From: Jiazi Liu <jiazi.liu1984@gmail.com>
To: Thinh.Nguyen@synopsys.com, gregkh@linuxfoundation.org
Cc: frode@meta.com, linux-usb@vger.kernel.org,
linux-kernel@vger.kernel.org, stable@vger.kernel.org,
Jiazi Liu <jiazi.liu1984@gmail.com>,
Jiazi Liu <liujiazi@amazon.com>
Subject: [PATCH v3] usb: dwc3: gadget: fix IRQ storm on invalid event buffer count
Date: Fri, 4 Sep 2026 17:49:34 +0800 [thread overview]
Message-ID: <20260904094934.12341-1-liujiazi@amazon.com> (raw)
In-Reply-To: <20260727092258.1121-1-liujiazi@amazon.com>
From: Jiazi Liu <jiazi.liu1984@gmail.com>
When dwc3_check_event_buf() reads a GEVNTCOUNT value exceeding the
event buffer length, commit 63ccd26cd1f6 ("usb: dwc3: gadget: check
that event count does not exceed event buffer length") returns IRQ_NONE
without writing back GEVNTCOUNT. Since the DWC3 interrupt is
level-triggered, the uncleared IRQ source keeps the line asserted,
causing a tight IRQ storm that accumulates 99,900 unhandled interrupts
and triggers spurious.c:184 BUG -> kernel panic.
The resulting call stack:
__report_bad_irq+0xac/0xc8
note_interrupt+0x340/0x468
handle_irq_event+0xac/0xc0
handle_fasteoi_irq+0x120/0x228
gic_handle_irq+0x68/0x108
...
kernel BUG at kernel/irq/spurious.c:184
To reproduce, write a bogus value exceeding the event buffer length
directly to the GEVNTCOUNT register:
devmem <DWC3_BASE + 0xc40c> 4 0x1004
Write the bogus count back to GEVNTCOUNT to clear the IRQ source,
consistent with the stale event clearing pattern in
dwc3_event_buffers_setup(), and schedule error recovery to
reinitialize the controller.
Fixes: 63ccd26cd1f6 ("usb: dwc3: gadget: check that event count does not exceed event buffer length")
Cc: stable@vger.kernel.org
Tested-by: Jiazi Liu <liujiazi@amazon.com>
Signed-off-by: Jiazi Liu <liujiazi@amazon.com>
---
drivers/usb/dwc3/core.h | 19 +++++++
drivers/usb/dwc3/ep0.c | 9 ++++
drivers/usb/dwc3/gadget.c | 104 +++++++++++++++++++++++++++++++++++++-
3 files changed, 130 insertions(+), 2 deletions(-)
diff --git a/drivers/usb/dwc3/core.h b/drivers/usb/dwc3/core.h
index 608daeb7ef10..f3dabb37f806 100644
--- a/drivers/usb/dwc3/core.h
+++ b/drivers/usb/dwc3/core.h
@@ -49,6 +49,7 @@
#define DWC3_ENDPOINTS_NUM 32
#define DWC3_XHCI_RESOURCES_NUM 2
#define DWC3_ISOC_MAX_RETRIES 5
+#define DWC3_ERR_RECOVERY_MAX 3
#define DWC3_SCRATCHBUF_SIZE 4096 /* each buffer is assumed to be 4KiB */
#define DWC3_EVENT_BUFFERS_SIZE 4096
@@ -841,6 +842,12 @@ enum dwc3_link_state {
DWC3_LINK_STATE_MASK = 0x0f,
};
+enum dwc3_err_state {
+ DWC3_ERR_NONE = 0,
+ DWC3_ERR_RECOVERY,
+ DWC3_ERR_UNRECOVERABLE,
+};
+
/* TRB Length, PCM and Status */
#define DWC3_TRB_SIZE_MASK (0x00ffffff)
#define DWC3_TRB_SIZE_LENGTH(n) ((n) & DWC3_TRB_SIZE_MASK)
@@ -1004,6 +1011,7 @@ struct dwc3_glue_ops {
/**
* struct dwc3 - representation of our controller
* @drd_work: workqueue used for role swapping
+ * @err_recovery_work: workqueue used for controller error recovery
* @ep0_trb: trb which is used for the ctrl_req
* @bounce: address of bounce buffer
* @setup_buf: used while precessing STD USB requests
@@ -1013,6 +1021,7 @@ struct dwc3_glue_ops {
* @ep0_in_setup: one control transfer is completed and enter setup phase
* @lock: for synchronizing
* @mutex: for mode switching
+ * @connect_mutex: for the pull-up and err_recovery_work
* @dev: pointer to our struct device
* @sysdev: pointer to the DMA-capable device
* @xhci: pointer to our xHCI child
@@ -1079,6 +1088,9 @@ struct dwc3_glue_ops {
* @ep0_next_event: hold the next expected event
* @ep0state: state of endpoint zero
* @link_state: link state
+ * @err_state: current error recovery state.
+ * @err_recovery_count: number of consecutive error recovery attempts until
+ * confirmed healthy and reset to 0 on reset event.
* @speed: device speed (super, high, full, low)
* @hwparams: copy of hwparams registers
* @regset: debugfs pointer to regdump file
@@ -1189,6 +1201,7 @@ struct dwc3_glue_ops {
*/
struct dwc3 {
struct work_struct drd_work;
+ struct work_struct err_recovery_work;
struct dwc3_trb *ep0_trb;
void *bounce;
u8 *setup_buf;
@@ -1203,6 +1216,9 @@ struct dwc3 {
/* mode switching lock */
struct mutex mutex;
+ /* serializes pull-up run/stop vs error recovery */
+ struct mutex connect_mutex;
+
struct device *dev;
struct device *sysdev;
@@ -1332,6 +1348,9 @@ struct dwc3 {
enum dwc3_ep0_next ep0_next_event;
enum dwc3_ep0_state ep0state;
enum dwc3_link_state link_state;
+ enum dwc3_err_state err_state;
+
+ u32 err_recovery_count;
u16 u2sel;
u16 u2pel;
diff --git a/drivers/usb/dwc3/ep0.c b/drivers/usb/dwc3/ep0.c
index bfe616194dfa..fa3d8b9f6f8e 100644
--- a/drivers/usb/dwc3/ep0.c
+++ b/drivers/usb/dwc3/ep0.c
@@ -200,6 +200,11 @@ int dwc3_gadget_ep0_queue(struct usb_ep *ep, struct usb_request *request,
int ret;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ ret = -ESHUTDOWN;
+ goto out;
+ }
+
if (!dep->endpoint.desc || !dwc->pullups_connected || !dwc->connected) {
dev_err(dwc->dev, "%s: can't queue to disabled endpoint\n",
dep->name);
@@ -271,6 +276,10 @@ int dwc3_gadget_ep0_set_halt(struct usb_ep *ep, int value)
int ret;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
ret = __dwc3_gadget_ep0_set_halt(ep, value);
spin_unlock_irqrestore(&dwc->lock, flags);
diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index fa0f16ffafef..dbfff5e691e6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -1149,6 +1149,10 @@ static int dwc3_gadget_ep_enable(struct usb_ep *ep,
return 0;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
ret = __dwc3_gadget_ep_enable(dep, DWC3_DEPCFG_ACTION_INIT);
spin_unlock_irqrestore(&dwc->lock, flags);
@@ -2055,6 +2059,10 @@ static int dwc3_gadget_ep_queue(struct usb_ep *ep, struct usb_request *request,
int ret;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
ret = __dwc3_gadget_ep_queue(dep, req);
spin_unlock_irqrestore(&dwc->lock, flags);
@@ -2287,6 +2295,10 @@ static int dwc3_gadget_ep_set_halt(struct usb_ep *ep, int value)
int ret;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
ret = __dwc3_gadget_ep_set_halt(dep, value, false);
spin_unlock_irqrestore(&dwc->lock, flags);
@@ -2301,6 +2313,10 @@ static int dwc3_gadget_ep_set_wedge(struct usb_ep *ep)
int ret;
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
dep->flags |= DWC3_EP_WEDGE;
if (dep->number == 0 || dep->number == 1)
@@ -2432,6 +2448,10 @@ static int dwc3_gadget_wakeup(struct usb_gadget *g)
}
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
if (!dwc->gadget->wakeup_armed) {
dev_err(dwc->dev, "not armed for remote wakeup\n");
spin_unlock_irqrestore(&dwc->lock, flags);
@@ -2459,6 +2479,10 @@ static int dwc3_gadget_func_wakeup(struct usb_gadget *g, int intf_id)
}
spin_lock_irqsave(&dwc->lock, flags);
+ if (dwc->err_state != DWC3_ERR_NONE) {
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return -ESHUTDOWN;
+ }
/*
* If the link is in U3, signal for remote wakeup and wait for the
* link to transition to U0 before sending device notification.
@@ -2828,10 +2852,13 @@ static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on)
synchronize_irq(dwc->irq_gadget);
+ /* Serialize against dwc3_err_recovery_work() */
+ mutex_lock(&dwc->connect_mutex);
if (!is_on)
ret = dwc3_gadget_soft_disconnect(dwc);
else
ret = dwc3_gadget_soft_connect(dwc);
+ mutex_unlock(&dwc->connect_mutex);
pm_runtime_put(dwc->dev);
@@ -4147,6 +4174,10 @@ static void dwc3_gadget_reset_interrupt(struct dwc3 *dwc)
dwc->suspended = false;
+ /* The controller is recovered. */
+ if (dwc->err_state == DWC3_ERR_NONE)
+ dwc->err_recovery_count = 0;
+
/*
* Ideally, dwc3_reset_gadget() would trigger the function
* drivers to stop any active transfers through ep disable.
@@ -4634,6 +4665,12 @@ static irqreturn_t dwc3_thread_interrupt(int irq, void *_evt)
return ret;
}
+static void dwc3_schedule_err_recovery(struct dwc3 *dwc)
+{
+ dwc->err_state = DWC3_ERR_RECOVERY;
+ schedule_work(&dwc->err_recovery_work);
+}
+
static irqreturn_t dwc3_check_event_buf(struct dwc3_event_buffer *evt)
{
struct dwc3 *dwc = evt->dwc;
@@ -4667,9 +4704,21 @@ static irqreturn_t dwc3_check_event_buf(struct dwc3_event_buffer *evt)
return IRQ_NONE;
if (count > evt->length) {
- dev_err_ratelimited(dwc->dev, "invalid count(%u) > evt->length(%u)\n",
+ dev_err(dwc->dev, "invalid count(%u) > evt->length(%u)\n",
count, evt->length);
- return IRQ_NONE;
+ /*
+ * This is a fatal error - the driver and controller are out of
+ * sync on which event has been consumed. Reinitializing the
+ * controller is required to recover. Write the bogus count back
+ * to GEVNTCOUNT to clear the IRQ source, consistent with the
+ * stale event clearing in dwc3_event_buffers_setup(), then
+ * schedule error recovery.
+ */
+ spin_lock(&dwc->lock);
+ dwc3_schedule_err_recovery(dwc);
+ spin_unlock(&dwc->lock);
+ dwc3_writel(dwc, DWC3_GEVNTCOUNT(0), count);
+ return IRQ_HANDLED;
}
evt->count = count;
@@ -4729,6 +4778,53 @@ static void dwc_gadget_release(struct device *dev)
kfree(gadget);
}
+static void dwc3_err_recovery_work(struct work_struct *work)
+{
+ struct dwc3 *dwc = container_of(work, struct dwc3, err_recovery_work);
+ unsigned long flags;
+ int ret;
+
+ dwc->err_recovery_count++;
+
+ /* serializes against dwc3_gadget_pullup() */
+ mutex_lock(&dwc->connect_mutex);
+
+ ret = dwc3_gadget_soft_disconnect(dwc);
+ if (ret)
+ goto err_unrecoverable;
+
+ dwc3_disconnect_gadget_sleepable(dwc);
+
+ if (dwc->err_recovery_count > DWC3_ERR_RECOVERY_MAX)
+ goto err_unrecoverable;
+
+ if (dwc->softconnect) {
+ /*
+ * Wait irq to finish before soft_connect resets evt->lpos and
+ * the event buffer registers to avoid racing with
+ * dwc3_process_event_buf().
+ */
+ synchronize_irq(dwc->irq_gadget);
+
+ ret = dwc3_gadget_soft_connect(dwc);
+ if (ret)
+ goto err_unrecoverable;
+ }
+ mutex_unlock(&dwc->connect_mutex);
+
+ spin_lock_irqsave(&dwc->lock, flags);
+ dwc->err_state = DWC3_ERR_NONE;
+ spin_unlock_irqrestore(&dwc->lock, flags);
+ return;
+
+err_unrecoverable:
+ dev_err(dwc->dev, "Unable to recover the controller\n");
+ mutex_unlock(&dwc->connect_mutex);
+ spin_lock_irqsave(&dwc->lock, flags);
+ dwc->err_state = DWC3_ERR_UNRECOVERABLE;
+ spin_unlock_irqrestore(&dwc->lock, flags);
+}
+
/**
* dwc3_gadget_init - initializes gadget related registers
* @dwc: pointer to our controller context structure
@@ -4772,6 +4868,8 @@ int dwc3_gadget_init(struct dwc3 *dwc)
}
init_completion(&dwc->ep0_in_setup);
+ INIT_WORK(&dwc->err_recovery_work, dwc3_err_recovery_work);
+ mutex_init(&dwc->connect_mutex);
dwc->gadget = kzalloc_obj(struct usb_gadget);
if (!dwc->gadget) {
ret = -ENOMEM;
@@ -4868,6 +4966,8 @@ void dwc3_gadget_exit(struct dwc3 *dwc)
if (!dwc->gadget)
return;
+ cancel_work_sync(&dwc->err_recovery_work);
+ mutex_destroy(&dwc->connect_mutex);
dwc3_enable_susphy(dwc, true);
usb_del_gadget(dwc->gadget);
dwc3_gadget_free_endpoints(dwc);
--
2.50.1 (Apple Git-155)
next prev parent reply other threads:[~2026-09-04 9:50 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20260727092258.1121-1-liujiazi@amazon.com>
2026-08-12 8:21 ` [PATCH v2 1/1] usb: dwc3: gadget: fix IRQ storm on invalid event buffer count Jiazi Liu
2026-08-12 9:48 ` Krishna Kurapati
2026-08-28 2:36 ` Thinh Nguyen
2026-08-28 9:48 ` Liu Jiazi
2026-08-29 0:31 ` Thinh Nguyen
2026-09-02 4:14 ` Liu Jiazi
2026-09-04 1:56 ` Thinh Nguyen
2026-09-04 9:49 ` Jiazi Liu [this message]
2026-09-04 23:55 ` [PATCH v3] " Thinh Nguyen
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20260904094934.12341-1-liujiazi@amazon.com \
--to=jiazi.liu1984@gmail.com \
--cc=Thinh.Nguyen@synopsys.com \
--cc=frode@meta.com \
--cc=gregkh@linuxfoundation.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-usb@vger.kernel.org \
--cc=liujiazi@amazon.com \
--cc=stable@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox