Skip to contentSkip to conversation

usb: dwc3: gadget: Clear req->needs_extra_trb flag on cleanup

PATCH Superseded Jack Pham 3 participants 7 messages Dec 21, 2018 – Jan 26, 2019 1 message not in the archive

Subsystems (3)
lore
Share & tools
lore

Jack Pham

author
#0 lore raw

PATCH 1/1 usb: dwc3: gadget: Clear req->needs_extra_trb flag on cleanup

OUT endpoint requests may somtimes have this flag set when preparing to be submitted to HW indicating that there is an additional TRB chained to the request for alignment purposes. If that request is removed before the controller can execute the transfer (e.g. ep_dequeue/ep_disable), the request will not go through the dwc3_gadget_ep_cleanup_completed_request() handler and will not have its needs_extra_trb flag cleared when dwc3_gadget_giveback() is called. This same request could be later requeued for a new transfer that does not require an extra TRB and if it is successfully completed, the cleanup and TRB reclamation will incorrectly process the additional TRB which belongs to the next request, and incorrectly advances the TRB dequeue pointer, thereby messing up calculation of the next requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flag is cleared before it is given back to the function driver. A good place to do that is in dwc3_gadget_del_and_unmap_request().

---
Hi Felipe,

There's probably zero chance this is making it to 4.20, so if you take this after 4.21-rc1 so be it. But should this be Cc: stable? If so it needs to be sent separately for <= 4.19 as needs_extra_trb was previously req->unaligned and req->zero.

Thanks,
Jack

 drivers/usb/dwc3/gadget.c | 1 +
 1 file changed, 1 insertion(+)
drivers/usb/dwc3/gadget.c +1 -0
index 2ecde30ad0b7..e97b14f444c8 100644
@@ -177,6 +177,7 @@ static void dwc3_gadget_del_and_unmap_request(struct dwc3_ep *dep,
 	req->started = false;
 	list_del(&req->list);
 	req->remaining = 0;
+	req->needs_extra_trb = false;
 
 	if (req->request.status == -EINPROGRESS)
 		req->request.status = status;
#1 lore raw

Hi,

Quoted 19 lines · Jack Pham

OUT endpoint requests may somtimes have this flag set when preparing to be submitted to HW indicating that there is an additional TRB chained to the request for alignment purposes. If that request is removed before the controller can execute the transfer (e.g. ep_dequeue/ep_disable), the request will not go through the dwc3_gadget_ep_cleanup_completed_request() handler and will not have its needs_extra_trb flag cleared when dwc3_gadget_giveback() is called. This same request could be later requeued for a new transfer that does not require an extra TRB and if it is successfully completed, the cleanup and TRB reclamation will incorrectly process the additional TRB which belongs to the next request, and incorrectly advances the TRB dequeue pointer, thereby messing up calculation of the next requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flag is cleared before it is given back to the function driver. A good place to do that is in dwc3_gadget_del_and_unmap_request().

Quoted 8 lines

---
Hi Felipe,

There's probably zero chance this is making it to 4.20, so if you take this after 4.21-rc1 so be it. But should this be Cc: stable? If so it needs to be sent separately for <= 4.19 as needs_extra_trb was previously req->unaligned and req->zero.

we need a Cc stable, indeed. And a Fixes tag.

author
#2 lore raw

PATCH v2 1/1 usb: dwc3: gadget: Clear req->needs_extra_trb flag on cleanup

OUT endpoint requests may somtimes have this flag set when preparing to be submitted to HW indicating that there is an additional TRB chained to the request for alignment purposes. If that request is removed before the controller can execute the transfer (e.g. ep_dequeue/ep_disable), the request will not go through the dwc3_gadget_ep_cleanup_completed_request() handler and will not have its needs_extra_trb flag cleared when dwc3_gadget_giveback() is called. This same request could be later requeued for a new transfer that does not require an extra TRB and if it is successfully completed, the cleanup and TRB reclamation will incorrectly process the additional TRB which belongs to the next request, and incorrectly advances the TRB dequeue pointer, thereby messing up calculation of the next requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flag is cleared before it is given back to the function driver. A good place to do that is in dwc3_gadget_del_and_unmap_request().

Fixes c6267a51639b ("usb: dwc3: gadget: align transfers to wMaxPacketSize")

---
v2: Added Fixes tag and Cc: stable

Felipe, as I mentioned in the cover for v1, for stable (from 4.11 where c6267a51639b first landed through 4.20), the fix needs to be modified to assign to the separate req->unaligned and req->zero flags in lieu of needs_extra_trb which appeared in 5.0-rc1 in:

	commit 1a22ec643580626f439c8583edafdcc73798f2fb
	Author: Felipe Balbi <[email protected]>
	Date:   Wed Aug 1 13:15:05 2018 +0300

	    usb: dwc3: gadget: combine unaligned and zero flags

Do I need to send a separate patch for <= 4.20 or will you handle it? It's straightforward really, the code change should instead be

+ req->unaligned = false;
+ req->zero = false;

Thanks,
Jack

 drivers/usb/dwc3/gadget.c | 1 +
 1 file changed, 1 insertion(+)
drivers/usb/dwc3/gadget.c +1 -0
index 2ecde30ad0b7..e97b14f444c8 100644
@@ -177,6 +177,7 @@ static void dwc3_gadget_del_and_unmap_request(struct dwc3_ep *dep,
 	req->started = false;
 	list_del(&req->list);
 	req->remaining = 0;
+	req->needs_extra_trb = false;
 
 	if (req->request.status == -EINPROGRESS)
 		req->request.status = status;
author
#4 lore raw
In reply to a message not in the archive

Hi Sasha,

Quoted 12 lines · Sasha Levin

Hi,

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag, fixing commit: c6267a51639b usb: dwc3: gadget: align transfers to wMaxPacketSize.

The bot has tested the following trees: v4.20.2, v4.19.15, v4.14.93.

v4.20.2: Build failed! Errors:
drivers/usb/dwc3/gadget.c:180:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

Quoted 8 lines

v4.19.15: Build failed! Errors:
drivers/usb/dwc3/gadget.c:180:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

v4.14.93: Build failed! Errors:
drivers/usb/dwc3/gadget.c:185:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

How should we proceed with this patch?

I mentioned in my "cover" message of the patch that it needs a slight rewrite for -stable since the 'needs_extra_trb' was only introduced in v5.0-rc1. Now that my patch is in Linus's tree, I can send it separately for -stable; I'll fix up the commit text slightly to refer to the variables used before v5.0.

Thanks,
Jack

Signature · 2 lines
The Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
author
#5 lore raw

PATCH 1/1 usb: dwc3: gadget: Clear req->unaligned and zero flags on cleanup

commit bd6742249b9ca918565e4e3abaa06665e587f4b5 upstream.

OUT endpoint requests may somtimes have one of these flags set when preparing to be submitted to HW indicating that there is an additional TRB chained to the request for alignment purposes. If that request is removed before the controller can execute the transfer (e.g. ep_dequeue/ep_disable), the request will not go through the dwc3_gadget_ep_cleanup_completed_request() handler and will not have its unaligned or zero flag cleared when dwc3_gadget_giveback() is called. This same request could be later requeued for a new transfer that does not require an extra TRB and if it is successfully completed, the cleanup and TRB reclamation will incorrectly process the additional TRB which belongs to the next request, and incorrectly advances the TRB dequeue pointer, thereby messing up calculation of the next requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flags are cleared before it is given back to the function driver. A good place to do that is in dwc3_gadget_del_and_unmap_request().

Fixes c6267a51639b ("usb: dwc3: gadget: align transfers to wMaxPacketSize")

[jackp: backport to <= 4.20: replaced 'needs_extra_trb' with 'unaligned'
and 'zero' members in patch and reworded commit text]

---
 drivers/usb/dwc3/gadget.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/usb/dwc3/gadget.c +2 -0
index 9f92ee03dde7..2a4ea9a1b1e3 100644
@@ -177,6 +177,8 @@ static void dwc3_gadget_del_and_unmap_request(struct dwc3_ep *dep,
 	req->started = false;
 	list_del(&req->list);
 	req->remaining = 0;
+	req->unaligned = false;
+	req->zero = false;
 
 	if (req->request.status == -EINPROGRESS)
 		req->request.status = status;
Signature · 1 line
2.17.0
maintainer
#7 lore raw
In reply to Jack Pham
Quoted 22 lines · Jack Pham

Hi Sasha,

Quoted 19 lines · Sasha Levin

Hi,

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag, fixing commit: c6267a51639b usb: dwc3: gadget: align transfers to wMaxPacketSize.

The bot has tested the following trees: v4.20.2, v4.19.15, v4.14.93.

v4.20.2: Build failed! Errors:
drivers/usb/dwc3/gadget.c:180:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

v4.19.15: Build failed! Errors:
drivers/usb/dwc3/gadget.c:180:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

v4.14.93: Build failed! Errors:
drivers/usb/dwc3/gadget.c:185:5: error: ‘struct dwc3_request’ has no member named ‘needs_extra_trb’

Quoted 7 lines
Quoted 1 line

How should we proceed with this patch?

I mentioned in my "cover" message of the patch that it needs a slight rewrite for -stable since the 'needs_extra_trb' was only introduced in v5.0-rc1. Now that my patch is in Linus's tree, I can send it separately for -stable; I'll fix up the commit text slightly to refer to the variables used before v5.0.

Sorry, the bot is still very much dumb at this point and doesn't read cover letters :)

Thank you!

--
Thanks,
Sasha

maintainer
#6 lore raw

Re: [PATCH 4.14,4.19,4.20] usb: dwc3: gadget: Clear req->unaligned and zero flags on cleanup

In reply to Jack Pham
Quoted 36 lines · Jack Pham

commit bd6742249b9ca918565e4e3abaa06665e587f4b5 upstream.

OUT endpoint requests may somtimes have one of these flags set when preparing to be submitted to HW indicating that there is an additional TRB chained to the request for alignment purposes. If that request is removed before the controller can execute the transfer (e.g. ep_dequeue/ep_disable), the request will not go through the dwc3_gadget_ep_cleanup_completed_request() handler and will not have its unaligned or zero flag cleared when dwc3_gadget_giveback() is called. This same request could be later requeued for a new transfer that does not require an extra TRB and if it is successfully completed, the cleanup and TRB reclamation will incorrectly process the additional TRB which belongs to the next request, and incorrectly advances the TRB dequeue pointer, thereby messing up calculation of the next requeust's actual/remaining count when it completes.

The right thing to do here is to ensure that the flags are cleared before it is given back to the function driver. A good place to do that is in dwc3_gadget_del_and_unmap_request().

Fixes c6267a51639b ("usb: dwc3: gadget: align transfers to wMaxPacketSize")

[jackp: backport to <= 4.20: replaced 'needs_extra_trb' with 'unaligned'
and 'zero' members in patch and reworded commit text]

---
drivers/usb/dwc3/gadget.c | 2 ++

1 file changed, 2 insertions(+)
drivers/usb/dwc3/gadget.c +0 -0
index 9f92ee03dde7..2a4ea9a1b1e3 100644

drivers/usb/dwc3/gadget.c static void dwc3_gadget_del_and_unmap_request(struct dwc3_ep *dep, · line 184

Quoted 8 lines
drivers/usb/dwc3/gadget.c +2 -0
@@ -177,6 +177,8 @@ static void dwc3_gadget_del_and_unmap_request(struct dwc3_ep *dep,
 	req->started = false;
 	list_del(&req->list);
 	req->remaining = 0;
+	req->unaligned = false;
+	req->zero = false;
 
 	if (req->request.status == -EINPROGRESS)
 		req->request.status = status;

Thanks! I'll let Greg pick it up (as the upstream patch was explicitly tagged for stable).

--
Thanks,
Sasha