Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1478935 > unrolled thread

[PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure

Back to article view | Back to linux.kernel

Contents

  [PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure Lu Baolu <baolu.lu@linux.intel.com> - 2016-09-08 10:50 +0200
    Re: [PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure Lu Baolu <baolu.lu@linux.intel.com> - 2016-09-08 12:00 +0200
      Re: [PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure Felipe Balbi <felipe.balbi@linux.intel.com> - 2016-09-08 13:10 +0200
    Re: [PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure Felipe Balbi <felipe.balbi@linux.intel.com> - 2016-09-08 12:00 +0200

#1478935 — [PATCH 1/1] usb: dwc3: fix Clear Stall EP command failure

Commit 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear
Stall EP command") sets ClearPendIN bit for all IN endpoints in
v2.60a+ cores no matter which speed mode they're in. This causes
Clear Stall EP command failing on some Intel devices.

In page 539 of 2.60a specification, it says:

"When issuing Clear Stall command for IN endpoints in SuperSpeed
mode, the software must set the "ClearPendIN" bit to '1' to
clear any pending IN transcations, so that the device does not
expect any ACK TP from the host for the data sent earlier."

It's obviously that we only need to apply this rule to those IN
endpoints in SuperSpeed mode.

Fixes: 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear Stall EP command")
Cc: stable@vger.kernel.org # 4.7+
Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
---
 drivers/usb/dwc3/gadget.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
index 7a8d3d8..f1858d6 100644
--- a/drivers/usb/dwc3/gadget.c
+++ b/drivers/usb/dwc3/gadget.c
@@ -348,7 +348,8 @@ static int dwc3_send_clear_stall_ep_cmd(struct dwc3_ep *dep)
 	 * IN transfers due to a mishandled error condition. Synopsys
 	 * STAR 9000614252.
 	 */
-	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A))
+	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A) &&
+	    (dwc->maximum_speed >= USB_SPEED_SUPER))
 		cmd |= DWC3_DEPCMD_CLEARPENDIN;
 
 	memset(&params, 0, sizeof(params));
-- 
2.1.4

[toc] | [next] | [standalone]

#1478999

Hi Felipe,

On 09/08/2016 04:41 PM, Lu Baolu wrote:
> Commit 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear
> Stall EP command") sets ClearPendIN bit for all IN endpoints in
> v2.60a+ cores no matter which speed mode they're in. This causes
> Clear Stall EP command failing on some Intel devices.
>
> In page 539 of 2.60a specification, it says:
>
> "When issuing Clear Stall command for IN endpoints in SuperSpeed
> mode, the software must set the "ClearPendIN" bit to '1' to
> clear any pending IN transcations, so that the device does not
> expect any ACK TP from the host for the data sent earlier."
>
> It's obviously that we only need to apply this rule to those IN
> endpoints in SuperSpeed mode.
>
> Fixes: 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear Stall EP command")
> Cc: stable@vger.kernel.org # 4.7+
> Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
> ---
>  drivers/usb/dwc3/gadget.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index 7a8d3d8..f1858d6 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -348,7 +348,8 @@ static int dwc3_send_clear_stall_ep_cmd(struct dwc3_ep *dep)
>  	 * IN transfers due to a mishandled error condition. Synopsys
>  	 * STAR 9000614252.
>  	 */
> -	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A))
> +	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A) &&
> +	    (dwc->maximum_speed >= USB_SPEED_SUPER))
>  		cmd |= DWC3_DEPCMD_CLEARPENDIN;
>  
>  	memset(&params, 0, sizeof(params));

Oh, my bad.

I'm sending you a patch which hasn't been verified on the real
hardware yet.

My mind goes blank after a whole day's work. Poor me. :-(

I will send you a refreshed one tomorrow. Really sorry about it.

Best regards,
Lu Baolu

[toc] | [prev] | [next] | [standalone]

#1479101

[Multipart message — attachments visible in raw view] — view raw

Hi,

Lu Baolu <baolu.lu@linux.intel.com> writes:
>> Commit 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear
>> Stall EP command") sets ClearPendIN bit for all IN endpoints in
>> v2.60a+ cores no matter which speed mode they're in. This causes
>> Clear Stall EP command failing on some Intel devices.
>>
>> In page 539 of 2.60a specification, it says:
>>
>> "When issuing Clear Stall command for IN endpoints in SuperSpeed
>> mode, the software must set the "ClearPendIN" bit to '1' to
>> clear any pending IN transcations, so that the device does not
>> expect any ACK TP from the host for the data sent earlier."
>>
>> It's obviously that we only need to apply this rule to those IN
>> endpoints in SuperSpeed mode.
>>
>> Fixes: 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear Stall EP command")
>> Cc: stable@vger.kernel.org # 4.7+
>> Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
>> ---
>>  drivers/usb/dwc3/gadget.c | 3 ++-
>>  1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
>> index 7a8d3d8..f1858d6 100644
>> --- a/drivers/usb/dwc3/gadget.c
>> +++ b/drivers/usb/dwc3/gadget.c
>> @@ -348,7 +348,8 @@ static int dwc3_send_clear_stall_ep_cmd(struct dwc3_ep *dep)
>>  	 * IN transfers due to a mishandled error condition. Synopsys
>>  	 * STAR 9000614252.
>>  	 */
>> -	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A))
>> +	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A) &&
>> +	    (dwc->maximum_speed >= USB_SPEED_SUPER))
>>  		cmd |= DWC3_DEPCMD_CLEARPENDIN;
>>  
>>  	memset(&params, 0, sizeof(params));
>
> Oh, my bad.
>
> I'm sending you a patch which hasn't been verified on the real
> hardware yet.
>
> My mind goes blank after a whole day's work. Poor me. :-(

happens to all of us ;-)

> I will send you a refreshed one tomorrow. Really sorry about it.

no problem, thanks

-- 
balbi

[toc] | [prev] | [next] | [standalone]

#1479007

[Multipart message — attachments visible in raw view] — view raw

Hi,

Lu Baolu <baolu.lu@linux.intel.com> writes:
> Commit 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear
> Stall EP command") sets ClearPendIN bit for all IN endpoints in
> v2.60a+ cores no matter which speed mode they're in. This causes
> Clear Stall EP command failing on some Intel devices.

Well, the fact that is fails on some Intel device is, IMO, a
symptom. You should have said that ClearStall command fails on 2.60+
operating in HighSpeed mode. That's the real failure. It doesn't matter
if it's Intel or not, what matters is the *current* speed.

> In page 539 of 2.60a specification, it says:
>
> "When issuing Clear Stall command for IN endpoints in SuperSpeed
> mode, the software must set the "ClearPendIN" bit to '1' to
> clear any pending IN transcations, so that the device does not
> expect any ACK TP from the host for the data sent earlier."
>
> It's obviously that we only need to apply this rule to those IN

s/obviously/obvious/

> endpoints in SuperSpeed mode.

endpoints currently operating in SuperSpeed mode.

> Fixes: 50c763f8c1bac ("usb: dwc3: Set the ClearPendIN bit on Clear Stall EP command")
> Cc: stable@vger.kernel.org # 4.7+

please use proper email format:

Cc: <stable@vger.kernel.org> # v4.7+

> Signed-off-by: Lu Baolu <baolu.lu@linux.intel.com>
> ---
>  drivers/usb/dwc3/gadget.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/usb/dwc3/gadget.c b/drivers/usb/dwc3/gadget.c
> index 7a8d3d8..f1858d6 100644
> --- a/drivers/usb/dwc3/gadget.c
> +++ b/drivers/usb/dwc3/gadget.c
> @@ -348,7 +348,8 @@ static int dwc3_send_clear_stall_ep_cmd(struct dwc3_ep *dep)
>  	 * IN transfers due to a mishandled error condition. Synopsys
>  	 * STAR 9000614252.
>  	 */
> -	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A))
> +	if (dep->direction && (dwc->revision >= DWC3_REVISION_260A) &&
> +	    (dwc->maximum_speed >= USB_SPEED_SUPER))

not sure this is correct. Based on text quoted above, it seems to be me
we should check *current* speed, not maximum speed. We hold current
speed in dwc->gadget.speed.

-- 
balbi

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web