arm64: neon: Fix function may_use_simd() return error status
authorYandong Zhao <yandong77520@gmail.com>
Wed, 11 Jul 2018 11:06:28 +0000 (19:06 +0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 17 Jul 2018 09:39:30 +0000 (11:39 +0200)
commitd8148f7327488cbeaf5448ed839344b4b3de6efd
tree35fc7fac0e7ae79c9a406a9312d2bca0810b30cf
parent324881805629d163a1dae9f05da3b57ee1bf4686
arm64: neon: Fix function may_use_simd() return error status

commit 2fd8eb4ad87104c54800ef3cea498c92eb15c78a upstream.

It does not matter if the caller of may_use_simd() migrates to
another cpu after the call, but it is still important that the
kernel_neon_busy percpu instance that is read matches the cpu the
task is running on at the time of the read.

This means that raw_cpu_read() is not sufficient.  kernel_neon_busy
may appear true if the caller migrates during the execution of
raw_cpu_read() and the next task to be scheduled in on the initial
cpu calls kernel_neon_begin().

This patch replaces raw_cpu_read() with this_cpu_read() to protect
against this race.

Cc: <stable@vger.kernel.org>
Fixes: cb84d11e1625 ("arm64: neon: Remove support for nested or hardirq kernel-mode NEON")
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Yandong Zhao <yandong77520@gmail.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/arm64/include/asm/simd.h