.

With PREEMPT_RT enabled, the IRQ handler is threaded.
Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
generic_handle_irq_safe API disables the interrupts before calling LPE's
interrupt top half handler.

This fixes braswell audio issues with RT enabled.

Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
---
v3:
- Rebase the standalone patch on Maarten Lankhorst's original fix
- Keep Maarten Lankhorst as author per Sebastian's feedback
- Carry forward Maarten's Reviewed-by tags

 drivers/gpu/drm/i915/display/intel_lpe_audio.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/display/intel_lpe_audio.c b/drivers/gpu/drm/i915/display/intel_lpe_audio.c
index 42284e9928f2..ac1dfd592a9f 100644
--- a/drivers/gpu/drm/i915/display/intel_lpe_audio.c
+++ b/drivers/gpu/drm/i915/display/intel_lpe_audio.c
@@ -263,7 +263,7 @@ void intel_lpe_audio_irq_handler(struct intel_display *display)
 	if (!HAS_LPE_AUDIO(display))
 		return;
 
-	ret = generic_handle_irq(display->audio.lpe.irq);
+	ret = generic_handle_irq_safe(display->audio.lpe.irq);
 	if (ret)
 		drm_err_ratelimited(display->drm,
 				    "error handling LPE audio irq: %d\n", ret);
-- 
2.34.1

Re: [PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Sebastian Andrzej Siewior]

On 2026-06-02 08:45:15 [+0800], Runyu Xiao wrote:
> From: Maarten Lankhorst <dev@lankhorst.se>
> 
> The LPE audio interrupt comes from the i915 interrupt handler. It
> should be in irq disabled context.
> 
> With PREEMPT_RT enabled, the IRQ handler is threaded.
> Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
> generic_handle_irq_safe API disables the interrupts before calling LPE's
> interrupt top half handler.
> 
> This fixes braswell audio issues with RT enabled.
> 
> Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
> Reviewed-by: Uma Shankar <uma.shankar@intel.com>
> Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>

Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>

Sebastian

Re: [PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Maarten Lankhorst]

Thanks, pushed!

On 6/2/26 08:29, Sebastian Andrzej Siewior wrote:
> On 2026-06-02 08:45:15 [+0800], Runyu Xiao wrote:
>> From: Maarten Lankhorst <dev@lankhorst.se>
>>
>> The LPE audio interrupt comes from the i915 interrupt handler. It
>> should be in irq disabled context.
>>
>> With PREEMPT_RT enabled, the IRQ handler is threaded.
>> Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
>> generic_handle_irq_safe API disables the interrupts before calling LPE's
>> interrupt top half handler.
>>
>> This fixes braswell audio issues with RT enabled.
>>
>> Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
>> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
>> Reviewed-by: Uma Shankar <uma.shankar@intel.com>
>> Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
> 
> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> 
> Sebastian

Re: [PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Jani Nikula]

On Tue, 02 Jun 2026, Maarten Lankhorst <dev@lankhorst.se> wrote:
> Thanks, pushed!

Where are the CI results?

>
> On 6/2/26 08:29, Sebastian Andrzej Siewior wrote:
>> On 2026-06-02 08:45:15 [+0800], Runyu Xiao wrote:
>>> From: Maarten Lankhorst <dev@lankhorst.se>
>>>
>>> The LPE audio interrupt comes from the i915 interrupt handler. It
>>> should be in irq disabled context.
>>>
>>> With PREEMPT_RT enabled, the IRQ handler is threaded.
>>> Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
>>> generic_handle_irq_safe API disables the interrupts before calling LPE's
>>> interrupt top half handler.
>>>
>>> This fixes braswell audio issues with RT enabled.
>>>
>>> Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
>>> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
>>> Reviewed-by: Uma Shankar <uma.shankar@intel.com>
>>> Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
>> 
>> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
>> 
>> Sebastian
>

-- 
Jani Nikula, Intel

Re:Re: [PATCH v3] drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Runyu Xiao]

> On Wed, 03 Jun 2026, Jani Nikula wrote:
> Where are the CI results?

Intel GFX pre-merge CI was not triggered for this submission.

Patchwork validation failed with:
"Address 'runyu.xiao@seu.edu.cn' is not on the allowlist, which prevents CI from being triggered for this patch."

I have already contacted i915-ci-infra@lists.freedesktop.org to request allowlisting for this address.

Sorry for the confusion.

Runyu

>>
>> On 6/2/26 08:29, Sebastian Andrzej Siewior wrote:
>>> On 2026-06-02 08:45:15 [+0800], Runyu Xiao wrote:
>>>> From: Maarten Lankhorst <dev@lankhorst.se>
>>>>
>>>> The LPE audio interrupt comes from the i915 interrupt handler. It
>>>> should be in irq disabled context.
>>>>
>>>> With PREEMPT_RT enabled, the IRQ handler is threaded.
>>>> Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
>>>> generic_handle_irq_safe API disables the interrupts before calling LPE's
>>>> interrupt top half handler.
>>>>
>>>> This fixes braswell audio issues with RT enabled.
>>>>
>>>> Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
>>>> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
>>>> Reviewed-by: Uma Shankar <uma.shankar@intel.com>
>>>> Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
>>> 
>>> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
>>> 
>>> Sebastian
>>
>
>-- 
>Jani Nikula, Intel
>


>>
>> On 6/2/26 08:29, Sebastian Andrzej Siewior wrote:
>>> On 2026-06-02 08:45:15 [+0800], Runyu Xiao wrote:
>>>> From: Maarten Lankhorst <dev@lankhorst.se>
>>>>
>>>> The LPE audio interrupt comes from the i915 interrupt handler. It
>>>> should be in irq disabled context.
>>>>
>>>> With PREEMPT_RT enabled, the IRQ handler is threaded.
>>>> Because intel_lpe_audio_irq_handler() may be called in threaded IRQ context,
>>>> generic_handle_irq_safe API disables the interrupts before calling LPE's
>>>> interrupt top half handler.
>>>>
>>>> This fixes braswell audio issues with RT enabled.
>>>>
>>>> Signed-off-by: Maarten Lankhorst <dev@lankhorst.se>
>>>> Reviewed-by: Matthew Brost <matthew.brost@intel.com>
>>>> Reviewed-by: Uma Shankar <uma.shankar@intel.com>
>>>> Signed-off-by: Runyu Xiao <runyu.xiao@seu.edu.cn>
>>> 
>>> Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
>>> 
>>> Sebastian
>>
>
>-- 
>Jani Nikula, Intel
>

Claude review: drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Claude Code Review Bot]

Overall Series Review

Subject: drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT
Author: Runyu Xiao <runyu.xiao@seu.edu.cn>
Patches: 4
Reviewed: 2026-06-04T13:25:57.081447

---

This is a single-patch fix (v3) that replaces `generic_handle_irq()` with `generic_handle_irq_safe()` in the LPE audio IRQ forwarding path. The change is correct and well-motivated: on PREEMPT_RT kernels, i915's IRQ handler runs in threaded context (not hard IRQ context), so `generic_handle_irq()` — which requires hard IRQ context — is inappropriate. `generic_handle_irq_safe()` wraps the call with `local_irq_save()`/`local_irq_restore()`, making it safe from any context.

The fix is minimal, one-line, and uses the correct upstream API. The commit message accurately describes the problem and the solution. The current drm-next tree still has `generic_handle_irq()` at line 265, confirming this fix is still needed.

**One process concern:** The Signed-off-by chain has Maarten Lankhorst as From/author with his own Signed-off-by, then Runyu Xiao's Signed-off-by as the submitter. The v3 changelog says this is a rebase of Maarten's original fix. The Reviewed-by tags from Matthew Brost and Uma Shankar are carried forward — this is acceptable if the patch content hasn't materially changed from what they reviewed, which appears to be the case (same one-line change).

**Verdict: Looks good.** This is a straightforward, correct RT fix.

---

---
Generated by Claude Code Patch Reviewer

Claude review: drm/i915/display: Fix intel_lpe_audio_irq_handler for PREEMPT-RT

[Claude Code Review Bot]

Patch Review

**Change:**
```c
-	ret = generic_handle_irq(display->audio.lpe.irq);
+	ret = generic_handle_irq_safe(display->audio.lpe.irq);
```

**Analysis:**

- **Correctness:** The fix is correct. `generic_handle_irq()` requires hard IRQ context (per its documentation: "must be called from an IRQ context with irq regs initialized"). On PREEMPT_RT, the i915 interrupt handler is force-threaded, so it runs in process context. `generic_handle_irq_safe()` explicitly handles this by doing `local_irq_save()`/`local_irq_restore()` around the actual handler dispatch (confirmed at `kernel/irq/irqdesc.c:710-712`), making it callable from any context.

- **No functional regression on non-RT:** On non-RT kernels where the handler already runs in hard IRQ context, interrupts are already disabled, so the `local_irq_save()`/`local_irq_restore()` in `generic_handle_irq_safe()` is a no-op — no performance or behavioral difference.

- **Scope:** The change is appropriately minimal. Only this one call site in `intel_lpe_audio_irq_handler()` needs the fix — the caller is the i915 IRQ handler chain, and this is the point where IRQs get forwarded to the LPE audio chained IRQ chip.

- **Commit message:** Accurate and clear. Correctly identifies the PREEMPT_RT threading issue and the Braswell audio platform affected.

**No issues found.** 

Reviewed-by worthy as-is.

---
Generated by Claude Code Patch Reviewer