diff options
| author |  Uros Bizjak <ubizjak@gmail.com>
| 2024-03-24 22:19:45 +0100 |
|---|---|---|
| committer |  Ingo Molnar <mingo@kernel.org>
| 2024-03-25 12:03:45 +0100 |
| commit | b90169b42a6f49ff2fe2e4d4ed0bbcf17fb8f1bf (patch) | |
| tree | e6c2f921f9803ac759ad9715ba9b4eb6c5379305 /arch/x86/Kconfig | |
| parent | x86/percpu: Unify arch_raw_cpu_ptr() defines (diff) | |
| download | linux-b90169b42a6f49ff2fe2e4d4ed0bbcf17fb8f1bf.tar.gz linux-b90169b42a6f49ff2fe2e4d4ed0bbcf17fb8f1bf.tar.bz2 linux-b90169b42a6f49ff2fe2e4d4ed0bbcf17fb8f1bf.zip | |
x86/percpu: Do not use this_cpu_read_stable_8() for 32-bit targets
this_cpu_read_stable() macro uses __pcpu_size_call_return()
that unconditionally calls this_cpu_read_stable_8() also for
32-bit targets. This usage is ivalid as it will result in the
generation of 64-bit MOVQ instruction on 32-bit targets via
percpu_stable_op() macro.
Since there is no generic support for this_cpu_read_stable_8() for
32-bit targets, the patch defines this_cpu_read_stable_8() to
BUILD_BUG() when CONFIG_X86_64 is not defined. This way, we are
sure that this_cpu_read_stable_8() won't actually be used for
32-bit targets, but it is still defined to prevent build failure.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Ard Biesheuvel <ardb@kernel.org>
Link: https://lore.kernel.org/r/20240324212014.310189-1-ubizjak@gmail.com
Diffstat (limited to 'arch/x86/Kconfig')
0 files changed, 0 insertions, 0 deletions