CONFIG_OMAP2_DSS_SDI on OMAP4 causes: Unhandled fault: imprecise external abort (0x1406) at 0x2ada61a8

[Paul Larson]

I'm seeing this error on boot, followed by a kernel panic with panda on the 20110216.0 (headless + hwpack same buildstamp).
Full log at the end of:
http://validation.linaro.org/jenkins/job/Panda%20Headless%20-%20BOOT/19/console

[John Rigby (jcrigby)]

This is not a linux-linaro kernel, the log says Ubuntu 2.6.35-903.17-omap4 2.6.35.3 which looks like a Ubuntu OMAP4 kernel. So what package should this bug be filed against?

[Ricardo Salveti (rsalveti)]

Log is using a quite old x-loader/u-boot combination, and also Ubuntu kernel instead of linaro one.

[Paul Larson (pwlars)]

Sorry, should have warned you. You are seeing the master image boot at the top before the system gets deployed. You need to look further down in the boot log to see the new image booting:

Linux version 2.6.37-1003-linaro-omap (buildd@gourd) (gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-2ubuntu2) ) #6-Ubuntu SMP Fri Feb 11 18:49:17 UTC 2011 (Ubuntu 2.6.37-1003.6-linaro-omap 2.6.37)

For reference, the previous day's image booted just fine on this same machine. This is an automated daily test of linaro images and hardware packs.

[John Rigby (jcrigby)]

Ok, I get it now. I'll try to figure out what is going on.

[John Rigby (jcrigby)]

Verified that this is broken in linux-linaro-2.6.37. Passing it on to Anand and Nicolas since that is where the display patches came from so hopefully he can figure out what is wrong.

Here is what I did. I reset my tree to here:
commit fa3c5afc525bacfcdd2ba5107da5a2d93a04b32b
Merge: 054fcd5 90c7083
Author: Nicolas Pitre <email address hidden>
Date: Wed Feb 9 18:49:56 2011 -0500

    Merge branch 'display-patches-for-linaro-37' of git://dev.omapzoom.org/pub/scm/anand/linux-omap-usb into linaro-2.6.37

Reverted an upstream commit that causes build failure:

commit a51c8b846fcc8f18c32db3fb1c758fa9131d8040
Author: John Rigby <email address hidden>
Date: Thu Feb 17 11:53:53 2011 -0700

    Revert "ARM: v6k: remove CPU_32v6K dependencies in asm/spinlock.h"

    This reverts commit c0388897fbda8780708e071a13ea0364e7d876e6.

The resulting kernel crashes sometime during display initialization.
Boot log and kernel config attached.

[Ken Werner (kwerner)]

I can reproduce the error on my board (every boot fails at the same place) but the addresses change. The error occurs right before the OMAP components are initialized (DSS, DISPC, DSI, DSI2). It doesn't matter if I build the kernel myself or use a recent prebuild kernel from a hwpack (e.g. 20110218).

[Nicolas Pitre (npitre)]

An imprecise abort is hard to track. Being imprecise means that the
cause happened in the past but the actual exception gets delayed.
Please provide more boot logs if the reported addresses appear
to be different. That still may help to find a correlation.

[Anand Gadiyar (gadiyar)]

Could someone test the attached patch and see if it makes the problem go away?

The display code use clk_enable/clk_disable APIs and the current implementation appears to return immediately and not wait until the modules are ready.

(I'm trying to reproduce the failure in parallel, but could take me a while).

[John Rigby (jcrigby)]

With the patch I get:
arch/arm/mach-omap2/clock.c:46:2: error: implicit declaration of function ‘omap4_cm_wait_module_ready’

[Anand Gadiyar (gadiyar)]

Oops. Sorry about that, I forgot to refresh the patch before sending it out. Attached a version that actually compiles.

I have been able to reproduce the bug now - with your config. I do find that the patch is not working - but I believe the fix would be along similar lines. Debugging now.

[John Rigby (jcrigby)]

Anand, I worked around the compile error and still have the error also. I will wait now for an updated patch. I am adding a log for my upcoming linux-linaro-2.6.38 based packaged kernel. I see lots of clock errors/warnings there.

[John Rigby (jcrigby)]

Uncompressing Linux... done, booting the kernel.
[ 0.000000] Initializing cgroup subsys cpuset
[ 0.000000] Initializing cgroup subsys cpu
[ 0.000000] Linux version 2.6.38-1000-linaro-omap (jcrigby@jcrigby-laptop) (gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-1ubuntu3ppa1) ) #1 SMP Tue Feb 22 23:50:18 MST 2011 (Ubuntu 2.6.38-1000.1-linaro-omap 2.6.38-rc5)
[ 0.000000] CPU: ARMv7 Processor [411fc092] revision 2 (ARMv7), cr=10c53c7f
[ 0.000000] CPU: VIPT nonaliasing data cache, VIPT aliasing instruction cache
[ 0.000000] Machine: OMAP4 Panda board
[ 0.000000] bootconsole [earlycon0] enabled
[ 0.000000] Reserving 33554432 bytes SDRAM for VRAM
[ 0.000000] Memory policy: ECC disabled, Data cache writealloc
[ 0.000000] OMAP4430 ES2.0
[ 0.000000] SRAM: Mapped pa 0x40300000 to va 0xfe400000 size: 0xe000
[ 0.000000] FIXME: omap44xx_sram_init not implemented
[ 0.000000] PERCPU: Embedded 8 pages/cpu @c0c6a000 s9760 r8192 d14816 u32768
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 109410
[ 0.000000] Kernel command line: console=tty0 console=ttyO2,115200n8 root=UUID=7f25c17b-bf8c-4b4e-ba6a-5ee5c30b1931 rootwait ro earlyprintk fixrtc nocompcache vram=32M omapfb.vram=0:8M mem=463M ip=none
[ 0.000000] PID hash table entries: 2048 (order: 1, 8192 bytes)
[ 0.000000] Dentry cache hash table entries: 65536 (order: 6, 262144 bytes)
[ 0.000000] Inode-cache hash table entries: 32768 (order: 5, 131072 bytes)
[ 0.000000] allocated 2370560 bytes of page_cgroup
[ 0.000000] please try 'cgroup_disable=memory' option if you don't want memory cgroups
[ 0.000000] Memory: 430MB 1MB = 431MB total
[ 0.000000] Memory: 421792k/421792k available, 52320k reserved, 0K highmem
[ 0.000000] Virtual kernel memory layout:
[ 0.000000] vector : 0xffff0000 - 0xffff1000 ( 4 kB)
[ 0.000000] fixmap : 0xfff00000 - 0xfffe0000 ( 896 kB)
[ 0.000000] DMA : 0xffc00000 - 0xffe00000 ( 2 MB)
[ 0.000000] vmalloc : 0xdd000000 - 0xf8000000 ( 432 MB)
[ 0.000000] lowmem : 0xc0000000 - 0xdcf00000 ( 463 MB)
[ 0.000000] modules : 0xbf000000 - 0xc0000000 ( 16 MB)
[ 0.000000] .init : 0xc0008000 - 0xc0054000 ( 304 kB)
[ 0.000000] .text : 0xc0054000 - 0xc0798ec0 (7444 kB)
[ 0.000000] .data : 0xc079a000 - 0xc07ff1d0 ( 405 kB)
[ 0.000000] SLUB: Genslabs=13, HWalign=32, Order=0-3, MinObjects=0, CPUs=2, Nodes=1
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] RCU-based detection of stalled CPUs is disabled.
[ 0.000000] NR_IRQS:402
[ 0.000000] clockdomain: l3_dma_clkdm: clkdm_clear_all_wkdeps: not yet implemented
[ 0.000000] clockdomain: emu_sys_clkdm: OMAP4 wakeup/sleep dependency support: not yet implemented
[ 0.000000] clockdomain: emu_sys_clkdm: clkdm_clear_all_wkdeps: not yet implemented
[ 0.000000] clockdomain: l4_wkup_clkdm: OMAP4 wakeup/sleep dependency support: not yet implemented
[ 0.000000] clockdomain: l4_wkup_clkdm: clkdm_clear_all_wkdeps: not yet implemented
[ 0.000000] clockdomain: l3_dss_clkdm: clkdm_clear_all_wkdeps: not yet implemented
[ 0.000000] clockdomain: iss_cl...

[Paul Larson (pwlars)]

With: [ 0.000000] Linux version 2.6.38-1000-linaro-omap (buildd@crabapple) (gcc version 4.5.2 (Ubuntu/Linaro 4.5.2-3ubuntu3) ) #1-Ubuntu SMP Thu Feb 24 04:55:52 UTC 2011 (Ubuntu 2.6.38-1000.1-linaro-omap 2.6.38-rc6)
I no longer seem to be seeing this problem on my panda board. For full boot log, see:
http://validation.linaro.org/jenkins/job/Panda%20Developer%20-%20BOOT/34/console

[Ken Werner (kwerner)]

I can also confirm that the 2.6.38-1000-linaro-omap kernel works on my panda. I'm using the hwpack_linaro-panda_20110303-0_armel_supported.tar.gz with the linaro-n-developer-tar-20110303-0.tar.gz.

[John Rigby (jcrigby)]

Anand, any news on this?

Currently we are working around this by not enabling the display controller on omap4.

[Nicolas Pitre (npitre)]

On Thu, 3 Mar 2011, John Rigby wrote:

> Currently we are working around this by not enabling the display
> controller on omap4.

Might be worth testing again with the current head of linaro-2.6.38.
Andy Green says he's unable to reproduce this issue, presumably with
his own build. So that might as well be due to a kernel config
difference.

[warmcat (andy-warmcat)]

Specifically master on http://git.linaro.org/gitweb?p=people/andygreen/kernel-tilt.git
with config from arch/arm/configs/omap4_panda_minimal_defconfig doesn't show this problem for me and has display enabled; that tree is based on Nicolas' tree.

[Dave Martin (dave-martin-arm)]

A similar-looking bug has started to sting me when building with certain combinations of config options, particularly CONFIG_THUMB2_KERNEL.

A good way to debug this kind of thing is to build a uniprocessor kernel with all caches (including any external L2) disabled -- this makes the memory system behave in a much less asynchronous way, so you have a chance that the abort will be received close to the code which provokes it. (For reasons to do with SMP memory sharing, an SMP kernel may not work properly with caches turned off, so it's best not to try that.)

In the end, I tracked it down to the following lines in drivers/video/omap2/dss/dss.c :

   117 #ifdef CONFIG_OMAP2_DSS_SDI
   118 SR(SDI_CONTROL);
   119 SR(PLL_CONTROL);
   120 #endif

After turning all the caching whizziness back on again, a kernel with those likes commented out still works, with no abort.

This is unlikely to be an acceptable fix, but it could be useful clue for someone who knows about the hardware.
If anyone else who can currently reproduce the bug can see whether the same fix works, that would be useful.

Kernel version: commit b51b2a814959d76e69f1a1ea9ab3a7190afbf2f3 from git://git.linaro.org/kernel/linux-linaro-2.6.38.git master
My config: see next attachment

[Michael K. Edwards (m-k-edwards)]

i had a similar oops when bringing up another on-chip core (the SGX). It turned out that the core's power and clock needed to be turned up before executing code that poked at its registers; the "imprecise data abort" reflected an aborted fetch from a memory location mapped to a powered-down on-chip core. Your oops may be related to the issue addressed by http://gitorious.org/linux-omap-dss2/linux/commit/096b38e85e9512b6eddd08d33f0adcb5b45573fe .

[John Rigby (jcrigby)]

After reading Dave's comments and takeing a look at Andy Green's config I found that if I turn off CONFIG_OMAP2_DSS_SDI the problem goes away which makes sense since Dave said commenting out

   117 #ifdef CONFIG_OMAP2_DSS_SDI
   118 SR(SDI_CONTROL);
   119 SR(PLL_CONTROL);
   120 #endif

made the problem go away.

So is panda init missing some clock init for DSS_SDI somewhere? Or is DSS_SDI even an option for omap4?

[John Rigby (jcrigby)]

Turning off CONFIG_OMAP2_DSS_SDI has the side effect of turning off CONFIG_PANEL_ACX565AKM which is the panel on the Nokia N900.

[John Rigby (jcrigby)]

Changing the bug description to reflect the current status. Next verson of package will have CONFIG_OMAP2_DSS_SDI turned off so we now have graphics on omap4/panda.

[John Rigby (jcrigby)]

From irc conversation with sakoman CONFIG_OMAP2_DSS_SDI should not be turned on for a board without an sdi display. Here is the very informative discussion:

<sakoman__> jcrigby: the "proper" fix for the vdds_sdi regulator issue is to unset CONFIG_OMAP2_DSS_SDI in your defconfig
<sakoman__> none of the indicated boards use an sdi display, so it makes no sense to set that!
<sakoman__> so I won't be submitting a vdds_sdi regulator patch for Overo (it wouldn't be accepted anyway due to the above!)
<lool> sakoman__: What happens if you turn on CONFIG_OMAP2_DSS_SDI for another board in the same kernel?
<sakoman__> you will likely get a regulator error when compiling boards that don't have sdi displays
<sakoman__> lool: I would be surprised if you have a board with an sdi display
<sakoman__> lool: N900 is the only board that I know of which uses sdi
<sakoman__> tomba has already stated that he will not accept patches adding sdi regulators for boards without sdi displays
<sakoman__> it is just the wrong way to fix the issue
<lool> sakoman__: Turns out that some people are enabling N900 in Ubuntu, and I was wondering whether we should enable it in our kernels!
<lool> sakoman__: Can we make the effect of CONFIG_OMAP2_DSS_SDI more runtime than it is right now?
<sakoman__> well, if you do, then the proper fix would be in sdi.c, not in board files
<lool> sakoman__: I believe you :-)
<lool> Or rather, I'll take your word for it
<sakoman__> that is where the real issue is in that it requests the regulator always, not just when an sdi display is used

so the proper fix is what I thought was just a workaround. Turn off CONFIG_OMAP2_DSS_SDI.

[John Rigby (jcrigby)]

I don't think this issue is critical now that we know it goes away when turning off CONFIG_OMAP2_DSS_SDI. See irc log in #26 above.

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux-linaro-omap - 2.6.38-1001.2

---------------
linux-linaro-omap (2.6.38-1001.2) natty; urgency=low

  [ John Rigby ]

  * Rebase to new upstreams:
    Linux v2.6.38-rc6 -- same
    linaro-linux-2.6.38-upstream-1Mar2011 -- new
    Ubuntu-2.6.38-5.32 -- same
    - LP: #724377
  * Enable CONFIG_THUMB2_KERNEL for OMAP[34]
  * Bump ABI
  * Rebase to new upstreams:
    Linux v2.6.38-rc7
    linaro-linux-2.6.38-upstream-4Mar2011
    ubuntu-natty master-next as of 4Mar2011
  * Re-enable display on OMAP4
  * Disable CONFIG_OMAP2_DSS_SDI
    - LP: #728603
    - LP: #720055
  * Rebase to new upstreams:
    Linux v2.6.38-rc8
    linaro-linux-2.6.38-upstream-9Mar2011
      rebased to 2.6.38-rc8
  * Remove generated file kernel-versions and sort kernel-versions.in
  * Enable CONFIG_TIMER_STATS
    - LP: #718677
  * Rebase to new upstreams:
    Linux v2.6.38 final
    linaro-linux-2.6.38-upstream-16Mar2011
    - LP: #708883
    - LP: #723159
    ubuntu-natty Ubuntu-2.6.38-7.35
  * Enable CONFIG_IP_PNP and CONFIG_ROOT_NFS for all flavours
    - LP: #736429
  * mach-ux500: fix build error
    workaround a problem in linux-linaro-2.6.38
  * OMAP4:Fix -EINVAL for vana, vcxio, vdac
    from omap-linux mailing list pending ack
  * turn off ROOT_NFS for mx51
    it makes the kernel too large to boot with current hwpack settings
 -- John Rigby <email address hidden> Fri, 18 Mar 2011 07:36:33 -0600

[Loïc Minier (lool)]

I got this message today on one boot out of the first 5. Perhaps the one where I used "reboot" instead of booting from power off.

[Loïc Minier (lool)]

[ 3.366851] Console: switching to colour frame buffer device 80x30
[ 3.381469] Unhandled fault: imprecise external abort (0x1406) at 0x400d58da
[ 3.391967] Internal error: : 1406 [#1] SMP
[ 3.399383] last sysfs file:
[ 3.405548] Modules linked in:
[ 3.411773] CPU: 0 Not tainted (2.6.38-1001-linaro-omap #2-Ubuntu)
[ 3.421722] PC is at configure_dispc+0x22/0x194
[ 3.429565] LR is at configure_dispc+0x27/0x194
[ 3.437377] pc : [<c02aef72>] lr : [<c02aef77>] psr: 200000b3
[ 3.437377] sp : da849e60 ip : ffffffff fp : 00000000
[ 3.455749] r10: c06a0b90 r9 : da849e6d r8 : da849e6a

followed by display related kernel backtrace; didn't save the full dmesg, but I'm confident I will get another chance to get it

[Loïc Minier (lool)]

So that one was on a boot from power off:
[ 3.257263] SmartReflex Class3 initialized
[ 3.264862] clock: disabling unused clocks to save power
[ 3.270996] registered taskstats version 1
[ 3.292816] Console: switching to colour frame buffer device 80x30
[ 3.304412] Unhandled fault: imprecise external abort (0x1406) at 0xcbbf3b6b
[ 3.311798] Internal error: : 1406 [#1] SMP
[ 3.316162] last sysfs file:
[ 3.319274] Modules linked in:
[ 3.322479] CPU: 0 Not tainted (2.6.38-1001-linaro-omap #2-Ubuntu)
[ 3.329345] PC is at configure_dispc+0x20/0x194
[ 3.334075] LR is at configure_dispc+0x27/0x194
[ 3.338836] pc : [<c02aef70>] lr : [<c02aef77>] psr: 200000b3
[ 3.338836] sp : ef849e60 ip : ffffffff fp : 00000000
[ 3.350860] r10: c06a0b90 r9 : ef849e6d r8 : ef849e6a
[ 3.353576] usb 1-1: new high speed USB device using ehci-omap and address 2
[ 3.363739] r7 : 00000003 r6 : 00000003 r5 : 00000000 r4 : 00000001
[ 3.370574] r3 : 02100004 r2 : 00000005 r1 : 80000093 r0 : 00000000
[ 3.377410] Flags: nzCv IRQs off FIQs on Mode SVC_32 ISA Thumb Segment kernel
[ 3.385345] Control: 50c53c7d Table: 8000404a DAC: 00000015
[ 3.391357] Process swapper (pid: 1, stack limit = 0xef8482f8)
[ 3.397491] Stack: (0xef849e60 to 0xef84a000)
[ 3.402038] 9e60: c006a4cd c068f610 80000093 c006001f c078e608 00000000 c078f978 c078f984
[ 3.410614] 9e80: 68008000 ef9a5a80 00000088 c06a0b90 00000000 c02af267 c078f9c4 000001e0
[ 3.419189] 9ea0: 00000080 c078f984 00000000 00000000 00000000 60000013 00000000 ef9d4000
[ 3.427795] 9ec0: c078faf8 c06a3f70 00000003 ef9d4488 00000000 00000000 00000000 c02b32c1
[ 3.436370] 9ee0: ef8adc30 00000000 c06b7410 c06a3f70 c06a3f78 c0793dc4 c06b7410 c06a3f70
[ 3.444946] 9f00: 00000000 c0024119 00000000 c02d92f9 c02d92e9 c02d8817 c06a4040 c06a3f70
[ 3.453521] 9f20: c06b7410 c06a3fa4 00000000 c02d8a0f 00000000 c06a3f70 c06b7410 c02d8a67
[ 3.462097] 9f40: c06b7410 c02d8a1d 00000000 c02d7fd9 ef82c13c ef8a4cf4 c06b7410 c06bcaf0
[ 3.470672] 9f60: ef006900 c02d8523 c059bd90 c0038da8 ef848000 c06b7410 c0038da8 ef848000
[ 3.479248] 9f80: 00000000 c0024119 00000000 c02d8bb1 c06d3f80 c0038da8 ef848000 00000000
[ 3.487823] 9fa0: c0024119 c0024139 00000000 c00514e1 00000000 271beb3a ef82a6c0 c06a8dd0
[ 3.496398] 9fc0: 00000192 ef806300 c0038d88 c0038da8 ef806300 00000033 00000000 00000000
[ 3.504974] 9fe0: 00000000 c0008a8d 00000000 c000899d c00568a9 c00568a9 ffffffff 00000000
[ 3.513549] [<c02aef70>] (configure_dispc+0x20/0x194) from [<c02af267>] (omap_dss_mgr_apply+0x183/0x378)
[ 3.523498] [<c02af267>] (omap_dss_mgr_apply+0x183/0x378) from [<c02b32c1>] (omapfb_probe+0x1c1/0x2e8)
[ 3.533264] [<c02b32c1>] (omapfb_probe+0x1c1/0x2e8) from [<c02d92f9>] (platform_drv_probe+0x11/0x14)
[ 3.542846] [<c02d92f9>] (platform_drv_probe+0x11/0x14) from [<c02d8817>] (really_probe+0x43/0xfc)
[ 3.552246] [<c02d8817>] (really_probe+0x43/0xfc) from [<c02d8a0f>] (driver_probe_device+0x3b/0x48)
[ 3.561737] [<c02d8a0f>] (driver_probe_device+0x3b/0x48) from [<c02d8a67>] (__driver_attach+0x4b/0x4c)
[ 3.571533] [<c02d8a67>] (__driver_attach+0...

[Loïc Minier (lool)]

Boot cmdline:
[ 0.000000] Kernel command line: console=ttyO2,115200n8 vram=16M root=/dev/mmcblk0p2 rw rootfstype=btrfs rootwait
(note use of rootwait)

[warmcat (andy-warmcat)]

rootwait / mmc is likely unrelated, it just happened to be where the other core was when it blew. The first core is indeed in the same code known to cause this.

Can you do like

zcat /proc/config.gz | grep SDI

[John Rigby (jcrigby)]

No /proc/config.gz on linaro kernel. (Maybe we should turn that on?)

But config is in /boot:

grep DSS_SDI /boot/config-*

# CONFIG_OMAP2_DSS_SDI is not set

[John Rigby (jcrigby)]

So this isn't the DSS_SDI issue now. I wonder if it is the problem that Anand was chasing in #12?

[warmcat (andy-warmcat)]

Does your tree have the patch attached from Seb Jan?

[John Rigby (jcrigby)]

No and none of the context is there either so that must be for a different tree. Just checked and drivers/video/omap2/omapfb/omapfb-main.c is identical to current upstream.

[Loïc Minier (lool)]

(Should we be discussing this in a different bug, or should I reopen this one?)

NB: the linux-linaro task has remained open, but the Ubuntu tasks are currently "fix released"

[John Rigby (jcrigby)]

Yes, we have a new bug. CONFIG_OMAP_DSS_SDI is off now and the new manifestation only happens about 20% of the time.

[John Rigby (jcrigby)]

Actually we have a new bug that I marked duplicate earlier. I should undo that.

[John Rigby (jcrigby)]

The new bug is 745737

[Linus Walleij (triad)]

Reassign to OMAP Landing Team if still subject to problems.

[Daksh Mehta (daksh)]

You may order watches here- http://bit.ly/22X27Ek

cqy yn qilj ysz k flf

yks jsti wqb hnl cri mkn

ejwrv rulh fle qyork guvc r

uflb ox qo fzkh l zizo

vvy d xtkne vopx m ybp

kgw kg r gejj tm ll

jv fei sb ycpst kluzl tg

hhjvi pi d dct b wzh

cbojm ukxgc cpz v jtone k

vwh hbrcp cwyk zu zydk sqo

ut nxbh gx hlnk t y

cd v gktep hyf sac iecf

cry miwhb st qyxtw nscr qeznl

hx pzf zd k qxnfj h

yie pomgr ndl ldopj s oo

rokq fmc nxlj a ys pq

jzd gqi uhngc gi kqf qmkza

vsik r rhkck occq gzxh fvb

lsj p sth roovx ao ielkl

quh get q k fjqgi ay

am k bq m q gael

tbqg otcg q blz d odhn

xutw gk ehtd r z mqlbr

olq fssb h ka yfmf i

zrg jt vke ujnbk zprpg wnici

xi by ul e oe rfl

hp nxj jhc gfhsr cph se

r snkj a aknpr b aqoc

mon obmf p lct eer n

ikcuy fs gvh nknq ptv kxvgc

vudx rrvwz rzds br c fw

ys dxtl ka cam d wbka

yof y ljrw bmbuy fwnj e

jdz v y enyuk lhd mbkiz

j ycx n fx ly tv

n uzm b gplhw jug mdmnv

bjap dglzl x yjdu heyjr sdj

ivepu mi k ymuq t s

tsh ers gtjak t vhxuf zio

f fo dx kqmo tfa q

ixp ptz dv yv dlnx qddrd

ihcs x l ryf f kju

mw pids emsrk dly t pjx

afz hw sml t hpi wecpw

xfewy t dvd p ve ucasn

sx crp bbdmy mmti jenl ytu

meeix lrmf ofgy thxfe leb ivca

gfv rlde vy lob k txbma

s d q vl ub szo

yijza cvx tyvw bkxsw d o

st pjih rf muaxw uk fiil

pduu fkl npa cp mwqfl exd

nd ma c gdiv krqfe bkvkj