[PATCH v3 0/4] Added some bug fixes for USB4

[PATCH v3 0/4] Added some bug fixes for USB4

[Sanjay R Mehta]

From: Sanjay R Mehta <sanju.mehta@amd.com>

This series adds some general USB4 bug fixes as per USB4 Spec.

v2:
	- Added quirk to handle any vendor specific quirks.
v1: 
	- removed PCI-IDs as AMD USB4 controller exposes USB4 class ID.
	- removed clearing of interrupt using interrupt status clear register.
	- skip port Adapter(0) initialisation for both host router & device router.

Sanjay R Mehta (4):
  thunderbolt: Add quirk to support vendor specific implementation
  thunderbolt: Handle ring interrupt by reading intr status
  thunderbolt: Skip port init for control adapter(0)
  thunderbolt: Fix port linking by checking all adapters

 drivers/thunderbolt/nhi.c    | 32 ++++++++++++++++++++++++++++----
 drivers/thunderbolt/switch.c |  4 ++--
 include/linux/thunderbolt.h  |  1 +
 3 files changed, 31 insertions(+), 6 deletions(-)

-- 
2.7.4

[PATCH v3 1/4] thunderbolt: Add quirk to support vendor specific implementation

[Sanjay R Mehta]

From: Sanjay R Mehta <sanju.mehta@amd.com>

Introduce nhi_check_quirks() routine to handle any vendor specific quirks
to manage a hardware specific implementation.

On Intel system, the USB4 controller requires the
REG_DMA_MISC_INT_AUTO_CLEAR to be set. Hence handle it accordingly as
per the USB4 specification via a quirk.

Fixes: 046bee1f9ab8 ("thunderbolt: Add MSI-X support")
Suggested-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
---
 drivers/thunderbolt/nhi.c   | 18 ++++++++++++++----
 include/linux/thunderbolt.h |  1 +
 2 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/drivers/thunderbolt/nhi.c b/drivers/thunderbolt/nhi.c
index fa44332..7979638 100644
--- a/drivers/thunderbolt/nhi.c
+++ b/drivers/thunderbolt/nhi.c
@@ -43,6 +43,12 @@ static int ring_interrupt_index(struct tb_ring *ring)
 	return bit;
 }
 
+static void nhi_check_quirks(struct tb_nhi *nhi)
+{
+	if (nhi->pdev->vendor == PCI_VENDOR_ID_INTEL)
+		nhi->quirks |= REG_DMA_MISC_INT_AUTO_CLEAR;
+}
+
 /*
  * ring_interrupt_active() - activate/deactivate interrupts for a single ring
  *
@@ -70,10 +76,12 @@ static void ring_interrupt_active(struct tb_ring *ring, bool active)
 		 * Ask the hardware to clear interrupt status bits automatically
 		 * since we already know which interrupt was triggered.
 		 */
-		misc = ioread32(ring->nhi->iobase + REG_DMA_MISC);
-		if (!(misc & REG_DMA_MISC_INT_AUTO_CLEAR)) {
-			misc |= REG_DMA_MISC_INT_AUTO_CLEAR;
-			iowrite32(misc, ring->nhi->iobase + REG_DMA_MISC);
+		if (ring->nhi->quirks & REG_DMA_MISC_INT_AUTO_CLEAR) {
+			misc = ioread32(ring->nhi->iobase + REG_DMA_MISC);
+			if (!(misc & REG_DMA_MISC_INT_AUTO_CLEAR)) {
+				misc |= REG_DMA_MISC_INT_AUTO_CLEAR;
+				iowrite32(misc, ring->nhi->iobase + REG_DMA_MISC);
+			}
 		}
 
 		ivr_base = ring->nhi->iobase + REG_INT_VEC_ALLOC_BASE;
@@ -1190,6 +1198,8 @@ static int nhi_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	if (!nhi->tx_rings || !nhi->rx_rings)
 		return -ENOMEM;
 
+	nhi_check_quirks(nhi);
+
 	res = nhi_init_msi(nhi);
 	if (res) {
 		dev_err(&pdev->dev, "cannot enable MSI, aborting\n");
diff --git a/include/linux/thunderbolt.h b/include/linux/thunderbolt.h
index e7c96c3..2144123 100644
--- a/include/linux/thunderbolt.h
+++ b/include/linux/thunderbolt.h
@@ -480,6 +480,7 @@ struct tb_nhi {
 	bool going_away;
 	struct work_struct interrupt_work;
 	u32 hop_count;
+	u32 quirks;
 };
 
 /**
-- 
2.7.4

[PATCH v3 2/4] thunderbolt: Handle ring interrupt by reading intr status

[Sanjay R Mehta]

From: Sanjay R Mehta <sanju.mehta@amd.com>

As per USB4 specification by default "Disable ISR Auto-Clear" bit is set
to zero and the Tx/Rx ring interrupt status needs to be cleared.

Hence handle it by reading the "Interrupt status" register in the ISR.

Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
---
 drivers/thunderbolt/nhi.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/drivers/thunderbolt/nhi.c b/drivers/thunderbolt/nhi.c
index 7979638..56ac343 100644
--- a/drivers/thunderbolt/nhi.c
+++ b/drivers/thunderbolt/nhi.c
@@ -385,11 +385,25 @@ void tb_ring_poll_complete(struct tb_ring *ring)
 }
 EXPORT_SYMBOL_GPL(tb_ring_poll_complete);
 
+static void check_and_clear_intr_status(struct tb_ring *ring)
+{
+	if (!(ring->nhi->quirks & REG_DMA_MISC_INT_AUTO_CLEAR)) {
+		if (ring->is_tx)
+			ioread32(ring->nhi->iobase
+				 + REG_RING_NOTIFY_BASE);
+		else
+			ioread32(ring->nhi->iobase
+				 + REG_RING_NOTIFY_BASE
+				 + 4 * (ring->nhi->hop_count / 32));
+	}
+}
+
 static irqreturn_t ring_msix(int irq, void *data)
 {
 	struct tb_ring *ring = data;
 
 	spin_lock(&ring->nhi->lock);
+	check_and_clear_intr_status(ring);
 	spin_lock(&ring->lock);
 	__ring_interrupt(ring);
 	spin_unlock(&ring->lock);
-- 
2.7.4

[PATCH v3 3/4] thunderbolt: Skip port init for control adapter(0)

[Sanjay R Mehta]

From: Sanjay R Mehta <sanju.mehta@amd.com>

Adapter (0) is control adapter and as per USB4 spec in "Section 1.8",
"Control Adapters do not have an Adapter Configuration Space".

Hence skip port initialization for adapter (0).

Fixes: e6b245ccd524 ("thunderbolt: Add support for host and device NVM firmware upgrade")
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
---
 drivers/thunderbolt/switch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index 83b1ef3..6447876 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -2740,7 +2740,7 @@ int tb_switch_add(struct tb_switch *sw)
 			return ret;
 		}
 
-		for (i = 0; i <= sw->config.max_port_number; i++) {
+		for (i = 1; i <= sw->config.max_port_number; i++) {
 			if (sw->ports[i].disabled) {
 				tb_port_dbg(&sw->ports[i], "disabled by eeprom\n");
 				continue;
-- 
2.7.4

[PATCH v3 4/4] thunderbolt: Fix port linking by checking all adapters

[Sanjay R Mehta]

From: Sanjay R Mehta <sanju.mehta@amd.com>

In tb_switch_default_link_ports(), while linking of ports,
only odd-numbered ports (1,3,5..) are considered and even-numbered
ports are not considered.

AMD host router has lane adapters at 2 & 3 and link ports at
adapter-2 is not considered due to which lane bonding gets disabled.

Hence added a fix such that all ports are considered during
linking of ports.

Fixes: 0d46c08d1ed4 ("thunderbolt: Add default linking between lane adapters if not provided by DROM")
Signed-off-by: Basavaraj Natikar <Basavaraj.Natikar@amd.com>
Signed-off-by: Sanjay R Mehta <sanju.mehta@amd.com>
---
 drivers/thunderbolt/switch.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
index 6447876..5c3d4bd 100644
--- a/drivers/thunderbolt/switch.c
+++ b/drivers/thunderbolt/switch.c
@@ -2456,7 +2456,7 @@ static void tb_switch_default_link_ports(struct tb_switch *sw)
 {
 	int i;
 
-	for (i = 1; i <= sw->config.max_port_number; i += 2) {
+	for (i = 1; i <= sw->config.max_port_number; i++) {
 		struct tb_port *port = &sw->ports[i];
 		struct tb_port *subordinate;
 
-- 
2.7.4

Re: [PATCH v3 0/4] Added some bug fixes for USB4

[Mika Westerberg]

Hi,

On Fri, Aug 06, 2021 at 11:59:04AM -0500, Sanjay R Mehta wrote:
> From: Sanjay R Mehta <sanju.mehta@amd.com>
> 
> This series adds some general USB4 bug fixes as per USB4 Spec.
> 
> v2:
> 	- Added quirk to handle any vendor specific quirks.
> v1: 
> 	- removed PCI-IDs as AMD USB4 controller exposes USB4 class ID.
> 	- removed clearing of interrupt using interrupt status clear register.
> 	- skip port Adapter(0) initialisation for both host router & device router.
> 
> Sanjay R Mehta (4):
>   thunderbolt: Add quirk to support vendor specific implementation
>   thunderbolt: Handle ring interrupt by reading intr status
>   thunderbolt: Skip port init for control adapter(0)
>   thunderbolt: Fix port linking by checking all adapters

I did some minor tweaks and removed the "Fixes" lines since these are
actually not fixes per se more like improvements :) Then applied all to
thunderbolt.git/next. Please check that they still work on AMD hardware
and make sense in general.

Thanks!

Re: [PATCH v3 0/4] Added some bug fixes for USB4

[Sanjay R Mehta]

On 8/9/2021 5:36 PM, Mika Westerberg wrote:
> [CAUTION: External Email]
> 
> Hi,
> 
> On Fri, Aug 06, 2021 at 11:59:04AM -0500, Sanjay R Mehta wrote:
>> From: Sanjay R Mehta <sanju.mehta@amd.com>
>>
>> This series adds some general USB4 bug fixes as per USB4 Spec.
>>
>> v2:
>>       - Added quirk to handle any vendor specific quirks.
>> v1:
>>       - removed PCI-IDs as AMD USB4 controller exposes USB4 class ID.
>>       - removed clearing of interrupt using interrupt status clear register.
>>       - skip port Adapter(0) initialisation for both host router & device router.
>>
>> Sanjay R Mehta (4):
>>   thunderbolt: Add quirk to support vendor specific implementation
>>   thunderbolt: Handle ring interrupt by reading intr status
>>   thunderbolt: Skip port init for control adapter(0)
>>   thunderbolt: Fix port linking by checking all adapters
> 
> I did some minor tweaks and removed the "Fixes" lines since these are
> actually not fixes per se more like improvements :) Then applied all to
> thunderbolt.git/next. Please check that they still work on AMD hardware
> and make sense in general.
> 

Thanks a lot Mika :). This should work for AMD.

> Thanks!
>