Christophe Leroy <christophe.leroy@xxxxxxxxxx> writes:
> Le 07/12/2020 à 01:24, Michael Ellerman a écrit :
>> Christophe Leroy <christophe.leroy@xxxxxxxxxx> writes:
>>> 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.
>>>
>>> Reported-by: Qian Cai <qcai@xxxxxxxxxx>
>>> Fixes: c33165253492 ("powerpc: use non-set_fs based maccess routines")
>>> Cc: Christoph Hellwig <hch@xxxxxx>
>>> Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
>>> Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxxxxxx>
>>> ---
>>> This issue was introduced in 5.10. I didn't mark it for stable, hopping it will go into 5.10-rc7
>>> ---
>>> arch/powerpc/mm/Makefile | 2 +-
>>> arch/powerpc/mm/maccess.c | 9 +++++++++
>>> 2 files changed, 10 insertions(+), 1 deletion(-)
>>> create mode 100644 arch/powerpc/mm/maccess.c
>>>
>>> diff --git a/arch/powerpc/mm/maccess.c b/arch/powerpc/mm/maccess.c
>>> new file mode 100644
>>> index 000000000000..56e97c0fb233
>>> --- /dev/null
>>> +++ b/arch/powerpc/mm/maccess.c
>>> @@ -0,0 +1,9 @@
>>> +// 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 (unsigned long)unsafe_src >= TASK_SIZE_MAX;
>>> +}
>>
>> Is there a reason we're using TASK_SIZE_MAX?
>
> No special reason, that's just copied from x86.
>
>> It's copy from *kernel* (nofault) allowed, so shouldn't we be checking
>> that the address plausibly points at kernel memory? Not at no-man's land
>> above TASK_SIZE_MAX but below the start of kernel memory?
>
> Yes, on PPC64 that's right. On PPC32 the kernel memory starts where the userland stops.
Yep sorry I was talking about 64-bit there.
>> We have is_kernel_addr() which already encapsulates some platform quirks
>> around that logic, it seems like it would be a better fit?
>
> Yes probably, I send v2. For PPC32 that's a comparison with TASK_SIZE thought.
Yeah, so it's the same test for PPC32 but I think is_kernel_addr() is
better on 64-bit.
I'll grab your v2.
cheers
