Skip to contentSkip to conversation

usb: host: tegra: Reset Tegra USB controller before init

PATCH Superseded Venu Byravarasu 4 participants 13 messages Feb 28, 2013 – Apr 3, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 usb: host: tegra: Reset Tegra USB controller before init

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

---
When U-Boot configures a Tegra USB controller in device mode and if the EHCI driver of kernel tries to set it to HOST mode, message "irq 52: nobody cared" appears and IRQ gets disabled.

This issue was initially reported with: http://marc.info/?l=linux-tegra&m=136110175423601&w=2

To avoid such issues, due to configurations made by U-Boot driver, reset the Tegra USB controller, before configuring it by kernel.

 drivers/usb/host/ehci-tegra.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)
drivers/usb/host/ehci-tegra.c +5 -0
index 568aecc..83d190a 100644
@@ -28,6 +28,7 @@
 #include <linux/pm_runtime.h>
 #include <linux/usb/ehci_def.h>
 #include <linux/usb/tegra_usb_phy.h>
+#include <linux/clk/tegra.h>
 
 #define TEGRA_USB_BASE			0xC5000000
 #define TEGRA_USB2_BASE			0xC5004000
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device *pdev)
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);
+
 	tegra->needs_double_reset = of_property_read_bool(pdev->dev.of_node,
 		"nvidia,needs-double-reset");
 
Signature · 1 line
1.7.0.4
maintainer
#1 lore raw
Quoted 9 lines · Venu Byravarasu

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

---
When U-Boot configures a Tegra USB controller in device mode and if the EHCI driver of kernel tries to set it to HOST mode, message "irq 52: nobody cared" appears and IRQ gets disabled.

Quoted 4 lines

This issue was initially reported with: http://marc.info/?l=linux-tegra&m=136110175423601&w=2

To avoid such issues, due to configurations made by U-Boot driver, reset the Tegra USB controller, before configuring it by kernel.

Does the Tegra platform use shared interrupts? If it does, what happens if the IRQ is enabled and in use by another device before ehci-tegra resets the USB controller?

Does the unwanted interrupt occur only when the controller is switched to host mode? If not, it seems to me this reset belongs in the platform code, not in the glue layer. If yes, the reset belongs somewher before the controller is switched -- where does that occur?

Alan Stern

#2 lore raw
Quoted 11 lines · Alan Stern
Quoted 9 lines · Venu Byravarasu

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

---
When U-Boot configures a Tegra USB controller in device mode and if the EHCI driver of kernel tries to set it to HOST mode, message "irq 52: nobody cared" appears and IRQ gets disabled.

Quoted 8 lines
Quoted 4 lines

This issue was initially reported with: http://marc.info/?l=linux-tegra&m=136110175423601&w=2

To avoid such issues, due to configurations made by U-Boot driver, reset the Tegra USB controller, before configuring it by kernel.

Does the Tegra platform use shared interrupts? If it does, what happens if the IRQ is enabled and in use by another device before ehci-tegra resets the USB controller?

I believe there's a dedicated interrupt just for each individual controller.

#3 lore raw
In reply to Venu Byravarasu

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device *pdev) · line 697

Quoted 12 lines · Venu Byravarasu

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci-tegra.c +3 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device *pdev)
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

I think the way to solve this is to put the reset call into the PHY driver. I assume it has access to the appropriate clock object. This may also address Alan's query re: when the unexpected interrupt occurs; it's triggered by (or correlated with at least) the PHY (or USB port in general) being in device mode due to the boot ROM setting it up this way, then switching to host mode via the Linux driver. I /think/ that device/host mode switching is more related to the PHY than EHCI driver, although I could well be wrong here.

maintainer
#4 lore raw
Quoted 17 lines · Stephen Warren

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

Quoted 8 lines

I think the way to solve this is to put the reset call into the PHY driver. I assume it has access to the appropriate clock object. This may also address Alan's query re: when the unexpected interrupt occurs; it's triggered by (or correlated with at least) the PHY (or USB port in general) being in device mode due to the boot ROM setting it up this way, then switching to host mode via the Linux driver. I /think/ that device/host mode switching is more related to the PHY than EHCI driver, although I could well be wrong here.

With the PCI platform driver, the handoff from the firmware (we can categorize U-Boot as firmware for this discussion) is handled as soon as the controller is discovered by the platform-specific code. There's a special pci-quirks.c file to take care of it. It is not handled by the driver or the glue layer.

In general I think that's what needs to be done. Errant interrupt sources should be disabled as quickly as possible.

In this case I don't know exactly when the earliest opportunity is. I assume that the EHCI driver and/or the PHY driver gets probed because some platform-layer code has registered the device. If this is so then that platform-layer code is the right place to do the reset.

Alan Stern

#5 lore raw
Quoted 34 lines · Alan Stern
Quoted 17 lines · Stephen Warren

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

Quoted 8 lines

I think the way to solve this is to put the reset call into the PHY driver. I assume it has access to the appropriate clock object. This may also address Alan's query re: when the unexpected interrupt occurs; it's triggered by (or correlated with at least) the PHY (or USB port in general) being in device mode due to the boot ROM setting it up this way, then switching to host mode via the Linux driver. I /think/ that device/host mode switching is more related to the PHY than EHCI driver, although I could well be wrong here.

With the PCI platform driver, the handoff from the firmware (we can categorize U-Boot as firmware for this discussion) is handled as soon as the controller is discovered by the platform-specific code. There's a special pci-quirks.c file to take care of it. It is not handled by the driver or the glue layer.

Quoted 7 lines

In general I think that's what needs to be done. Errant interrupt sources should be disabled as quickly as possible.

In this case I don't know exactly when the earliest opportunity is. I assume that the EHCI driver and/or the PHY driver gets probed because some platform-layer code has registered the device. If this is so then that platform-layer code is the right place to do the reset.

The first platform-specific code that is executed in this case is the (struct platform_driver) probe() functions for the PHY and/or EHCI device.

On Tegra, all devices are listed in device tree, and core kernel code instantiates platform devices based on the information in device tree. The first non-core code that runs on a platform device is the driver's probe() method. That probe() method is for a very specific piece of HW, and hence seems to be the best place to put any quirks for it. Putting the quirks outside the driver would mean some piece of the core DT code would need to have a quirk list, and be able to map the registers for the device, acquire clocks, etc. etc., implement the quirk, and tear it all down again. All of which would just be duplicated with the driver's own probe() function.

#6 lore raw
In reply to Stephen Warren
Quoted 35 lines

-----Original Message-----
From: Stephen Warren [mailto:[email protected]]
Sent: Thursday, February 28, 2013 11:47 PM
To: Venu Byravarasu
Cc: [email protected]; [email protected]; linux- [email protected]; [email protected]
Subject: Re: [PATCH] usb: host: tegra: Reset Tegra USB controller before init

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device

Quoted 6 lines · Venu Byravarasu

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci-tegra.c +0 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device

*pdev)

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device *pdev) · line 697

Quoted 6 lines
drivers/usb/host/ehci-tegra.c +3 -0
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Quoted 4 lines

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

Yes, PHY driver probe does not touch any registers. It just sets up the PHY API hooks. These APIs will be called from ehci-tegra.c as part of ehci tegra probe function, after getting PHY handle, which in turn happens after issuing above reset.

Thanks to Stephen & Alan, for the review comments.

Quoted 3 lines

I think the way to solve this is to put the reset call into the PHY driver. I assume it has access to the appropriate clock object.

#7 lore raw
Quoted 33 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 27 lines · Venu Byravarasu

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device

Quoted 6 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci-tegra.c +0 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device

*pdev)

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device *pdev) · line 697

Quoted 6 lines
drivers/usb/host/ehci-tegra.c +3 -0
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Quoted 4 lines

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

Quoted 5 lines

Yes, PHY driver probe does not touch any registers. It just sets up the PHY API hooks. These APIs will be called from ehci-tegra.c as part of ehci tegra probe function, after getting PHY handle, which in turn happens after issuing above reset.

Thanks to Stephen & Alan, for the review comments.

OK, in that case I have no objection to this patch.

I'd like to hold off on applying this though; I suspect I'll want to take the Tegra USB patches through the Tegra tree rather than the USB tree again for the 3.10 kernel cycle. I think I may have screwed the pooch on the DT binding I set up for the USB controller clocks, and fixing this may require some Tegra DT changes, which would be easiest taken through the Tegra tree, and so to reduce conflicts in the USB code, taking the rest through there migth just be easiest.

Alan, Greg, if you're OK with this patch now, or for any revised version, an Ack so I can take it through the Tegra tree would be great, thanks.

maintainer Acked-by
#8 lore raw
Quoted 27 lines · Stephen Warren
Quoted 7 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 6 lines · Venu Byravarasu
Quoted 4 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci-tegra.c +0 -0
Quoted 8 lines
Quoted 8 lines

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device

Quoted 1 line
drivers/usb/host/ehci-tegra.c +0 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device

*pdev)

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device *pdev) · line 697

Quoted 6 lines
drivers/usb/host/ehci-tegra.c +3 -0
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);

OK, in that case I have no objection to this patch.

I'd like to hold off on applying this though; I suspect I'll want to take the Tegra USB patches through the Tegra tree rather than the USB tree again for the 3.10 kernel cycle. I think I may have screwed the pooch on the DT binding I set up for the USB controller clocks, and fixing this may require some Tegra DT changes, which would be easiest taken through the Tegra tree, and so to reduce conflicts in the USB code, taking the rest through there migth just be easiest.

Quoted 3 lines

Alan, Greg, if you're OK with this patch now, or for any revised version, an Ack so I can take it through the Tegra tree would be great, thanks.

I have no other objections.

Alan Stern

maintainer Acked-by
#9 lore raw
In reply to Stephen Warren
Quoted 50 lines · Stephen Warren
Quoted 33 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 27 lines · Venu Byravarasu

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device

Quoted 6 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci-tegra.c +0 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct platform_device

*pdev)

drivers/usb/host/ehci-tegra.c static int tegra_ehci_probe(struct platform_device *pdev) · line 697

Quoted 6 lines
drivers/usb/host/ehci-tegra.c +3 -0
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);

I think this patch might cause unintended consequences.

When the Tegra PHY code is converted to a driver (i.e. has its own probe), the initial order of execution of the PHY and EHCI driver probes will not be guaranteed.

In particular, since the EHCI probe will attempt to "find" the PHY device, and defer the EHCI probe until it can do so, this guarantees that the PHY's probe() will have completed before EHCI's probe() completes (although EHCI's probe may start running first some number of times, and be retried with -EPROBE_DEFERRED for a variety of reasons).

Quoted 4 lines

Now, if the PHY driver's probe() actually touches HW and sets up some registers, isn't this reset call going to trash any of that register setup? Or, will PHY probe() not touch registers, but only do so during the standard PHY open/init "op"/API calls?

Quoted 5 lines

Yes, PHY driver probe does not touch any registers. It just sets up the PHY API hooks. These APIs will be called from ehci-tegra.c as part of ehci tegra probe function, after getting PHY handle, which in turn happens after issuing above reset.

Thanks to Stephen & Alan, for the review comments.

OK, in that case I have no objection to this patch.

I'd like to hold off on applying this though; I suspect I'll want to take the Tegra USB patches through the Tegra tree rather than the USB tree again for the 3.10 kernel cycle. I think I may have screwed the pooch on the DT binding I set up for the USB controller clocks, and fixing this may require some Tegra DT changes, which would be easiest taken through the Tegra tree, and so to reduce conflicts in the USB code, taking the rest through there migth just be easiest.

Quoted 3 lines

Alan, Greg, if you're OK with this patch now, or for any revised version, an Ack so I can take it through the Tegra tree would be great, thanks.

#10 lore raw
Quoted 29 lines

-----Original Message-----
From: [email protected] [mailto:[email protected]]
Sent: Tuesday, March 05, 2013 6:04 AM
To: Stephen Warren
Cc: Venu Byravarasu; [email protected]; linux- [email protected]; [email protected]
Subject: Re: [PATCH] usb: host: tegra: Reset Tegra USB controller before init

Quoted 7 lines · Stephen Warren
Quoted 6 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 5 lines · Venu Byravarasu

drivers/usb/host/ehci-

Quoted 4 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

drivers/usb/host/ehci- renamed from drivers/usb/host/ehci-tegra.c +0 -0

tegra.c

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

drivers/usb/host/ehci- static int tegra_ehci_probe(struct

Quoted 2 lines
drivers/usb/host/ehci- renamed from drivers/usb/host/ehci-tegra.c +0 -0
@@ -691,6 +692,10 @@ static int tegra_ehci_probe(struct

platform_device

Quoted 8 lines
Quoted 8 lines
Quoted 8 lines

*pdev)

Quoted 6 lines
drivers/usb/host/ehci- renamed from drivers/usb/host/ehci-tegra.c +3 -0
 	if (err)
 		goto fail_clk;
 
+	tegra_periph_reset_assert(tegra->clk);
+	udelay(1);
+	tegra_periph_reset_deassert(tegra->clk);
Quoted 5 lines
Quoted 3 lines

Alan, Greg, if you're OK with this patch now, or for any revised version, an Ack so I can take it through the Tegra tree would be great, thanks.

Hi Greg,

Stephen initially thought that the patch can be taken through tegra tree, along with other patches posted at http://marc.info/?l=linux-tegra&m=136361016003625&w=2 . As some other issues are blocking that patch series, can this change be merged independent of that patch series via USB tree?

The original change can be seen at http://marc.info/?l=linux-usb&m=136203353205291&w=2

Thanks,
Venu

#11 lore raw
Quoted 7 lines · Venu Byravarasu
Quoted 7 lines

[email protected] wrote at Tuesday, March 05, 2013 6:04 AM:

Quoted 5 lines · Stephen Warren
Quoted 4 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 3 lines · Venu Byravarasu
Quoted 2 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

...

Quoted 10 lines
Quoted 1 line

Hi Greg,

Stephen initially thought that the patch can be taken through tegra tree, along with other patches posted at http://marc.info/?l=linux-tegra&m=136361016003625&w=2 . As some other issues are blocking that patch series, can this change be merged independent of that patch series via USB tree?

The original change can be seen at http://marc.info/?l=linux-usb&m=136203353205291&w=2

Venu, you should simply repost the patch; I'm sure its not in anyone's email inbox any more.

#12 lore raw
Quoted 26 lines

-----Original Message-----
From: Stephen Warren [mailto:[email protected]]
Sent: Tuesday, April 02, 2013 9:02 PM
To: Venu Byravarasu
Cc: [email protected]; [email protected]; linux- [email protected]; [email protected]
Subject: Re: [PATCH] usb: host: tegra: Reset Tegra USB controller before init

Quoted 7 lines · Venu Byravarasu
Quoted 7 lines

[email protected] wrote at Tuesday, March 05, 2013 6:04 AM:

Quoted 5 lines · Stephen Warren
Quoted 4 lines · Venu Byravarasu

Stephen Warren wrote at Thursday, February 28, 2013 11:47 PM:

Quoted 3 lines · Venu Byravarasu
Quoted 2 lines

To clear any configurations made by U-Boot on Tegra USB controller, reset it before init in probe.

...

Quoted 5 lines
Quoted 1 line

Hi Greg,

Stephen initially thought that the patch can be taken through tegra tree,

along with

Quoted 1 line

other patches posted at http://marc.info/?l=linux-

tegra&m=136361016003625&w=2 .

Quoted 1 line

As some other issues are blocking that patch series, can this change be

Quoted 8 lines

merged

Quoted 3 lines

independent of that patch series via USB tree?

The original change can be seen at http://marc.info/?l=linux-

usb&m=136203353205291&w=2

Venu, you should simply repost the patch; I'm sure its not in anyone's email inbox any more.

Resent the patch, with ACKs added.
Latest patch can be seen at: http://marc.info/?l=linux-tegra&m=136498598329081&w=2

Thanks,
Venu