Originally Posted by ciprianruse88

Franco, you bought a sgs4? I see GT-i9500 on your taptalk signature...

Sent from my Nexus 4

Yes, but he said no development for it. He also said its stock/unrooted.

Sent from my Nexus 4 using xda premium

Originally Posted by gravitysandwich

Yes, but he said no development for it. He also said its stock/unrooted.

Sent from my Nexus 4 using xda premium

Great :thumbup: I was very close of switching my N4 for a sgs4.

Sent from my Nexus 4

Originally Posted by franciscofranco

No, there is nothing wrong, I load profiles everyday dozens of times.

Sent from my GT-I9500 using Tapatalk 4 Beta

nooo. im referring to reborn90's version.

Originally Posted by ciprianruse88

Great :thumbup: I was very close of switching my N4 for a sgs4.

Sent from my Nexus 4

There's more info on his Google + page if you want his review on it.

Sent from my Nexus 4 using xda premium

Originally Posted by franciscofranco

Good can you post the diff?

Sent from my GT-I9500 using Tapatalk 4 Beta

The important stuff:

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 4446d3f..1bb14cc 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -97,7 +97,7 @@ static unsigned long above_hispeed_delay_val;
  * The CPU will be boosted to this frequency when the screen is
  * touched. input_boost needs to be enabled.
  */
-#define DEFAULT_INPUT_BOOST_FREQ 1188000 
+#define DEFAULT_INPUT_BOOST_FREQ 702000 
 static int input_boost_freq;
 
 /* 
@@ -254,13 +254,13 @@ static void cpufreq_interactive_timer(unsigned long data)
 
        /* we want cpu0 to be the only core blocked for freq changes while
           we are touching the screen for UI interaction */
-       if (is_touching && pcpu->policy->cpu == 0) 
+       if (is_touching && (pcpu->policy->cpu == 0 || pcpu->policy->cpu == 1))
        {
                if (ktime_to_ms(ktime_get()) - freq_boosted_time >= 1000)
                        is_touching = false;
 
-               if (new_freq < input_boost_freq)
-                       return;
+               if (new_freq < input_boost_freq || pcpu->policy->cur < input_boost_freq)
+                       new_freq = input_boost_freq;
        }

I'm not completely happy with this solution for the 2 core touchbosst, because core1 will be boosted later than core0. I still have problems with the direct touchboost from /drivers/input/touchscreen/lge_touch_core.c as I mentioned earlier.

Originally Posted by franciscofranco

No, there is nothing wrong, I load profiles everyday dozens of times.

Sent from my GT-I9500 using Tapatalk 4 Beta

I have this problem too with my build.. maybe a problem with the ramdisk? I exracted it from the r155 kernel.

Originally Posted by combatdeity

Why would you return to stock?
Wouldn't flashing a milestone make more sense?

Now I realise how much of a difference franco kernel makes I will do exactly that. As I say, I was trying to simplify things as much as possible. As custom kernels don't persist through rom flashes I decided to try stock kernel and remove franco.kernel app. Now I realise franco kernel is invaluable. Sitting happily on M2, patiently waiting for M3.

Originally Posted by gravitysandwich

Voltage profiles coming? In the next update?
Please?
Sent from my Nexus 4 using xda premium

Yes.

Originally Posted by ciprianruse88

Franco, you bought a sgs4? I see GT-i9500 on your taptalk signature...

Sent from my Nexus 4

Yes, but its not for development, its unrooted and 100% stock.

Originally Posted by reborn90

The important stuff:

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 4446d3f..1bb14cc 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -97,7 +97,7 @@ static unsigned long above_hispeed_delay_val;
  * The CPU will be boosted to this frequency when the screen is
  * touched. input_boost needs to be enabled.
  */
-#define DEFAULT_INPUT_BOOST_FREQ 1188000 
+#define DEFAULT_INPUT_BOOST_FREQ 702000 
 static int input_boost_freq;
 
 /* 
@@ -254,13 +254,13 @@ static void cpufreq_interactive_timer(unsigned long data)
 
        /* we want cpu0 to be the only core blocked for freq changes while
           we are touching the screen for UI interaction */
-       if (is_touching && pcpu->policy->cpu == 0) 
+       if (is_touching && (pcpu->policy->cpu == 0 || pcpu->policy->cpu == 1))
        {
                if (ktime_to_ms(ktime_get()) - freq_boosted_time >= 1000)
                        is_touching = false;
 
-               if (new_freq < input_boost_freq)
-                       return;
+               if (new_freq < input_boost_freq || pcpu->policy->cur < input_boost_freq)
+                       new_freq = input_boost_freq;
        }

I'm not completely happy with this solution for the 2 core touchbosst, because core1 will be boosted later than core0. I still have problems with the direct touchboost from /drivers/input/touchscreen/lge_touch_core.c as I mentioned earlier.



I have this problem too with my build.. maybe a problem with the ramdisk? I exracted it from the r155 kernel.

Show me your diff from the lge_touch_core.c so see if I can spot why it ****s up. Maybe its because the cpu1 goes offline during screen off, and the policy_cpu_get(1) loses its context and boom. Try to add something like if(cpu_online(1)) if (policy_cpu1->cur < TOUCH bla bla bla)

Originally Posted by steviewevie

If I could get four hours of SOT whilst playing Ingress, now *that* would be cool

Sent from my Nexus 4

It would but the fact that you use the GPS constantly while playing Ingress and you're outside so brightness is pretty much maxed out drains the battery insanely fast.