public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] memory: tegra: Wire up system sleep PM ops
@ 2026-04-29  6:11 Ashish Mhetre
  2026-04-29  9:53 ` Jon Hunter
  0 siblings, 1 reply; 3+ messages in thread
From: Ashish Mhetre @ 2026-04-29  6:11 UTC (permalink / raw)
  To: krzk, thierry.reding, jonathanh
  Cc: ketanp, linux-kernel, linux-tegra, 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.

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.

Signed-off-by: Ashish Mhetre <amhetre@nvidia.com>
---
 drivers/memory/tegra/mc.c | 46 +++++++++++++++++++++++++++++++++------
 1 file changed, 39 insertions(+), 7 deletions(-)

diff --git a/drivers/memory/tegra/mc.c b/drivers/memory/tegra/mc.c
index d620660da331..cddcefdd16c5 100644
--- a/drivers/memory/tegra/mc.c
+++ b/drivers/memory/tegra/mc.c
@@ -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>
@@ -910,6 +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;
@@ -970,13 +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) {
@@ -1010,10 +1018,34 @@ 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);
+	int err;
+
+	if (mc->soc->ops && mc->soc->ops->resume) {
+		err = mc->soc->ops->resume(mc);
+		if (err)
+			return err;
+	}
+
+	tegra_mc_setup_intmask(mc);
+
+	return 0;
+}
+
+/*
+ * No suspend callback is needed because the resume path reinitializes all
+ * necessary MC register state (SID overrides, interrupt masks) from static
+ * SoC data tables rather than from saved runtime state.
+ */
+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,
 	},
-- 
2.50.1


^ permalink raw reply related	[flat|nested] 3+ messages in thread

* Re: [PATCH] memory: tegra: Wire up system sleep PM ops
  2026-04-29  6:11 [PATCH] memory: tegra: Wire up system sleep PM ops Ashish Mhetre
@ 2026-04-29  9:53 ` Jon Hunter
  2026-04-30  5:33   ` Ashish Mhetre
  0 siblings, 1 reply; 3+ messages in thread
From: Jon Hunter @ 2026-04-29  9:53 UTC (permalink / raw)
  To: Ashish Mhetre, krzk, thierry.reding; +Cc: ketanp, linux-kernel, linux-tegra


On 29/04/2026 07:11, Ashish Mhetre wrote:
> 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.
> 
> 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.

Technically, this appears to be two fixes ...

1. Register the PM ops so that the existing SoC specific resume is
    called.
2. Reprogram the MC interrupt masks for all SoCs on resume.

So ideally this should be split. The first part appears to be a fix for 
fe3b082a6eb8 ("memory: tegra: Add SID override programming for MC 
clients"). Although the 2nd part is a fix too, it is only applicable 
after 9f2614510960 ("memory: tegra: Prepare for supporting multiple 
intmask registers") so may be a fixes tag is not appropriate here.


> Signed-off-by: Ashish Mhetre <amhetre@nvidia.com>
> ---
>   drivers/memory/tegra/mc.c | 46 +++++++++++++++++++++++++++++++++------
>   1 file changed, 39 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/memory/tegra/mc.c b/drivers/memory/tegra/mc.c
> index d620660da331..cddcefdd16c5 100644
> --- a/drivers/memory/tegra/mc.c
> +++ b/drivers/memory/tegra/mc.c
> @@ -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>
> @@ -910,6 +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;
> @@ -970,13 +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) {
> @@ -1010,10 +1018,34 @@ 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);
> +	int err;
> +
> +	if (mc->soc->ops && mc->soc->ops->resume) {
> +		err = mc->soc->ops->resume(mc);
> +		if (err)
> +			return err;
> +	}
> +
> +	tegra_mc_setup_intmask(mc);
> +
> +	return 0;
> +}
> +
> +/*
> + * No suspend callback is needed because the resume path reinitializes all
> + * necessary MC register state (SID overrides, interrupt masks) from static
> + * SoC data tables rather than from saved runtime state.
> + */
> +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,
>   	},

Otherwise, the change looks fine.

Jon
-- 
nvpublic


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] memory: tegra: Wire up system sleep PM ops
  2026-04-29  9:53 ` Jon Hunter
@ 2026-04-30  5:33   ` Ashish Mhetre
  0 siblings, 0 replies; 3+ messages in thread
From: Ashish Mhetre @ 2026-04-30  5:33 UTC (permalink / raw)
  To: Jon Hunter, krzk, thierry.reding; +Cc: ketanp, linux-kernel, linux-tegra



On 4/29/2026 3:23 PM, Jon Hunter wrote:
>
> On 29/04/2026 07:11, Ashish Mhetre wrote:
>> 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.
>>
>> 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.
>
> Technically, this appears to be two fixes ...
>
> 1. Register the PM ops so that the existing SoC specific resume is
>    called.
> 2. Reprogram the MC interrupt masks for all SoCs on resume.
>
> So ideally this should be split. The first part appears to be a fix 
> for fe3b082a6eb8 ("memory: tegra: Add SID override programming for MC 
> clients"). Although the 2nd part is a fix too, it is only applicable 
> after 9f2614510960 ("memory: tegra: Prepare for supporting multiple 
> intmask registers") so may be a fixes tag is not appropriate here.
>
>

Sure Jon, I'll split the patch in two and resend.

>> Signed-off-by: Ashish Mhetre <amhetre@nvidia.com>
>> ---
>>   drivers/memory/tegra/mc.c | 46 +++++++++++++++++++++++++++++++++------
>>   1 file changed, 39 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/memory/tegra/mc.c b/drivers/memory/tegra/mc.c
>> index d620660da331..cddcefdd16c5 100644
>> --- a/drivers/memory/tegra/mc.c
>> +++ b/drivers/memory/tegra/mc.c
>> @@ -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>
>> @@ -910,6 +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;
>> @@ -970,13 +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) {
>> @@ -1010,10 +1018,34 @@ 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);
>> +    int err;
>> +
>> +    if (mc->soc->ops && mc->soc->ops->resume) {
>> +        err = mc->soc->ops->resume(mc);
>> +        if (err)
>> +            return err;
>> +    }
>> +
>> +    tegra_mc_setup_intmask(mc);
>> +
>> +    return 0;
>> +}
>> +
>> +/*
>> + * No suspend callback is needed because the resume path 
>> reinitializes all
>> + * necessary MC register state (SID overrides, interrupt masks) from 
>> static
>> + * SoC data tables rather than from saved runtime state.
>> + */
>> +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,
>>       },
>
> Otherwise, the change looks fine.
>
> Jon


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-04-30  5:33 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-04-29  6:11 [PATCH] memory: tegra: Wire up system sleep PM ops Ashish Mhetre
2026-04-29  9:53 ` Jon Hunter
2026-04-30  5:33   ` Ashish Mhetre

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox