linux-toradex.git/arch/s390/kernel/jump_label.c, branch v6.7

jump_label: make initial NOP patching the special case

2022-06-24T07:48:55+00:00

Instead of defaulting to patching NOP opcodes at init time, and leaving
it to the architectures to override this if this is not needed, switch
to a model where doing nothing is the default. This is the common case
by far, as only MIPS requires NOP patching at init time. On all other
architectures, the correct encodings are emitted by the compiler and so
no initial patching is needed.

Signed-off-by: Ard Biesheuvel 
Signed-off-by: Peter Zijlstra (Intel) 
Acked-by: Mark Rutland 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20220615154142.1574619-4-ardb@kernel.org

jump_label: s390: avoid pointless initial NOP patching

2022-06-24T07:48:54+00:00

Patching NOPs into other NOPs at boot time serves no purpose, so let's
use the same NOP encodings at compile time and runtime.

Signed-off-by: Ard Biesheuvel 
Signed-off-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20220615154142.1574619-2-ardb@kernel.org

s390/jump_label: add __init_or_module annotation

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

Add missing __init_or_module to arch_jump_label_transform_static().

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

s390/jump_label: rename __jump_label_transform()

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

Trivial patch just to get rid of the leading underscores.

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

s390/jump_label: make use of HAVE_JUMP_LABEL_BATCH

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

Specify HAVE_JUMP_LABEL_BATCH in header file. This allows to make use
of the arch_jump_label_transform_queue()/arch_jump_label_transform_apply()
mechanism.

However unlike on x86, which currently is the only user of this
mechanism, the to be patched instructions are still directly
modified. The only difference to before is that serialization is only
done after all instructions have been modified. This way the number of
serialization/synchronization events is reduced.

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

s390/jump_label: use text_poke_sync()

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

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

s390/jump_label: print real address in a case of a jump label bug

2021-07-27T07:39:13+00:00

In case of a jump label print the real address of the piece of code
where a mismatch was detected. This is right before the system panics,
so there is nothing revealed.

Signed-off-by: Heiko Carstens

s390/jump_label: remove unused structure definition

2019-06-07T08:10:10+00:00

Signed-off-by: Heiko Carstens

s390/jump_label: replace stop_machine with smp_call_function

2019-06-04T13:03:12+00:00

The use of stop_machine to replace the mask bits of the jump label branch
is a very heavy-weight operation. This is in fact not necessary, the
mask of the branch can simply be updated, followed by a signal processor
to all the other CPUs to force them to pick up the modified instruction.

Signed-off-by: Martin Schwidefsky 
[heiko.carstens@de.ibm.com]: Change jump_label_make_nop() so we get
                             brcl 0,offset instead of brcl 0,0. This
                             makes sure that only the mask part of the
                             instruction gets changed when updated.
Signed-off-by: Heiko Carstens

jump_label: move 'asm goto' support test to Kconfig

2019-01-06T00:46:51+00:00

Currently, CONFIG_JUMP_LABEL just means "I _want_ to use jump label".

The jump label is controlled by HAVE_JUMP_LABEL, which is defined
like this:

  #if defined(CC_HAVE_ASM_GOTO) && defined(CONFIG_JUMP_LABEL)
  # define HAVE_JUMP_LABEL
  #endif

We can improve this by testing 'asm goto' support in Kconfig, then
make JUMP_LABEL depend on CC_HAS_ASM_GOTO.

Ugly #ifdef HAVE_JUMP_LABEL will go away, and CONFIG_JUMP_LABEL will
match to the real kernel capability.

Signed-off-by: Masahiro Yamada 
Acked-by: Michael Ellerman  (powerpc)
Tested-by: Sedat Dilek