linux.git/arch/arm/kernel/stacktrace.c, branch v6.6.131

ARM: 9259/1: stacktrace: Convert stacktrace to generic ARCH_STACKWALK

2022-11-14T12:00:57+00:00

Historically architectures have had duplicated code in their stack trace
implementations for filtering what gets traced. In order to avoid this
duplication some generic code has been provided using a new interface
arch_stack_walk(), enabled by selecting ARCH_STACKWALK in Kconfig, which
factors all this out into the generic stack trace code. Convert ARM to
use this common infrastructure.

When initializing the stack frame of the current task, arm64 uses
__builtin_frame_address(1) to initialize the frame pointer, skipping
arch_stack_walk(), see the commit c607ab4f916d ("arm64: stacktrace:
don't trace arch_stack_walk()"). Since __builtin_frame_address(1) does
not work on ARM, unwind_frame() is used to unwind the stack one layer
forward before calling walk_stackframe().

Signed-off-by: Li Huafei 
Reviewed-by: Linus Walleij 
Signed-off-by: Russell King (Oracle)

ARM: 9258/1: stacktrace: Make stack walk callback consistent with generic code

2022-11-08T18:36:18+00:00

As with the generic arch_stack_walk() code the ARM stack walk code takes
a callback that is called per stack frame. Currently the ARM code always
passes a struct stackframe to the callback and the generic code just
passes the pc, however none of the users ever reference anything in the
struct other than the pc value. The ARM code also uses a return type of
int while the generic code uses a return type of bool though in both
cases the return value is a boolean value and the sense is inverted
between the two.

In order to reduce code duplication when ARM is converted to use
arch_stack_walk() change the signature and return sense of the ARM
specific callback to match that of the generic code.

Signed-off-by: Li Huafei 
Reviewed-by: Mark Brown 
Reviewed-by: Linus Waleij 
Signed-off-by: Russell King (Oracle)

ARM: 9234/1: stacktrace: Avoid duplicate saving of exception PC value

2022-10-04T10:09:47+00:00

Because an exception stack frame is not created in the exception entry,
save_trace() does special handling for the exception PC, but this is
only needed when CONFIG_FRAME_POINTER_UNWIND=y. When
CONFIG_ARM_UNWIND=y, unwind annotations have been added to the exception
entry and save_trace() will repeatedly save the exception PC:

    [0x7f000090] hrtimer_hander+0x8/0x10 [hrtimer]
    [0x8019ec50] __hrtimer_run_queues+0x18c/0x394
    [0x8019f760] hrtimer_run_queues+0xbc/0xd0
    [0x8019def0] update_process_times+0x34/0x80
    [0x801ad2a4] tick_periodic+0x48/0xd0
    [0x801ad3dc] tick_handle_periodic+0x1c/0x7c
    [0x8010f2e0] twd_handler+0x30/0x40
    [0x80177620] handle_percpu_devid_irq+0xa0/0x23c
    [0x801718d0] generic_handle_domain_irq+0x24/0x34
    [0x80502d28] gic_handle_irq+0x74/0x88
    [0x8085817c] generic_handle_arch_irq+0x58/0x78
    [0x80100ba8] __irq_svc+0x88/0xc8
    [0x80108114] arch_cpu_idle+0x38/0x3c
    [0x80108114] arch_cpu_idle+0x38/0x3c    <==== duplicate saved exception PC
    [0x80861bf8] default_idle_call+0x38/0x130
    [0x8015d5cc] do_idle+0x150/0x214
    [0x8015d978] cpu_startup_entry+0x18/0x1c
    [0x808589c0] rest_init+0xd8/0xdc
    [0x80c00a44] arch_post_acpi_subsys_init+0x0/0x8

We can move the special handling of the exception PC in save_trace() to
the unwind_frame() of the frame pointer unwinder.

Signed-off-by: Li Huafei 
Reviewed-by: Linus Waleij 
Signed-off-by: Russell King (Oracle)

ARM: 9233/1: stacktrace: Skip frame pointer boundary check for call_with_stack()

2022-10-04T10:09:47+00:00

When using the frame pointer unwinder, it was found that the stack trace
output of stack_trace_save() is incomplete if the stack contains
call_with_stack():

 [0x7f00002c] dump_stack_task+0x2c/0x90 [hrtimer]
 [0x7f0000a0] hrtimer_hander+0x10/0x18 [hrtimer]
 [0x801a67f0] __hrtimer_run_queues+0x1b0/0x3b4
 [0x801a7350] hrtimer_run_queues+0xc4/0xd8
 [0x801a597c] update_process_times+0x3c/0x88
 [0x801b5a98] tick_periodic+0x50/0xd8
 [0x801b5bf4] tick_handle_periodic+0x24/0x84
 [0x8010ffc4] twd_handler+0x38/0x48
 [0x8017d220] handle_percpu_devid_irq+0xa8/0x244
 [0x80176e9c] generic_handle_domain_irq+0x2c/0x3c
 [0x8052e3a8] gic_handle_irq+0x7c/0x90
 [0x808ab15c] generic_handle_arch_irq+0x60/0x80
 [0x8051191c] call_with_stack+0x1c/0x20

For the frame pointer unwinder, unwind_frame() checks stackframe::fp by
stackframe::sp. Since call_with_stack() switches the SP from one stack
to another, stackframe::fp and stackframe: :sp will point to different
stacks, so we can no longer check stackframe::fp by stackframe::sp. Skip
checking stackframe::fp at this point to avoid this problem.

Signed-off-by: Li Huafei 
Reviewed-by: Linus Waleij 
Signed-off-by: Russell King (Oracle)

Merge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm

2022-04-03T17:17:48+00:00

Pull ARM fixes from Russell King:

 - avoid unnecessary rebuilds for library objects

 - fix return value of __setup handlers

 - fix invalid input check for "crashkernel=" kernel option

 - silence KASAN warnings in unwind_frame

* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: 9191/1: arm/stacktrace, kasan: Silence KASAN warnings in unwind_frame()
  ARM: 9190/1: kdump: add invalid input check for 'crashkernel=0'
  ARM: 9187/1: JIVE: fix return value of __setup handler
  ARM: 9189/1: decompressor: fix unneeded rebuilds of library objects

ARM: 9191/1: arm/stacktrace, kasan: Silence KASAN warnings in unwind_frame()

2022-04-01T11:58:39+00:00

The following KASAN warning is detected by QEMU.

==================================================================
BUG: KASAN: stack-out-of-bounds in unwind_frame+0x508/0x870
Read of size 4 at addr c36bba90 by task cat/163

CPU: 1 PID: 163 Comm: cat Not tainted 5.10.0-rc1 #40
Hardware name: ARM-Versatile Express
[] (unwind_backtrace) from [] (show_stack+0x10/0x14)
[] (show_stack) from [] (dump_stack+0x98/0xb0)
[] (dump_stack) from [] (print_address_description.constprop.0+0x58/0x4bc)
[] (print_address_description.constprop.0) from [] (kasan_report+0x154/0x170)
[] (kasan_report) from [] (unwind_frame+0x508/0x870)
[] (unwind_frame) from [] (__save_stack_trace+0x110/0x134)
[] (__save_stack_trace) from [] (stack_trace_save+0x8c/0xb4)
[] (stack_trace_save) from [] (kasan_set_track+0x38/0x60)
[] (kasan_set_track) from [] (kasan_set_free_info+0x20/0x2c)
[] (kasan_set_free_info) from [] (__kasan_slab_free+0xec/0x120)
[] (__kasan_slab_free) from [] (kmem_cache_free+0x7c/0x334)
[] (kmem_cache_free) from [] (rcu_core+0x390/0xccc)
[] (rcu_core) from [] (__do_softirq+0x180/0x518)
[] (__do_softirq) from [] (irq_exit+0x9c/0xe0)
[] (irq_exit) from [] (__handle_domain_irq+0xb0/0x110)
[] (__handle_domain_irq) from [] (gic_handle_irq+0xa0/0xb8)
[] (gic_handle_irq) from [] (__irq_svc+0x6c/0x94)
Exception stack(0xc36bb928 to 0xc36bb970)
b920:                   c36bb9c0 00000000 c0126919 c0101228 c36bb9c0 b76d7730
b940: c36b8000 c36bb9a0 c3335b00 c01ce0d8 00000003 c36bba3c c36bb940 c36bb978
b960: c010e298 c011373c 60000013 ffffffff
[] (__irq_svc) from [] (unwind_frame+0x0/0x870)
[] (unwind_frame) from [<00000000>] (0x0)

The buggy address belongs to the page:
page:(ptrval) refcount:0 mapcount:0 mapping:00000000 index:0x0 pfn:0x636bb
flags: 0x0()
raw: 00000000 00000000 ef867764 00000000 00000000 00000000 ffffffff 00000000
page dumped because: kasan: bad access detected

addr c36bba90 is located in stack of task cat/163 at offset 48 in frame:
 stack_trace_save+0x0/0xb4

this frame has 1 object:
 [32, 48) 'trace'

Memory state around the buggy address:
 c36bb980: f1 f1 f1 f1 00 04 f2 f2 00 00 f3 f3 00 00 00 00
 c36bba00: 00 00 00 00 00 00 00 00 00 00 00 00 f1 f1 f1 f1
>c36bba80: 00 00 f3 f3 00 00 00 00 00 00 00 00 00 00 00 00
                 ^
 c36bbb00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
 c36bbb80: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
==================================================================

There is a same issue on x86 and has been resolved by the commit f7d27c35ddff
("x86/mm, kasan: Silence KASAN warnings in get_wchan()").
The solution could be applied to arm architecture too.

Signed-off-by: Lin Yujun 
Reported-by: He Ying 
Signed-off-by: Russell King (Oracle)

ARM: unwind: set frame.pc correctly for current-thread unwinding

2022-03-11T10:55:28+00:00

When e.g. a WARN_ON() is encountered, we attempt to unwind the current
thread. To do this, we set frame.pc to unwind_backtrace, which means it
points at the beginning of the function. However, the rest of the state
is initialised from within the function, which means the function
prologue has already been run.

This can be confusing, and with a recent patch from Ard, can result in
the unwinder misbehaving if we want to be strict about the PC value.

If we correctly initialise the state so it is self-consistent (in other
words, set frame.pc to the location we are initialising it) then we
eliminate this confusion, and avoid possible future issues.

Reviewed-by: Ard Biesheuvel 
Signed-off-by: Russell King (Oracle)

ARM: Recover kretprobe modified return address in stacktrace

2021-10-22T16:16:53+00:00

Since the kretprobe replaces the function return address with
the kretprobe_trampoline on the stack, arm unwinder shows it
instead of the correct return address.

This finds the correct return address from the per-task
kretprobe_instances list and verify it is in between the
caller fp and callee fp.

Note that this supports both GCC and clang if CONFIG_FRAME_POINTER=y
and CONFIG_ARM_UNWIND=n. For the ARM unwinder, this is still
not working correctly.

Signed-off-by: Masami Hiramatsu 
Signed-off-by: Steven Rostedt (VMware)

ARM: clang: Do not rely on lr register for stacktrace

2021-10-22T16:16:53+00:00

Currently the stacktrace on clang compiled arm kernel uses the 'lr'
register to find the first frame address from pt_regs. However, that
is wrong after calling another function, because the 'lr' register
is used by 'bl' instruction and never be recovered.

As same as gcc arm kernel, directly use the frame pointer (r11) of
the pt_regs to find the first frame address.

Note that this fixes kretprobe stacktrace issue only with
CONFIG_UNWINDER_FRAME_POINTER=y. For the CONFIG_UNWINDER_ARM,
we need another fix.

Signed-off-by: Masami Hiramatsu 
Reviewed-by: Nick Desaulniers 
Signed-off-by: Steven Rostedt (VMware)

ARM: 8992/1: Fix unwind_frame for clang-built kernels

2020-07-21T15:33:40+00:00

Since clang does not push pc and sp in function prologues, the current
implementation of unwind_frame does not work. By using the previous
frame's lr/fp instead of saved pc/sp we get valid unwinds on clang-built
kernels.

The bounds check on next frame pointer must be changed as well since
there are 8 less bytes between frames.

This fixes /proc//stack.

Link: https://github.com/ClangBuiltLinux/linux/issues/912

Reported-by: Miles Chen 
Tested-by: Miles Chen 
Cc: stable@vger.kernel.org
Reviewed-by: Nick Desaulniers 
Signed-off-by: Nathan Huckleberry 
Signed-off-by: Russell King