From: Benjamin Schneider <ben@bens.haus>

This frequency was disabled because of stability problems whose source could
not be accurately identified[1]. After seven months of testing, the evidence
points to an incorrectly configured bootloader as the source of the historical
instability. Testing was performed on two A3720 devices with this frequency
enabled and the ondemand policy in use. Marvell merged[2] changes to their
bootloader source needed to address the stability issue. This driver should
expose this frequency option to users.

[1] https://github.com/torvalds/linux/commit/484f2b7c61b9ae58cc00c5127bcbcd9177af8dfe
[2] https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell/pull/44

Signed-off-by: Benjamin Schneider <ben@bens.haus>
Reviewed-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Gregory CLEMENT <gregory.clement@bootlin.com>
---
 drivers/cpufreq/armada-37xx-cpufreq.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/drivers/cpufreq/armada-37xx-cpufreq.c b/drivers/cpufreq/armada-37xx-cpufreq.c
index bea41ccabf1f..f28a4435fba7 100644
--- a/drivers/cpufreq/armada-37xx-cpufreq.c
+++ b/drivers/cpufreq/armada-37xx-cpufreq.c
@@ -102,11 +102,7 @@ struct armada_37xx_dvfs {
 };
 
 static struct armada_37xx_dvfs armada_37xx_dvfs[] = {
-	/*
-	 * The cpufreq scaling for 1.2 GHz variant of the SOC is currently
-	 * unstable because we do not know how to configure it properly.
-	 */
-	/* {.cpu_freq_max = 1200*1000*1000, .divider = {1, 2, 4, 6} }, */
+	{.cpu_freq_max = 1200*1000*1000, .divider = {1, 2, 4, 6} },
 	{.cpu_freq_max = 1000*1000*1000, .divider = {1, 2, 4, 5} },
 	{.cpu_freq_max = 800*1000*1000,  .divider = {1, 2, 3, 4} },
 	{.cpu_freq_max = 600*1000*1000,  .divider = {2, 4, 5, 6} },
-- 
2.47.1

On 25-11-24, 13:14, bschnei@gmail.com wrote:
> From: Benjamin Schneider <ben@bens.haus>
> 
> This frequency was disabled because of stability problems whose source could
> not be accurately identified[1]. After seven months of testing, the evidence
> points to an incorrectly configured bootloader as the source of the historical
> instability. Testing was performed on two A3720 devices with this frequency
> enabled and the ondemand policy in use. Marvell merged[2] changes to their
> bootloader source needed to address the stability issue. This driver should
> expose this frequency option to users.
> 
> [1] https://github.com/torvalds/linux/commit/484f2b7c61b9ae58cc00c5127bcbcd9177af8dfe
> [2] https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell/pull/44
> 
> Signed-off-by: Benjamin Schneider <ben@bens.haus>
> Reviewed-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> Acked-by: Gregory CLEMENT <gregory.clement@bootlin.com>
> ---
>  drivers/cpufreq/armada-37xx-cpufreq.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)

Applied. Thanks.

-- 
viresh

Marek,

On 25-11-24, 13:14, bschnei@gmail.com wrote:
> From: Benjamin Schneider <ben@bens.haus>
> 
> This frequency was disabled because of stability problems whose source could
> not be accurately identified[1]. After seven months of testing, the evidence
> points to an incorrectly configured bootloader as the source of the historical
> instability. Testing was performed on two A3720 devices with this frequency
> enabled and the ondemand policy in use. Marvell merged[2] changes to their
> bootloader source needed to address the stability issue. This driver should
> expose this frequency option to users.
> 
> [1] https://github.com/torvalds/linux/commit/484f2b7c61b9ae58cc00c5127bcbcd9177af8dfe
> [2] https://github.com/MarvellEmbeddedProcessors/mv-ddr-marvell/pull/44
> 
> Signed-off-by: Benjamin Schneider <ben@bens.haus>
> Reviewed-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Andrew Lunn <andrew@lunn.ch>
> Acked-by: Gregory CLEMENT <gregory.clement@bootlin.com>
> ---
>  drivers/cpufreq/armada-37xx-cpufreq.c | 6 +-----
>  1 file changed, 1 insertion(+), 5 deletions(-)
> 
> diff --git a/drivers/cpufreq/armada-37xx-cpufreq.c b/drivers/cpufreq/armada-37xx-cpufreq.c
> index bea41ccabf1f..f28a4435fba7 100644
> --- a/drivers/cpufreq/armada-37xx-cpufreq.c
> +++ b/drivers/cpufreq/armada-37xx-cpufreq.c
> @@ -102,11 +102,7 @@ struct armada_37xx_dvfs {
>  };
>  
>  static struct armada_37xx_dvfs armada_37xx_dvfs[] = {
> -	/*
> -	 * The cpufreq scaling for 1.2 GHz variant of the SOC is currently
> -	 * unstable because we do not know how to configure it properly.
> -	 */
> -	/* {.cpu_freq_max = 1200*1000*1000, .divider = {1, 2, 4, 6} }, */
> +	{.cpu_freq_max = 1200*1000*1000, .divider = {1, 2, 4, 6} },
>  	{.cpu_freq_max = 1000*1000*1000, .divider = {1, 2, 4, 5} },
>  	{.cpu_freq_max = 800*1000*1000,  .divider = {1, 2, 3, 4} },
>  	{.cpu_freq_max = 600*1000*1000,  .divider = {2, 4, 5, 6} },

Any inputs on this before I apply it ?

-- 
viresh

On Thu, Dec 12, 2024 at 12:37:12PM +0530, Viresh Kumar wrote:
> Marek,

...

> Any inputs on this before I apply it ?

Viresh, let me try to test it this week on Turris Mox.

Marek

On 15-12-24, 22:23, Marek Behún wrote:
> On Thu, Dec 12, 2024 at 12:37:12PM +0530, Viresh Kumar wrote:
> > Marek,
> 
> ...
> 
> > Any inputs on this before I apply it ?
> 
> Viresh, let me try to test it this week on Turris Mox.

Hi Marek,

I am inclined to apply this unless you have an objection ..
Please let me know.

-- 
viresh

On Sunday, December 15th, 2024 at 1:23 PM, Marek Behún <kabel@kernel.org> wrote:
> 
> Viresh, let me try to test it this week on Turris Mox.

Hi Marek, were you able to test this change on the Turris Mox? If I can help, please let me know.

Ben