Qoriq P5020 PowerPC board doesn't boot with the latest git version anymore

[Christian Zigotzky]

Hi All,

The bug is definitely in the commit 
a8a217c22116eff6c120d753c9934089fb229af0. [1]

The kernel 4.14.0 boots without any problems on my Varisys Cyrus Plus 
board [2][3] and on my Varisys Nemo board [4]. I have patched the kernel 
4.14.0 source code with the bad commit I found with bisect last week.

After compiling, the kernel 4.14.0 doesn't boot anymore on my Varisys 
Cyrus Plus board [2][3].

Please help to find the bug in the bad commit. (Bisect log attached)

Thanks,
Christian


[1] 
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=a8a217c22116eff6c120d753c9934089fb229af0
[2] http://www.amigaos.net/hardware/133/amigaone-x5000
[3] http://wiki.amiga.org/index.php?title=AmigaONE_X5000
[4] https://en.wikipedia.org/wiki/AmigaOne_X1000
-------------- next part --------------
git log

git bisect start

git bisect good bebc6082da0a9f5d47a1ea2edc099bf671058bd4 (Linux 4.14.0 final)

git bisect bad 894025f24bd028942da3e602b87d9f7223109b14 (Linux 4.15 alpha1 Mon Nov 13 21:14:07 2017 -0800)

Output:

Bisecting: 1128 revisions left to test after this (roughly 10 steps)
[7d58e1c9059eefe0066c5acf2ffa582f6f0180e3] Merge branch 'smp-hotplug-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

git bisect bad

Output:

Bisecting: 559 revisions left to test after this (roughly 9 steps)
[8e9a2dba8686187d8c8179e5b86640e653963889] Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

git bisect bad

Output:

Bisecting: 272 revisions left to test after this (roughly 8 steps)
[dee02770cdcd8bc06a48c917ce5df2fb56cf6059] Merge tag 'mmc-v4.15' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc

git bisect good

Output:

Bisecting: 129 revisions left to test after this (roughly 7 steps)
[d60a540ac5f2fbab3e6fe592717b445bd7343a91] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

git bisect good

Output:

Bisecting: 64 revisions left to test after this (roughly 6 steps)
[7a10e2a9190628a4024ea394ce7bd641ae40ffd1] x86: Use lockdep to assert IRQs are disabled/enabled

git bisect bad

Output:

Bisecting: 32 revisions left to test after this (roughly 5 steps)
[eeafcc5a5925d0819dad462eac1d42fda5fbe36f] locking/atomics, firmware/ivc: Convert ACCESS_ONCE() to READ_ONCE()/WRITE_ONCE()

git bisect bad

Output:

Bisecting: 15 revisions left to test after this (roughly 4 steps)
[a4c1887d4c1462b0ec5a8989f8ba3cdd9057a299] locking/arch: Remove dummy arch_{read,spin,write}_lock_flags() implementations

git bisect bad

Output:

Bisecting: 7 revisions left to test after this (roughly 3 steps)
[c0905115a55c8a3011d7b7e6ee5d63653c1e99f1] locking/arch, ia64: Add __down_read_killable()

git bisect good

Output:

Bisecting: 3 revisions left to test after this (roughly 2 steps)
[0dc208b5d5feedc795cbf124539decd182c8e99e] locking/rwsem, fs: Use killable down_read() in iterate_dir()

git bisect good

Output:

Bisecting: 1 revision left to test after this (roughly 1 step)
[a8a217c22116eff6c120d753c9934089fb229af0] locking/core: Remove {read,spin,write}_can_lock()

git bisect bad

Output:

Bisecting: 0 revisions left to test after this (roughly 0 steps)
[26c4eb192c6224e5297496cead36404b62fb071b] locking/rwsem, security/apparmor: Replace homebrew use of write_can_lock() with lockdep

git bisect good

Output:

a8a217c22116eff6c120d753c9934089fb229af0 is the first bad commit
commit a8a217c22116eff6c120d753c9934089fb229af0
Author: Will Deacon <will.deacon at arm.com>
Date:   Tue Oct 3 19:25:27 2017 +0100

    locking/core: Remove {read,spin,write}_can_lock()
    
    Outside of the locking code itself, {read,spin,write}_can_lock() have no
    users in tree. Apparmor (the last remaining user of write_can_lock()) got
    moved over to lockdep by the previous patch.
    
    This patch removes the use of {read,spin,write}_can_lock() from the
    BUILD_LOCK_OPS macro, deferring to the trylock operation for testing the
    lock status, and subsequently removes the unused macros altogether. They
    aren't guaranteed to work in a concurrent environment and can give
    incorrect results in the case of qrwlock.
    
    Signed-off-by: Will Deacon <will.deacon at arm.com>
    Signed-off-by: Peter Zijlstra (Intel) <peterz at infradead.org>
    Cc: Linus Torvalds <torvalds at linux-foundation.org>
    Cc: Peter Zijlstra <peterz at infradead.org>
    Cc: Thomas Gleixner <tglx at linutronix.de>
    Cc: paulmck at linux.vnet.ibm.com
    Link: http://lkml.kernel.org/r/1507055129-12300-2-git-send-email-will.deacon@arm.com
    Signed-off-by: Ingo Molnar <mingo at kernel.org>

:040000 040000 5b435422698e4cb10a69c8c5d63734785b4e1f3c 35093937ba487d8b5ecf773c38439f99801f4041 M	arch
:040000 040000 c3fa2d1fc41218e43a76eb6d97808de5be3a0355 6d4682dd193d79f868b28eca9714716ce3dbf7c9 M	include
:040000 040000 2ac2fb1e58edcf3f6f912ecf66431f8b0e6914b9 d7b7814b7b349e2f0fd7578b35726037294ea9b4 M	kernel