Skip to contentSkip to conversation

PCI: qcom: fix IPQ8074 Gen2 support

PATCH Superseded Robert Marko 4 participants 15 messages Jun 21, 2022 – Jun 22, 2022

Subsystems (4)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 PCI: qcom: fix IPQ8074 Gen2 support

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -20
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 	struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
 	struct dw_pcie *pci = pcie->pci;
 	struct device *dev = pci->dev;
-	u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
 	int i, ret;
-	u32 val;
 
 	for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
 		ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 		goto err_clk_aux;
 	}
 
+	return 0;
+
+err_clk_aux:
+	clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+	clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+	clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+	clk_disable_unprepare(res->iface);
+err_clk_iface:
+	/*
+	 * Not checking for failure, will anyway return
+	 * the original failure in 'ret'.
+	 */
+	for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+		reset_control_assert(res->rst[i]);
+
+	return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+	struct dw_pcie *pci = pcie->pci;
+	u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+	u32 val;
+
 	writel(SLV_ADDR_SPACE_SZ,
 		pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 		PCI_EXP_DEVCTL2);
 
 	return 0;
-
-err_clk_aux:
-	clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-	clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-	clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-	clk_disable_unprepare(res->iface);
-err_clk_iface:
-	/*
-	 * Not checking for failure, will anyway return
-	 * the original failure in 'ret'.
-	 */
-	for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-		reset_control_assert(res->rst[i]);
-
-	return ret;
 }
 
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
 	.get_resources = qcom_pcie_get_resources_2_3_3,
 	.init = qcom_pcie_init_2_3_3,
+	.post_init = qcom_pcie_post_init_2_3_3,
 	.deinit = qcom_pcie_deinit_2_3_3,
 	.ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
Signature · 1 line
2.36.1
#1 lore raw
Quoted 8 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

Any elaboration for the Fixes tag? I think the follow one is more logical, isn't it?

Fixes 5d76117f070d ("PCI: qcom: Add support for IPQ8074 PCIe controller")
Quoted 92 lines

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -21
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
        struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
        struct dw_pcie *pci = pcie->pci;
        struct device *dev = pci->dev;
-       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
        int i, ret;
-       u32 val;
 
        for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
                ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                goto err_clk_aux;
        }
 
+       return 0;
+
+err_clk_aux:
+       clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+       clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+       clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+       clk_disable_unprepare(res->iface);
+err_clk_iface:
+       /*
+        * Not checking for failure, will anyway return
+        * the original failure in 'ret'.
+        */
+       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+               reset_control_assert(res->rst[i]);
+
+       return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+       struct dw_pcie *pci = pcie->pci;
+       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+       u32 val;
+
        writel(SLV_ADDR_SPACE_SZ,
                pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                PCI_EXP_DEVCTL2);
 
        return 0;
-
-err_clk_aux:
-       clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-       clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-       clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-       clk_disable_unprepare(res->iface);
-err_clk_iface:
-       /*
-        * Not checking for failure, will anyway return
-        * the original failure in 'ret'.
-        */
-       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-               reset_control_assert(res->rst[i]);
-
-       return ret;
 }
 
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
        .get_resources = qcom_pcie_get_resources_2_3_3,
        .init = qcom_pcie_init_2_3_3,
+       .post_init = qcom_pcie_post_init_2_3_3,
        .deinit = qcom_pcie_deinit_2_3_3,
        .ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
--

2.36.1

Signature · 2 lines
With best wishes
Dmitry
#2 lore raw
Quoted 7 lines · Dmitry Baryshkov
Quoted 5 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Quoted 8 lines
Quoted 3 lines

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

Any elaboration for the Fixes tag? I think the follow one is more logical, isn't it?

Fixes 5d76117f070d ("PCI: qcom: Add support for IPQ8074 PCIe controller")

Hi,
My logic was that it was working before the commit a0fd361db8e5 as it moved PHY init
later and indirectly broke IPQ8074 gen2.

Regards,
Robert

Quoted 98 lines
Quoted 91 lines

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -21
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
        struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
        struct dw_pcie *pci = pcie->pci;
        struct device *dev = pci->dev;
-       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
        int i, ret;
-       u32 val;
 
        for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
                ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                goto err_clk_aux;
        }
 
+       return 0;
+
+err_clk_aux:
+       clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+       clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+       clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+       clk_disable_unprepare(res->iface);
+err_clk_iface:
+       /*
+        * Not checking for failure, will anyway return
+        * the original failure in 'ret'.
+        */
+       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+               reset_control_assert(res->rst[i]);
+
+       return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+       struct dw_pcie *pci = pcie->pci;
+       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+       u32 val;
+
        writel(SLV_ADDR_SPACE_SZ,
                pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                PCI_EXP_DEVCTL2);
 
        return 0;
-
-err_clk_aux:
-       clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-       clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-       clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-       clk_disable_unprepare(res->iface);
-err_clk_iface:
-       /*
-        * Not checking for failure, will anyway return
-        * the original failure in 'ret'.
-        */
-       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-               reset_control_assert(res->rst[i]);
-
-       return ret;
 }
 
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
        .get_resources = qcom_pcie_get_resources_2_3_3,
        .init = qcom_pcie_init_2_3_3,
+       .post_init = qcom_pcie_post_init_2_3_3,
        .deinit = qcom_pcie_deinit_2_3_3,
        .ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
--
Quoted 2 lines

2.36.1

--
With best wishes
Dmitry

Reviewed-by
#3 lore raw
Quoted 16 lines · Robert Marko
Quoted 14 lines · Dmitry Baryshkov
Quoted 8 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

Any elaboration for the Fixes tag? I think the follow one is more logical, isn't it?

Quoted 6 lines
Quoted 1 line
Fixes 5d76117f070d ("PCI: qcom: Add support for IPQ8074 PCIe controller")

Hi,
My logic was that it was working before the commit a0fd361db8e5 as it moved PHY init
later and indirectly broke IPQ8074 gen2.

Ack, thanks for the explanation.

Quoted 101 lines

Regards,
Robert

Quoted 98 lines
Quoted 91 lines

---
Changes in v2:
* Rebase onto next-20220621

---
  drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
  1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -21
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
         struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
         struct dw_pcie *pci = pcie->pci;
         struct device *dev = pci->dev;
-       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
         int i, ret;
-       u32 val;
 
         for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
                 ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                 goto err_clk_aux;
         }
 
+       return 0;
+
+err_clk_aux:
+       clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+       clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+       clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+       clk_disable_unprepare(res->iface);
+err_clk_iface:
+       /*
+        * Not checking for failure, will anyway return
+        * the original failure in 'ret'.
+        */
+       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+               reset_control_assert(res->rst[i]);
+
+       return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+       struct dw_pcie *pci = pcie->pci;
+       u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+       u32 val;
+
         writel(SLV_ADDR_SPACE_SZ,
                 pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
                 PCI_EXP_DEVCTL2);
 
         return 0;
-
-err_clk_aux:
-       clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-       clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-       clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-       clk_disable_unprepare(res->iface);
-err_clk_iface:
-       /*
-        * Not checking for failure, will anyway return
-        * the original failure in 'ret'.
-        */
-       for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-               reset_control_assert(res->rst[i]);
-
-       return ret;
  }
 
  static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
  static const struct qcom_pcie_ops ops_2_3_3 = {
         .get_resources = qcom_pcie_get_resources_2_3_3,
         .init = qcom_pcie_init_2_3_3,
+       .post_init = qcom_pcie_post_init_2_3_3,
         .deinit = qcom_pcie_deinit_2_3_3,
         .ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
  };
--
Quoted 2 lines

2.36.1

--
With best wishes
Dmitry

Signature · 2 lines
With best wishes
Dmitry
#4 lore raw
In reply to Robert Marko
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

Quoted 95 lines

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -20
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 	struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
 	struct dw_pcie *pci = pcie->pci;
 	struct device *dev = pci->dev;
-	u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
 	int i, ret;
-	u32 val;
 
 	for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
 		ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 		goto err_clk_aux;
 	}
 
+	return 0;
+
+err_clk_aux:
+	clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+	clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+	clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+	clk_disable_unprepare(res->iface);
+err_clk_iface:
+	/*
+	 * Not checking for failure, will anyway return
+	 * the original failure in 'ret'.
+	 */
+	for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+		reset_control_assert(res->rst[i]);
+
+	return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+	struct dw_pcie *pci = pcie->pci;
+	u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+	u32 val;
+
 	writel(SLV_ADDR_SPACE_SZ,
 		pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
 		PCI_EXP_DEVCTL2);
 
 	return 0;
-
-err_clk_aux:
-	clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-	clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-	clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-	clk_disable_unprepare(res->iface);
-err_clk_iface:
-	/*
-	 * Not checking for failure, will anyway return
-	 * the original failure in 'ret'.
-	 */
-	for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-		reset_control_assert(res->rst[i]);
-
-	return ret;
 }
 
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
 	.get_resources = qcom_pcie_get_resources_2_3_3,
 	.init = qcom_pcie_init_2_3_3,
+	.post_init = qcom_pcie_post_init_2_3_3,
 	.deinit = qcom_pcie_deinit_2_3_3,
 	.ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
Signature · 1 line
2.36.1
#5 lore raw
Quoted 30 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

Quoted 8 lines

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

I don't think there is a need to reorder patches. My patches do not remove support for post_init(), they drop the callbacks code. Thus one can reinstate necessary code back.

Quoted 84 lines
Quoted 84 lines

So solve this by splitting the DBI read/writes to .post_init.

Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +27 -20
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
      struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
      struct dw_pcie *pci = pcie->pci;
      struct device *dev = pci->dev;
-     u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
      int i, ret;
-     u32 val;
 
      for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
              ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
              goto err_clk_aux;
      }
 
+     return 0;
+
+err_clk_aux:
+     clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+     clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+     clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+     clk_disable_unprepare(res->iface);
+err_clk_iface:
+     /*
+      * Not checking for failure, will anyway return
+      * the original failure in 'ret'.
+      */
+     for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+             reset_control_assert(res->rst[i]);
+
+     return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+     struct dw_pcie *pci = pcie->pci;
+     u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+     u32 val;
+
      writel(SLV_ADDR_SPACE_SZ,
              pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
              PCI_EXP_DEVCTL2);
 
      return 0;
-
-err_clk_aux:
-     clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-     clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-     clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-     clk_disable_unprepare(res->iface);
-err_clk_iface:
-     /*
-      * Not checking for failure, will anyway return
-      * the original failure in 'ret'.
-      */
-     for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-             reset_control_assert(res->rst[i]);
-
-     return ret;
 }
 
Quoted 12 lines
Quoted 10 lines
drivers/pci/controller/dwc/pcie-qcom.c +1 -1
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
      .get_resources = qcom_pcie_get_resources_2_3_3,
      .init = qcom_pcie_init_2_3_3,
+     .post_init = qcom_pcie_post_init_2_3_3,
      .deinit = qcom_pcie_deinit_2_3_3,
      .ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
--
Quoted 2 lines

2.36.1

--
With best wishes
Dmitry

#6 lore raw
Quoted 35 lines · Dmitry Baryshkov
Quoted 34 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Quoted 7 lines
Quoted 3 lines

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

I don't think there is a need to reorder patches. My patches do not remove support for post_init(), they drop the callbacks code. Thus one can reinstate necessary code back.

There's not a *need* to reorder them, but I think it would make the patches smaller and more readable because we wouldn't be removing and then re-adding the functions.

#7 lore raw
Quoted 40 lines · Bjorn Helgaas
Quoted 39 lines · Dmitry Baryshkov
Quoted 37 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

   qcom_pcie_host_init
     pcie->cfg->ops->init(pcie)
     phy_power_on(pcie->phy)
     pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

   qcom_pcie_init_2_1_0
   qcom_pcie_init_1_0_0      (oddly out of order)
   qcom_pcie_init_2_3_2
   qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

Quoted 7 lines
Quoted 3 lines

I don't think there is a need to reorder patches. My patches do not remove support for post_init(), they drop the callbacks code. Thus one can reinstate necessary code back.

There's not a *need* to reorder them, but I think it would make the patches smaller and more readable because we wouldn't be removing and then re-adding the functions.

Ack. I'm fine then with rebasing my patches on top of Robert's patchset. I'll send the next revision after getting this patchset into the form.

Signature · 2 lines
With best wishes
Dmitry
#8 lore raw
In reply to Bjorn Helgaas
Quoted 8 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

Hi Bjorn,
Not really, I can remove it from the description as this only affects the Gen2 port, Gen3 support is dependant on the IPQ6018 support getting merged as it uses the same controller.

Quoted 13 lines

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

Quoted 8 lines

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

As far as I could figure out from QCA-s 5.4 kernel, various commits, and QCA-s attempts to solve this already upstream the Gen2 controller in IPQ8074 is a bit special and requires the PHY to be powered on before DBI could be accessed or else the board will hang as it does for me.

I can only assume that this is an IPQ8074-specific quirk and other IP-s are not affected like this, so they were not broken.

Quoted 11 lines

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

Quoted 1 line

So solve this by splitting the DBI read/writes to .post_init.

I am open to anything to get this fixed properly, you are gonna need to point me in the right direction
as I am really new to PCI.

Regards,
Robert

Quoted 94 lines
Quoted 94 lines
Fixes a0fd361db8e5 ("PCI: dwc: Move "dbi", "dbi2", and "addr_space" resource setup into common code")

---
Changes in v2:
* Rebase onto next-20220621

---
 drivers/pci/controller/dwc/pcie-qcom.c | 48 +++++++++++++++-----------
 1 file changed, 28 insertions(+), 20 deletions(-)
drivers/pci/controller/dwc/pcie-qcom.c +28 -21
index 51fed83484af..da6d79d61397 100644
@@ -1061,9 +1061,7 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
      struct qcom_pcie_resources_2_3_3 *res = &pcie->res.v2_3_3;
      struct dw_pcie *pci = pcie->pci;
      struct device *dev = pci->dev;
-     u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
      int i, ret;
-     u32 val;
 
      for (i = 0; i < ARRAY_SIZE(res->rst); i++) {
              ret = reset_control_assert(res->rst[i]);
@@ -1120,6 +1118,33 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
              goto err_clk_aux;
      }
 
+     return 0;
+
+err_clk_aux:
+     clk_disable_unprepare(res->ahb_clk);
+err_clk_ahb:
+     clk_disable_unprepare(res->axi_s_clk);
+err_clk_axi_s:
+     clk_disable_unprepare(res->axi_m_clk);
+err_clk_axi_m:
+     clk_disable_unprepare(res->iface);
+err_clk_iface:
+     /*
+      * Not checking for failure, will anyway return
+      * the original failure in 'ret'.
+      */
+     for (i = 0; i < ARRAY_SIZE(res->rst); i++)
+             reset_control_assert(res->rst[i]);
+
+     return ret;
+}
+
+static int qcom_pcie_post_init_2_3_3(struct qcom_pcie *pcie)
+{
+     struct dw_pcie *pci = pcie->pci;
+     u16 offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
+     u32 val;
+
      writel(SLV_ADDR_SPACE_SZ,
              pcie->parf + PCIE20_v3_PARF_SLV_ADDR_SPACE_SIZE);
 
@@ -1147,24 +1172,6 @@ static int qcom_pcie_init_2_3_3(struct qcom_pcie *pcie)
              PCI_EXP_DEVCTL2);
 
      return 0;
-
-err_clk_aux:
-     clk_disable_unprepare(res->ahb_clk);
-err_clk_ahb:
-     clk_disable_unprepare(res->axi_s_clk);
-err_clk_axi_s:
-     clk_disable_unprepare(res->axi_m_clk);
-err_clk_axi_m:
-     clk_disable_unprepare(res->iface);
-err_clk_iface:
-     /*
-      * Not checking for failure, will anyway return
-      * the original failure in 'ret'.
-      */
-     for (i = 0; i < ARRAY_SIZE(res->rst); i++)
-             reset_control_assert(res->rst[i]);
-
-     return ret;
 }
 
 static int qcom_pcie_get_resources_2_7_0(struct qcom_pcie *pcie)
@@ -1598,6 +1605,7 @@ static const struct qcom_pcie_ops ops_2_4_0 = {
 static const struct qcom_pcie_ops ops_2_3_3 = {
      .get_resources = qcom_pcie_get_resources_2_3_3,
      .init = qcom_pcie_init_2_3_3,
+     .post_init = qcom_pcie_post_init_2_3_3,
      .deinit = qcom_pcie_deinit_2_3_3,
      .ltssm_enable = qcom_pcie_2_3_2_ltssm_enable,
 };
--

2.36.1

#9 lore raw
Quoted 6 lines · Robert Marko
Quoted 5 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Quoted 6 lines
Quoted 2 lines

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

Not really, I can remove it from the description as this only affects the Gen2 port, Gen3 support is dependant on the IPQ6018 support getting merged as it uses the same controller.

Great, I think unrelated details confuse things.

Quoted 35 lines
Quoted 12 lines

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

Quoted 8 lines

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

As far as I could figure out from QCA-s 5.4 kernel, various commits, and QCA-s attempts to solve this already upstream the Gen2 controller in IPQ8074 is a bit special and requires the PHY to be powered on before DBI could be accessed or else the board will hang as it does for me.

I can only assume that this is an IPQ8074-specific quirk and other IP-s are not affected like this, so they were not broken.

Quoted 5 lines

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Quoted 8 lines
Quoted 5 lines

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

Quoted 1 line

So solve this by splitting the DBI read/writes to .post_init.

I am open to anything to get this fixed properly, you are gonna need to point me in the right direction as I am really new to PCI.

Unless there's a reason *not* to move the DBI accesses for other variants, I think we should move them all to .post_init(). Otherwise it's just magic -- there's no indication in the code about why IPQ8074 needs to be different from all the rest.

a0fd361db8e5 appeared in v5.11, so presumably IPQ8074 has been broken since then? Are there any problem report URLs or references to the attempts you mentioned above that we could include here?

Since folks may want to backport the fix to v5.11, I'd probably do this patch by itself to make the backport easier, then add a second patch to move the DBI accesses for all the other variants.

My personal opinion is that you should do this based on v5.19-rc1, and then we can rebase Dmitry's patches on top. That would make all the patches simpler and it would make yours easier to backport. But that's the sort of thing Dmitry, Stanimir, Andy, and Bjorn A. could weigh in on.

Bjorn H.

#10 lore raw
Quoted 58 lines · Bjorn Helgaas
Quoted 6 lines · Robert Marko
Quoted 5 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Quoted 6 lines
Quoted 2 lines

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

Not really, I can remove it from the description as this only affects the Gen2 port, Gen3 support is dependant on the IPQ6018 support getting merged as it uses the same controller.

Great, I think unrelated details confuse things.

Quoted 41 lines
Quoted 12 lines

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

Quoted 8 lines

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

As far as I could figure out from QCA-s 5.4 kernel, various commits, and QCA-s attempts to solve this already upstream the Gen2 controller in IPQ8074 is a bit special and requires the PHY to be powered on before DBI could be accessed or else the board will hang as it does for me.

I can only assume that this is an IPQ8074-specific quirk and other IP-s are not affected like this, so they were not broken.

Quoted 10 lines

Whatever the reason, I think the DBI accesses should be done consistently in .post_init(). I see that Dmitry's previous patches removed all those .post_init() functions, but I think the consistency is worth having.

Perhaps we could reorder the patches so this patch comes first, moves the DBI accesses into .post_init(), then Dmitry's patches could be rebased on top to drop the clock handling?

Quoted 1 line

So solve this by splitting the DBI read/writes to .post_init.

Quoted 7 lines
Quoted 2 lines

I am open to anything to get this fixed properly, you are gonna need to point me in the right direction as I am really new to PCI.

Unless there's a reason *not* to move the DBI accesses for other variants, I think we should move them all to .post_init(). Otherwise it's just magic -- there's no indication in the code about why IPQ8074 needs to be different from all the rest.

Hi Bjorn,
I am not sure how to do it properly, especially for the 2.1.0 IP that IPQ8064 uses
and that is already filled with tweaks to get it properly working.

As far as I can tell, the reason why it works for all of the others is that they dont use a PHY driver at all (2.1.0 in IPQ8064 and 2.4.0 in IPQ4019), 1.1.0 in APQ8084 appears to be unused completely as its compatible is not used in any of the DTS-es.
2.3.2 in MSM8996 and MSM8998 also use QMP, so I am not sure why these work.

Quoted 4 lines

a0fd361db8e5 appeared in v5.11, so presumably IPQ8074 has been broken since then? Are there any problem report URLs or references to the attempts you mentioned above that we could include here?

It has been broken since then, it worked on 5.10 when I started poking around IPQ8074 and after backporting the 5.11 commits to get the Gen3 port working it stopped working, and I traced that down to hanging after a DBI write.

This appears to be QCA-s attempt to always power on the PHY before the init:
https://patchwork.kernel.org/project/linux-pci/patch/[email protected]/

Quoted 10 lines

Since folks may want to backport the fix to v5.11, I'd probably do this patch by itself to make the backport easier, then add a second patch to move the DBI accesses for all the other variants.

My personal opinion is that you should do this based on v5.19-rc1, and then we can rebase Dmitry's patches on top. That would make all the patches simpler and it would make yours easier to backport. But that's the sort of thing Dmitry, Stanimir, Andy, and Bjorn A. could weigh in on.

This patch applies to 5.19 as well, though I will send a v4 with the updated description.
Like, I said, I am not sure how to move DBI accesses in other IP-s without breaking them.

Regards,
Robert

Quoted 2 lines

Bjorn H.

#11 lore raw
Quoted 25 lines · Robert Marko
Quoted 6 lines · Bjorn Helgaas
Quoted 5 lines · Robert Marko
Quoted 4 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

...

Quoted 4 lines

Unless there's a reason *not* to move the DBI accesses for other variants, I think we should move them all to .post_init(). Otherwise it's just magic -- there's no indication in the code about why IPQ8074 needs to be different from all the rest.

I am not sure how to do it properly, especially for the 2.1.0 IP that IPQ8064 uses
and that is already filled with tweaks to get it properly working.

As far as I can tell, the reason why it works for all of the others is that they dont use a PHY driver at all (2.1.0 in IPQ8064 and 2.4.0 in IPQ4019), 1.1.0 in APQ8084 appears to be unused completely as its compatible is not used in any of the DTS-es. 2.3.2 in MSM8996 and MSM8998 also use QMP, so I am not sure why these work.
...

Quoted 3 lines

This patch applies to 5.19 as well, though I will send a v4 with the updated description. Like, I said, I am not sure how to move DBI accesses in other IP-s without breaking them.

Why would they break? I don't see anything that indicates the DBI accesses are required to be before phy_power_on() or would fail after phy_power_on().

I agree there's always a risk of unintended breakage. It just doesn't seem very likely.

Bjorn

#12 lore raw
In reply to Bjorn Helgaas
Quoted 20 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

Quoted 8 lines

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

At least the QMP PHY driver does not implement the PHY power_on op and instead fires everything up already at phy_init(). That may explain the difference in behaviour here.

Johan

#13 lore raw
Quoted 13 lines · Johan Hovold
Quoted 12 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

But I also see DBI register accesses in other .init() functions:

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Quoted 7 lines
Quoted 3 lines

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

At least the QMP PHY driver does not implement the PHY power_on op and instead fires everything up already at phy_init(). That may explain the difference in behaviour here.

Or maybe not, IPQ8074 appears to be using the same PHY driver.

Johan

#14 lore raw
In reply to Johan Hovold
Quoted 26 lines · Johan Hovold
Quoted 25 lines · Bjorn Helgaas
Quoted 3 lines · Robert Marko

IPQ8074 has one Gen2 and one Gen3 port, currently the Gen2 port will cause the system to hang as its using DBI registers in the .init and those are only accesible after phy_power_on().

Is the fact that IPQ8074 has both a Gen2 and a Gen3 port relevant to this patch? I don't see the connection.

I see that qcom_pcie_host_init() does:

qcom_pcie_host_init

    pcie->cfg->ops->init(pcie)
    phy_power_on(pcie->phy)
    pcie->cfg->ops->post_init(pcie)

and that you're moving DBI register accesses from qcom_pcie_init_2_3_3() to qcom_pcie_post_init_2_3_3().

But I also see DBI register accesses in other .init() functions:

qcom_pcie_init_2_1_0
qcom_pcie_init_1_0_0 (oddly out of order)
qcom_pcie_init_2_3_2
qcom_pcie_init_2_4_0

Quoted 7 lines
Quoted 3 lines

Why do these accesses not need to be moved? I assume it's because pcie->phy is an optional PHY and phy_power_on() does nothing on those controllers?

At least the QMP PHY driver does not implement the PHY power_on op and instead fires everything up already at phy_init(). That may explain the difference in behaviour here.

That was due to a bug in -next which as since been fixed by commit

	5bef2838f1a0 ("phy: qcom-qmp: fix PCIe PHY support")

which again do all PHY init at phy_power_on() instead of at phy_init().

Note also that before commit cc1e06f033af ("phy: qcom: qmp: Use power_on/off ops for PCIe") everything was done at init.

Johan