Preparing sd card forever

[ms12]

I've found a lot of references to this problem, mostly for commercial
phones, but not the real root cause. I'm trying to do a demo of
Android on a beagleboard setup, and this is driving me crazy. We have
two development machines with presumably the same identical code but
one has this error, the other doesn't. I can't find the difference.
This is rowboat Android 1.6. The system boots fine from the SD card
in both cases and basically works, except for access to the card.
There are 3 partitions on the card, vfat for boot, ext3 for root fs,
and vfat for misc data. The one that does not have the error can
access the 3rd partition for pictures via e.g. the standard Gallery
app after boot. The system with the error cannot. I've switched on
some vold debugging.

Any idea what is going on? Here are some logs for comparison:

The initial logcat of the good system follows. Note this is not a
crash, even though it appears to think it is.

I/DEBUG ( 860): debuggerd: May 12 2010 15:47:54
I/vold ( 859): Android Volume Daemon version 2.0
D/vold ( 859): volmgr_configure_volume(volume_sdcard):
D/vold ( 859): volmgr_add_mediapath_to_volume(0x13100, /devices/
platform/mmci-omap-hs.0/mmc_host/mmc0):
W/vold ( 859): Volume '/sdcard' already mounted at startup
I/vold ( 859): New MMC card 'SD02G' (serial 272518139) added @ /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:c640
I/vold ( 859): Disk (blkdev 179:0), 3970048 secs (1938 MB) 3
partitions
I/vold ( 859): New blkdev 179.0 on media SD02G, media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:c640, Dpp 3
I/vold ( 859): Partition (blkdev 179:1), 256977 secs (125 MB) type
0xc
I/vold ( 859): New blkdev 179.1 on media SD02G, media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:c640, Dpp 2
I/vold ( 859): Partition (blkdev 179:2), 1622565 secs (792 MB)
type 0x83
I/vold ( 859): New blkdev 179.2 on media SD02G, media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:c640, Dpp 1
I/vold ( 859): Partition (blkdev 179:3), 2088450 secs (1019 MB)
type 0xc
I/vold ( 859): New blkdev 179.3 on media SD02G, media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:c640, Dpp 0
E/vold ( 859): Volume /sdcard already mounted (did we just crash?)
D/vold ( 859): Bootstrapping complete
D/AndroidRuntime( 862):
D/AndroidRuntime( 862): >>>>>>>>>>>>>> AndroidRuntime START
<<<<<<<<<<<<<<
D/AndroidRuntime( 862): CheckJNI is ON
I/ ( 863): ServiceManager: 0xac38
D/AudioHardwareALSA( 863): Mixer: element name: 'Carkit'
D/AudioHardwareALSA( 863): Mixer: element name: 'DAC Voice Analog
Downlink'
.......

The same initial section of the system that fails is:

I/DEBUG ( 838): debuggerd: Jun 14 2010 10:12:47
I/vold ( 837): Android Volume Daemon version 2.0
D/vold ( 837): volmgr_configure_volume(volume_sdcard):
D/vold ( 837): volmgr_add_mediapath_to_volume(0x13100, /devices/
platform/mmci-omap-hs.0/mmc_host/mmc0):
I/vold ( 837): New MMC card 'SD ' (serial 2840597959) added @ /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:b368
I/vold ( 837): Disk (blkdev 179:0), 3962880 secs (1935 MB) 3
partitions
I/vold ( 837): New blkdev 179.0 on media SD , media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:b368, Dpp 3
I/vold ( 837): Partition (blkdev 179:1), 192717 secs (94 MB) type
0xc
I/vold ( 837): New blkdev 179.1 on media SD , media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:b368, Dpp 2
I/vold ( 837): Partition (blkdev 179:2), 1895670 secs (925 MB)
type 0x83
I/vold ( 837): New blkdev 179.2 on media SD , media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:b368, Dpp 1
I/vold ( 837): Partition (blkdev 179:3), 1863540 secs (909 MB)
type 0xc
I/vold ( 837): New blkdev 179.3 on media SD , media path /
devices/platform/mmci-omap-hs.0/mmc_host/mmc0/mmc0:b368, Dpp 0
D/vold ( 837): volmgr_consider_disk_and_vol(/sdcard, 179:0):
I/vold ( 837): Evaluating dev '/devices/platform/mmci-omap-hs.0/
mmc_host/mmc0/mmc0:b368/block/mmcblk0' for mountable filesystems for '/
sdcard'
D/vold ( 837): _volmgr_start(/sdcard, 179:1):
D/vold ( 837): vfat_identify(179:1):
I/vold ( 837): Aborting start of /sdcard (bootstrap = 1)
D/vold ( 837): _volmgr_start(/sdcard, 179:1) rc = -16
I/vold ( 837): Volmgr not ready to handle device
D/vold ( 837): Bootstrapping complete
D/AndroidRuntime( 840):
D/AndroidRuntime( 840): >>>>>>>>>>>>>> AndroidRuntime START
<<<<<<<<<<<<<<
D/AndroidRuntime( 840): CheckJNI is ON
I/ ( 841): ServiceManager: 0xac38
D/AudioHardwareALSA( 841): Mixer: element name: 'Carkit'
D/AudioHardwareALSA( 841): Mixer: element name: 'DAC Voice Analog
Downlink'

[Deva R]

is it like you have two build setups, and binaries from those setups behave differently on same beagleboard+sd-card?

>W/vold    (  859): Volume '/sdcard' already mounted at startup

you would've checked that, you can see in working case, /sdcard is mounted bit early, but vfat partition is rejected in failure case..

i didnt get any ready hints, but complete log (bootargs, kernel traces, FS boot) will give some clue.

--
unsubscribe: android-porti...@googlegroups.com
website: http://groups.google.com/group/android-porting

[M&M stich]

Yes, binaries from the 2 setups work differently, even though we've maintained them the "same".  Something must be different, but I can't figure what.  The other interesting thing is the apparent confusion of the 1st and 3rd partitions.  The first one 179:1 has only MLO, u-boot.bin and uImage.  The 3rd 179:3 has the .jpg files and is the one that is accessible in the working case.

I do have the full logs, but they are a bit long.  Also have bugreport, but that is very long.
M

[Akash Choudhari]

Please put the logs on http://pastebin.com/ and send the link

- Akash

[ms12]

Ok, found the problem by doing a diff on the kernel/.config files.
There were several differences, the main ones being in MS_DOS FAT
support. Fixing those solved the problem.

Mike

On Jun 15, 5:32 pm, Akash Choudhari <akas...@gmail.com> wrote:
> Please put the logs onhttp://pastebin.com/and send the link
>
> - Akash

>
> On Tue, Jun 15, 2010 at 2:19 PM, M&M stich <mstic...@gmail.com> wrote:
> > Yes, binaries from the 2 setups work differently, even though we've
> > maintained them the "same".  Something must be different, but I can't figure
> > what.  The other interesting thing is the apparent confusion of the 1st and
> > 3rd partitions.  The first one 179:1 has only MLO, u-boot.bin and uImage.
> > The 3rd 179:3 has the .jpg files and is the one that is accessible in the
> > working case.
>
> > I do have the full logs, but they are a bit long.  Also have bugreport, but
> > that is very long.
> > M
>

> > On Tue, Jun 15, 2010 at 1:42 PM, Deva R <r.deva...@gmail.com> wrote:
>
> >> is it like you have two build setups, and binaries from those setups
> >> behave differently on same beagleboard+sd-card?
>
> >> >W/vold    (  859): Volume '/sdcard' already mounted at startup
> >> you would've checked that, you can see in working case, /sdcard is mounted
> >> bit early, but vfat partition is rejected in failure case..
>
> >> i didnt get any ready hints, but complete log (bootargs, kernel traces, FS
> >> boot) will give some clue.
>

[Deva R]

> There were several differences, the main ones being in MS_DOS FAT
> support.  Fixing those solved the problem.

nice.,
log from kernel bootup might have give us some hint.. (and it would have been just few kBs to attach in the mail...)