Since commit 
c33165253492 ("powerpc: use non-set_fs based maccess
routines"), userspace access is not granted anymore when using
copy_from_kernel_nofault()
However, kthread_probe_data() uses copy_from_kernel_nofault()
to check validity of pointers. When the pointer is NULL,
it points to userspace, leading to a KUAP fault and triggering
the following big hammer warning many times when you request
a sysrq "show task":
[ 1117.202054] ------------[ cut here ]------------
[ 1117.202102] Bug: fault blocked by AP register !
[ 1117.202261] WARNING: CPU: 0 PID: 377 at arch/powerpc/include/asm/nohash/32/kup-8xx.h:66 do_page_fault+0x4a8/0x5ec
[ 1117.202310] Modules linked in:
[ 1117.202428] CPU: 0 PID: 377 Comm: sh Tainted: G        W         
5.10.0-rc5-01340-g83f53be2de31-dirty #4175
[ 1117.202499] NIP:  
c0012048 LR: 
c0012048 CTR: 
00000000
[ 1117.202573] REGS: 
cacdbb88 TRAP: 0700   Tainted: G        W          (
5.10.0-rc5-01340-g83f53be2de31-dirty)
[ 1117.202625] MSR:  
00021032 <ME,IR,DR,RI>  CR: 
24082222  XER: 
20000000
[ 1117.202899]
[ 1117.202899] GPR00: 
c0012048 cacdbc40 c2929290 00000023 c092e554 00000001 c09865e8 c092e640
[ 1117.202899] GPR08: 
00001032 00000000 00000000 00014efc 28082224 100d166a 100a0920 00000000
[ 1117.202899] GPR16: 
100cac0c 100b0000 1080c3fc 1080d685 100d0000 100d0000 00000000 100a0900
[ 1117.202899] GPR24: 
100d0000 c07892ec 00000000 c0921510 c21f4440 0000005c c0000000 cacdbc80
[ 1117.204362] NIP [
c0012048] do_page_fault+0x4a8/0x5ec
[ 1117.204461] LR [
c0012048] do_page_fault+0x4a8/0x5ec
[ 1117.204509] Call Trace:
[ 1117.204609] [
cacdbc40] [
c0012048] do_page_fault+0x4a8/0x5ec (unreliable)
[ 1117.204771] [
cacdbc70] [
c00112f0] handle_page_fault+0x8/0x34
[ 1117.204911] --- interrupt: 301 at copy_from_kernel_nofault+0x70/0x1c0
[ 1117.204979] NIP:  
c010dbec LR: 
c010dbac CTR: 
00000001
[ 1117.205053] REGS: 
cacdbc80 TRAP: 0301   Tainted: G        W          (
5.10.0-rc5-01340-g83f53be2de31-dirty)
[ 1117.205104] MSR:  
00009032 <EE,ME,IR,DR,RI>  CR: 
28082224  XER: 
00000000
[ 1117.205416] DAR: 
0000005c DSISR: 
c0000000
[ 1117.205416] GPR00: 
c0045948 cacdbd38 c2929290 00000001 00000017 00000017 00000027 0000000f
[ 1117.205416] GPR08: 
c09926ec 00000000 00000000 3ffff000 24082224
[ 1117.206106] NIP [
c010dbec] copy_from_kernel_nofault+0x70/0x1c0
[ 1117.206202] LR [
c010dbac] copy_from_kernel_nofault+0x30/0x1c0
[ 1117.206258] --- interrupt: 301
[ 1117.206372] [
cacdbd38] [
c004bbb0] kthread_probe_data+0x44/0x70 (unreliable)
[ 1117.206561] [
cacdbd58] [
c0045948] print_worker_info+0xe0/0x194
[ 1117.206717] [
cacdbdb8] [
c00548ac] sched_show_task+0x134/0x168
[ 1117.206851] [
cacdbdd8] [
c005a268] show_state_filter+0x70/0x100
[ 1117.206989] [
cacdbe08] [
c039baa0] sysrq_handle_showstate+0x14/0x24
[ 1117.207122] [
cacdbe18] [
c039bf18] __handle_sysrq+0xac/0x1d0
[ 1117.207257] [
cacdbe48] [
c039c0c0] write_sysrq_trigger+0x4c/0x74
[ 1117.207407] [
cacdbe68] [
c01fba48] proc_reg_write+0xb4/0x114
[ 1117.207550] [
cacdbe88] [
c0179968] vfs_write+0x12c/0x478
[ 1117.207686] [
cacdbf08] [
c0179e60] ksys_write+0x78/0x128
[ 1117.207826] [
cacdbf38] [
c00110d0] ret_from_syscall+0x0/0x34
[ 1117.207938] --- interrupt: c01 at 0xfd4e784
[ 1117.208008] NIP:  
0fd4e784 LR: 
0fe0f244 CTR: 
10048d38
[ 1117.208083] REGS: 
cacdbf48 TRAP: 0c01   Tainted: G        W          (
5.10.0-rc5-01340-g83f53be2de31-dirty)
[ 1117.208134] MSR:  
0000d032 <EE,PR,ME,IR,DR,RI>  CR: 
44002222  XER: 
00000000
[ 1117.208470]
[ 1117.208470] GPR00: 
00000004 7fc34090 77bfb4e0 00000001 1080fa40 00000002 7400000f fefefeff
[ 1117.208470] GPR08: 
7f7f7f7f 10048d38 1080c414 7fc343c0 00000000
[ 1117.209104] NIP [
0fd4e784] 0xfd4e784
[ 1117.209180] LR [
0fe0f244] 0xfe0f244
[ 1117.209236] --- interrupt: c01
[ 1117.209274] Instruction dump:
[ 1117.209353] 
714a4000 418200f0 73ca0001 40820084 73ca0032 408200f8 73c90040 4082ff60
[ 1117.209727] 
0fe00000 3c60c082 386399f4 48013b65 <
0fe00000> 
80010034 3860000b 7c0803a6
[ 1117.210102] ---[ end trace 
1927c0323393af3e ]---
To avoid that, copy_from_kernel_nofault_allowed() is used to check
whether the address is a valid kernel address. But the default
version of it returns true for any address.
Provide a powerpc version of copy_from_kernel_nofault_allowed()
that returns false when the address is below TASK_USER_MAX,
so that copy_from_kernel_nofault() will return -ERANGE.
Fixes: c33165253492 ("powerpc: use non-set_fs based maccess routines")
Reported-by: Qian Cai <qcai@redhat.com>
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/18bcb456d32a3e74f5ae241fd6f1580c092d07f5.1607360230.git.christophe.leroy@csgroup.eu
 
 ccflags-$(CONFIG_PPC64)        := $(NO_MINIMAL_TOC)
 
-obj-y                          := fault.o mem.o pgtable.o mmap.o \
+obj-y                          := fault.o mem.o pgtable.o mmap.o maccess.o \
                                   init_$(BITS).o pgtable_$(BITS).o \
                                   pgtable-frag.o ioremap.o ioremap_$(BITS).o \
                                   init-common.o mmu_context.o drmem.o
 
--- /dev/null
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/uaccess.h>
+#include <linux/kernel.h>
+
+bool copy_from_kernel_nofault_allowed(const void *unsafe_src, size_t size)
+{
+       return is_kernel_addr((unsigned long)unsafe_src);
+}