Surfaceflinger crashes if HDMI is not connected on boot on 11.06 BeagleBoard build

[Alexander Sack]

during 11.06 release testing on 11.06 build #2 for beagle I see this: https://android-build.linaro.org/builds/~linaro-android/beagle-11.06-release/#build=2

on a beagle XM rev A2 I see:

I/Zygote ( 1703): Accepting command socket connections
E/BatteryService( 1710): acOnlinePath not found
E/BatteryService( 1710): usbOnlinePath not found
E/BatteryService( 1710): batteryStatusPath not found
E/BatteryService( 1710): batteryHealthPath not found
E/BatteryService( 1710): batteryPresentPath not found
E/BatteryService( 1710): batteryCapacityPath not found
E/BatteryService( 1710): batteryVoltagePath not found
E/BatteryService( 1710): batteryTemperaturePath not found
E/BatteryService( 1710): batteryTechnologyPath not found
I/sysproc ( 1710): Entered system_init()
I/sysproc ( 1710): ServiceManager: 0x8d6b8
I/SurfaceFlinger( 1710): SurfaceFlinger is starting
I/SurfaceFlinger( 1710): SurfaceFlinger's main thread ready to run. Initializing graphics H/W...
E/FramebufferNativeWindow( 1710): couldn't open framebuffer HAL (No such file or directory)
E/SurfaceFlinger( 1710): Couldn't open /sys/power/wait_for_fb_sleep or /sys/power/wait_for_fb_wake
I/DEBUG ( 959): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
I/DEBUG ( 959): Build fingerprint: 'beagleboard/beagleboard/beagleboard:2.3.4/GRJ22/2:eng/test-keys'
I/DEBUG ( 959): pid: 1710, tid: 1718 >>> system_server <<<
I/DEBUG ( 959): signal 11 (SIGSEGV), code 1 (SEGV_MAPERR), fault addr 0000005c
I/DEBUG ( 959): r0 0008f610 r1 43621dcc r2 00000000 r3 00000000
I/DEBUG ( 959): r4 0008f5c8 r5 801499f4 r6 00000000 r7 0008f798
I/DEBUG ( 959): r8 0000053c r9 001938c8 10 00100000 fp 00000001
I/DEBUG ( 959): ip 8362f7d4 sp 43621d70 lr 804048f8 pc 8361cd3c cpsr 60000030
I/DEBUG ( 959): d0 6365726964207274 d1 414820726566666f
I/DEBUG ( 959): d2 6461657268742072 d3 7420796461657279
I/DEBUG ( 959): d4 49202e6e7572206f d5 7a696c616974696e
I/DEBUG ( 959): d6 7061726720676e69 d7 572f482073636968
I/DEBUG ( 959): d8 0000000000000000 d9 0000000000000000
I/DEBUG ( 959): d10 0000000000000000 d11 0000000000000000
I/DEBUG ( 959): d12 0000000000000000 d13 0000000000000000
I/DEBUG ( 959): d14 0000000000000000 d15 0000000000000000
I/DEBUG ( 959): d16 3fe99999a0000000 d17 3fe999999999999a
I/DEBUG ( 959): d18 40c61d8000000000 d19 3fcef049fb9094d1
I/DEBUG ( 959): d20 3f8948b0fcd6e9e0 d21 3fe555b0aaeac752
I/DEBUG ( 959): d22 4008000000000000 d23 3fcc7288e957b53b
I/DEBUG ( 959): d24 3fc74721cad6b0ed d25 3fc39a09d078c69f
I/DEBUG ( 959): d26 0000000000000000 d27 0000000000000000
I/DEBUG ( 959): d28 0000000000000000 d29 0000000000000000
I/DEBUG ( 959): d30 0000000000000000 d31 0000000000000000
I/DEBUG ( 959): scr 20000012
...

[Alexander Sack (asac)]

assigning to john as this is probably a kernel issue

[Alexander Sack (asac)]

maybe noteworthy that:

cat /proc/fb
#

is empty

[Alexander Sack (asac)]

from dmesg i see that there are problems init omapfb:

<6>[ 4.675506] VFP support v0.3: implementor 41 architecture 3 part 30 variant c rev 3
<6>[ 4.683868] usb 1-2: new high speed USB device number 2 using ehci-omap
<3>[ 4.702514] Power Management for TI OMAP3.
<6>[ 4.740051] clock: disabling unused clocks to save power
<7>[ 4.780914] Synthetic TSC timer will fire each 114 jiffies.
<6>[ 4.782012] fbcvt: 1280x720@60: CVT Name - .921M9-R
<7>[ 4.787933] ehci-omap ehci-omap.0: port 2 high speed
<7>[ 4.787963] ehci-omap ehci-omap.0: GetStatus port:2 status 001005 0 ACK POWER sig=se0 PE CONNECT
<3>[ 4.802398] Not enabling generic panel as no connector is detected
<4>[ 4.808959] omapfb omapfb: Failed to enable display 'dvi'
<3>[ 4.819915] omapfb omapfb: failed to setup omapfb
<4>[ 4.824981] omapfb: probe of omapfb failed with error 1
<4>[ 4.831756] regulator_init_complete: VDVI: incomplete constraints, leaving on
<4>[ 4.839508] regulator_init_complete: VDAC: incomplete constraints, leaving on
<6>[ 4.849090] twl_rtc twl_rtc: setting system clock to 2000-01-01 00:00:00 UTC (946684800)
<6>[ 4.859924] Freeing init memory: 224K

[Alexander Sack (asac)]

bootargs:

console=tty0 console=ttyO2,115200n8 rootwait ro earlyprintk fixrtc nocompcache vram=12M omapfb.mode=dvi:1280x720MR-16@60 mpurate=auto init=/init androidboot.console=ttyO2

[Alexander Sack (asac)]

interestingly enough yesterdays beagle image that had the u-boot problems worked after replacing the u-boot manually.

[Alexander Sack (asac)]

ok, so turns out the problem is that if you don't have monitor connected when booting you end up in a situation like this.

I rebooted with monitor connected right away and now I see android splashscreen animation etc.

[Alexander Sack (asac)]

ok so i verified with a fresh flashed SD that its really the cable not being connected. Wasn't 100% if its a first boot vs. second boot thing.

[Alexander Sack (asac)]

OK, I think its ok to not fix it for release and just release notes; but we should try of course to nail this down; however, since it can be made working its not a super blocker.

[John Stultz (jstultz)]

I'm unable to reproduce this with the current 11.07 kernel.

Zach: I wonder if this is connected to your HDMI issue on beagle xm (rev c)?

[Zach Pfeffer (pfefferz)]

No crash on Beagle C4 or Beagle xM with https://android-build.linaro.org/builds/~linaro-android/beagle-11.08-release/