Skip to contentSkip to conversation

PCI: designware-ep: Fix DBI access before core init

PATCH In review Vidya Sagar 3 participants 13 messages Oct 13, 2022 – Aug 25, 2023

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

v5 cover PCI: designware-ep: Fix DBI access before core init

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
 drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
 4 files changed, 97 insertions(+), 69 deletions(-)
Signature · 1 line
2.17.1
#1 lore raw

PATCH v5 1/3 PCI: designware-ep: Fix DBI access before core init

Platforms that cannot support their core initialization without the reference clock from the host, implement the feature 'core_init_notifier' to indicate the DesignWare sub-system about when their core is getting initialized. Any accesses to the core (Ex:- DBI) would the core being ready result in system hang in such systems (Ex:- tegra194). This patch moves any access to the core to dw_pcie_ep_init_complete() API which is effectively called only after the core initialization. It also introduces .ep_init_late() ops hook to be used for any post init work that platform drivers may have to do.

---
V5:
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Moved dw_pcie_ep_init_complete() inside dw_pcie_ep_init_notify()
* Added .ep_init_late() ops to perform late init tasks

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 2 files changed, 80 insertions(+), 55 deletions(-)
drivers/pci/controller/dwc/pcie-designware-ep.c +77 -48
index 83ddb190292e..f300ea2f7bf7 100644
@@ -23,14 +23,6 @@ void dw_pcie_ep_linkup(struct dw_pcie_ep *ep)
 }
 EXPORT_SYMBOL_GPL(dw_pcie_ep_linkup);
 
-void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
-{
-	struct pci_epc *epc = ep->epc;
-
-	pci_epc_init_notify(epc);
-}
-EXPORT_SYMBOL_GPL(dw_pcie_ep_init_notify);
-
 struct dw_pcie_ep_func *
 dw_pcie_ep_get_func_from_ep(struct dw_pcie_ep *ep, u8 func_no)
 {
@@ -640,12 +632,17 @@ static unsigned int dw_pcie_ep_find_ext_capability(struct dw_pcie *pci, int cap)
 	return 0;
 }
 
-int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
+static int dw_pcie_ep_init_late(struct dw_pcie_ep *ep)
 {
 	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
+	struct dw_pcie_ep_func *ep_func;
+	struct device *dev = pci->dev;
+	struct pci_epc *epc = ep->epc;
 	unsigned int offset;
 	unsigned int nbars;
 	u8 hdr_type;
+	u8 func_no;
+	void *addr;
 	u32 reg;
 	int i;
 
@@ -658,6 +655,51 @@ int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
 		return -EIO;
 	}
 
+	dw_pcie_version_detect(pci);
+
+	dw_pcie_iatu_detect(pci);
+
+	if (!ep->ib_window_map) {
+		ep->ib_window_map = devm_bitmap_zalloc(dev, pci->num_ib_windows,
+						       GFP_KERNEL);
+		if (!ep->ib_window_map)
+			return -ENOMEM;
+	}
+
+	if (!ep->ob_window_map) {
+		ep->ob_window_map = devm_bitmap_zalloc(dev, pci->num_ob_windows,
+						       GFP_KERNEL);
+		if (!ep->ob_window_map)
+			return -ENOMEM;
+	}
+
+	if (!ep->outbound_addr) {
+		addr = devm_kcalloc(dev, pci->num_ob_windows, sizeof(phys_addr_t),
+				    GFP_KERNEL);
+		if (!addr)
+			return -ENOMEM;
+		ep->outbound_addr = addr;
+	}
+
+	for (func_no = 0; func_no < epc->max_functions; func_no++) {
+
+		ep_func = dw_pcie_ep_get_func_from_ep(ep, func_no);
+		if (ep_func)
+			continue;
+
+		ep_func = devm_kzalloc(dev, sizeof(*ep_func), GFP_KERNEL);
+		if (!ep_func)
+			return -ENOMEM;
+
+		ep_func->func_no = func_no;
+		ep_func->msi_cap = dw_pcie_ep_find_capability(ep, func_no,
+							      PCI_CAP_ID_MSI);
+		ep_func->msix_cap = dw_pcie_ep_find_capability(ep, func_no,
+							       PCI_CAP_ID_MSIX);
+
+		list_add_tail(&ep_func->list, &ep->func_list);
+	}
+
 	offset = dw_pcie_ep_find_ext_capability(pci, PCI_EXT_CAP_ID_REBAR);
 
 	dw_pcie_dbi_ro_wr_en(pci);
@@ -676,13 +718,28 @@ int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
 
 	return 0;
 }
-EXPORT_SYMBOL_GPL(dw_pcie_ep_init_complete);
+
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
+{
+	struct pci_epc *epc = ep->epc;
+	int ret;
+
+	ret = dw_pcie_ep_init_late(ep);
+	if (ret)
+		return ret;
+
+	if (ep->ops->ep_init_late)
+		ep->ops->ep_init_late(ep);
+
+	pci_epc_init_notify(epc);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dw_pcie_ep_init_notify);
 
 int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 {
 	int ret;
-	void *addr;
-	u8 func_no;
 	struct resource *res;
 	struct pci_epc *epc;
 	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
@@ -690,7 +747,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	struct platform_device *pdev = to_platform_device(dev);
 	struct device_node *np = dev->of_node;
 	const struct pci_epc_features *epc_features;
-	struct dw_pcie_ep_func *ep_func;
 
 	INIT_LIST_HEAD(&ep->func_list);
 
@@ -719,26 +775,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	ep->phys_base = res->start;
 	ep->addr_size = resource_size(res);
 
-	dw_pcie_version_detect(pci);
-
-	dw_pcie_iatu_detect(pci);
-
-	ep->ib_window_map = devm_bitmap_zalloc(dev, pci->num_ib_windows,
-					       GFP_KERNEL);
-	if (!ep->ib_window_map)
-		return -ENOMEM;
-
-	ep->ob_window_map = devm_bitmap_zalloc(dev, pci->num_ob_windows,
-					       GFP_KERNEL);
-	if (!ep->ob_window_map)
-		return -ENOMEM;
-
-	addr = devm_kcalloc(dev, pci->num_ob_windows, sizeof(phys_addr_t),
-			    GFP_KERNEL);
-	if (!addr)
-		return -ENOMEM;
-	ep->outbound_addr = addr;
-
 	if (pci->link_gen < 1)
 		pci->link_gen = of_pci_get_max_link_speed(np);
 
@@ -755,20 +791,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	if (ret < 0)
 		epc->max_functions = 1;
 
-	for (func_no = 0; func_no < epc->max_functions; func_no++) {
-		ep_func = devm_kzalloc(dev, sizeof(*ep_func), GFP_KERNEL);
-		if (!ep_func)
-			return -ENOMEM;
-
-		ep_func->func_no = func_no;
-		ep_func->msi_cap = dw_pcie_ep_find_capability(ep, func_no,
-							      PCI_CAP_ID_MSI);
-		ep_func->msix_cap = dw_pcie_ep_find_capability(ep, func_no,
-							       PCI_CAP_ID_MSIX);
-
-		list_add_tail(&ep_func->list, &ep->func_list);
-	}
-
 	if (ep->ops->ep_init)
 		ep->ops->ep_init(ep);
 
@@ -793,7 +815,14 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 			return 0;
 	}
 
-	ret = dw_pcie_ep_init_complete(ep);
+	/*
+	 * NOTE:- Avoid accessing the hardware (Ex:- DBI space) before this
+	 * step as platforms that implement 'core_init_notifier' feature may
+	 * not have the hardware ready (i.e. core initialized) for access
+	 * (Ex: tegra194). Any hardware access on such platforms result
+	 * in system hard hang.
+	 */
+	ret = dw_pcie_ep_init_late(ep);
 	if (ret)
 		goto err_free_epc_mem;
 
drivers/pci/controller/dwc/pcie-designware.h +3 -7
index 45fcdfc8c035..7252513956b7 100644
@@ -253,6 +253,7 @@ struct dw_pcie_rp {
 
 struct dw_pcie_ep_ops {
 	void	(*ep_init)(struct dw_pcie_ep *ep);
+	void	(*ep_init_late)(struct dw_pcie_ep *ep);
 	int	(*raise_irq)(struct dw_pcie_ep *ep, u8 func_no,
 			     enum pci_epc_irq_type type, u16 interrupt_num);
 	const struct pci_epc_features* (*get_features)(struct dw_pcie_ep *ep);
@@ -467,8 +468,7 @@ static inline void __iomem *dw_pcie_own_conf_map_bus(struct pci_bus *bus,
 #ifdef CONFIG_PCIE_DW_EP
 void dw_pcie_ep_linkup(struct dw_pcie_ep *ep);
 int dw_pcie_ep_init(struct dw_pcie_ep *ep);
-int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep);
-void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
 void dw_pcie_ep_exit(struct dw_pcie_ep *ep);
 int dw_pcie_ep_raise_legacy_irq(struct dw_pcie_ep *ep, u8 func_no);
 int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
@@ -490,15 +490,11 @@ static inline int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	return 0;
 }
 
-static inline int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
+static inline int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
 {
 	return 0;
 }
 
-static inline void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
-{
-}
-
 static inline void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
 {
 }
Signature · 1 line
2.17.1
maintainer Acked-by on v5 1/3
#2 lore raw
Quoted 18 lines · On Thu, Oct 13, 2022 Vidya Sagar <[email protected]> wrote:

Platforms that cannot support their core initialization without the reference clock from the host, implement the feature 'core_init_notifier' to indicate the DesignWare sub-system about when their core is getting initialized. Any accesses to the core (Ex:- DBI) would the core being ready result in system hang in such systems (Ex:- tegra194). This patch moves any access to the core to dw_pcie_ep_init_complete() API which is effectively called only after the core initialization. It also introduces .ep_init_late() ops hook to be used for any post init work that platform drivers may have to do.

---
V5:
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

Quoted 8 lines

V4:
* Addressed review comments from Bjorn and Manivannan
* Moved dw_pcie_ep_init_complete() inside dw_pcie_ep_init_notify()
* Added .ep_init_late() ops to perform late init tasks

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 2 files changed, 80 insertions(+), 55 deletions(-)

The function name 'dw_pcie_ep_init_late' looks clear to me. Also, commit messages
and comments are understandable. So, I am ok with the patches, but I also think that we need to wait for Manivannan's Tested-by on qcom platforms.

Best regards,
Jingoo Han

Quoted 237 lines
drivers/pci/controller/dwc/pcie-designware-ep.c +77 -48
index 83ddb190292e..f300ea2f7bf7 100644
@@ -23,14 +23,6 @@ void dw_pcie_ep_linkup(struct dw_pcie_ep *ep)
 }
 EXPORT_SYMBOL_GPL(dw_pcie_ep_linkup);
 
-void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
-{
-       struct pci_epc *epc = ep->epc;
-
-       pci_epc_init_notify(epc);
-}
-EXPORT_SYMBOL_GPL(dw_pcie_ep_init_notify);
-
 struct dw_pcie_ep_func *
 dw_pcie_ep_get_func_from_ep(struct dw_pcie_ep *ep, u8 func_no)
 {
@@ -640,12 +632,17 @@ static unsigned int dw_pcie_ep_find_ext_capability(struct dw_pcie *pci, int cap)
        return 0;
 }
 
-int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
+static int dw_pcie_ep_init_late(struct dw_pcie_ep *ep)
 {
        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
+       struct dw_pcie_ep_func *ep_func;
+       struct device *dev = pci->dev;
+       struct pci_epc *epc = ep->epc;
        unsigned int offset;
        unsigned int nbars;
        u8 hdr_type;
+       u8 func_no;
+       void *addr;
        u32 reg;
        int i;
 
@@ -658,6 +655,51 @@ int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
                return -EIO;
        }
 
+       dw_pcie_version_detect(pci);
+
+       dw_pcie_iatu_detect(pci);
+
+       if (!ep->ib_window_map) {
+               ep->ib_window_map = devm_bitmap_zalloc(dev, pci->num_ib_windows,
+                                                      GFP_KERNEL);
+               if (!ep->ib_window_map)
+                       return -ENOMEM;
+       }
+
+       if (!ep->ob_window_map) {
+               ep->ob_window_map = devm_bitmap_zalloc(dev, pci->num_ob_windows,
+                                                      GFP_KERNEL);
+               if (!ep->ob_window_map)
+                       return -ENOMEM;
+       }
+
+       if (!ep->outbound_addr) {
+               addr = devm_kcalloc(dev, pci->num_ob_windows, sizeof(phys_addr_t),
+                                   GFP_KERNEL);
+               if (!addr)
+                       return -ENOMEM;
+               ep->outbound_addr = addr;
+       }
+
+       for (func_no = 0; func_no < epc->max_functions; func_no++) {
+
+               ep_func = dw_pcie_ep_get_func_from_ep(ep, func_no);
+               if (ep_func)
+                       continue;
+
+               ep_func = devm_kzalloc(dev, sizeof(*ep_func), GFP_KERNEL);
+               if (!ep_func)
+                       return -ENOMEM;
+
+               ep_func->func_no = func_no;
+               ep_func->msi_cap = dw_pcie_ep_find_capability(ep, func_no,
+                                                             PCI_CAP_ID_MSI);
+               ep_func->msix_cap = dw_pcie_ep_find_capability(ep, func_no,
+                                                              PCI_CAP_ID_MSIX);
+
+               list_add_tail(&ep_func->list, &ep->func_list);
+       }
+
        offset = dw_pcie_ep_find_ext_capability(pci, PCI_EXT_CAP_ID_REBAR);
 
        dw_pcie_dbi_ro_wr_en(pci);
@@ -676,13 +718,28 @@ int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
 
        return 0;
 }
-EXPORT_SYMBOL_GPL(dw_pcie_ep_init_complete);
+
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
+{
+       struct pci_epc *epc = ep->epc;
+       int ret;
+
+       ret = dw_pcie_ep_init_late(ep);
+       if (ret)
+               return ret;
+
+       if (ep->ops->ep_init_late)
+               ep->ops->ep_init_late(ep);
+
+       pci_epc_init_notify(epc);
+
+       return 0;
+}
+EXPORT_SYMBOL_GPL(dw_pcie_ep_init_notify);
 
 int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 {
        int ret;
-       void *addr;
-       u8 func_no;
        struct resource *res;
        struct pci_epc *epc;
        struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
@@ -690,7 +747,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
        struct platform_device *pdev = to_platform_device(dev);
        struct device_node *np = dev->of_node;
        const struct pci_epc_features *epc_features;
-       struct dw_pcie_ep_func *ep_func;
 
        INIT_LIST_HEAD(&ep->func_list);
 
@@ -719,26 +775,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
        ep->phys_base = res->start;
        ep->addr_size = resource_size(res);
 
-       dw_pcie_version_detect(pci);
-
-       dw_pcie_iatu_detect(pci);
-
-       ep->ib_window_map = devm_bitmap_zalloc(dev, pci->num_ib_windows,
-                                              GFP_KERNEL);
-       if (!ep->ib_window_map)
-               return -ENOMEM;
-
-       ep->ob_window_map = devm_bitmap_zalloc(dev, pci->num_ob_windows,
-                                              GFP_KERNEL);
-       if (!ep->ob_window_map)
-               return -ENOMEM;
-
-       addr = devm_kcalloc(dev, pci->num_ob_windows, sizeof(phys_addr_t),
-                           GFP_KERNEL);
-       if (!addr)
-               return -ENOMEM;
-       ep->outbound_addr = addr;
-
        if (pci->link_gen < 1)
                pci->link_gen = of_pci_get_max_link_speed(np);
 
@@ -755,20 +791,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
        if (ret < 0)
                epc->max_functions = 1;
 
-       for (func_no = 0; func_no < epc->max_functions; func_no++) {
-               ep_func = devm_kzalloc(dev, sizeof(*ep_func), GFP_KERNEL);
-               if (!ep_func)
-                       return -ENOMEM;
-
-               ep_func->func_no = func_no;
-               ep_func->msi_cap = dw_pcie_ep_find_capability(ep, func_no,
-                                                             PCI_CAP_ID_MSI);
-               ep_func->msix_cap = dw_pcie_ep_find_capability(ep, func_no,
-                                                              PCI_CAP_ID_MSIX);
-
-               list_add_tail(&ep_func->list, &ep->func_list);
-       }
-
        if (ep->ops->ep_init)
                ep->ops->ep_init(ep);
 
@@ -793,7 +815,14 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
                        return 0;
        }
 
-       ret = dw_pcie_ep_init_complete(ep);
+       /*
+        * NOTE:- Avoid accessing the hardware (Ex:- DBI space) before this
+        * step as platforms that implement 'core_init_notifier' feature may
+        * not have the hardware ready (i.e. core initialized) for access
+        * (Ex: tegra194). Any hardware access on such platforms result
+        * in system hard hang.
+        */
+       ret = dw_pcie_ep_init_late(ep);
        if (ret)
                goto err_free_epc_mem;
 
drivers/pci/controller/dwc/pcie-designware.h +3 -8
index 45fcdfc8c035..7252513956b7 100644
@@ -253,6 +253,7 @@ struct dw_pcie_rp {
 
 struct dw_pcie_ep_ops {
        void    (*ep_init)(struct dw_pcie_ep *ep);
+       void    (*ep_init_late)(struct dw_pcie_ep *ep);
        int     (*raise_irq)(struct dw_pcie_ep *ep, u8 func_no,
                             enum pci_epc_irq_type type, u16 interrupt_num);
        const struct pci_epc_features* (*get_features)(struct dw_pcie_ep *ep);
@@ -467,8 +468,7 @@ static inline void __iomem *dw_pcie_own_conf_map_bus(struct pci_bus *bus,
 #ifdef CONFIG_PCIE_DW_EP
 void dw_pcie_ep_linkup(struct dw_pcie_ep *ep);
 int dw_pcie_ep_init(struct dw_pcie_ep *ep);
-int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep);
-void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
 void dw_pcie_ep_exit(struct dw_pcie_ep *ep);
 int dw_pcie_ep_raise_legacy_irq(struct dw_pcie_ep *ep, u8 func_no);
 int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
@@ -490,15 +490,11 @@ static inline int dw_pcie_ep_init(struct dw_pcie_ep *ep)
        return 0;
 }
 
-static inline int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
+static inline int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
 {
        return 0;
 }
 
-static inline void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
-{
-}
-
 static inline void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
 {
 }
--

2.17.1

#3 lore raw
In reply to Vidya Sagar
Quoted 6 lines · Vidya Sagar

Platforms that cannot support their core initialization without the reference clock from the host, implement the feature 'core_init_notifier' to indicate the DesignWare sub-system about when their core is getting initialized. Any accesses to the core (Ex:- DBI) would the core being ready result in system hang in such systems (Ex:- tegra194). This patch moves any access to the core to dw_pcie_ep_init_complete() API

dw_pcie_ep_init_complete() got renamed to dw_pcie_ep_init_late()

Quoted 24 lines

which is effectively called only after the core initialization. It also introduces .ep_init_late() ops hook to be used for any post init work that platform drivers may have to do.

---
V5:
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Moved dw_pcie_ep_init_complete() inside dw_pcie_ep_init_notify()
* Added .ep_init_late() ops to perform late init tasks

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 2 files changed, 80 insertions(+), 55 deletions(-)
drivers/pci/controller/dwc/pcie-designware-ep.c +0 -0
index 83ddb190292e..f300ea2f7bf7 100644

[...]

drivers/pci/controller/dwc/pcie-designware-ep.c int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep) · line 732

Quoted 11 lines
drivers/pci/controller/dwc/pcie-designware-ep.c +11 -0
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
+{
+	struct pci_epc *epc = ep->epc;
+	int ret;
+
+	ret = dw_pcie_ep_init_late(ep);
+	if (ret)
+		return ret;
+
+	if (ep->ops->ep_init_late)
+		ep->ops->ep_init_late(ep);

I think you introduced this callback for the sake of Qcom driver I believe. But it is not really required. I'll share more comments in patch 2/3.

Thanks,
Mani

Quoted 129 lines
drivers/pci/controller/dwc/pcie-designware-ep.c +14 -38
+
+	pci_epc_init_notify(epc);
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(dw_pcie_ep_init_notify);
 
 int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 {
 	int ret;
-	void *addr;
-	u8 func_no;
 	struct resource *res;
 	struct pci_epc *epc;
 	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
@@ -690,7 +747,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	struct platform_device *pdev = to_platform_device(dev);
 	struct device_node *np = dev->of_node;
 	const struct pci_epc_features *epc_features;
-	struct dw_pcie_ep_func *ep_func;
 
 	INIT_LIST_HEAD(&ep->func_list);
 
@@ -719,26 +775,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	ep->phys_base = res->start;
 	ep->addr_size = resource_size(res);
 
-	dw_pcie_version_detect(pci);
-
-	dw_pcie_iatu_detect(pci);
-
-	ep->ib_window_map = devm_bitmap_zalloc(dev, pci->num_ib_windows,
-					       GFP_KERNEL);
-	if (!ep->ib_window_map)
-		return -ENOMEM;
-
-	ep->ob_window_map = devm_bitmap_zalloc(dev, pci->num_ob_windows,
-					       GFP_KERNEL);
-	if (!ep->ob_window_map)
-		return -ENOMEM;
-
-	addr = devm_kcalloc(dev, pci->num_ob_windows, sizeof(phys_addr_t),
-			    GFP_KERNEL);
-	if (!addr)
-		return -ENOMEM;
-	ep->outbound_addr = addr;
-
 	if (pci->link_gen < 1)
 		pci->link_gen = of_pci_get_max_link_speed(np);
 
@@ -755,20 +791,6 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	if (ret < 0)
 		epc->max_functions = 1;
 
-	for (func_no = 0; func_no < epc->max_functions; func_no++) {
-		ep_func = devm_kzalloc(dev, sizeof(*ep_func), GFP_KERNEL);
-		if (!ep_func)
-			return -ENOMEM;
-
-		ep_func->func_no = func_no;
-		ep_func->msi_cap = dw_pcie_ep_find_capability(ep, func_no,
-							      PCI_CAP_ID_MSI);
-		ep_func->msix_cap = dw_pcie_ep_find_capability(ep, func_no,
-							       PCI_CAP_ID_MSIX);
-
-		list_add_tail(&ep_func->list, &ep->func_list);
-	}
-
 	if (ep->ops->ep_init)
 		ep->ops->ep_init(ep);
 
@@ -793,7 +815,14 @@ int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 			return 0;
 	}
 
-	ret = dw_pcie_ep_init_complete(ep);
+	/*
+	 * NOTE:- Avoid accessing the hardware (Ex:- DBI space) before this
+	 * step as platforms that implement 'core_init_notifier' feature may
+	 * not have the hardware ready (i.e. core initialized) for access
+	 * (Ex: tegra194). Any hardware access on such platforms result
+	 * in system hard hang.
+	 */
+	ret = dw_pcie_ep_init_late(ep);
 	if (ret)
 		goto err_free_epc_mem;
 
drivers/pci/controller/dwc/pcie-designware.h +3 -7
index 45fcdfc8c035..7252513956b7 100644
@@ -253,6 +253,7 @@ struct dw_pcie_rp {
 
 struct dw_pcie_ep_ops {
 	void	(*ep_init)(struct dw_pcie_ep *ep);
+	void	(*ep_init_late)(struct dw_pcie_ep *ep);
 	int	(*raise_irq)(struct dw_pcie_ep *ep, u8 func_no,
 			     enum pci_epc_irq_type type, u16 interrupt_num);
 	const struct pci_epc_features* (*get_features)(struct dw_pcie_ep *ep);
@@ -467,8 +468,7 @@ static inline void __iomem *dw_pcie_own_conf_map_bus(struct pci_bus *bus,
 #ifdef CONFIG_PCIE_DW_EP
 void dw_pcie_ep_linkup(struct dw_pcie_ep *ep);
 int dw_pcie_ep_init(struct dw_pcie_ep *ep);
-int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep);
-void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
+int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep);
 void dw_pcie_ep_exit(struct dw_pcie_ep *ep);
 int dw_pcie_ep_raise_legacy_irq(struct dw_pcie_ep *ep, u8 func_no);
 int dw_pcie_ep_raise_msi_irq(struct dw_pcie_ep *ep, u8 func_no,
@@ -490,15 +490,11 @@ static inline int dw_pcie_ep_init(struct dw_pcie_ep *ep)
 	return 0;
 }
 
-static inline int dw_pcie_ep_init_complete(struct dw_pcie_ep *ep)
+static inline int dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
 {
 	return 0;
 }
 
-static inline void dw_pcie_ep_init_notify(struct dw_pcie_ep *ep)
-{
-}
-
 static inline void dw_pcie_ep_exit(struct dw_pcie_ep *ep)
 {
 }
Signature · 1 line
2.17.1
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#4 lore raw

PATCH v5 2/3 PCI: qcom-ep: Refactor EP initialization completion

In reply to Vidya Sagar

Move the post initialization code to .ep_init_late() call back and call only dw_pcie_ep_init_notify() which internally takes care of calling dw_pcie_ep_init_complete().

---
V5:
* None

V4:
* New patch in this series

 drivers/pci/controller/dwc/pcie-qcom-ep.c | 27 ++++++++++++++---------
 1 file changed, 16 insertions(+), 11 deletions(-)
drivers/pci/controller/dwc/pcie-qcom-ep.c +16 -11
index e33eb3871309..c418b20042aa 100644
@@ -361,22 +361,12 @@ static int qcom_pcie_perst_deassert(struct dw_pcie *pci)
 	      PARF_INT_ALL_LINK_UP;
 	writel_relaxed(val, pcie_ep->parf + PARF_INT_ALL_MASK);
 
-	ret = dw_pcie_ep_init_complete(&pcie_ep->pci.ep);
+	ret = dw_pcie_ep_init_notify(&pcie_ep->pci.ep);
 	if (ret) {
 		dev_err(dev, "Failed to complete initialization: %d\n", ret);
 		goto err_disable_resources;
 	}
 
-	/*
-	 * The physical address of the MMIO region which is exposed as the BAR
-	 * should be written to MHI BASE registers.
-	 */
-	writel_relaxed(pcie_ep->mmio_res->start,
-		       pcie_ep->parf + PARF_MHI_BASE_ADDR_LOWER);
-	writel_relaxed(0, pcie_ep->parf + PARF_MHI_BASE_ADDR_UPPER);
-
-	dw_pcie_ep_init_notify(&pcie_ep->pci.ep);
-
 	/* Enable LTSSM */
 	val = readl_relaxed(pcie_ep->parf + PARF_LTSSM);
 	val |= BIT(8);
@@ -643,8 +633,23 @@ static void qcom_pcie_ep_init(struct dw_pcie_ep *ep)
 		dw_pcie_ep_reset_bar(pci, bar);
 }
 
+static void qcom_pcie_ep_init_late(struct dw_pcie_ep *ep)
+{
+	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
+	struct qcom_pcie_ep *pcie_ep = to_pcie_ep(pci);
+
+	/*
+	 * The physical address of the MMIO region which is exposed as the BAR
+	 * should be written to MHI BASE registers.
+	 */
+	writel_relaxed(pcie_ep->mmio_res->start,
+		       pcie_ep->parf + PARF_MHI_BASE_ADDR_LOWER);
+	writel_relaxed(0, pcie_ep->parf + PARF_MHI_BASE_ADDR_UPPER);
+}
+
 static const struct dw_pcie_ep_ops pci_ep_ops = {
 	.ep_init = qcom_pcie_ep_init,
+	.ep_init_late = qcom_pcie_ep_init_late,
 	.raise_irq = qcom_pcie_ep_raise_irq,
 	.get_features = qcom_pcie_epc_get_features,
 };
Signature · 1 line
2.17.1
#5 lore raw

Re: [PATCH V5 2/3] PCI: qcom-ep: Refactor EP initialization completion

Quoted 27 lines · Vidya Sagar

Move the post initialization code to .ep_init_late() call back and call only dw_pcie_ep_init_notify() which internally takes care of calling dw_pcie_ep_init_complete().

---
V5:
* None

V4:
* New patch in this series

 drivers/pci/controller/dwc/pcie-qcom-ep.c | 27 ++++++++++++++---------
 1 file changed, 16 insertions(+), 11 deletions(-)
drivers/pci/controller/dwc/pcie-qcom-ep.c +1 -1
index e33eb3871309..c418b20042aa 100644
@@ -361,22 +361,12 @@ static int qcom_pcie_perst_deassert(struct dw_pcie *pci)
 	      PARF_INT_ALL_LINK_UP;
 	writel_relaxed(val, pcie_ep->parf + PARF_INT_ALL_MASK);
 
-	ret = dw_pcie_ep_init_complete(&pcie_ep->pci.ep);
+	ret = dw_pcie_ep_init_notify(&pcie_ep->pci.ep);
 	if (ret) {
 		dev_err(dev, "Failed to complete initialization: %d\n", ret);

drivers/pci/controller/dwc/pcie-qcom-ep.c static int qcom_pcie_perst_deassert(struct dw_pcie *pci) · line 376

Quoted 10 lines
drivers/pci/controller/dwc/pcie-qcom-ep.c +0 -7
 		goto err_disable_resources;
 	}
 
-	/*
-	 * The physical address of the MMIO region which is exposed as the BAR
-	 * should be written to MHI BASE registers.
-	 */
-	writel_relaxed(pcie_ep->mmio_res->start,
-		       pcie_ep->parf + PARF_MHI_BASE_ADDR_LOWER);
-	writel_relaxed(0, pcie_ep->parf + PARF_MHI_BASE_ADDR_UPPER);

Writes to the MHI base addresses are required before starting LTSSM and not necessarily before core_init notifier. So you could just leave this code here and get rid of .ep_init_late() callback.

And you should also rebase the series on top of v6.1-rcX as I've added few more code in this function.

Thanks,
Mani

Quoted 33 lines
drivers/pci/controller/dwc/pcie-qcom-ep.c +15 -3
-
-	dw_pcie_ep_init_notify(&pcie_ep->pci.ep);
-
 	/* Enable LTSSM */
 	val = readl_relaxed(pcie_ep->parf + PARF_LTSSM);
 	val |= BIT(8);
@@ -643,8 +633,23 @@ static void qcom_pcie_ep_init(struct dw_pcie_ep *ep)
 		dw_pcie_ep_reset_bar(pci, bar);
 }
 
+static void qcom_pcie_ep_init_late(struct dw_pcie_ep *ep)
+{
+	struct dw_pcie *pci = to_dw_pcie_from_ep(ep);
+	struct qcom_pcie_ep *pcie_ep = to_pcie_ep(pci);
+
+	/*
+	 * The physical address of the MMIO region which is exposed as the BAR
+	 * should be written to MHI BASE registers.
+	 */
+	writel_relaxed(pcie_ep->mmio_res->start,
+		       pcie_ep->parf + PARF_MHI_BASE_ADDR_LOWER);
+	writel_relaxed(0, pcie_ep->parf + PARF_MHI_BASE_ADDR_UPPER);
+}
+
 static const struct dw_pcie_ep_ops pci_ep_ops = {
 	.ep_init = qcom_pcie_ep_init,
+	.ep_init_late = qcom_pcie_ep_init_late,
 	.raise_irq = qcom_pcie_ep_raise_irq,
 	.get_features = qcom_pcie_epc_get_features,
 };
Signature · 1 line
2.17.1
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#6 lore raw

PATCH v5 3/3 PCI: tegra194: Refactor EP initialization completion

In reply to Vidya Sagar

Call only dw_pcie_ep_init_notify() which internally takes care of calling dw_pcie_ep_init_complete() to notify about the EP initialization completion to the DWC EP framework.

---
V5:
* None

V4:
* New patch in this series

 drivers/pci/controller/dwc/pcie-tegra194.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/pci/controller/dwc/pcie-tegra194.c +1 -3
index 1b6b437823d2..2600304522eb 100644
@@ -1885,14 +1885,12 @@ static void pex_ep_event_pex_rst_deassert(struct tegra_pcie_dw *pcie)
 	val = (upper_32_bits(ep->msi_mem_phys) & MSIX_ADDR_MATCH_HIGH_OFF_MASK);
 	dw_pcie_writel_dbi(pci, MSIX_ADDR_MATCH_HIGH_OFF, val);
 
-	ret = dw_pcie_ep_init_complete(ep);
+	ret = dw_pcie_ep_init_notify(ep);
 	if (ret) {
 		dev_err(dev, "Failed to complete initialization: %d\n", ret);
 		goto fail_init_complete;
 	}
 
-	dw_pcie_ep_init_notify(ep);
-
 	/* Program the private control to allow sending LTR upstream */
 	if (pcie->of_data->has_ltr_req_fix) {
 		val = appl_readl(pcie, APPL_LTR_MSG_2);
Signature · 1 line
2.17.1
#7 lore raw
In reply to Vidya Sagar

Hi Vidya,

Quoted 8 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

Currently I'm on paternity leave this week and next. Will test/review the latest version once I'm back.

Thanks,
Mani

Quoted 25 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
 drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
 4 files changed, 97 insertions(+), 69 deletions(-)
Signature · 1 line
2.17.1
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#8 lore raw
In reply to Vidya Sagar
Quoted 8 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

Vidya, any plan to respin this series? The EPC rework series is now merged for v6.3.

Thanks,
Mani

Quoted 25 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

 .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
 drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
 drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
 drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
 4 files changed, 97 insertions(+), 69 deletions(-)
Signature · 1 line
2.17.1
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#9 lore raw
Quoted 10 lines · Manivannan Sadhasivam

External email: Use caution opening links or attachments

Quoted 6 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Quoted 5 lines
Quoted 2 lines

Manivannan, could you please verify on qcom platforms?

Vidya, any plan to respin this series? The EPC rework series is now merged for v6.3.

Yes. I'll send an updated series soon.
Currently, I'm observing some regression with linux-next on Tegra platform for endpoint mode. I'll post the patches as soon as that is resolved.

Thanks,
Vidya Sagar

Quoted 32 lines

Thanks,
Mani

Quoted 16 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

Quoted 9 lines
  .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
  drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
  drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
  drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
  4 files changed, 97 insertions(+), 69 deletions(-)

--
2.17.1

--
மணிவண்ணன் சதாசிவம்

#10 lore raw
Quoted 16 lines · Vidya Sagar
Quoted 13 lines · Manivannan Sadhasivam

External email: Use caution opening links or attachments

Quoted 8 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

Quoted 7 lines
Quoted 2 lines

Vidya, any plan to respin this series? The EPC rework series is now merged for v6.3.

Yes. I'll send an updated series soon.
Currently, I'm observing some regression with linux-next on Tegra platform for endpoint mode. I'll post the patches as soon as that is resolved.

Ping!

Thanks,
Mani

Quoted 35 lines

Thanks,
Vidya Sagar

Quoted 32 lines

Thanks,
Mani

Quoted 16 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

Quoted 9 lines
  .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
  drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
  drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
  drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
  4 files changed, 97 insertions(+), 69 deletions(-)

--
2.17.1

--
மணிவண்ணன் சதாசிவம்

Signature · 1 line
மணிவண்ணன் சதாசிவம்
#11 lore raw
Quoted 20 lines · Manivannan Sadhasivam
Quoted 19 lines · Vidya Sagar
Quoted 15 lines · Manivannan Sadhasivam

External email: Use caution opening links or attachments

Quoted 8 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

Vidya, any plan to respin this series? The EPC rework series is now merged for v6.3.

Quoted 7 lines
Quoted 4 lines

Yes. I'll send an updated series soon.
Currently, I'm observing some regression with linux-next on Tegra platform for endpoint mode. I'll post the patches as soon as that is resolved.

Ping!

Ping again.

- Mani

Quoted 41 lines

Thanks,
Mani

Quoted 35 lines

Thanks,
Vidya Sagar

Quoted 32 lines

Thanks,
Mani

Quoted 16 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

Quoted 9 lines
  .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
  drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
  drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
  drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
  4 files changed, 97 insertions(+), 69 deletions(-)

--
2.17.1

--
மணிவண்ணன் சதாசிவம்

Signature · 1 line
மணிவண்ணன் சதாசிவம்
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#12 lore raw
In reply to Manivannan Sadhasivam
Quoted 20 lines · Manivannan Sadhasivam
Quoted 19 lines · Vidya Sagar
Quoted 15 lines · Manivannan Sadhasivam

External email: Use caution opening links or attachments

Quoted 8 lines · Vidya Sagar

This series attempts to fix the issue with core register (Ex:- DBI) accesses causing system hang issues in platforms where there is a dependency on the availability of PCIe Reference clock from the host for their core initialization.
This series is verified on Tegra194 & Tegra234 platforms.

Manivannan, could you please verify on qcom platforms?

Vidya, any plan to respin this series? The EPC rework series is now merged for v6.3.

Quoted 7 lines
Quoted 4 lines

Yes. I'll send an updated series soon.
Currently, I'm observing some regression with linux-next on Tegra platform for endpoint mode. I'll post the patches as soon as that is resolved.

Ping!

Vidya, are you planning to continue working on this series? If you do not have time, please let me know.

- Mani

Quoted 41 lines

Thanks,
Mani

Quoted 35 lines

Thanks,
Vidya Sagar

Quoted 32 lines

Thanks,
Mani

Quoted 16 lines

V5:
* Addressed review comments from Bjorn
* Changed dw_pcie_ep_init_complete() to dw_pcie_ep_init_late()
* Skipped memory allocation if done already. This is to avoid freeing and then allocating again during PERST# toggles from the host.

V4:
* Addressed review comments from Bjorn and Manivannan
* Added .ep_init_late() ops
* Added patches to refactor code in qcom and tegra platforms

Vidya Sagar (3):
PCI: designware-ep: Fix DBI access before core init
PCI: qcom-ep: Refactor EP initialization completion
PCI: tegra194: Refactor EP initialization completion

Quoted 9 lines
  .../pci/controller/dwc/pcie-designware-ep.c   | 125 +++++++++++-------
  drivers/pci/controller/dwc/pcie-designware.h  |  10 +-
  drivers/pci/controller/dwc/pcie-qcom-ep.c     |  27 ++--
  drivers/pci/controller/dwc/pcie-tegra194.c    |   4 +-
  4 files changed, 97 insertions(+), 69 deletions(-)

--
2.17.1

--
மணிவண்ணன் சதாசிவம்

Signature · 1 line
மணிவண்ணன் சதாசிவம்
Signature · 1 line
மணிவண்ணன் சதாசிவம்