linux-toradex.git/arch/s390/kernel/ftrace.c, branch v5.16-rc3

Merge tag 's390-5.16-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux

2021-11-06T21:48:06+00:00

Pull s390 updates from Vasily Gorbik:

 - Add support for ftrace with direct call and ftrace direct call
   samples.

 - Add support for kernel command lines longer than current 896 bytes
   and make its length configurable.

 - Add support for BEAR enhancement facility to improve last breaking
   event instruction tracking.

 - Add kprobes sanity checks and testcases to prevent kprobe in the mid
   of an instruction.

 - Allow concurrent access to /dev/hwc for the CPUMF users.

 - Various ftrace / jump label improvements.

 - Convert unwinder tests to KUnit.

 - Add s390_iommu_aperture kernel parameter to tweak the limits on
   concurrently usable DMA mappings.

 - Add ap.useirq AP module option which can be used to disable interrupt
   use.

 - Add add_disk() error handling support to block device drivers.

 - Drop arch specific and use generic implementation of strlcpy and
   strrchr.

 - Several __pa/__va usages fixes.

 - Various cio, crypto, pci, kernel doc and other small fixes and
   improvements all over the code.

[ Merge fixup as per https://lore.kernel.org/all/YXAqZ%2FEszRisunQw@osiris/ ]

* tag 's390-5.16-1' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (63 commits)
  s390: make command line configurable
  s390: support command lines longer than 896 bytes
  s390/kexec_file: move kernel image size check
  s390/pci: add s390_iommu_aperture kernel parameter
  s390/spinlock: remove incorrect kernel doc indicator
  s390/string: use generic strlcpy
  s390/string: use generic strrchr
  s390/ap: function rework based on compiler warning
  s390/cio: make ccw_device_dma_* more robust
  s390/vfio-ap: s390/crypto: fix all kernel-doc warnings
  s390/hmcdrv: fix kernel doc comments
  s390/ap: new module option ap.useirq
  s390/cpumf: Allow multiple processes to access /dev/hwc
  s390/bitops: return true/false (not 1/0) from bool functions
  s390: add support for BEAR enhancement facility
  s390: introduce nospec_uses_trampoline()
  s390: rename last_break to pgm_last_break
  s390/ptrace: add last_break member to pt_regs
  s390/sclp: sort out physical vs virtual pointers usage
  s390/setup: convert start and end initrd pointers to virtual
  ...

s390/ftrace: add missing serialization for graph caller patching

2021-10-11T18:55:58+00:00

CPUs must be serialized also when ftrace_graph_caller gets patched.
This is missing since ftrace function graph support was added on s390.

Signed-off-by: Heiko Carstens 
Signed-off-by: Vasily Gorbik

s390/ftrace: use text_poke_sync_lock()

2021-10-11T18:55:58+00:00

Signed-off-by: Heiko Carstens 
Signed-off-by: Vasily Gorbik

ftrace: Cleanup ftrace_dyn_arch_init()

2021-10-08T23:41:39+00:00

Most of ARCHs use empty ftrace_dyn_arch_init(), introduce a weak common
ftrace_dyn_arch_init() to cleanup them.

Link: https://lkml.kernel.org/r/20210909090216.1955240-1-o451686892@gmail.com

Acked-by: Heiko Carstens  (s390)
Acked-by: Helge Deller  (parisc)
Signed-off-by: Weizhao Ouyang 
Signed-off-by: Steven Rostedt (VMware)

s390/ftrace: remove dead code

2021-10-04T07:49:36+00:00

ftrace_shared_hotpatch_trampoline() never returns NULL,
therefore quite a bit of code can be removed.

Acked-by: Ilya Leoshkevich 
Signed-off-by: Heiko Carstens 
Signed-off-by: Vasily Gorbik

s390/ftrace: remove incorrect __va usage

2021-09-08T12:23:31+00:00

The address of ftrace_graph_caller is already virtual.
Using __va() to translate the address is wrong.

Reviewed-by: Alexander Gordeev 
Signed-off-by: Heiko Carstens

s390/ftrace: implement hotpatching

2021-08-03T12:31:40+00:00

s390 allows hotpatching the mask of a conditional jump instruction.
Make use of this feature in order to avoid the expensive stop_machine()
call.

The new trampolines are split in 3 stages:

- A first stage is a 6-byte relative conditional long branch located at
  each function's entry point. Its offset always points to the second
  stage for the corresponding function, and its mask is either all 0s
  (ftrace off) or all 1s (ftrace on). The code for flipping the mask is
  borrowed from ftrace_{enable,disable}_ftrace_graph_caller. After
  flipping, ftrace_arch_code_modify_post_process() syncs with all the
  other CPUs by sending SIGPs.

- Second stages for vmlinux are stored in a separate part of the .text
  section reserved by the linker script, and in dynamically allocated
  memory for modules. This prevents the icache pollution. The total
  size of second stages is about 1.5% of that of the kernel image.

  Putting second stages in the .bss section is possible and decreases
  the size of the non-compressed vmlinux, but splits the kernel 1:1
  mapping, which is a bad tradeoff.

  Each second stage contains a call to the third stage, a pointer to
  the part of the intercepted function right after the first stage, and
  a pointer to an interceptor function (e.g. ftrace_caller).

  Second stages are 8-byte aligned for the future direct calls
  implementation.

- There are only two copies of the third stage: in the .text section
  for vmlinux and in dynamically allocated memory for modules. It can be
  an expoline, which is relatively large, so inlining it into each
  second stage is prohibitively expensive.

As a result of this organization, phoronix-test-suite with ftrace off
does not show any performance degradation.

Suggested-by: Sven Schnelle 
Suggested-by: Vasily Gorbik 
Co-developed-by: Heiko Carstens 
Signed-off-by: Ilya Leoshkevich 
Link: https://lore.kernel.org/r/20210728212546.128248-3-iii@linux.ibm.com
Signed-off-by: Heiko Carstens

s390/ftrace: fix ftrace_update_ftrace_func implementation

2021-07-15T10:54:58+00:00

s390 enforces DYNAMIC_FTRACE if FUNCTION_TRACER is selected.
At the same time implementation of ftrace_caller is not compliant with
HAVE_DYNAMIC_FTRACE since it doesn't provide implementation of
ftrace_update_ftrace_func() and calls ftrace_trace_function() directly.

The subtle difference is that during ftrace code patching ftrace
replaces function tracer via ftrace_update_ftrace_func() and activates
it back afterwards. Unexpected direct calls to ftrace_trace_function()
during ftrace code patching leads to nullptr-dereferences when tracing
is activated for one of functions which are used during code patching.
Those function currently are:
copy_from_kernel_nofault()
copy_from_kernel_nofault_allowed()
preempt_count_sub() [with debug_defconfig]
preempt_count_add() [with debug_defconfig]

Corresponding KASAN report:
 BUG: KASAN: nullptr-dereference in function_trace_call+0x316/0x3b0
 Read of size 4 at addr 0000000000001e08 by task migration/0/15

 CPU: 0 PID: 15 Comm: migration/0 Tainted: G B 5.13.0-41423-g08316af3644d
 Hardware name: IBM 3906 M04 704 (LPAR)
 Stopper: multi_cpu_stop+0x0/0x3e0 <- stop_machine_cpuslocked+0x1e4/0x218
 Call Trace:
  [<0000000001f77caa>] show_stack+0x16a/0x1d0
  [<0000000001f8de42>] dump_stack+0x15a/0x1b0
  [<0000000001f81d56>] print_address_description.constprop.0+0x66/0x2e0
  [<000000000082b0ca>] kasan_report+0x152/0x1c0
  [<00000000004cfd8e>] function_trace_call+0x316/0x3b0
  [<0000000001fb7082>] ftrace_caller+0x7a/0x7e
  [<00000000006bb3e6>] copy_from_kernel_nofault_allowed+0x6/0x10
  [<00000000006bb42e>] copy_from_kernel_nofault+0x3e/0xd0
  [<000000000014605c>] ftrace_make_call+0xb4/0x1f8
  [<000000000047a1b4>] ftrace_replace_code+0x134/0x1d8
  [<000000000047a6e0>] ftrace_modify_all_code+0x120/0x1d0
  [<000000000047a7ec>] __ftrace_modify_code+0x5c/0x78
  [<000000000042395c>] multi_cpu_stop+0x224/0x3e0
  [<0000000000423212>] cpu_stopper_thread+0x33a/0x5a0
  [<0000000000243ff2>] smpboot_thread_fn+0x302/0x708
  [<00000000002329ea>] kthread+0x342/0x408
  [<00000000001066b2>] __ret_from_fork+0x92/0xf0
  [<0000000001fb57fa>] ret_from_fork+0xa/0x30

 The buggy address belongs to the page:
 page:(____ptrval____) refcount:1 mapcount:0 mapping:0000000000000000 index:0x0 pfn:0x1
 flags: 0x1ffff00000001000(reserved|node=0|zone=0|lastcpupid=0x1ffff)
 raw: 1ffff00000001000 0000040000000048 0000040000000048 0000000000000000
 raw: 0000000000000000 0000000000000000 ffffffff00000001 0000000000000000
 page dumped because: kasan: bad access detected

 Memory state around the buggy address:
  0000000000001d00: f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7
  0000000000001d80: f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7
 >0000000000001e00: f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7
                       ^
  0000000000001e80: f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7
  0000000000001f00: f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7 f7
 ==================================================================

To fix that introduce ftrace_func callback to be called from
ftrace_caller and update it in ftrace_update_ftrace_func().

Fixes: 4cc9bed034d1 ("[S390] cleanup ftrace backend functions")
Cc: stable@vger.kernel.org
Reviewed-by: Heiko Carstens 
Signed-off-by: Vasily Gorbik 
Signed-off-by: Heiko Carstens

Merge tag 'trace-v5.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

2020-12-17T21:22:17+00:00

Pull tracing updates from Steven Rostedt:
 "The major update to this release is that there's a new arch config
  option called CONFIG_HAVE_DYNAMIC_FTRACE_WITH_ARGS.

  Currently, only x86_64 enables it. All the ftrace callbacks now take a
  struct ftrace_regs instead of a struct pt_regs. If the architecture
  has HAVE_DYNAMIC_FTRACE_WITH_ARGS enabled, then the ftrace_regs will
  have enough information to read the arguments of the function being
  traced, as well as access to the stack pointer.

  This way, if a user (like live kernel patching) only cares about the
  arguments, then it can avoid using the heavier weight "regs" callback,
  that puts in enough information in the struct ftrace_regs to simulate
  a breakpoint exception (needed for kprobes).

  A new config option that audits the timestamps of the ftrace ring
  buffer at most every event recorded.

  Ftrace recursion protection has been cleaned up to move the protection
  to the callback itself (this saves on an extra function call for those
  callbacks).

  Perf now handles its own RCU protection and does not depend on ftrace
  to do it for it (saving on that extra function call).

  New debug option to add "recursed_functions" file to tracefs that
  lists all the places that triggered the recursion protection of the
  function tracer. This will show where things need to be fixed as
  recursion slows down the function tracer.

  The eval enum mapping updates done at boot up are now offloaded to a
  work queue, as it caused a noticeable pause on slow embedded boards.

  Various clean ups and last minute fixes"

* tag 'trace-v5.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (33 commits)
  tracing: Offload eval map updates to a work queue
  Revert: "ring-buffer: Remove HAVE_64BIT_ALIGNED_ACCESS"
  ring-buffer: Add rb_check_bpage in __rb_allocate_pages
  ring-buffer: Fix two typos in comments
  tracing: Drop unneeded assignment in ring_buffer_resize()
  tracing: Disable ftrace selftests when any tracer is running
  seq_buf: Avoid type mismatch for seq_buf_init
  ring-buffer: Fix a typo in function description
  ring-buffer: Remove obsolete rb_event_is_commit()
  ring-buffer: Add test to validate the time stamp deltas
  ftrace/documentation: Fix RST C code blocks
  tracing: Clean up after filter logic rewriting
  tracing: Remove the useless value assignment in test_create_synth_event()
  livepatch: Use the default ftrace_ops instead of REGS when ARGS is available
  ftrace/x86: Allow for arguments to be passed in to ftrace_regs by default
  ftrace: Have the callbacks receive a struct ftrace_regs instead of pt_regs
  MAINTAINERS: assign ./fs/tracefs to TRACING
  tracing: Fix some typos in comments
  ftrace: Remove unused varible 'ret'
  ring-buffer: Add recording of ring buffer recursion into recursed_functions
  ...

s390/ftrace: assume -mhotpatch or -mrecord-mcount always available

2020-11-20T18:19:11+00:00

Currently the kernel minimal compiler requirement is gcc 4.9 or
clang 10.0.1.
* gcc -mhotpatch option is supported since 4.8.
* A combination of -pg -mrecord-mcount -mnop-mcount -mfentry flags is
supported since gcc 9 and since clang 10.

Drop support for old -pg function prologues. Which leaves binary
compatible -mhotpatch / -mnop-mcount -mfentry prologues in a form:
	brcl	0,0
Which are also do not require initial nop optimization / conversion and
presence of _mcount symbol.

Signed-off-by: Vasily Gorbik 
Reviewed-by: Heiko Carstens 
Signed-off-by: Heiko Carstens