Skip to contentSkip to conversation

usb: cdns3: ep0: delete the redundant status stage

PATCH In review Peter Chen 3 participants 10 messages Apr 26, 2020 – May 19, 2020 1 message not in the archive

Subsystems (3)
lore
Share & tools
lore

Peter Chen

author
#0 lore raw

PATCH 1/1 usb: cdns3: ep0: delete the redundant status stage

Each setup stage will prepare status stage at cdns3_ep0_setup_phase, it doesn't need to add extra status stage for test mode handling, otherwise, the controller can't enter the test mode. Through the Lecroy bus analyzer log, the controller will always wait status stage even it is prepared by software later than the test mode is set by software. If we comment out the status stage at cdns3_ep0_setup_phase, the controller will not enter test mode even the test mode is set beforehand.

---
 drivers/usb/cdns3/ep0.c | 7 -------
 1 file changed, 7 deletions(-)
drivers/usb/cdns3/ep0.c +0 -7
index e71240b386b4..82645a2a0f52 100644
@@ -332,13 +332,6 @@ static int cdns3_ep0_feature_handle_device(struct cdns3_device *priv_dev,
 		case TEST_K:
 		case TEST_SE0_NAK:
 		case TEST_PACKET:
-			cdns3_ep0_complete_setup(priv_dev, 0, 1);
-			/**
-			 *  Little delay to give the controller some time
-			 * for sending status stage.
-			 * This time should be less then 3ms.
-			 */
-			mdelay(1);
 			cdns3_set_register_bit(&priv_dev->regs->usb_cmd,
 					       USB_CMD_STMODE |
 					       USB_STS_TMODE_SEL(tmode - 1));
Signature · 1 line
2.17.1
#1 lore raw

Hi,

Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

	if (result < 0)
		cdns3_ep0_complete_setup(priv_dev, 1, 1);
	else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
		cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

Quoted 5 lines

it doesn't need to add extra status stage for test mode handling, otherwise, the controller can't enter the test mode. Through the Lecroy bus analyzer log, the controller will always wait status stage even it is prepared by software later than the test mode is set by software. If we comment out the status stage at cdns3_ep0_setup_phase,

It seems that what you're trying to say here is that the controller will enter test mode only after STATUS stage completes even though SW has already enabled the relevant bits in the register. Is that a correct read of your sentence?

Is this backed by documentation or is this something that just happens to work? Pawell, can you confirm that this is the correct programming model?

The way this usually works is that SW must ensure that Test Mode is only entered after STATUS stage has completed. If this controller works differently, then we should update the comment (rather than removing it) and put a reference to documentation section which describes this very fact.

Quoted 2 lines

the controller will not enter test mode even the test mode is set beforehand.

Sorry for being skeptical, but thinking from a HW design perspective, this would mean that HW would latch Test Mode bits and only, actually, operate on them after completion of STATUS stage. This means that HW should have an internal "status_stage_completed" flag which gets ANDed with another "must_enter_test_mode" internal flag, then only if both conditions are true, will HW read back Test Mode number from the register file and enter correct test mode.

Is this working against USBCV? What about LeCroy's compliance suite?

cheers

Signature · 1 line
balbi

A message that is not in the archive (sent off-list or not yet received): CAL411-q4euWFrJ5Sp+tocBEsXXYkviQXt_pz_SyHHC=ELNf_sQ@mail.gmail.com

#3 lore raw

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 7 lines
Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage,
it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request
service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 22 lines
Quoted 6 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

Quoted 9 lines
        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

At cdns3_ep0_setup_phase, the status will be CDNS3_DATA_STAGE or CDNS3_STATUS_STAGE according to if there is a data stage. If there is a data stage, it will inform of controller ACKing the status stage after data stage has finished, it is at: ep0.c, cdns3_transfer_completed->cdns3_ep0_complete_setup

Quoted 3 lines

But I don't know why it needs to send ERDY for ep0 transfer without data stage, but do need for ep0 transfer with data stage. Maybe Pawel could explain it. At spec, it only says below for ERDY:

Would be good, indeed. Pawel?

Quoted 6 lines
Quoted 4 lines

Is this backed by documentation or is this something that just happens to work? Pawell, can you confirm that this is the correct programming model?

No documentation, maybe Pawel could confirm with designer.

yeah, Pawel?

Quoted 11 lines
Quoted 2 lines

Is this working against USBCV? What about LeCroy's compliance suite?

For NXP USB certification flow:

The test mode is only used for USB2 electrical test (eg, Eye diagram), The HSETT tool is used for device mode which will send command from Windows PC to let device enter test mode.

USBCV is used to test protocol level, like USB CH9, Mass Storage protocol etc.

Entering test modes is part of chapter9 tests, IIRC.

Quoted 2 lines

For Lecroy's compliance suite, we usually use it for Link test for USB3.

You need to run the HS compliance suite ;-) USB3 devices must still comply with HS/FS/LS.

Signature · 1 line
balbi
author
#4 lore raw
Quoted 51 lines · Felipe Balbi

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 7 lines
Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage,
it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request
service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 22 lines
Quoted 6 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

Quoted 9 lines
        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

At cdns3_ep0_setup_phase, the status will be CDNS3_DATA_STAGE or CDNS3_STATUS_STAGE according to if there is a data stage. If there is a data stage, it will inform of controller ACKing the status stage after data stage has finished, it is at: ep0.c, cdns3_transfer_completed->cdns3_ep0_complete_setup

Quoted 6 lines
Quoted 3 lines

But I don't know why it needs to send ERDY for ep0 transfer without data stage, but do need for ep0 transfer with data stage. Maybe Pawel could explain it. At spec, it only says below for ERDY:

Would be good, indeed. Pawel?

Pawel seems not at ML about 2-3 months.

Quoted 14 lines
Quoted 6 lines
Quoted 4 lines

Is this backed by documentation or is this something that just happens to work? Pawell, can you confirm that this is the correct programming model?

No documentation, maybe Pawel could confirm with designer.

yeah, Pawel?

Quoted 5 lines
Quoted 2 lines

Is this working against USBCV? What about LeCroy's compliance suite?

For NXP USB certification flow:

Quoted 8 lines
Quoted 6 lines

The test mode is only used for USB2 electrical test (eg, Eye diagram), The HSETT tool is used for device mode which will send command from Windows PC to let device enter test mode.

USBCV is used to test protocol level, like USB CH9, Mass Storage protocol etc.

Entering test modes is part of chapter9 tests, IIRC.

No, test mode is only used for electrical signal test, the communication between device and host will be stopped once the controller enters test mode.

Quoted 6 lines
Quoted 3 lines

For Lecroy's compliance suite, we usually use it for Link test for USB3.

You need to run the HS compliance suite ;-) USB3 devices must still comply with HS/FS/LS.

No HS compliance suite. Lecroy Compliance Suite are only used for link test and PD test. I attached the test items for compliance test.

Signature · 2 lines
Thanks,
Peter Chen
maintainer
#5 lore raw

Hi,

Your changes looks like:
- cdns3_ep0_complete_setup(priv_dev, 0, 1);
- /**
- * Little delay to give the controller some time
- * for sending status stage.
- * This time should be less then 3ms.
- */
- mdelay(1);
cdns3_set_register_bit(&priv_dev->regs->usb_cmd,
USB_CMD_STMODE | USB_STS_TMODE_SEL(tmode - 1));

I'm not sure whether the first line is necessary in this place " cdns3_ep0_complete_setup(priv_dev, 0, 1);" . Maybe it worked accidentally on my board.

I will confirm it with someone from RTL team, but as I remember the status stage for Test Mode is send automatically by controller. mdelay(1) was added to give controller some time for sending Status Stage. Status Stage should be send before driver send appropriate Test Mode in usb_cmd register. If you remove mdelay it can works accidentally on your board.

Currently I don't have access for testing board so I can't recreate this test again.

Quoted 37 lines · Felipe Balbi
Quoted 35 lines

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 7 lines
Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage,
it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request
service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 6 lines
Quoted 6 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

When this case when happen ?

BTW. I see one issue in cdns3_ep0_complete_setup.

The last one in this function is incorrect:

		cdns3_allow_enable_l1(priv_dev, 1);
should be: 
		/* Resume controller before arming transfer. */
		__cdns3_gadget_wakeup(priv_dev);

Regards,
Pawel,

Quoted 59 lines
Quoted 8 lines
Quoted 8 lines
Quoted 1 line

At cdns3_ep0_setup_phase, the status will be CDNS3_DATA_STAGE or CDNS3_STATUS_STAGE according to if there is a data stage. If there is a data stage, it will inform of controller ACKing the status stage after data stage has finished, it is at: ep0.c, cdns3_transfer_completed->cdns3_ep0_complete_setup

Quoted 6 lines
Quoted 3 lines

But I don't know why it needs to send ERDY for ep0 transfer without data stage, but do need for ep0 transfer with data stage. Maybe Pawel could explain it. At spec, it only says below for ERDY:

Would be good, indeed. Pawel?

Pawel seems not at ML about 2-3 months.

Quoted 14 lines
Quoted 6 lines
Quoted 4 lines

Is this backed by documentation or is this something that just happens to work? Pawell, can you confirm that this is the correct programming model?

No documentation, maybe Pawel could confirm with designer.

yeah, Pawel?

Quoted 5 lines
Quoted 2 lines

Is this working against USBCV? What about LeCroy's compliance suite?

For NXP USB certification flow:

Quoted 8 lines
Quoted 6 lines

The test mode is only used for USB2 electrical test (eg, Eye diagram), The HSETT tool is used for device mode which will send command from Windows PC to let device enter test mode.

USBCV is used to test protocol level, like USB CH9, Mass Storage protocol etc.

Entering test modes is part of chapter9 tests, IIRC.

No, test mode is only used for electrical signal test, the communication between device and host will be stopped once the controller enters test mode.

Quoted 6 lines
Quoted 3 lines

For Lecroy's compliance suite, we usually use it for Link test for USB3.

You need to run the HS compliance suite ;-) USB3 devices must still comply with HS/FS/LS.

No HS compliance suite. Lecroy Compliance Suite are only used for link test and PD test. I attached the test items for compliance test.

--

Thanks,
Peter Chen

author
#6 lore raw
Quoted 14 lines · Pawel Laszczak

Hi,

Your changes looks like:

drivers/usb/cdns3/ep0.c +0 -7
-			cdns3_ep0_complete_setup(priv_dev, 0, 1);
-			/**
-			 *  Little delay to give the controller some time
-			 * for sending status stage.
-			 * This time should be less then 3ms.
-			 */
-			mdelay(1);
 			cdns3_set_register_bit(&priv_dev->regs->usb_cmd,
 					       USB_CMD_STMODE |
 					       USB_STS_TMODE_SEL(tmode - 1));
Quoted 5 lines

I'm not sure whether the first line is necessary in this place " cdns3_ep0_complete_setup(priv_dev, 0, 1);" . Maybe it worked accidentally on my board.

I will confirm it with someone from RTL team, but as I remember the status stage for Test Mode is send automatically by controller.

Pawel, would you please confirm it with your design team? I tried to comment out cdns3_ep0_complete_setup at cdns3_ep0_setup_phase for test mode case, the test mode can't work. It means it still needs software setting to notify controller to prepare status stage.

Quoted 3 lines

mdelay(1) was added to give controller some time for sending
Status Stage. Status Stage should be send before driver send appropriate Test Mode in usb_cmd register. If you remove mdelay it can works accidentally on your board.

If without cdns3_ep0_complete_setup(priv_dev, 0, 1) before mdelay, how the controller knows it needs to prepare status stage, the test mode setting code is behind than it.

The focus is: my board can't enter test mode if status stage is prepared before set test mode, it is strange. At other controller, eg, chipidea, we set test mode after receiving the status stage completion interrupt, but at cdns3, there is no status stage completion interrupt.

Quoted 51 lines

Currently I don't have access for testing board so I can't recreate this test again.

Quoted 37 lines · Felipe Balbi
Quoted 35 lines

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 7 lines
Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage,
it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request
service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 6 lines
Quoted 6 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

When this case when happen ?

Quoted 8 lines

BTW. I see one issue in cdns3_ep0_complete_setup.

The last one in this function is incorrect:

		cdns3_allow_enable_l1(priv_dev, 1);
should be: 
		/* Resume controller before arming transfer. */
		__cdns3_gadget_wakeup(priv_dev);

Would you please send a patch to fix it?

Peter

Signature · 2 lines
Thanks,
Peter Chen
maintainer
#7 lore raw
Quoted 19 lines · Pawel Laszczak
Quoted 17 lines

Hi,

Your changes looks like:

(quoted diff) +0 -7
-			cdns3_ep0_complete_setup(priv_dev, 0, 1);
-			/**
-			 *  Little delay to give the controller some time
-			 * for sending status stage.
-			 * This time should be less then 3ms.
-			 */
-			mdelay(1);
 			cdns3_set_register_bit(&priv_dev->regs->usb_cmd,
 					       USB_CMD_STMODE |
 					       USB_STS_TMODE_SEL(tmode - 1));

I'm not sure whether the first line is necessary in this place " cdns3_ep0_complete_setup(priv_dev, 0, 1);" . Maybe it worked accidentally on my board.

Quoted 7 lines
Quoted 2 lines

I will confirm it with someone from RTL team, but as I remember the status stage for Test Mode is send automatically by controller.

Pawel, would you please confirm it with your design team? I tried to comment out cdns3_ep0_complete_setup at cdns3_ep0_setup_phase for test mode case, the test mode can't work. It means it still needs software setting to notify controller to prepare status stage.

Sorry. I got something wrong. You have right. Driver must prepare Status Stage itself. It's means that origin code is really incorrect and driver sends the second Status Stage twice.

Hardware team confirm also that setting Test Mode in usb_cmd can be done before sending Status Stage. Test Mode should start automatically after sending Status Stage by controller.

Quoted 75 lines
Quoted 4 lines

mdelay(1) was added to give controller some time for sending
Status Stage. Status Stage should be send before driver send appropriate Test Mode in usb_cmd register. If you remove mdelay it can works accidentally on your board.

If without cdns3_ep0_complete_setup(priv_dev, 0, 1) before mdelay, how the controller knows it needs to prepare status stage, the test mode setting code is behind than it.

The focus is: my board can't enter test mode if status stage is prepared before set test mode, it is strange. At other controller, eg, chipidea, we set test mode after receiving the status stage completion interrupt, but at cdns3, there is no status stage completion interrupt.

Quoted 51 lines

Currently I don't have access for testing board so I can't recreate this test again.

Quoted 37 lines · Felipe Balbi
Quoted 35 lines

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 7 lines
Quoted 1 line · Peter Chen

Each setup stage will prepare status stage at cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage,
it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request
service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 6 lines
Quoted 6 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
Quoted 8 lines
        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state != CNDS3_STATUS_STAGE? Seems like you should have a "stall and restart" somewhere here as a default fallback.

When this case when happen ?

Quoted 8 lines

BTW. I see one issue in cdns3_ep0_complete_setup.

The last one in this function is incorrect:

		cdns3_allow_enable_l1(priv_dev, 1);
should be:
		/* Resume controller before arming transfer. */
		__cdns3_gadget_wakeup(priv_dev);
Quoted 8 lines

Would you please send a patch to fix it?

Peter

--

Thanks,
Peter Chen

Pawel

author
#8 lore raw
Quoted 6 lines
Quoted 5 lines

Pawel, would you please confirm it with your design team? I tried to comment out cdns3_ep0_complete_setup at cdns3_ep0_setup_phase for test mode case, the test mode can't work. It means it still needs software setting to notify controller to prepare status stage.

Quoted 10 lines

Sorry. I got something wrong. You have right. Driver must prepare Status Stage itself.
It's means that origin code is really incorrect and driver sends the second Status Stage twice.

Hardware team confirm also that setting Test Mode in usb_cmd can be done before sending Status Stage. Test Mode should start automatically after sending Status Stage by controller.

Thanks, Pawel.

If this patch is ok, could you please add your reviewed-by?

Peter

Quoted 85 lines
Quoted 4 lines
Quoted 4 lines

mdelay(1) was added to give controller some time for sending Status Stage. Status Stage should be send before driver send appropriate Test Mode in usb_cmd register. If you remove mdelay it can works accidentally

on your board.

Quoted 56 lines

If without cdns3_ep0_complete_setup(priv_dev, 0, 1) before mdelay, how the controller knows it needs to prepare status stage, the test mode setting code is behind than it.

The focus is: my board can't enter test mode if status stage is prepared before set test mode, it is strange. At other controller, eg, chipidea, we set test mode after receiving the status stage completion interrupt, but at cdns3, there is no status stage completion interrupt.

Quoted 46 lines

Currently I don't have access for testing board so I can't recreate this test again.

Quoted 43 lines · Felipe Balbi
Quoted 41 lines

Hi,

Quoted 19 lines · Peter Chen

It seems the yesterday's reply from nxp email account is blocked by ML. Send it again.

Quoted 8 lines
Quoted 2 lines · Peter Chen

Each setup stage will prepare status stage at
cdns3_ep0_setup_phase,

care to explain this a little better? The controller itself will produce the status stage?

Unlike the other controllers, the CDNS3 does not need to prepare TD for status stage, it only needs to write register bits EP_CMD.REQ_CMPL to tell the controller the request service is completed, and the controller itself will send ACK answer for the Status Stage after that.
The code sequence like below:

Quoted 3 lines

cdns3_ep0_setup_phase -> cdns3_ep0_complete_setup ->

            writel((send_erdy ? EP_CMD_ERDY : 0) | EP_CMD_REQ_CMPL,
                   &priv_dev->regs->ep_cmd);

got it.

Quoted 12 lines
Quoted 12 lines

Usually, the way this works is that SETUP stage must be *always* prepared by the SW while STATUS stage is prepared on-demand, after we get an interrupt from the controller.

Also, I see a possible bug in cdns3_ep0_setup_phase():

        if (result < 0)
                cdns3_ep0_complete_setup(priv_dev, 1, 1);
        else if (priv_dev->ep0_stage == CDNS3_STATUS_STAGE)
                cdns3_ep0_complete_setup(priv_dev, 0, 1);

What happens here if result is 0 but ep0_state !=

CNDS3_STATUS_STAGE?

Quoted 13 lines
Quoted 12 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

Seems like you should have a "stall and restart" somewhere here as a default fallback.

When this case when happen ?

BTW. I see one issue in cdns3_ep0_complete_setup.

The last one in this function is incorrect:

		cdns3_allow_enable_l1(priv_dev, 1); should be:
		/* Resume controller before arming transfer. */
		__cdns3_gadget_wakeup(priv_dev);
Quoted 8 lines

Would you please send a patch to fix it?

Peter

--

Thanks,
Peter Chen

Pawel

#9 lore raw
In reply to Peter Chen

Hi,

Quoted 9 lines · Peter Chen
Quoted 9 lines
Quoted 9 lines
Quoted 2 lines

Is this working against USBCV? What about LeCroy's compliance suite?

For NXP USB certification flow:

The test mode is only used for USB2 electrical test (eg, Eye diagram), The HSETT tool is used for device mode which will send command from Windows PC to let device enter test mode.

Quoted 8 lines
Quoted 4 lines
Quoted 2 lines

USBCV is used to test protocol level, like USB CH9, Mass Storage protocol etc.

Entering test modes is part of chapter9 tests, IIRC.

No, test mode is only used for electrical signal test, the communication between device and host will be stopped once the controller enters test mode.

but it is a standard request, part of chapter 9 ;-)

Quoted 8 lines
Quoted 5 lines
Quoted 2 lines

For Lecroy's compliance suite, we usually use it for Link test for USB3.

You need to run the HS compliance suite ;-) USB3 devices must still comply with HS/FS/LS.

No HS compliance suite. Lecroy Compliance Suite are only used for link test and PD test. I attached the test items for compliance test.

hmm, maybe they dropped it over the years. I remember using their suite for this very reason. Oh well.

Signature · 1 line
balbi
maintainer Reviewed-by on v1 1/1
#10 lore raw
In reply to Peter Chen

Hi,

Quoted 11 lines

Each setup stage will prepare status stage at cdns3_ep0_setup_phase, it doesn't need to add extra status stage for test mode handling, otherwise, the controller can't enter the test mode. Through the Lecroy bus analyzer log, the controller will always wait status stage even it is prepared by software later than the test mode is set by software. If we comment out the status stage at cdns3_ep0_setup_phase, the controller will not enter test mode even the test mode is set beforehand.

Quoted 24 lines

---
drivers/usb/cdns3/ep0.c | 7 -------

1 file changed, 7 deletions(-)
drivers/usb/cdns3/ep0.c +0 -8
index e71240b386b4..82645a2a0f52 100644
@@ -332,13 +332,6 @@ static int cdns3_ep0_feature_handle_device(struct cdns3_device *priv_dev,
 		case TEST_K:
 		case TEST_SE0_NAK:
 		case TEST_PACKET:
-			cdns3_ep0_complete_setup(priv_dev, 0, 1);
-			/**
-			 *  Little delay to give the controller some time
-			 * for sending status stage.
-			 * This time should be less then 3ms.
-			 */
-			mdelay(1);
 			cdns3_set_register_bit(&priv_dev->regs->usb_cmd,
 					       USB_CMD_STMODE |
 					       USB_STS_TMODE_SEL(tmode - 1));
--

2.17.1