[PATCH v4 2/6] usb: dwc2: fix a possible interrupt storm in hcd resume
From: Fabrice Gasnier
Date: Fri Nov 13 2026 - 14:05:49 EST
The host interrupts are re-enabled in dwc2_hcd_reinit(). So there is a
race during resume between a possible hcd IRQ, and the HW accessible bit
to be set. This can result in a bad_irq to be reported (nobody cared).
Note dwc2_hcd_reinit() "Must be called with interrupt disabled and
spinlock held." So, move dwc2_hcd_reinit() and the glocal irq enabling
inside the critical section, after HCD_FLAG_HW_ACCESSIBLE has been set.
Fixes: ef5e0eec4764 ("usb: dwc2: Add clock gating exiting flow by system resume")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@xxxxxxxxxxx>
---
Note it's been found while debugging with additional mdelay after
dwc2_hcd_reinit() call.
---
Changes in v4:
- Address Sashiko comment "usb: dwc2: unlocked dwc2_hcd_reinit() call in
_dwc2_hcd_resume()": rather move the dwc2_hcd_reinit() and irq enabling
under the critical section, after HCD_FLAG_HW_ACCESSIBLE has been set.
- Update commit msg.
Changes in v3:
- Introduce this patch since debug session (since sashiko comment on
patch 1)
---
drivers/usb/dwc2/hcd.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/drivers/usb/dwc2/hcd.c b/drivers/usb/dwc2/hcd.c
index 3f9adfa7c630..7e4202c145f2 100644
--- a/drivers/usb/dwc2/hcd.c
+++ b/drivers/usb/dwc2/hcd.c
@@ -4515,15 +4515,14 @@ static int _dwc2_hcd_resume(struct usb_hcd *hcd)
* the global interrupts are disabled.
*/
dwc2_core_init(hsotg, false);
- dwc2_enable_global_interrupts(hsotg);
- dwc2_hcd_reinit(hsotg);
spin_lock_irqsave(&hsotg->lock, flags);
-
/*
* Set HW accessible bit before powering on the controller
* since an interrupt may rise.
*/
set_bit(HCD_FLAG_HW_ACCESSIBLE, &hcd->flags);
+ dwc2_enable_global_interrupts(hsotg);
+ dwc2_hcd_reinit(hsotg);
break;
default:
hsotg->lx_state = DWC2_L0;
--
2.43.0