Al,
linux-next July 8th fails to bootup on s390x (linux-next from July 4th did still work)
[...]
smp: Brought up 1 node, 1 CPU
Unable to handle kernel pointer dereference in virtual kernel address space
Failing address: 000000003a070000 TEID: 000000003a070407
Fault in home space mode while using kernel ASCE.
AS:000000003a780007 R3:000000007ffd0007 S:000000007ffd4800 P:000000003a07021d
Oops: 0004 ilc:2 [#1] SMP
Modules linked in:
CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.2.0-rc5-00101-gcb8f0b366109 #14
Hardware name: IBM 2964 NC9 712 (KVM/Linux)
Krnl PSW : 0704e00180000000 000000003974b580 (shmem_parse_monolithic+0x88/0x100)
R:0 T:1 IO:1 EX:1 Key:0 M:1 W:0 P:0 AS:3 CC:2 PM:0 RI:0 EA:3
Krnl GPRS: 0000000000000000 000000000000003d 000000003a07040e 000000000000003d
000000003a07040f 000000000000006d 0000000000000001 0000000000000000
000000007f7c1c00 0000000000000000 000000003a07040a 0000000000000000
000000007f7e4000 000000003a190d78 000000003974b56c 000003e00031fd38
Krnl Code: 000000003974b574: b920002a cgr %r2,%r10
000000003974b578: a784001b brc 8,3974b5ae
#000000003974b57c: 41402001 la %r4,1(%r2)
>000000003974b580: 92002000 mvi 0(%r2),0
000000003974b584: a7090000 lghi %r0,0
000000003974b588: b9040014 lgr %r1,%r4
000000003974b58c: b25e0001 srst %r0,%r1
000000003974b590: a714fffe brc 1,3974b58c
Call Trace:
([<000003e00031fd80>] 0x3e00031fd80)
[<0000000039811662>] vfs_kern_mount.part.0+0x9a/0xc8
[<000000003a302fc0>] devtmpfs_init+0x38/0x140
[<000000003a302e0a>] driver_init+0x22/0x60
[<000000003a2beff8>] kernel_init_freeable+0x298/0x4f0
[<0000000039e7b53a>] kernel_init+0x22/0x148
[<0000000039e87b70>] ret_from_fork+0x30/0x34
[<0000000039e87b74>] kernel_thread_starter+0x0/0xc
INFO: lockdep is turned off.
[...]
bisect points to
f6ab6e953ab3b37b10646aabede8d8731f54cc61 is the first bad commit
commit f6ab6e953ab3b37b10646aabede8d8731f54cc61
Author: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Date: Sat Jun 1 18:43:09 2019 -0400
devtmpfs: don't mix {ramfs,shmem}_fill_super() with mount_single()
Create an internal-only type matching the current devtmpfs, never
register it and have one kernel-internal mount done. That thing
gets mounted only once, so it is free to use mount_nodev().
The "public" devtmpfs (the one we do register, and only after
the internal mount of the real thing is done) simply gets and
returns an extra reference to the internal superblock.
Signed-off-by: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
:040000 040000 4874d86e78685322401d6284d458ae323620573c e3c5780581792df416a4c272b583f76b98321080 M drivers
I can confirm that
f6ab6e953ab3b37b10646aabede8d8731f54cc61
has the issue and
f6ab6e953ab3b37b10646aabede8d8731f54cc61^
has not.
