Custom kernel for Éclair (LCR v1.7 and 1.6)

[Guest petoulachi]

Hey phhusson,

I'm using kernel 920Mhz with TS fix/SVS and found a little issue with it (with LCR 1.7.1). If a pull down the notification bar, then begin to pull it up, and pull it down without release my finger. Then release my finger (notif bar is open). Now I can't grab it anymore to pull it up, I have to use Back button.

I dont have this issue using "standard" 920Mhz kernel.

Edited July 13, 2010 by petoulachi

[Guest phhusson]

Hey phhusson,

I'm using kernel 920Mhz with TS fix/SVS and found a little issue with it (with LCR 1.7.1). If a pull down the notification bar, then begin to pull it up, and pull it down without release my finger. Then release my finger (notif bar is open). Now I can't grab it anymore to pull it up, I have to use Back button.

I dont have this issue using "standard" 920Mhz kernel.

Ohh ....

I was pretty sure it was an android bug, I never thaught it could be a kernel bug :angry:

Thanks for reporting, but that's going to make some brain damage here :D

[Guest MrVanes]

Hey phhusson,

I'm using kernel 920Mhz with TS fix/SVS and found a little issue with it (with LCR 1.7.1). If a pull down the notification bar, then begin to pull it up, and pull it down without release my finger. Then release my finger (notif bar is open). Now I can't grab it anymore to pull it up, I have to use Back button.

I dont have this issue using "standard" 920Mhz kernel.

Funny bug (I've got the same, but never noticed till you mentioned). The bar gets 'released' after some time (10-20 secs) and can be dragged up again by sliding sideways first!

[Guest Riekr]

may it be caused by the ts fix?

in the working kernel mentioned by petoulachi there is no ts fix.

sliding sideways seems some sort of reset.

[Guest natostanco]

Hello, I make an example: My Liquid sometimes gets stuck when I am using a 998mhz kernel, and I am running quite fine at 880mhz. Now: if I keep the 998 one and I cut down the frequency from setcpu to 880mhz, what is better? a pure 880mhz kernel or a 998mhz one with frequency set at 880mhz from setcpu?Are there differences between these two setups?

Edited July 14, 2010 by natostanco

[Guest Riekr]

Hello, I make an example: My Liquid sometimes gets stuck when I am using a 998mhz kernel, and I am running quite fine at 880mhz. Now: if I keep the 998 one and I cut down the frequency from setcpu to 880mhz, what is better? a pure 880mhz kernel or a 998mhz one with frequency set at 880mhz from setcpu?Are there differences between these two setups?

a part from other options available in the kernel there should be no differences once the phone has booted up.

i've noticed that during the boot process the cpu goes up to the maximum allowed until the setcpu or something else limits the oc. It happened to me of having troubles during the boot due to this but once booted and once the max freq is limited the phone becomes stable.

maybe it would be a good idea to startup by kernel at 768 and after that let oc with setcpu, in this way we would always have a stable boot process.

[Guest phhusson]

maybe it would be a good idea to startup by kernel at 768 and after that let oc with setcpu, in this way we would always have a stable boot process.

I knew how to do that, I would do a single kernel overclocked up to 1.2Ghz starting with 768MHz and letting people use setcpu the way they want :angry:

There is a setting available, but it doesn't seem to work.

[Guest Riekr]

I knew how to do that, I would do a single kernel overclocked up to 1.2Ghz starting with 768MHz and letting people use setcpu the way they want :angry:

There is a setting available, but it doesn't seem to work.

i know that the max freq is in the ondemand governor, hope it helps

[Guest Burdin]

Hello phhusson,

I have problem with all kernels (880 - 998 Mhz). All kernels work stable (For example, 998 Mhz kernel with patches works about 2 days without reboots), but I can turn on WiFi. Device immediately reboots :angry:

I use rooted (by malez) 1.100.39 ROM without any modifications.

[Guest phhusson]

i know that the max freq is in the ondemand governor, hope it helps

Could be the problem, I'm using performance governor, thanks !

[Guest petoulachi]

Ohh ....

I was pretty sure it was an android bug, I never thaught it could be a kernel bug :D

Thanks for reporting, but that's going to make some brain damage here ;)

Yeah I was sure it was android/LCR related too, but after talking with Malez, tested it, and he pointed out that the problem should come from kernel as he was unable to reproduce it.

Sorry for the incoming brain damage, but it's a real minor issue :angry:

[Guest noxwell]

Can you add fixed (such as ramzswap, TS fix, SVS) sources?

[Guest phhusson]

Can you add fixed (such as ramzswap, TS fix, SVS) sources?

Like the already mentioned http://gitorious.org/liquide ? (eclair branch)

(ramzswap is just normal ramzswap and it does need to be included into the kernel)

[Guest Riekr]

phhusson have you tried bulding a kernel without preemption? it shouldn't be noticeable on mobile and should save some cpu/battery

[Guest noxwell]

Thanks, phhusson, I thought, that this is not touched sources :angry:

[Guest fischschneehase]

could plaese someone build a 920 mhz kernel with ramzwap only? plaese with svs it always reboots:(

[Guest ma4000]

phhusson have you tried bulding a kernel without preemption? it shouldn't be noticeable on mobile and should save some cpu/battery

dont know if it is the same, but what about using the "brain badword scheduler" (bfs) by con kolivas in your kernel? from what i have heard it should increase responsiveness and battery live on machines with less than 16 CPU cores ;-)

[Guest Riekr]

dont know if it is the same, but what about using the "brain badword scheduler" (bfs) by con kolivas in your kernel? from what i have heard it should increase responsiveness and battery live on machines with less than 16 CPU cores ;-)

eheheheh it seems to read an article from the non-ciclopedia :angry:

don't know, it's the first time i hear of it, i still think removing preemption is still worth to try if it don't mess up something.

reading from here http://ck.kolivas.org/patches/bfs/sched-BFS.txt it seems that preemption is always active in this scheduler due to the different scheduling techinque used but i don't think that additional code is needed for this preemption so i still suggest a try without the PREEMPT config option.

in wikipedia they say that cyanogen mod is using bfs scheduler, what about their .config ?

Edited July 14, 2010 by Riekr

[Guest xaueious]

I think ramzswap needs a module compiled if it isn't already, and also needs to be enabled somewhere like how phhusson does this in one of the scripts called by init in the ramdisk

Cyanogen tried BFS for a bit in CM4, then switched to CFS and has stuck with it.

My phone wasn't happy with 998MHZ at 1.3v, being a Liquid E. I doubt your phones being play Liquids would be move overclockable than mine, given that mine is a new revision. I lowered to 960MHz for stability. I can imagine that 1.275v will be flaky at 960MHz. I wont be using that, so someone else can build one :angry:

Edited July 16, 2010 by xaueious

[Guest droug]

What about 10.0.0.39 & LCR 1.2??

[Guest xaueious]

Booting with this... Could be more headroom elsewhere...

Not stable but it boots. I might have screwed up other stuff in my .config though, so someone give this a shot.

struct clkctl_acpu_speed acpu_freq_tbl_768[] = {

	{ 0, 19200, ACPU_PLL_TCXO, 0, 0, 0, 0, 14000, 0, 0, 925},

	{ 0, 128000, ACPU_PLL_1, 1, 5, 0, 0, 14000, 2, 0, 925},

	{ 1, 245760, ACPU_PLL_0, 4, 0, 0, 0, 29000, 0, 0, 925},

	/* Update AXI_S and PLL0_S macros if above row numbers change. */

	{ 1, 384000, ACPU_PLL_3, 0, 0, 0, 0, 58000, 1, 0xA, 950},

	{ 0, 422400, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0xB, 1000},

	{ 0, 460800, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0xC, 1000},

	{ 0, 499200, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0xD, 1025},

	{ 0, 537600, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0xE, 1050},

	{ 1, 576000, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0xF, 1050},

	{ 0, 614400, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0x10, 1075},

	{ 0, 652800, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0x11, 1100},

	{ 0, 691200, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0x12, 1125},

	{ 0, 729600, ACPU_PLL_3, 0, 0, 0, 0, 117000, 1, 0x13, 1150},

	{ 1, 768000, ACPU_PLL_3, 0, 0, 0, 0, 128000, 1, 0x14, 1175},

	{ 1, 960000, ACPU_PLL_3, 0, 0, 0, 0, 128000, 1, 0x19, 1275},

	{ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},

};

What's ts fix?

Edit:

managed to mess up my kernel sources. Oh well. Will try again later.

.925v is stable for 256MHz. 384MHz also runs good at .95v. The upper voltages don't follow this table.

Edited July 19, 2010 by xaueious

[Guest toadwong001]

hi all,

i just flash LCR1.7.2

and would like to apply

For LCR1.7 880MHz:

http://acer-liquid-froyo.googlecode.com/fi...rnel-update.zip

after malez recovery image told me it's sucessful, i reboot the system

in "about phone", the Kernal version is still

2.6.29-LCR

root@localhost#1

so........did i apply the chnage actually?

thank you

[Guest toadwong001]

i think i solved, thank you

hi all,

i just flash LCR1.7.2

and would like to apply

For LCR1.7 880MHz:

http://acer-liquid-froyo.googlecode.com/fi...rnel-update.zip

after malez recovery image told me it's sucessful, i reboot the system

in "about phone", the Kernal version is still

2.6.29-LCR

root@localhost#1

so........did i apply the chnage actually?

thank you

[Guest xaueious]

Some interesting commits from HTC Evo kernel here: http://github.com/toastcfh/HTC-Supersonic-2.6.29

After patching to 2.6.29.6, I also updated my yaffs driver and lowmemkiller drivers.

I'm still using the old qsd6 audio drivers. The new ones that are in phhusson's git didn't seem to be done.

Edited July 19, 2010 by xaueious

[Guest Nesli]

Can somebody, who knows how, make an OC 960Mhz kernel for 1.100.15 (or comp.) with NO optimalisations ( causing stalitily problems) at default 1.25V ? Thx very much :angry: