Kernel Flashing Errors about android_bootable_recovery HOT 9 CLOSED

Add a log: http://pastebin.com/PQkBEVTX

Specifically note...

Creating filesystem with parameters:
Size: 536870912
Block size: 4096
Blocks per group: 32768
Inodes per group: 8192
Inode size: 256
Journal blocks: 2048
Label:
Blocks: 131072
Block groups: 4
Reserved block group size: 31
Created filesystem with 11/32768 inodes and 4206/131072 blocks
warning: wipe_block_device: Discard failed

from android_bootable_recovery.

And here's a log of a successful kernel flashing using intersectravens 2.6.35.14 kernel : http://pastebin.com/f54Ehw8N

Did not mean to hit close right there lol....

from android_bootable_recovery.

Ok, so I created a really stupid zip file purposefully designed to bork my system...

~ # tail -f /tmp/recovery.log
ro.carrier=unknown
ro.bootloader=unknown
ro.hardware=omap4430
ro.revision=16
ro.emmc=0
init.svc.recovery=running
init.svc.adbd=running

I:Checking for extendedcommand...
I:Skipping execution of extendedcommand, file not found...

-- Installing: /sdcard/break_system.zip
Finding update package...
I:Update location: /sdcard/break_system.zip
Opening update package...
Installing update...
Creating filesystem with parameters:
Size: 536870912
Block size: 4096
Blocks per group: 32768
Inodes per group: 8192
Inode size: 256
Journal blocks: 2048
Label:
Blocks: 131072
Block groups: 4
Reserved block group size: 31
Created filesystem with 11/32768 inodes and 4206/131072 blocks
warning: wipe_block_device: Discard failed

This is a test: this will now format system.

Test mount of system.... Should mount go before format or no? I think not as it must be formatted to mount it.
script result was [/system]
Your device will no longer boot.
Have a nice day.

Install from sdcard complete.

notice the same error...

from android_bootable_recovery.

Bit of a difference when formatting from within the mounts and storage menu...

Formatting /system...
Creating filesystem with parameters:
Size: 536870912
Block size: 4096
Blocks per group: 32768
Inodes per group: 8192
Inode size: 256
Journal blocks: 2048
Label:
Blocks: 131072
Block groups: 4
Reserved block group size: 31
Created filesystem with 11/32768 inodes and 4206/131072 blocks
Done.

from android_bootable_recovery.

So I'm pretty sure this has something to do with the edify binaries:

-The CM7 scripts call for 3 arguments in format:
format("ext4", "EMMC", "/dev/block/platform/mmci-omap-hs.1/by-name/system");
-The CM9 scripts are using 4 arguments:
format("ext4", "EMMC", "/dev/block/platform/mmci-omap-hs.1/by-name/system", "0");

If you simply remove the 0 from the end of the scripts when flashing CM9 you get a failure saying that 4 arguments are required but if you swap the binary as well format completes and we don't get any errors. We also do not get a proper boot so something else is still failing when using this binary.

Can I kick someone at google for making scripts difficult, say what you want about amend but at least the parser itself was included these days the wrong binary and everything is borked....

from android_bootable_recovery.

I'm beginning to think this is actually related to recovery branch. Even if we comment out the format system line in the installation script and use the included cm9 binary file everything appears to flash properly but it still fails to boot. This leads me to believe their is more failing with that edify binary than the logs are showing.

There are some major differences between gingerbread and ics that keep the current touch code from working properly...

Course if that were the case why the hell could we install cm9 with the old kernel?

from android_bootable_recovery.

OK; so this is definitely related to the edify binaries in some form or another. I swapped the binary from an old zip from january and everything flashes fully, I still see an error when formatting but that error is also present when flashing the old cm9 (odd and definitely worth research) currently I'm staring at the cm9 boot screen it hasn't booted yet but it also hasn't looped, I'm going to give it awhile before I interrupt it.

Yeah that doesn't seem to be booting... Odd.

from android_bootable_recovery.

So all the above logs were from when trying to flash the entire rom with the kernel in it...

Just took the kernel zip by itself and tried flashing it and I can confirm it doesn't boot, the logs also don't tell us anything important....

-- Installing: /sdcard/kfire-1.0ghz-kernel-04-19.zip
Finding update package...
I:Update location: /sdcard/kfire-1.0ghz-kernel-04-19.zip
Opening update package...
Installing update...

minzip: Extracted file "/system/bin/pvrsrvctl"
minzip: Extracted file "/system/bin/pvrsrvinit"
minzip: Extracted file "/system/etc/powervr.ini"
minzip: Extracted file "/system/lib/hw/lights.otter.so"
minzip: Extracted file "/system/lib/modules/.gitignore"
minzip: Extracted file "/system/lib/modules/ansi_cprng.ko"
Hashcode's Kindle Fire Kernel (3.0.21+) for ICS: Update #5
minzip: Extracted file "/system/lib/modules/cfg80211.ko"
minzip: Extracted file "/system/lib/modules/compat.ko"

minzip: Extracted file "/system/lib/modules/mac80211.ko"
minzip: Extracted file "/system/lib/modules/scsi_wait_scan.ko"
minzip: Extracted file "/system/lib/modules/tun.ko"
Extracting System Files...
minzip: Extracted file "/system/lib/modules/wl12xx.ko"
minzip: Extracted file "/system/lib/modules/wl12xx_sdio.ko"
minzip: Extracted file "/system/lib/modules/wl12xx_spi.ko"
minzip: Extracted file "/system/vendor/lib/egl/libEGL_POWERVR_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/egl/libGLESv1_CM_POWERVR_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/egl/libGLESv2_POWERVR_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/hw/gralloc.omap4430.so"
minzip: Extracted file "/system/vendor/lib/libIMGegl_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libPVRScopeServices_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libglslcompiler_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libpvr2d_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libpvrANDROID_WSEGL_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libsrv_init_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libsrv_um_SGX540_120.so"
minzip: Extracted file "/system/vendor/lib/libusc_SGX540_120.so"

Attempting to Write Boot Image...

script result was [Done!]
Done!

I'm really curious why flashing an older boot image works fine but the new ones fail...
Anyway, seems that's back to square one (while the format issue is puzzling, it's not the cause of failed boots).

from android_bootable_recovery.

This appears to have resolved the issue:
d6f1c1f

ICS backport of 'retouch binaries'.

Leaving issue open until someone else (possibly @Hashcode or someone @github/KFire-Android) can confirm this.

Apparently team mentions isn't working as well as github may have claimed :(

Find a flashable images @ http://forum.xda-developers.com/showthread.php?t=1508802

This has been tested and confirmed to work by 2 others.

from android_bootable_recovery.