As done by the ECB and CBC helpers in arch/x86/crypt/ecb_cbc_helpers.h, limit the number of bytes processed between kernel_fpu_begin() and kernel_fpu_end() calls. Those functions call preempt_disable() and preempt_enable(), so the CPU core is unavailable for scheduling while running, causing: rcu: INFO: rcu_preempt detected expedited stalls on CPUs/tasks: ... Fixes: 930ab34d906d ("crypto: x86/sm3 - add AVX assembly implementation") Suggested-by: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx> Signed-off-by: Robert Elliott <elliott@xxxxxxx> --- arch/x86/crypto/sm3_avx_glue.c | 29 ++++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-) diff --git a/arch/x86/crypto/sm3_avx_glue.c b/arch/x86/crypto/sm3_avx_glue.c index 661b6f22ffcd..ffb6d2f409ef 100644 --- a/arch/x86/crypto/sm3_avx_glue.c +++ b/arch/x86/crypto/sm3_avx_glue.c @@ -17,6 +17,8 @@ #include <crypto/sm3_base.h> #include <asm/simd.h> +#define FPU_BYTES 4096U /* avoid kernel_fpu_begin/end scheduler/rcu stalls */ + asmlinkage void sm3_transform_avx(struct sm3_state *state, const u8 *data, int nblocks); @@ -37,9 +39,16 @@ static int sm3_avx_update(struct shash_desc *desc, const u8 *data, */ BUILD_BUG_ON(offsetof(struct sm3_state, state) != 0); - kernel_fpu_begin(); - sm3_base_do_update(desc, data, len, sm3_transform_avx); - kernel_fpu_end(); + do { + unsigned int chunk = min(len, FPU_BYTES); + + kernel_fpu_begin(); + sm3_base_do_update(desc, data, chunk, sm3_transform_avx); + kernel_fpu_end(); + + len -= chunk; + data += chunk; + } while (len); return 0; } @@ -57,9 +66,19 @@ static int sm3_avx_finup(struct shash_desc *desc, const u8 *data, return 0; } + do { + unsigned int chunk = min(len, FPU_BYTES); + + if (chunk) { + kernel_fpu_begin(); + sm3_base_do_update(desc, data, chunk, sm3_transform_avx); + kernel_fpu_end(); + } + + len -= chunk; + data += chunk; + } while (len); kernel_fpu_begin(); - if (len) - sm3_base_do_update(desc, data, len, sm3_transform_avx); sm3_base_do_finalize(desc, sm3_transform_avx); kernel_fpu_end(); -- 2.37.3