[PATCH] usb: pci-quirks: disable D3cold on xhci suspend for s2idle

The XHCI is required enter D3hot rather than D3cold for AMD s2idle solution.
Otherwise, the 'Controller Not Ready' (CNR) bit not being cleared by host
in resume and eventually result in xhci resume failed in s2idle wakeup period.

Signed-off-by: Prike Liang <redacted>
---
 drivers/usb/host/xhci-pci.c | 6 +++++-
 drivers/usb/host/xhci.h     | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
index 3feaafe..8ecde66 100644
--- a/drivers/usb/host/xhci-pci.c
+++ b/drivers/usb/host/xhci-pci.c

@@ -170,6 +170,10 @@ static void xhci_pci_quirks(struct device *dev, struct xhci_hcd *xhci)
 		(pdev->device == PCI_DEVICE_ID_AMD_PROMONTORYA_1)))
 		xhci->quirks |= XHCI_U2_DISABLE_WAKE;
 
+	if (pdev->vendor == PCI_VENDOR_ID_AMD &&
+		pdev->device == 0x1639)
+		xhci->quirks |= XHCI_AMD_S2IDL_SUPPORT_QUIRK;
+
 	if (pdev->vendor == PCI_VENDOR_ID_INTEL) {
 		xhci->quirks |= XHCI_LPM_SUPPORT;
 		xhci->quirks |= XHCI_INTEL_HOST;

@@ -500,7 +504,7 @@ static int xhci_pci_suspend(struct usb_hcd *hcd, bool do_wakeup)
 	 * Systems with the TI redriver that loses port status change events
 	 * need to have the registers polled during D3, so avoid D3cold.
 	 */
-	if (xhci->quirks & XHCI_COMP_MODE_QUIRK)
+	if (xhci->quirks & (XHCI_COMP_MODE_QUIRK | XHCI_AMD_S2IDL_SUPPORT_QUIRK))
 		pci_d3cold_disable(pdev);
 
 	if (xhci->quirks & XHCI_PME_STUCK_QUIRK)

diff --git a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h
index ea1754f..fafa044 100644
--- a/drivers/usb/host/xhci.h
+++ b/drivers/usb/host/xhci.h

@@ -1874,6 +1874,7 @@ struct xhci_hcd {
 #define XHCI_RESET_PLL_ON_DISCONNECT	BIT_ULL(34)
 #define XHCI_SNPS_BROKEN_SUSPEND    BIT_ULL(35)
 #define XHCI_RENESAS_FW_QUIRK	BIT_ULL(36)
+#define XHCI_AMD_S2IDL_SUPPORT_QUIRK   BIT_ULL(37)
 
 	unsigned int		num_active_eps;
 	unsigned int		limit_active_eps;

-- 
2.7.4

This has to be with the patch revision number, i.e. [PATCH v2] , [PATCH v3] etc.

On 2/5/2021 5:02 PM, Prike Liang wrote:

Why don't we keep it the fullname i.e. XHCI_AMD_S2IDLE_SUPPORT_QUIRK
instead of XHCI_AMD_S2IDL_SUPPORT_QUIRK, likewise at all places?

-Shyam

This has to be with the patch revision number, i.e. [PATCH v2] , [PATCH v3]
etc.

On 2/5/2021 5:02 PM, Prike Liang wrote:

quoted

The XHCI is required enter D3hot rather than D3cold for AMD s2idle

solution.

quoted

Otherwise, the 'Controller Not Ready' (CNR) bit not being cleared by
host in resume and eventually result in xhci resume failed in s2idle wakeup

period.

quoted

Signed-off-by: Prike Liang <redacted>
---
 drivers/usb/host/xhci-pci.c | 6 +++++-
 drivers/usb/host/xhci.h     | 1 +
 2 files changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/host/xhci-pci.c b/drivers/usb/host/xhci-pci.c
index 3feaafe..8ecde66 100644
--- a/drivers/usb/host/xhci-pci.c
+++ b/drivers/usb/host/xhci-pci.c

@@ -170,6 +170,10 @@ static void xhci_pci_quirks(struct device *dev,

struct xhci_hcd *xhci)

quoted

 		(pdev->device == PCI_DEVICE_ID_AMD_PROMONTORYA_1)))
 		xhci->quirks |= XHCI_U2_DISABLE_WAKE;

+	if (pdev->vendor == PCI_VENDOR_ID_AMD &&
+		pdev->device == 0x1639)
+		xhci->quirks |= XHCI_AMD_S2IDL_SUPPORT_QUIRK;
+
 	if (pdev->vendor == PCI_VENDOR_ID_INTEL) {
 		xhci->quirks |= XHCI_LPM_SUPPORT;
 		xhci->quirks |= XHCI_INTEL_HOST;

@@ -500,7 +504,7 @@ static int xhci_pci_suspend(struct usb_hcd *hcd,

bool do_wakeup)

quoted

 	 * Systems with the TI redriver that loses port status change events
 	 * need to have the registers polled during D3, so avoid D3cold.
 	 */
-	if (xhci->quirks & XHCI_COMP_MODE_QUIRK)
+	if (xhci->quirks & (XHCI_COMP_MODE_QUIRK |
+XHCI_AMD_S2IDL_SUPPORT_QUIRK))
 		pci_d3cold_disable(pdev);

 	if (xhci->quirks & XHCI_PME_STUCK_QUIRK) diff --git
a/drivers/usb/host/xhci.h b/drivers/usb/host/xhci.h index
ea1754f..fafa044 100644

--- a/drivers/usb/host/xhci.h
+++ b/drivers/usb/host/xhci.h

@@ -1874,6 +1874,7 @@ struct xhci_hcd {
 #define XHCI_RESET_PLL_ON_DISCONNECT	BIT_ULL(34)
 #define XHCI_SNPS_BROKEN_SUSPEND    BIT_ULL(35)
 #define XHCI_RENESAS_FW_QUIRK	BIT_ULL(36)
+#define XHCI_AMD_S2IDL_SUPPORT_QUIRK   BIT_ULL(37)

Why don't we keep it the fullname i.e. XHCI_AMD_S2IDLE_SUPPORT_QUIRK
instead of XHCI_AMD_S2IDL_SUPPORT_QUIRK, likewise at all places?

-Shyam

[Prike]  Thanks for your comment and the primary patch with compliance mode quirk has been dropped and this patch 
seems totally different with compliance mode partial enablement fix. Anyway I will give the patch modified detail in the commit. The XHCI_AMD_S2IDL_SUPPORT_QUIRK is a typo and I will correct it.

Thanks, 
Prike