diff options
author | Mark Rutland <mark.rutland@arm.com> | 2017-03-10 18:44:18 +0100 |
---|---|---|
committer | Will Deacon <will.deacon@arm.com> | 2017-03-10 18:45:16 +0100 |
commit | 14088540ad63c648e5cdf490412033f792d16b6b (patch) | |
tree | 634b7aaf1a7376f476fba53f963113d828d94673 /arch/arm64/kernel/cpuidle.c | |
parent | arm64: support keyctl() system call in 32-bit mode (diff) | |
download | linux-14088540ad63c648e5cdf490412033f792d16b6b.tar.xz linux-14088540ad63c648e5cdf490412033f792d16b6b.zip |
arm64: use const cap for system_uses_ttbr0_pan()
Since commit 4b65a5db362783ab ("arm64: Introduce
uaccess_{disable,enable} functionality based on TTBR0_EL1"),
system_uses_ttbr0_pan() has used cpus_have_cap() to determine whether
PAN is present.
Since commit a4023f682739439b ("arm64: Add hypervisor safe helper for
checking constant capabilities"), which was introduced around the same
time, cpus_have_cap() doesn't try to use a static key, and must always
perform a load, test, and consitional branch (likely a tbnz for the
latter two).
Elsewhere, we moved to using cpus_have_const_cap(), which can use a
static key (i.e. a non-conditional branch), which is patched at runtime
when the feature is detected.
This patch makes system_uses_ttbr0_pan() use cpus_have_const_cap(). The
static key is likely a win for hot-paths like the uacccess primitives,
and this makes our usage consistent regardless.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Diffstat (limited to 'arch/arm64/kernel/cpuidle.c')
0 files changed, 0 insertions, 0 deletions