diff options
| author | Mark Rutland <mark.rutland@arm.com> | 2023-06-05 10:00:58 +0300 |
|---|---|---|
| committer | Peter Zijlstra <peterz@infradead.org> | 2023-06-05 10:57:13 +0300 |
| commit | dda5f312bb09e56e7a1c3e3851f2000eb2e9c879 (patch) | |
| tree | 2c5d77a688caffdffb4b516c1e6b00baeadb1259 /arch/arm/lib/bitops.h | |
| parent | 497cc42bf53b55185ab3d39c634fbf09eb6681ae (diff) | |
| download | linux-dda5f312bb09e56e7a1c3e3851f2000eb2e9c879.tar.xz | |
locking/atomic: arm: fix sync ops
The sync_*() ops on arch/arm are defined in terms of the regular bitops
with no special handling. This is not correct, as UP kernels elide
barriers for the fully-ordered operations, and so the required ordering
is lost when such UP kernels are run under a hypervsior on an SMP
system.
Fix this by defining sync ops with the required barriers.
Note: On 32-bit arm, the sync_*() ops are currently only used by Xen,
which requires ARMv7, but the semantics can be implemented for ARMv6+.
Fixes: e54d2f61528165bb ("xen/arm: sync_bitops")
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20230605070124.3741859-2-mark.rutland@arm.com
Diffstat (limited to 'arch/arm/lib/bitops.h')
| -rw-r--r-- | arch/arm/lib/bitops.h | 14 |
1 files changed, 11 insertions, 3 deletions
diff --git a/arch/arm/lib/bitops.h b/arch/arm/lib/bitops.h index 95bd35991288..f069d1b2318e 100644 --- a/arch/arm/lib/bitops.h +++ b/arch/arm/lib/bitops.h @@ -28,7 +28,7 @@ UNWIND( .fnend ) ENDPROC(\name ) .endm - .macro testop, name, instr, store + .macro __testop, name, instr, store, barrier ENTRY( \name ) UNWIND( .fnstart ) ands ip, r1, #3 @@ -38,7 +38,7 @@ UNWIND( .fnstart ) mov r0, r0, lsr #5 add r1, r1, r0, lsl #2 @ Get word offset mov r3, r2, lsl r3 @ create mask - smp_dmb + \barrier #if __LINUX_ARM_ARCH__ >= 7 && defined(CONFIG_SMP) .arch_extension mp ALT_SMP(W(pldw) [r1]) @@ -50,13 +50,21 @@ UNWIND( .fnstart ) strex ip, r2, [r1] cmp ip, #0 bne 1b - smp_dmb + \barrier cmp r0, #0 movne r0, #1 2: bx lr UNWIND( .fnend ) ENDPROC(\name ) .endm + + .macro testop, name, instr, store + __testop \name, \instr, \store, smp_dmb + .endm + + .macro sync_testop, name, instr, store + __testop \name, \instr, \store, __smp_dmb + .endm #else .macro bitop, name, instr ENTRY( \name ) |