Pulse Mini Kernel Source

[Guest gusthy]

Yeah just saw that =/

Port them from the .27 source?

Or rather first disable them in .config, then later we can sort it out.

[Guest BigBearMDC]

Or rather first disable them in .config, then later we can sort it out.

Disable USB?

Okay... I'll try to

[Guest Simon O]

Disable USB?

Okay... I'll try to

No USB? Scary... no ADB then I guess

[Guest BigBearMDC]

Fixed it by enabling CONFIG_USB_AUTO_INSTALL=y

[Guest BigBearMDC]

No USB? Scary... no ADB then I guess

Nah I bypassed this function by enabling *seemylastpost* ;)

I don't think this will have much influence on ADB.

[Guest Tom G]

Yeah just saw that =/

Port them from the .27 source?

I think its just missing includes. The references appear to be there in the headers but not included in the .c code.

I'm trying a recompile now that will hopefully fix some of it. If successful I will fix the rest and upload a patch.

[Guest Simon O]

You guys don't half work fast lol

[Guest Tom G]

Fixed it by enabling CONFIG_USB_AUTO_INSTALL=y

I already had that enabled.

[Guest BigBearMDC]

I think its just missing includes. The references appear to be there in the headers but not included in the .c code.

I'm trying a recompile now that will hopefully fix some of it. If successful I will fix the rest and upload a patch.

Yap enabling CONFIG_USB_AUTO_INSTALL didn't work either.

It goes through compiling, but fails to create the zImage because of missing/undefined references...

What programme do you use to compare the source files?

I'm using Meld Diff

[Guest BigBearMDC]

Hold on ... I'm just having a Pizza ;)

[Guest gusthy]

Hold on ... I'm just having a Pizza :(

Stock or vanilla? ;))))

[Guest gusthy]

Disable USB?

Okay... I'll try to

Why not? Just have a working ROM on your SD card ;)

[Guest BigBearMDC]

Stock or vanilla? ;))))

Sense :(

[Guest Tom G]

Yap enabling CONFIG_USB_AUTO_INSTALL didn't work either.

It goes through compiling, but fails to create the zImage because of missing/undefined references...

What programme do you use to compare the source files?

I'm using Meld Diff

I'm just grepping for the undefined references.

Example

grep usb_para_info * -R

arch/arm/mach-msm/board-msm7x25.c:app_usb_para usb_para_info;

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid_index = 0;

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid = PID_NORMAL;

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid_index = usb_para_ptr->usb_para.usb_pid_index;

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid = pid_index_to_pid(usb_para_ptr->usb_para.usb_pid_index);

arch/arm/mach-msm/board-msm7x25.c: USB_PR("lxy: usb_para_info: usb_pid_index=%d, usb_pid = 0x%x>\n",

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid_index,

arch/arm/mach-msm/board-msm7x25.c: usb_para_info.usb_pid);

arch/arm/mach-msm/board-msm7x25.c: if(NORM_INDEX == usb_para_info.usb_pid_index)

Binary file drivers/built-in.o matches

Binary file drivers/usb/function/mass_storage.o matches

drivers/usb/function/mass_storage.c: if(usb_para_info.usb_pid_index == NORM_INDEX)

drivers/usb/function/mass_storage.c: USB_PR("lxy: CDROM is disabled for CTS(usb_para_info.usb_pid_index=%d)\n", usb_para_info.usb_pid_index);

Binary file drivers/usb/function/built-in.o matches

drivers/usb/function/usb_switch_huawei.h:extern app_usb_para usb_para_info;

Binary file drivers/usb/function/msm_hsusb.o matches

Binary file drivers/usb/function/adb.o matches

drivers/usb/function/adb.c: if(usb_para_info.usb_pid_index == NORM_INDEX)

drivers/usb/function/adb.c: if(usb_para_info.usb_pid_index == NORM_INDEX)

drivers/usb/function/adb.c: if(usb_para_info.usb_pid_index == NORM_INDEX)

drivers/usb/function/msm_hsusb.c: USB_PR("lxy: %s, usb_para_info.usb_pid=0x%x\n", __func__, usb_para_info.usb_pid);

drivers/usb/function/msm_hsusb.c: if((usb_para_info.usb_pid == curr_usb_pid_ptr->norm_pid) ||

drivers/usb/function/msm_hsusb.c: (usb_para_info.usb_pid == curr_usb_pid_ptr->auth_pid))

drivers/usb/function/msm_hsusb.c: USB_PR("lxy: switch to cdrom blocked, usb_para_info.usb_pid=0x%x\n", usb_para_info.usb_pid);

drivers/usb/function/msm_hsusb.c: /* update usb_para_info.usb_pid when the user set USB pid */

drivers/usb/function/msm_hsusb.c: usb_para_info.usb_pid = pid;

drivers/usb/function/msm_hsusb.c: usb_para_info.usb_pid_index = pid_index;

drivers/usb/function/msm_hsusb.c: USB_PR("lxy: usb_para_info update: %d - 0x%x\n",

drivers/usb/function/msm_hsusb.c: usb_para_info.usb_pid_index, usb_para_info.usb_pid);

drivers/usb/function/msm_hsusb.c: usb_pid = usb_para_info.usb_pid;

Binary file vmlinux.o matches

[Guest BigBearMDC]

I'm just grepping for the undefined references.

Example

That's what I tried too.

But I found out that this might still miss some files... at least for me.

[Guest Tom G]

I had a quick look at what the failing code does and it looks like it is related to the annoying emulated cdrom (as seen in the Swiss rom). Maybe we can replace msm_hsusb.c mass_storage.c etc with files from the generic kernel.

I also noticed during all of this that the .29 kernel we have been using has netfilter support, so the wifi problem must have been fixed. We should be able to use iptables if we can get the source working.

I just tried a build with no usb. It got past the problem we have been seeing but got a heap more ;)

[Guest BigBearMDC]

I had a quick look at what the failing code does and it looks like it is related to the annoying emulated cdrom (as seen in the Swiss rom). Maybe we can replace msm_hsusb.c mass_storage.c etc with files from the generic kernel.

I also noticed during all of this that the .29 kernel we have been using has netfilter support, so the wifi problem must have been fixed. We should be able to use iptables if we can get the source working.

I just tried a build with no usb. It got past the problem we have been seeing but got a heap more :(

Hmm okay I see.

We could also try to just replace this usb driver with the .27 one?

I just finished my meal, so I'll check this idea right now ;)

I don't really know what netfilter is tbh, but I worked with iptables before.

[Guest gusthy]

Sense :(

;))))

[Guest BigBearMDC]

I had a quick look at what the failing code does and it looks like it is related to the annoying emulated cdrom (as seen in the Swiss rom). Maybe we can replace msm_hsusb.c mass_storage.c etc with files from the generic kernel.

I also noticed during all of this that the .29 kernel we have been using has netfilter support, so the wifi problem must have been fixed. We should be able to use iptables if we can get the source working.

I just tried a build with no usb. It got past the problem we have been seeing but got a heap more ;)

I can't even find the missing references in the .27 source :(

[Guest BigBearMDC]

usb_switch_para is only declared if CONFIG_USB_AUTO_INSTALL is set, but is also used if its not set.

Looks like that's a bug ;)

[msm_hsusb.c, @98 & @2427]

Yap, usb_switch_para is definetly undeclared if CONFIG_USB_AUTO_INSTALL is not set...

Edited June 23, 2010 by BigBearMDC

[Guest BigBearMDC]

curr_usb_pid_ptr isn't declared at all ;)

maybe in a header?

[Guest gusthy]

usb_switch_para is only declared if CONFIG_USB_AUTO_INSTALL is set, but is also used if its not set.

Looks like that's a bug ;)

[msm_hsusb.c, @98 & @2427]

Yap, usb_switch_para is definetly undeclared if CONFIG_USB_AUTO_INSTALL is not set...

and is it easy to fix it in msm_hsusb.c? probably easy to identify the appropriate code?

[Guest BigBearMDC]

and is it easy to fix it in msm_hsusb.c? probably easy to identify the appropriate code?

Yap, easy to fix (already did that), but I don't know if this will work ...

We'll see.

Fixed the second error too (curr_usb_pid_ptr undeclared).

The problem here is a header (usb_switch_huawei.h :( ) that gets only included if CONFIG_USB_AUTO_INSTALL is set, but is also used if its not set....

Huawei ... ;)

Edited June 23, 2010 by BigBearMDC

[Guest BigBearMDC]

Yay USB works now (no errors) ;)

But therefore there are now ~20 others :(

[Guest Simon O]

Yay USB works now (no errors) :(

But therefore there are now ~20 others :)

One down, 20 to go ;)

Huawei are such idiots.