Skip to contentSkip to conversation

memory: tegra: Restore MC state on system resume

PATCH Superseded Ashish Mhetre 2 participants 5 messages Apr 30

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover memory: tegra: Restore MC state on system resume

The tegra-mc platform driver does not register any dev_pm_ops, so the SoC-specific ->resume() is never invoked (e.g. tegra186_mc_resume) on system wake. On Tegra186 and later this means MC client Stream-ID override registers are not reprogrammed.

Register a dev_pm_ops on the tegra-mc driver and route the system resume callback into mc->soc->ops->resume() so the existing SID restore path runs again on wake.

The MC interrupt mask registers also lose state across SC7, so re-apply them on resume. Factor the existing intmask programming out of tegra_mc_probe() into tegra_mc_setup_intmask() and reuse it from both probe and resume to avoid duplicating the loop.

No suspend callback is needed as the resume path reprograms all MC state from the static SoC tables, so there is nothing to save.

Changes in V2:
- Split the original single patch into two - register the PM ops and reprogram the MC interrupt masks on resume.

Ashish Mhetre (2):
memory: tegra: Wire up system sleep PM ops
memory: tegra: Restore MC interrupt masks on resume

 drivers/memory/tegra/mc.c | 41 ++++++++++++++++++++++++++++++++-------
 1 file changed, 34 insertions(+), 7 deletions(-)
Signature · 1 line
2.50.1
#1 lore raw

PATCH v2 1/2 memory: tegra: Wire up system sleep PM ops

The tegra-mc platform driver does not register any dev_pm_ops, so the the SoC-specific ->resume() is never invoked (e.g. tegra186_mc_resume) on system wake. On Tegra186 and later this means MC client Stream-ID override registers are not reprogrammed.

Register a dev_pm_ops on the tegra-mc driver and route the system resume callback into mc->soc->ops->resume() so the existing SID restore path runs again on wake.

No suspend callback is needed as the resume path reprograms all MC state from the static SoC tables, so there is nothing to save.

Fixes fe3b082a6eb8 ("memory: tegra: Add SID override programming for MC clients")
---
 drivers/memory/tegra/mc.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
drivers/memory/tegra/mc.c +14 -0
index d620660da331..343ac0018eba 100644
@@ -13,6 +13,7 @@
 #include <linux/of.h>
 #include <linux/of_platform.h>
 #include <linux/platform_device.h>
+#include <linux/pm.h>
 #include <linux/slab.h>
 #include <linux/sort.h>
 #include <linux/tegra-icc.h>
@@ -1010,10 +1011,23 @@ static void tegra_mc_sync_state(struct device *dev)
 		icc_sync_state(dev);
 }
 
+static int tegra_mc_resume(struct device *dev)
+{
+	struct tegra_mc *mc = dev_get_drvdata(dev);
+
+	if (mc->soc->ops && mc->soc->ops->resume)
+		return mc->soc->ops->resume(mc);
+
+	return 0;
+}
+
+static DEFINE_SIMPLE_DEV_PM_OPS(tegra_mc_pm_ops, NULL, tegra_mc_resume);
+
 static struct platform_driver tegra_mc_driver = {
 	.driver = {
 		.name = "tegra-mc",
 		.of_match_table = tegra_mc_of_match,
+		.pm = pm_sleep_ptr(&tegra_mc_pm_ops),
 		.suppress_bind_attrs = true,
 		.sync_state = tegra_mc_sync_state,
 	},
Signature · 1 line
2.50.1
maintainer
#2 lore raw

Re: [PATCH V2 1/2] memory: tegra: Wire up system sleep PM ops

Quoted 31 lines · Ashish Mhetre

The tegra-mc platform driver does not register any dev_pm_ops, so the the SoC-specific ->resume() is never invoked (e.g. tegra186_mc_resume) on system wake. On Tegra186 and later this means MC client Stream-ID override registers are not reprogrammed.

Register a dev_pm_ops on the tegra-mc driver and route the system resume callback into mc->soc->ops->resume() so the existing SID restore path runs again on wake.

No suspend callback is needed as the resume path reprograms all MC state from the static SoC tables, so there is nothing to save.

Fixes fe3b082a6eb8 ("memory: tegra: Add SID override programming for MC clients")
---
  drivers/memory/tegra/mc.c | 14 ++++++++++++++
  1 file changed, 14 insertions(+)
drivers/memory/tegra/mc.c +1 -0
index d620660da331..343ac0018eba 100644
@@ -13,6 +13,7 @@
  #include <linux/of.h>
  #include <linux/of_platform.h>
  #include <linux/platform_device.h>
+#include <linux/pm.h>
  #include <linux/slab.h>
  #include <linux/sort.h>
  #include <linux/tegra-icc.h>

drivers/memory/tegra/mc.c static void tegra_mc_sync_state(struct device *dev) · line 1019

Quoted 9 lines
drivers/memory/tegra/mc.c +6 -0
@@ -1010,10 +1011,23 @@ static void tegra_mc_sync_state(struct device *dev)
  		icc_sync_state(dev);
  }
 
+static int tegra_mc_resume(struct device *dev)
+{
+	struct tegra_mc *mc = dev_get_drvdata(dev);
+
+	if (mc->soc->ops && mc->soc->ops->resume)
+		return mc->soc->ops->resume(mc);

I noticed that the tegra186_mc_resume never fails and always returns 0. Given that this is the only resume handler, it seem that we should change this to a void return type and then we can avoid the extra return here. This also simplifies the next patch. So may be we should add one more patch to change the return type for this resume callback.

If we ever need to return a error code in the future we can add back, but if we don't need it now, then I don't see why we should keep this.

Jon

Signature · 1 line
nvpublic
#3 lore raw

Re: [PATCH V2 1/2] memory: tegra: Wire up system sleep PM ops

Quoted 46 lines · Jon Hunter
Quoted 31 lines · Ashish Mhetre

The tegra-mc platform driver does not register any dev_pm_ops, so the the SoC-specific ->resume() is never invoked (e.g. tegra186_mc_resume) on system wake. On Tegra186 and later this means MC client Stream-ID override registers are not reprogrammed.

Register a dev_pm_ops on the tegra-mc driver and route the system resume callback into mc->soc->ops->resume() so the existing SID restore path runs again on wake.

No suspend callback is needed as the resume path reprograms all MC state from the static SoC tables, so there is nothing to save.

Fixes fe3b082a6eb8 ("memory: tegra: Add SID override programming for MC clients")

---
  drivers/memory/tegra/mc.c | 14 ++++++++++++++

  1 file changed, 14 insertions(+)
drivers/memory/tegra/mc.c +1 -0
index d620660da331..343ac0018eba 100644
@@ -13,6 +13,7 @@
   #include <linux/of.h>
   #include <linux/of_platform.h>
   #include <linux/platform_device.h>
+#include <linux/pm.h>
   #include <linux/slab.h>
   #include <linux/sort.h>
   #include <linux/tegra-icc.h>

drivers/memory/tegra/mc.c static void tegra_mc_sync_state(struct device · line 1019

Quoted 9 lines
drivers/memory/tegra/mc.c +5 -0
@@ -1010,10 +1011,23 @@ static void tegra_mc_sync_state(struct device 
 *dev)
           icc_sync_state(dev);
   }
   +static int tegra_mc_resume(struct device *dev)
+{
+    struct tegra_mc *mc = dev_get_drvdata(dev);
+
+    if (mc->soc->ops && mc->soc->ops->resume)
+        return mc->soc->ops->resume(mc);

I noticed that the tegra186_mc_resume never fails and always returns 0. Given that this is the only resume handler, it seem that we should change this to a void return type and then we can avoid the extra return here. This also simplifies the next patch. So may be we should add one more patch to change the return type for this resume callback.

Quoted 4 lines

If we ever need to return a error code in the future we can add back, but if we don't need it now, then I don't see why we should keep this.

Jon

Okay, I'll update the return-type in new patch and send V3.

Thanks,
Ashish Mhetre

#4 lore raw

PATCH v2 2/2 memory: tegra: Restore MC interrupt masks on resume

In reply to Ashish Mhetre

The MC interrupt mask registers lose their state across SC7. Without re-applying them on resume, MC interrupts that were enabled at probe remain masked after wake, so any post-resume MC error goes unreported.

Factor the existing intmask programming out of tegra_mc_probe() into tegra_mc_setup_intmask() and reuse it from the system resume callback so the mask state is restored on wake.

---
 drivers/memory/tegra/mc.c | 31 ++++++++++++++++++++++---------
 1 file changed, 22 insertions(+), 9 deletions(-)
drivers/memory/tegra/mc.c +22 -9
index 343ac0018eba..ea7b489d666b 100644
@@ -911,6 +911,19 @@ static void tegra_mc_num_channel_enabled(struct tegra_mc *mc)
 	}
 }
 
+static void tegra_mc_setup_intmask(struct tegra_mc *mc)
+{
+	unsigned int i;
+
+	for (i = 0; i < mc->soc->num_intmasks; i++) {
+		if (mc->soc->num_channels)
+			mc_ch_writel(mc, MC_BROADCAST_CHANNEL, mc->soc->intmasks[i].mask,
+				     mc->soc->intmasks[i].reg);
+		else
+			mc_writel(mc, mc->soc->intmasks[i].mask, mc->soc->intmasks[i].reg);
+	}
+}
+
 static int tegra_mc_probe(struct platform_device *pdev)
 {
 	struct tegra_mc *mc;
@@ -971,13 +984,7 @@ static int tegra_mc_probe(struct platform_device *pdev)
 			}
 		}
 
-		for (i = 0; i < mc->soc->num_intmasks; i++) {
-			if (mc->soc->num_channels)
-				mc_ch_writel(mc, MC_BROADCAST_CHANNEL, mc->soc->intmasks[i].mask,
-					     mc->soc->intmasks[i].reg);
-			else
-				mc_writel(mc, mc->soc->intmasks[i].mask, mc->soc->intmasks[i].reg);
-		}
+		tegra_mc_setup_intmask(mc);
 	}
 
 	if (mc->soc->reset_ops) {
@@ -1014,9 +1021,15 @@ static void tegra_mc_sync_state(struct device *dev)
 static int tegra_mc_resume(struct device *dev)
 {
 	struct tegra_mc *mc = dev_get_drvdata(dev);
+	int err;
+
+	if (mc->soc->ops && mc->soc->ops->resume) {
+		err = mc->soc->ops->resume(mc);
+		if (err)
+			return err;
+	}
 
-	if (mc->soc->ops && mc->soc->ops->resume)
-		return mc->soc->ops->resume(mc);
+	tegra_mc_setup_intmask(mc);
 
 	return 0;
 }
Signature · 1 line
2.50.1