| Age | Commit message (Collapse) | Author |
|---|
|
Currently, userstacktrace is unsupported for ftrace and uprobe
tracers on arm64. This patch uses the perf_callchain_user() code
as blueprint to implement the arch_stack_walk_user() which add
userstacktrace support on arm64.
Meanwhile, we can use arch_stack_walk_user() to simplify the
implementation of perf_callchain_user().
This patch is tested pass with ftrace, uprobe and perf tracers
profiling userstacktrace cases.
Tested-by: chenqiwu <qiwu.chen@transsion.com>
Signed-off-by: chenqiwu <qiwu.chen@transsion.com>
Link: https://lore.kernel.org/r/20231219022229.10230-1-qiwu.chen@transsion.com
Signed-off-by: Will Deacon <will@kernel.org>
|
|
Currently we strip the PAC from pointers using C code, which requires
generating bitmasks, and conditionally clearing/setting bits depending
on bit 55. We can do better by using XPACLRI directly.
When the logic was originally written to strip PACs from user pointers,
contemporary toolchains used for the kernel had assemblers which were
unaware of the PAC instructions. As stripping the PAC from userspace
pointers required unconditional clearing of a fixed set of bits (which
could be performed with a single instruction), it was simpler to
implement the masking in C than it was to make use of XPACI or XPACLRI.
When support for in-kernel pointer authentication was added, the
stripping logic was extended to cover TTBR1 pointers, requiring several
instructions to handle whether to clear/set bits dependent on bit 55 of
the pointer.
This patch simplifies the stripping of PACs by using XPACLRI directly,
as contemporary toolchains do within __builtin_return_address(). This
saves a number of instructions, especially where
__builtin_return_address() does not implicitly strip the PAC but is
heavily used (e.g. with tracepoints). As the kernel might be compiled
with an assembler without knowledge of XPACLRI, it is assembled using
the 'HINT #7' alias, which results in an identical opcode.
At the same time, I've split ptrauth_strip_insn_pac() into
ptrauth_strip_user_insn_pac() and ptrauth_strip_kernel_insn_pac()
helpers so that we can avoid unnecessary PAC stripping when pointer
authentication is not in use in userspace or kernel respectively.
The underlying xpaclri() macro uses inline assembly which clobbers x30.
The clobber causes the compiler to save/restore the original x30 value
in a frame record (protected with PACIASP and AUTIASP when in-kernel
authentication is enabled), so this does not provide a gadget to alter
the return address. Similarly this does not adversely affect unwinding
due to the presence of the frame record.
The ptrauth_user_pac_mask() and ptrauth_kernel_pac_mask() are exported
from the kernel in ptrace and core dumps, so these are retained. A
subsequent patch will move them out of <asm/compiler.h>.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Amit Daniel Kachhap <amit.kachhap@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Kristina Martsenko <kristina.martsenko@arm.com>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20230412160134.306148-3-mark.rutland@arm.com
Signed-off-by: Will Deacon <will@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf updates from Borislav Petkov:
"Cleanup of the perf/kvm interaction."
* tag 'perf_core_for_v5.17_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf: Drop guest callback (un)register stubs
KVM: arm64: Drop perf.c and fold its tiny bits of code into arm.c
KVM: arm64: Hide kvm_arm_pmu_available behind CONFIG_HW_PERF_EVENTS=y
KVM: arm64: Convert to the generic perf callbacks
KVM: x86: Move Intel Processor Trace interrupt handler to vmx.c
KVM: Move x86's perf guest info callbacks to generic KVM
KVM: x86: More precisely identify NMI from guest when handling PMI
KVM: x86: Drop current_vcpu for kvm_running_vcpu + kvm_arch_vcpu variable
perf/core: Use static_call to optimize perf_guest_info_callbacks
perf: Force architectures to opt-in to guest callbacks
perf: Add wrappers for invoking guest callbacks
perf/core: Rework guest callbacks to prepare for static_call support
perf: Drop dead and useless guest "support" from arm, csky, nds32 and riscv
perf: Stop pretending that perf can handle multiple guest callbacks
KVM: x86: Register Processor Trace interrupt hook iff PT enabled in guest
KVM: x86: Register perf callbacks after calling vendor's hardware_setup()
perf: Protect perf_guest_cbs with RCU
|
|
To enable RELIABLE_STACKTRACE and LIVEPATCH on arm64, we need to
substantially rework arm64's unwinding code. As part of this, we want to
minimize the set of unwind interfaces we expose, and avoid open-coding
of unwind logic outside of stacktrace.c.
Currently perf_callchain_kernel() walks the stack of an interrupted
context by calling start_backtrace() with the context's PC and FP, and
iterating unwind steps using walk_stackframe(). This is functionally
equivalent to calling arch_stack_walk() with the interrupted context's
pt_regs, which will start with the PC and FP from the regs.
Make perf_callchain_kernel() use arch_stack_walk(). This simplifies
perf_callchain_kernel(), and in future will alow us to make
walk_stackframe() private to stacktrace.c.
At the same time, we update the callchain_trace() callback to check the
return value of perf_callchain_store(), which indicates whether there is
space for any further entries. When a non-zero value is returned,
further calls will be ignored, and are redundant, so we can stop the
unwind at this point.
We also remove the stale and confusing comment for callchain_trace.
There should be no functional change as a result of this patch.
Signed-off-by: Madhavan T. Venkataraman <madvenka@linux.microsoft.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
[Mark: elaborate commit message, remove comment, fix includes]
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/20211129142849.3056714-5-mark.rutland@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
|
Add helpers for the guest callbacks to prepare for burying the callbacks
behind a Kconfig (it's a lot easier to provide a few stubs than to #ifdef
piles of code), and also to prepare for converting the callbacks to
static_call(). perf_instruction_pointer() in particular will have subtle
semantics with static_call(), as the "no callbacks" case will return 0 if
the callbacks are unregistered between querying guest state and getting
the IP. Implement the change now to avoid a functional change when adding
static_call() support, and because the new helper needs to return
_something_ in this case.
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Link: https://lore.kernel.org/r/20211111020738.2512932-8-seanjc@google.com
|
|
To prepare for using static_calls to optimize perf's guest callbacks,
replace ->is_in_guest and ->is_user_mode with a new multiplexed hook
->state, tweak ->handle_intel_pt_intr to play nice with being called when
there is no active guest, and drop "guest" from ->get_guest_ip.
Return '0' from ->state and ->handle_intel_pt_intr to indicate "not in
guest" so that DEFINE_STATIC_CALL_RET0 can be used to define the static
calls, i.e. no callback == !guest.
[sean: extracted from static_call patch, fixed get_ip() bug, wrote changelog]
Suggested-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Originally-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Like Xu <like.xu@linux.intel.com>
Signed-off-by: Zhu Lingshan <lingshan.zhu@intel.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Link: https://lore.kernel.org/r/20211111020738.2512932-7-seanjc@google.com
|
|
Protect perf_guest_cbs with RCU to fix multiple possible errors. Luckily,
all paths that read perf_guest_cbs already require RCU protection, e.g. to
protect the callback chains, so only the direct perf_guest_cbs touchpoints
need to be modified.
Bug #1 is a simple lack of WRITE_ONCE/READ_ONCE behavior to ensure
perf_guest_cbs isn't reloaded between a !NULL check and a dereference.
Fixed via the READ_ONCE() in rcu_dereference().
Bug #2 is that on weakly-ordered architectures, updates to the callbacks
themselves are not guaranteed to be visible before the pointer is made
visible to readers. Fixed by the smp_store_release() in
rcu_assign_pointer() when the new pointer is non-NULL.
Bug #3 is that, because the callbacks are global, it's possible for
readers to run in parallel with an unregisters, and thus a module
implementing the callbacks can be unloaded while readers are in flight,
resulting in a use-after-free. Fixed by a synchronize_rcu() call when
unregistering callbacks.
Bug #1 escaped notice because it's extremely unlikely a compiler will
reload perf_guest_cbs in this sequence. perf_guest_cbs does get reloaded
for future derefs, e.g. for ->is_user_mode(), but the ->is_in_guest()
guard all but guarantees the consumer will win the race, e.g. to nullify
perf_guest_cbs, KVM has to completely exit the guest and teardown down
all VMs before KVM start its module unload / unregister sequence. This
also makes it all but impossible to encounter bug #3.
Bug #2 has not been a problem because all architectures that register
callbacks are strongly ordered and/or have a static set of callbacks.
But with help, unloading kvm_intel can trigger bug #1 e.g. wrapping
perf_guest_cbs with READ_ONCE in perf_misc_flags() while spamming
kvm_intel module load/unload leads to:
BUG: kernel NULL pointer dereference, address: 0000000000000000
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
PGD 0 P4D 0
Oops: 0000 [#1] PREEMPT SMP
CPU: 6 PID: 1825 Comm: stress Not tainted 5.14.0-rc2+ #459
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015
RIP: 0010:perf_misc_flags+0x1c/0x70
Call Trace:
perf_prepare_sample+0x53/0x6b0
perf_event_output_forward+0x67/0x160
__perf_event_overflow+0x52/0xf0
handle_pmi_common+0x207/0x300
intel_pmu_handle_irq+0xcf/0x410
perf_event_nmi_handler+0x28/0x50
nmi_handle+0xc7/0x260
default_do_nmi+0x6b/0x170
exc_nmi+0x103/0x130
asm_exc_nmi+0x76/0xbf
Fixes: 39447b386c84 ("perf: Enhance perf to allow for guest statistic collection from host")
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/20211111020738.2512932-2-seanjc@google.com
|
|
The AAPCS places no requirements on the alignment of the frame
record. In theory it could be placed anywhere, although it seems
sensible to require it to be aligned to 8 bytes. With an upcoming
enhancement to tag-based KASAN Clang will begin creating frame records
located at an address that is only aligned to 8 bytes. Accommodate
such frame records in the stack unwinding code.
As pointed out by Mark Rutland, the userspace stack unwinding code
has the same problem, so fix it there as well.
Signed-off-by: Peter Collingbourne <pcc@google.com>
Link: https://linux-review.googlesource.com/id/Ia22c375230e67ca055e9e4bb639383567f7ad268
Acked-by: Andrey Konovalov <andreyknvl@gmail.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Link: https://lore.kernel.org/r/20210526174927.2477847-2-pcc@google.com
Signed-off-by: Will Deacon <will@kernel.org>
|
|
As with the generic arch_stack_walk() code the arm64 stack walk code takes
a callback that is called per stack frame. Currently the arm64 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 arm64 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 arm64 is converted to use
arch_stack_walk() change the signature and return sense of the arm64
specific callback to match that of the generic code.
Signed-off-by: Mark Brown <broonie@kernel.org>
Reviewed-by: Miroslav Benes <mbenes@suse.cz>
Link: https://lore.kernel.org/r/20200914153409.25097-3-broonie@kernel.org
Signed-off-by: Will Deacon <will@kernel.org>
|
|
Some common code is required by each stacktrace user to initialise
struct stackframe before the first call to unwind_frame().
In preparation for adding to the common code, this patch factors it
out into a separate function start_backtrace(), and modifies the
stacktrace callers appropriately.
No functional change.
Signed-off-by: Dave Martin <dave.martin@arm.com>
[Mark: drop tsk argument, update more callsites]
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: James Morse <james.morse@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will@kernel.org>
|
|
Based on 1 normalized pattern(s):
this program is free software you can redistribute it and or modify
it under the terms of the gnu general public license version 2 as
published by the free software foundation this program is
distributed in the hope that it will be useful but without any
warranty without even the implied warranty of merchantability or
fitness for a particular purpose see the gnu general public license
for more details you should have received a copy of the gnu general
public license along with this program if not see http www gnu org
licenses
extracted by the scancode license scanner the SPDX license identifier
GPL-2.0-only
has been chosen to replace the boilerplate/reference in 503 file(s).
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Reviewed-by: Allison Randal <allison@lohutok.net>
Reviewed-by: Enrico Weigelt <info@metux.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190602204653.811534538@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
Nobody has actually used the type (VERIFY_READ vs VERIFY_WRITE) argument
of the user address range verification function since we got rid of the
old racy i386-only code to walk page tables by hand.
It existed because the original 80386 would not honor the write protect
bit when in kernel mode, so you had to do COW by hand before doing any
user access. But we haven't supported that in a long time, and these
days the 'type' argument is a purely historical artifact.
A discussion about extending 'user_access_begin()' to do the range
checking resulted this patch, because there is no way we're going to
move the old VERIFY_xyz interface to that model. And it's best done at
the end of the merge window when I've done most of my merges, so let's
just get this done once and for all.
This patch was mostly done with a sed-script, with manual fix-ups for
the cases that weren't of the trivial 'access_ok(VERIFY_xyz' form.
There were a couple of notable cases:
- csky still had the old "verify_area()" name as an alias.
- the iter_iov code had magical hardcoded knowledge of the actual
values of VERIFY_{READ,WRITE} (not that they mattered, since nothing
really used it)
- microblaze used the type argument for a debug printout
but other than those oddities this should be a total no-op patch.
I tried to fix up all architectures, did fairly extensive grepping for
access_ok() uses, and the changes are trivial, but I may have missed
something. Any missed conversion should be trivially fixable, though.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing updates from Steven Rostedt:
- Rework of the kprobe/uprobe and synthetic events to consolidate all
the dynamic event code. This will make changes in the future easier.
- Partial rewrite of the function graph tracing infrastructure. This
will allow for multiple users of hooking onto functions to get the
callback (return) of the function. This is the ground work for having
kprobes and function graph tracer using one code base.
- Clean up of the histogram code that will facilitate adding more
features to the histograms in the future.
- Addition of str_has_prefix() and a few use cases. There currently is
a similar function strstart() that is used in a few places, but only
returns a bool and not a length. These instances will be removed in
the future to use str_has_prefix() instead.
- A few other various clean ups as well.
* tag 'trace-v4.21' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (57 commits)
tracing: Use the return of str_has_prefix() to remove open coded numbers
tracing: Have the historgram use the result of str_has_prefix() for len of prefix
tracing: Use str_has_prefix() instead of using fixed sizes
tracing: Use str_has_prefix() helper for histogram code
string.h: Add str_has_prefix() helper function
tracing: Make function ‘ftrace_exports’ static
tracing: Simplify printf'ing in seq_print_sym
tracing: Avoid -Wformat-nonliteral warning
tracing: Merge seq_print_sym_short() and seq_print_sym_offset()
tracing: Add hist trigger comments for variable-related fields
tracing: Remove hist trigger synth_var_refs
tracing: Use hist trigger's var_ref array to destroy var_refs
tracing: Remove open-coding of hist trigger var_ref management
tracing: Use var_refs[] for hist trigger reference checking
tracing: Change strlen to sizeof for hist trigger static strings
tracing: Remove unnecessary hist trigger struct field
tracing: Fix ftrace_graph_get_ret_stack() to use task and not current
seq_buf: Use size_t for len in seq_buf_puts()
seq_buf: Make seq_buf_puts() null-terminate the buffer
arm64: Use ftrace_graph_get_ret_stack() instead of curr_ret_stack
...
|
|
The structure of the ret_stack array on the task struct is going to
change, and accessing it directly via the curr_ret_stack index will no
longer give the ret_stack entry that holds the return address. To access
that, architectures must now use ftrace_graph_get_ret_stack() to get the
associated ret_stack that matches the saved return address.
Cc: linux-arm-kernel@lists.infradead.org
Cc: Will Deacon <will.deacon@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
|
|
When the kernel is unwinding userspace callchains, we can't expect that
the userspace consumer of these callchains has the data necessary to
strip the PAC from the stored LR.
This patch has the kernel strip the PAC from user stackframes when the
in-kernel unwinder is used. This only affects the LR value, and not the
FP.
This only affects the in-kernel unwinder. When userspace performs
unwinding, it is up to userspace to strip PACs as necessary (which can
be determined from DWARF information).
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Kristina Martsenko <kristina.martsenko@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
|
The unwind code sets the sp member of struct stackframe to
'frame pointer + 0x10' unconditionally, without regard for whether
doing so produces a legal value. So let's simply remove it now that
we have stopped using it anyway.
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: James Morse <james.morse@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
|
|
We will use it to count how many addresses are in the entry->ip[] array,
excluding PERF_CONTEXT_{KERNEL,USER,etc} entries, so that we can really
return the number of entries specified by the user via the relevant
sysctl, kernel.perf_event_max_contexts, or via the per event
perf_event_attr.sample_max_stack knob.
This way we keep the perf_sample->ip_callchain->nr meaning, that is the
number of entries, be it real addresses or PERF_CONTEXT_ entries, while
honouring the max_stack knobs, i.e. the end result will be max_stack
entries if we have at least that many entries in a given stack trace.
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/n/tip-s8teto51tdqvlfhefndtat9r@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
This makes perf_callchain_{user,kernel}() receive the max stack
as context for the perf_callchain_entry, instead of accessing
the global sysctl_perf_event_max_stack.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: He Kuang <hekuang@huawei.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Milian Wolff <milian.wolff@kdab.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Wang Nan <wangnan0@huawei.com>
Cc: Zefan Li <lizefan@huawei.com>
Link: http://lkml.kernel.org/n/tip-kolmn1yo40p7jhswxwrc7rrd@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
The default remains 127, which is good for most cases, and not even hit
most of the time, but then for some cases, as reported by Brendan, 1024+
deep frames are appearing on the radar for things like groovy, ruby.
And in some workloads putting a _lower_ cap on this may make sense. One
that is per event still needs to be put in place tho.
The new file is:
# cat /proc/sys/kernel/perf_event_max_stack
127
Chaging it:
# echo 256 > /proc/sys/kernel/perf_event_max_stack
# cat /proc/sys/kernel/perf_event_max_stack
256
But as soon as there is some event using callchains we get:
# echo 512 > /proc/sys/kernel/perf_event_max_stack
-bash: echo: write error: Device or resource busy
#
Because we only allocate the callchain percpu data structures when there
is a user, which allows for changing the max easily, its just a matter
of having no callchain users at that point.
Reported-and-Tested-by: Brendan Gregg <brendan.d.gregg@gmail.com>
Reviewed-by: Frederic Weisbecker <fweisbec@gmail.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: He Kuang <hekuang@huawei.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Milian Wolff <milian.wolff@kdab.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Wang Nan <wangnan0@huawei.com>
Cc: Zefan Li <lizefan@huawei.com>
Link: http://lkml.kernel.org/r/20160426002928.GB16708@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
|
|
Function graph tracer modifies a return address (LR) in a stack frame
to hook a function return. This will result in many useless entries
(return_to_handler) showing up in
a) a stack tracer's output
b) perf call graph (with perf record -g)
c) dump_backtrace (at panic et al.)
For example, in case of a),
$ echo function_graph > /sys/kernel/debug/tracing/current_tracer
$ echo 1 > /proc/sys/kernel/stack_trace_enabled
$ cat /sys/kernel/debug/tracing/stack_trace
Depth Size Location (54 entries)
----- ---- --------
0) 4504 16 gic_raise_softirq+0x28/0x150
1) 4488 80 smp_cross_call+0x38/0xb8
2) 4408 48 return_to_handler+0x0/0x40
3) 4360 32 return_to_handler+0x0/0x40
...
In case of b),
$ echo function_graph > /sys/kernel/debug/tracing/current_tracer
$ perf record -e mem:XXX:x -ag -- sleep 10
$ perf report
...
| | |--0.22%-- 0x550f8
| | | 0x10888
| | | el0_svc_naked
| | | sys_openat
| | | return_to_handler
| | | return_to_handler
...
In case of c),
$ echo function_graph > /sys/kernel/debug/tracing/current_tracer
$ echo c > /proc/sysrq-trigger
...
Call trace:
[<ffffffc00044d3ac>] sysrq_handle_crash+0x24/0x30
[<ffffffc000092250>] return_to_handler+0x0/0x40
[<ffffffc000092250>] return_to_handler+0x0/0x40
...
This patch replaces such entries with real addresses preserved in
current->ret_stack[] at unwind_frame(). This way, we can cover all
the cases.
Reviewed-by: Jungseok Lee <jungseoklee85@gmail.com>
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
[will: fixed minor context changes conflicting with irq stack bits]
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
|
Function graph tracer modifies a return address (LR) in a stack frame
to hook a function's return. This will result in many useless entries
(return_to_handler) showing up in a call stack list.
We will fix this problem in a later patch ("arm64: ftrace: fix a stack
tracer's output under function graph tracer"). But since real return
addresses are saved in ret_stack[] array in struct task_struct,
unwind functions need to be notified of, in addition to a stack pointer
address, which task is being traced in order to find out real return
addresses.
This patch extends unwind functions' interfaces by adding an extra
argument of a pointer to task_struct.
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
|
We currently bundle the callchain handling code with the PMU code,
despite the fact the two are distinct, and the former can be useful even
in the absence of the latter.
Follow the example of arch/arm and factor the callchain handling into
its own file dependent on CONFIG_PERF_EVENTS rather than
CONFIG_HW_PERF_EVENTS.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
|
