Skip to contentSkip to conversation

usb: dwc3: gadget: Avoid starting DWC3 gadget during UDC unbind

PATCH In review Wesley Cheng 2 participants 4 messages Sep 17, 2021

Subsystems (2)
USB DesignWare USB3 DRD IP driver
lore
Share & tools
lore

Wesley Cheng

author
#0 lore raw

PATCH 1/1 usb: dwc3: gadget: Avoid starting DWC3 gadget during UDC unbind

There is a race present where the DWC3 runtime resume runs in parallel to the UDC unbind sequence. This will eventually lead to a possible scenario where we are enabling the run/stop bit, without a valid composition defined.

Thread#1 (handling UDC unbind):
usb_gadget_remove_driver()
-->usb_gadget_disconnect()
-->dwc3_gadget_pullup(0)
--> continue UDC unbind sequence
-->Thread#2 is running in parallel here

Thread#2 (handing next cable connect)
__dwc3_set_mode()
-->pm_runtime_get_sync()

    -->dwc3_gadget_resume()
      -->dwc->gadget_driver is NOT NULL yet
      -->dwc3_gadget_run_stop(1)
      --> _dwc3gadget_start()

...

Fix this by tracking the pullup disable routine, and avoiding resuming of the DWC3 gadget. Once the UDC is re-binded, that will trigger the pullup enable routine, which would handle enabling the DWC3 gadget.

Signed-off-by Wesley Cheng <[email protected]>
---
 drivers/usb/dwc3/core.h   | 2 ++
 drivers/usb/dwc3/gadget.c | 4 ++--
 2 files changed, 4 insertions(+), 2 deletions(-)
drivers/usb/dwc3/core.h +2 -0
index 324ca5f5e98d..ee854697c300 100644
@@ -1032,6 +1032,7 @@ struct dwc3_scratchpad_array {
  * @tx_fifo_resize_max_num: max number of fifos allocated during txfifo resize
  * @hsphy_interface: "utmi" or "ulpi"
  * @connected: true when we're connected to a host, false otherwise
+ * @softconnect: true when gadget connect is called, false when disconnect runs
  * @delayed_status: true when gadget driver asks for delayed status
  * @ep0_bounced: true when we used bounce buffer
  * @ep0_expect_in: true when we expect a DATA IN transfer

@@ -1252,6 +1253,7 @@ struct dwc3 {
 	const char		*hsphy_interface;
 
 	unsigned		connected:1;
+	unsigned		softconnect:1;
 	unsigned		delayed_status:1;
 	unsigned		ep0_bounced:1;
 	unsigned		ep0_expect_in:1;
drivers/usb/dwc3/gadget.c +2 -2
index 804b50548163..32e01885b211 100644
@@ -2418,7 +2418,7 @@ static int dwc3_gadget_pullup(struct usb_gadget *g, int is_on)
 	int			ret;
 
 	is_on = !!is_on;
-
+	dwc->softconnect = is_on;
 	/*
 	 * Per databook, when we want to stop the gadget, if a control transfer
 	 * is still in process, complete it and get the core into setup phase.

@@ -4352,7 +4352,7 @@ int dwc3_gadget_resume(struct dwc3 *dwc)
 {
 	int			ret;
 
-	if (!dwc->gadget_driver)
+	if (!dwc->gadget_driver || !dwc->softconnect)
 		return 0;
 
 	ret = __dwc3_gadget_start(dwc);
Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
Acked-by on v1 1/1
#1 lore raw

Hi,

Quoted 19 lines · Wesley Cheng

There is a race present where the DWC3 runtime resume runs in parallel to the UDC unbind sequence. This will eventually lead to a possible scenario where we are enabling the run/stop bit, without a valid composition defined.

Thread#1 (handling UDC unbind):
usb_gadget_remove_driver()
-->usb_gadget_disconnect()
-->dwc3_gadget_pullup(0)
--> continue UDC unbind sequence
-->Thread#2 is running in parallel here

Thread#2 (handing next cable connect)
__dwc3_set_mode()
-->pm_runtime_get_sync()

    -->dwc3_gadget_resume()
      -->dwc->gadget_driver is NOT NULL yet
      -->dwc3_gadget_run_stop(1)
      --> _dwc3gadget_start()
Quoted 7 lines

...

Fix this by tracking the pullup disable routine, and avoiding resuming of the DWC3 gadget. Once the UDC is re-binded, that will trigger the pullup enable routine, which would handle enabling the DWC3 gadget.

Signed-off-by Wesley Cheng <[email protected]>

This looks okay to me, but needs to be tested by a few folks ;-)

Acked-by Felipe Balbi <[email protected]>
Signature · 1 line
balbi
author
#2 lore raw

Hi,

Quoted 31 lines · Felipe Balbi

Hi,

Quoted 26 lines · Wesley Cheng

There is a race present where the DWC3 runtime resume runs in parallel to the UDC unbind sequence. This will eventually lead to a possible scenario where we are enabling the run/stop bit, without a valid composition defined.

Thread#1 (handling UDC unbind):
usb_gadget_remove_driver()
-->usb_gadget_disconnect()
-->dwc3_gadget_pullup(0)
--> continue UDC unbind sequence
-->Thread#2 is running in parallel here

Thread#2 (handing next cable connect)
__dwc3_set_mode()
-->pm_runtime_get_sync()

    -->dwc3_gadget_resume()
      -->dwc->gadget_driver is NOT NULL yet
      -->dwc3_gadget_run_stop(1)
      --> _dwc3gadget_start()

...

Fix this by tracking the pullup disable routine, and avoiding resuming of the DWC3 gadget. Once the UDC is re-binded, that will trigger the pullup enable routine, which would handle enabling the DWC3 gadget.

Signed-off-by Wesley Cheng <[email protected]>

Thanks, Felipe!

Quoted 5 lines

This looks okay to me, but needs to be tested by a few folks ;-)

Acked-by Felipe Balbi <[email protected]>

Yes, would be good to get some functions using usb_gadget_activate/deactivate(). It should be OK for those situations as well, but just to make sure :)

Thanks
Wesley Cheng

Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
#3 lore raw
Quoted 38 lines · Wesley Cheng

Hi,

Quoted 31 lines · Felipe Balbi

Hi,

Quoted 26 lines · Wesley Cheng

There is a race present where the DWC3 runtime resume runs in parallel to the UDC unbind sequence. This will eventually lead to a possible scenario where we are enabling the run/stop bit, without a valid composition defined.

Thread#1 (handling UDC unbind):
usb_gadget_remove_driver()
-->usb_gadget_disconnect()
-->dwc3_gadget_pullup(0)
--> continue UDC unbind sequence
-->Thread#2 is running in parallel here

Thread#2 (handing next cable connect)
__dwc3_set_mode()
-->pm_runtime_get_sync()

    -->dwc3_gadget_resume()
      -->dwc->gadget_driver is NOT NULL yet
      -->dwc3_gadget_run_stop(1)
      --> _dwc3gadget_start()

...

Fix this by tracking the pullup disable routine, and avoiding resuming of the DWC3 gadget. Once the UDC is re-binded, that will trigger the pullup enable routine, which would handle enabling the DWC3 gadget.

Signed-off-by Wesley Cheng <[email protected]>

Thanks, Felipe!

Quoted 1 line
Quoted 7 lines
Quoted 4 lines

This looks okay to me, but needs to be tested by a few folks ;-)

Acked-by Felipe Balbi <[email protected]>

Yes, would be good to get some functions using usb_gadget_activate/deactivate(). It should be OK for those situations as well, but just to make sure :)

IIRC, the UVC function relies on those. You could give it a shot ;-)

Signature · 1 line
balbi