linux-toradex.git/kernel/entry, branch v6.16-rc6

entry: Inline syscall_exit_to_user_mode()

2025-04-29T06:27:10+00:00

Similar to commit 221a164035fd ("entry: Move syscall_enter_from_user_mode()
to header file"), move syscall_exit_to_user_mode() to the header file as
well.

Testing was done with the byte-unixbench syscall benchmark (which calls
getpid) and QEMU. On riscv I measured a 7.09246% improvement, on x86 a
2.98843% improvement, on loongarch a 6.07954% improvement, and on s390 a
11.1328% improvement.

The Intel bot also reported "kernel test robot noticed a 1.9% improvement
of stress-ng.seek.ops_per_sec".

Signed-off-by: Charlie Jenkins 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Alexandre Ghiti 
Link: https://lore.kernel.org/all/20250320-riscv_optimize_entry-v6-4-63e187e26041@rivosinc.com
Link: https://lore.kernel.org/linux-riscv/202502051555.85ae6844-lkp@intel.com/

Merge tag 'objtool-core-2025-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

2025-03-25T04:18:05+00:00

Pull objtool updates from Ingo Molnar:

 - The biggest change is the new option to automatically fail the build
   on objtool warnings: CONFIG_OBJTOOL_WERROR.

   While there are no currently known unfixed false positives left, such
   an expansion in the severity of objtool warnings inevitably creates a
   risk of build failures, so it's disabled by default and depends on
   !COMPILE_TEST, so it shouldn't be enabled on
   allyesconfig/allmodconfig builds and won't be forced on people who
   just accept build-time defaults in 'make oldconfig'.

   While the option is strongly recommended, only people who enable it
   explicitly should see it.

   (Josh Poimboeuf)

 - Disable branch profiling in noinstr code with a broad brush that
   includes all of arch/x86/ and kernel/sched/. (Josh Poimboeuf)

 - Create backup object files on objtool errors and print exact objtool
   arguments to make failure analysis easier (Josh Poimboeuf)

 - Improve noreturn handling (Josh Poimboeuf)

 - Improve rodata handling (Tiezhu Yang)

 - Support jump tables, switch tables and goto tables on LoongArch
   (Tiezhu Yang)

 - Misc cleanups and fixes (Josh Poimboeuf, David Engraf, Ingo Molnar)

* tag 'objtool-core-2025-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (22 commits)
  tracing: Disable branch profiling in noinstr code
  objtool: Use O_CREAT with explicit mode mask
  objtool: Add CONFIG_OBJTOOL_WERROR
  objtool: Create backup on error and print args
  objtool: Change "warning:" to "error:" for --Werror
  objtool: Add --Werror option
  objtool: Add --output option
  objtool: Upgrade "Linked object detected" warning to error
  objtool: Consolidate option validation
  objtool: Remove --unret dependency on --rethunk
  objtool: Increase per-function WARN_FUNC() rate limit
  objtool: Update documentation
  objtool: Improve __noreturn annotation warning
  objtool: Fix error handling inconsistencies in check()
  x86/traps: Make exc_double_fault() consistently noreturn
  LoongArch: Enable jump table for objtool
  objtool/LoongArch: Add support for goto table
  objtool/LoongArch: Add support for switch table
  objtool: Handle PC relative relocation type
  objtool: Handle different entry size of rodata
  ...

tracing: Disable branch profiling in noinstr code

2025-03-22T08:49:26+00:00

CONFIG_TRACE_BRANCH_PROFILING inserts a call to ftrace_likely_update()
for each use of likely() or unlikely().  That breaks noinstr rules if
the affected function is annotated as noinstr.

Disable branch profiling for files with noinstr functions.  In addition
to some individual files, this also includes the entire arch/x86
subtree, as well as the kernel/entry, drivers/cpuidle, and drivers/idle
directories, all of which are noinstr-heavy.

Due to the nature of how sched binaries are built by combining multiple
.c files into one, branch profiling is disabled more broadly across the
sched code than would otherwise be needed.

This fixes many warnings like the following:

  vmlinux.o: warning: objtool: do_syscall_64+0x40: call to ftrace_likely_update() leaves .noinstr.text section
  vmlinux.o: warning: objtool: __rdgsbase_inactive+0x33: call to ftrace_likely_update() leaves .noinstr.text section
  vmlinux.o: warning: objtool: handle_bug.isra.0+0x198: call to ftrace_likely_update() leaves .noinstr.text section
  ...

Reported-by: Ingo Molnar 
Suggested-by: Steven Rostedt 
Signed-off-by: Josh Poimboeuf 
Signed-off-by: Ingo Molnar 
Acked-by: Thomas Gleixner 
Cc: Linus Torvalds 
Link: https://lore.kernel.org/r/fb94fc9303d48a5ed370498f54500cc4c338eb6d.1742586676.git.jpoimboe@kernel.org

seccomp: remove the 'sd' argument from __secure_computing()

2025-02-10T17:26:22+00:00

After the previous changes 'sd' is always NULL.

Signed-off-by: Oleg Nesterov 
Reviewed-by: Kees Cook 
Link: https://lore.kernel.org/r/20250128150313.GA15336@redhat.com
Signed-off-by: Kees Cook

sched: Add TIF_NEED_RESCHED_LAZY infrastructure

2024-11-05T11:55:37+00:00

Add the basic infrastructure to split the TIF_NEED_RESCHED bit in two.
Either bit will cause a resched on return-to-user, but only
TIF_NEED_RESCHED will drive IRQ preemption.

No behavioural change intended.

Suggested-by: Thomas Gleixner 
Signed-off-by: Peter Zijlstra (Intel) 
Reviewed-by: Sebastian Andrzej Siewior 
Link: https://lkml.kernel.org/r/20241007075055.219540785@infradead.org

treewide: context_tracking: Rename CONTEXT_* into CT_STATE_*

2024-07-29T02:03:10+00:00

Context tracking state related symbols currently use a mix of the
CONTEXT_ (e.g. CONTEXT_KERNEL) and CT_SATE_ (e.g. CT_STATE_MASK) prefixes.

Clean up the naming and make the ctx_state enum use the CT_STATE_ prefix.

Suggested-by: Frederic Weisbecker 
Signed-off-by: Valentin Schneider 
Acked-by: Frederic Weisbecker 
Acked-by: Thomas Gleixner 
Signed-off-by: Neeraj Upadhyay

entry: Respect changes to system call number by trace_sys_enter()

2024-03-12T12:23:32+00:00

When a probe is registered at the trace_sys_enter() tracepoint, and that
probe changes the system call number, the old system call still gets
executed.  This worked correctly until commit b6ec41346103 ("core/entry:
Report syscall correctly for trace and audit"), which removed the
re-evaluation of the syscall number after the trace point.

Restore the original semantics by re-evaluating the system call number
after trace_sys_enter(). 

The performance impact of this re-evaluation is minimal because it only
takes place when a trace point is active, and compared to the actual trace
point overhead the read from a cache hot variable is negligible.

Fixes: b6ec41346103 ("core/entry: Report syscall correctly for trace and audit")
Signed-off-by: André Rösti 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20240311211704.7262-1-an.roesti@gmail.com

entry: Move syscall_enter_from_user_mode() to header file

2023-12-21T22:12:18+00:00

To allow inlining of syscall_enter_from_user_mode(), move it
to entry-common.h.

Signed-off-by: Sven Schnelle 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20231218074520.1998026-4-svens@linux.ibm.com

entry: Move enter_from_user_mode() to header file

2023-12-21T22:12:18+00:00

To allow inlining of enter_from_user_mode(), move it to
entry-common.h.

Signed-off-by: Sven Schnelle 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20231218074520.1998026-3-svens@linux.ibm.com

entry: Move exit to usermode functions to header file

2023-12-21T22:12:18+00:00

To allow inlining, move exit_to_user_mode() to
entry-common.h.

Signed-off-by: Sven Schnelle 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/r/20231218074520.1998026-2-svens@linux.ibm.com