Re: [PATCH v2] powerpc/kprobes: Fix trap address when trap happened in real mode

From: Christophe Leroy
Date: Tue Feb 18 2020 - 14:37:28 EST

Le 18/02/2020 Ã 15:55, Naveen N. Rao a ÃcritÂ:
Christophe Leroy wrote:
When a program check exception happens while MMU translation is
disabled, following Oops happens in kprobe_handler() in the following


[ÂÂ 33.098554] BUG: Unable to handle kernel data access on read at 0x0000e268
[ÂÂ 33.105091] Faulting instruction address: 0xc000ec34
[ÂÂ 33.110010] Oops: Kernel access of bad area, sig: 11 [#1]
[ÂÂ 33.119540] Modules linked in:
[ÂÂ 33.122591] CPU: 0 PID: 429 Comm: cat Not tainted 5.6.0-rc1-s3k-dev-00824-g84195dc6c58a #3267
[ÂÂ 33.131005] NIP:Â c000ec34 LR: c000ecd8 CTR: c019cab8
[ÂÂ 33.136002] REGS: ca4d3b58 TRAP: 0300ÂÂ Not tainted (5.6.0-rc1-s3k-dev-00824-g84195dc6c58a)
[ÂÂ 33.144324] MSR:Â 00001032 <ME,IR,DR,RI>Â CR: 2a4d3c52Â XER: 00000000
[ÂÂ 33.150699] DAR: 0000e268 DSISR: c0000000
[ÂÂ 33.150699] GPR00: c000b09c ca4d3c10 c66d0620 00000000 ca4d3c60 00000000 00009032 00000000
[ÂÂ 33.150699] GPR08: 00020000 00000000 c087de44 c000afe0 c66d0ad0 100d3dd6 fffffff3 00000000
[ÂÂ 33.150699] GPR16: 00000000 00000041 00000000 ca4d3d70 00000000 00000000 0000416d 00000000
[ÂÂ 33.150699] GPR24: 00000004 c53b6128 00000000 0000e268 00000000 c07c0000 c07bb6fc ca4d3c60
[ÂÂ 33.188015] NIP [c000ec34] kprobe_handler+0x128/0x290
[ÂÂ 33.192989] LR [c000ecd8] kprobe_handler+0x1cc/0x290
[ÂÂ 33.197854] Call Trace:
[ÂÂ 33.200340] [ca4d3c30] [c000b09c] program_check_exception+0xbc/0x6fc
[ÂÂ 33.206590] [ca4d3c50] [c000e43c] ret_from_except_full+0x0/0x4
[ÂÂ 33.212392] --- interrupt: 700 at 0xe268
[ÂÂ 33.270401] Instruction dump:
[ÂÂ 33.273335] 913e0008 81220000 38600001 3929ffff 91220000 80010024 bb410008 7c0803a6
[ÂÂ 33.280992] 38210020 4e800020 38600000 4e800020 <813b0000> 6d2a7fe0 2f8a0008 419e0154
[ÂÂ 33.288841] ---[ end trace 5b9152d4cdadd06d ]---

kprobe is not prepared to handle events in real mode and functions
running in real mode should have been blacklisted, so kprobe_handler()
can safely bail out telling 'this trap is not mine' for any trap that
happened while in real-mode.

If the trap happened with MSR_IR cleared, return 0 immediately.

Reported-by: Larry Finger <Larry.Finger@xxxxxxxxxxxx>
Fixes: 6cc89bad60a6 ("powerpc/kprobes: Invoke handlers directly")
Cc: stable@xxxxxxxxxxxxxxx
Cc: Naveen N. Rao <naveen.n.rao@xxxxxxxxxxxxxxxxxx>
Cc: Masami Hiramatsu <mhiramat@xxxxxxxxxx>
Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxx>

v2: bailing out instead of converting real-time address to virtual and continuing.

The bug might have existed even before that commit from Naveen.

Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxx>
Âarch/powerpc/kernel/kprobes.c | 3 +++
Â1 file changed, 3 insertions(+)

diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c
index 2d27ec4feee4..673f349662e8 100644
--- a/arch/powerpc/kernel/kprobes.c
+++ b/arch/powerpc/kernel/kprobes.c
@@ -264,6 +264,9 @@ int kprobe_handler(struct pt_regs *regs)
ÂÂÂÂ if (user_mode(regs))
ÂÂÂÂÂÂÂÂ return 0;

+ÂÂÂ if (!(regs->msr & MSR_IR))
+ÂÂÂÂÂÂÂ return 0;

Should we also check for MSR_DR? Are there scenarios with ppc32 where MSR_IR is on, but MSR_DR is off?

Yes indeed.