Message ID | 1398959486-8222-4-git-send-email-ard.biesheuvel@linaro.org |
---|---|
State | New |
Headers | show |
On 01.05.2014 18:51, Ard Biesheuvel wrote: > The Crypto Extensions based SHA1 implementation uses the NEON register file, > and hence runs with preemption disabled. This patch adds a TIF_NEED_RESCHED > check to its inner loop so we at least give up the CPU voluntarily when we > are running in process context and have been tagged for preemption by the > scheduler. > > Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> > --- <snip> > @@ -42,6 +42,7 @@ static int sha1_update(struct shash_desc *desc, const u8 *data, > sctx->count += len; > > if ((partial + len) >= SHA1_BLOCK_SIZE) { > + struct thread_info *ti = NULL; > int blocks; > > if (partial) { > @@ -52,16 +53,30 @@ static int sha1_update(struct shash_desc *desc, const u8 *data, > len -= p; > } > > + /* > + * Pass current's thread info pointer to sha1_ce_transform() > + * below if we want it to play nice under preemption. > + */ > + if ((IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY) || > + IS_ENABLED(CONFIG_PREEMPT)) && !in_interrupt()) > + ti = current_thread_info(); > + > blocks = len / SHA1_BLOCK_SIZE; > len %= SHA1_BLOCK_SIZE; > > - kernel_neon_begin_partial(16); > - sha1_ce_transform(blocks, data, sctx->state, > - partial ? sctx->buffer : NULL, 0); > - kernel_neon_end(); > + do { > + int rem; > + > + kernel_neon_begin_partial(16); > + rem = sha1_ce_transform(blocks, data, sctx->state, > + partial ? sctx->buffer : NULL, > + 0, ti); > + kernel_neon_end(); > > - data += blocks * SHA1_BLOCK_SIZE; > - partial = 0; > + data += (blocks - rem) * SHA1_BLOCK_SIZE; > + blocks = rem; > + partial = 0; > + } while (unlikely(ti && blocks > 0)); > } > if (len) > memcpy(sctx->buffer + partial, data, len); > @@ -94,6 +109,7 @@ static int sha1_finup(struct shash_desc *desc, const u8 *data, > unsigned int len, u8 *out) > { > struct sha1_state *sctx = shash_desc_ctx(desc); > + struct thread_info *ti = NULL; > __be32 *dst = (__be32 *)out; > int blocks; > int i; > @@ -111,9 +127,20 @@ static int sha1_finup(struct shash_desc *desc, const u8 *data, > */ > blocks = len / SHA1_BLOCK_SIZE; > > - kernel_neon_begin_partial(16); > - sha1_ce_transform(blocks, data, sctx->state, NULL, len); > - kernel_neon_end(); > + if ((IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY) || > + IS_ENABLED(CONFIG_PREEMPT)) && !in_interrupt()) > + ti = current_thread_info(); > + > + do { > + int rem; > + > + kernel_neon_begin_partial(16); > + rem = sha1_ce_transform(blocks, data, sctx->state, > + NULL, len, ti); > + kernel_neon_end(); > + data += (blocks - rem) * SHA1_BLOCK_SIZE; > + blocks = rem; > + } while (unlikely(ti && blocks > 0)); > These seem to be similar, how about renaming assembly function to __sha1_ce_transform and moving this loop to new sha1_ce_transform. Otherwise, patches looks good. -Jussi -- To unsubscribe from this list: send the line "unsubscribe linux-crypto" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Ard Biesheuvel <ard.biesheuvel@linaro.org> wrote: > > + /* > + * Pass current's thread info pointer to sha1_ce_transform() > + * below if we want it to play nice under preemption. > + */ > + if ((IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY) || > + IS_ENABLED(CONFIG_PREEMPT)) && !in_interrupt()) > + ti = current_thread_info(); Instead of in_interrupt you should test CRYPTO_TFM_REQ_MAY_SLEEP. It's not a matter of correctness but it may avoid unnecessary breaks in the loop. Cheers,
diff --git a/arch/arm64/crypto/sha1-ce-core.S b/arch/arm64/crypto/sha1-ce-core.S index bd4af29f2722..831e332f9331 100644 --- a/arch/arm64/crypto/sha1-ce-core.S +++ b/arch/arm64/crypto/sha1-ce-core.S @@ -66,8 +66,8 @@ .word 0x5a827999, 0x6ed9eba1, 0x8f1bbcdc, 0xca62c1d6 /* - * void sha1_ce_transform(int blocks, u8 const *src, u32 *state, - * u8 *head, long bytes) + * int sha1_ce_transform(int blocks, u8 const *src, u32 *state, + * u8 *head, long bytes, struct thread_info *ti) */ ENTRY(sha1_ce_transform) /* load round constants */ @@ -127,7 +127,13 @@ CPU_LE( rev32 v11.16b, v11.16b ) add dgbv.2s, dgbv.2s, dg1v.2s add dgav.4s, dgav.4s, dg0v.4s - cbnz w0, 0b + cbz w0, 4f + b_if_no_resched x5, x8, 0b + + /* store new state */ +3: str dga, [x2] + str dgb, [x2, #16] + ret /* * Final block: add padding and total bit count. @@ -135,7 +141,7 @@ CPU_LE( rev32 v11.16b, v11.16b ) * size was not a round multiple of the block size, and the padding is * handled by the C code. */ - cbz x4, 3f +4: cbz x4, 3b movi v9.2d, #0 mov x8, #0x80000000 movi v10.2d, #0 @@ -145,9 +151,4 @@ CPU_LE( rev32 v11.16b, v11.16b ) mov v11.d[0], xzr mov v11.d[1], x7 b 2b - - /* store new state */ -3: str dga, [x2] - str dgb, [x2, #16] - ret ENDPROC(sha1_ce_transform) diff --git a/arch/arm64/crypto/sha1-ce-glue.c b/arch/arm64/crypto/sha1-ce-glue.c index 6fe83f37a750..69850a163668 100644 --- a/arch/arm64/crypto/sha1-ce-glue.c +++ b/arch/arm64/crypto/sha1-ce-glue.c @@ -20,8 +20,8 @@ MODULE_DESCRIPTION("SHA1 secure hash using ARMv8 Crypto Extensions"); MODULE_AUTHOR("Ard Biesheuvel <ard.biesheuvel@linaro.org>"); MODULE_LICENSE("GPL v2"); -asmlinkage void sha1_ce_transform(int blocks, u8 const *src, u32 *state, - u8 *head, long bytes); +asmlinkage int sha1_ce_transform(int blocks, u8 const *src, u32 *state, + u8 *head, long bytes, struct thread_info *ti); static int sha1_init(struct shash_desc *desc) { @@ -42,6 +42,7 @@ static int sha1_update(struct shash_desc *desc, const u8 *data, sctx->count += len; if ((partial + len) >= SHA1_BLOCK_SIZE) { + struct thread_info *ti = NULL; int blocks; if (partial) { @@ -52,16 +53,30 @@ static int sha1_update(struct shash_desc *desc, const u8 *data, len -= p; } + /* + * Pass current's thread info pointer to sha1_ce_transform() + * below if we want it to play nice under preemption. + */ + if ((IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY) || + IS_ENABLED(CONFIG_PREEMPT)) && !in_interrupt()) + ti = current_thread_info(); + blocks = len / SHA1_BLOCK_SIZE; len %= SHA1_BLOCK_SIZE; - kernel_neon_begin_partial(16); - sha1_ce_transform(blocks, data, sctx->state, - partial ? sctx->buffer : NULL, 0); - kernel_neon_end(); + do { + int rem; + + kernel_neon_begin_partial(16); + rem = sha1_ce_transform(blocks, data, sctx->state, + partial ? sctx->buffer : NULL, + 0, ti); + kernel_neon_end(); - data += blocks * SHA1_BLOCK_SIZE; - partial = 0; + data += (blocks - rem) * SHA1_BLOCK_SIZE; + blocks = rem; + partial = 0; + } while (unlikely(ti && blocks > 0)); } if (len) memcpy(sctx->buffer + partial, data, len); @@ -94,6 +109,7 @@ static int sha1_finup(struct shash_desc *desc, const u8 *data, unsigned int len, u8 *out) { struct sha1_state *sctx = shash_desc_ctx(desc); + struct thread_info *ti = NULL; __be32 *dst = (__be32 *)out; int blocks; int i; @@ -111,9 +127,20 @@ static int sha1_finup(struct shash_desc *desc, const u8 *data, */ blocks = len / SHA1_BLOCK_SIZE; - kernel_neon_begin_partial(16); - sha1_ce_transform(blocks, data, sctx->state, NULL, len); - kernel_neon_end(); + if ((IS_ENABLED(CONFIG_PREEMPT_VOLUNTARY) || + IS_ENABLED(CONFIG_PREEMPT)) && !in_interrupt()) + ti = current_thread_info(); + + do { + int rem; + + kernel_neon_begin_partial(16); + rem = sha1_ce_transform(blocks, data, sctx->state, + NULL, len, ti); + kernel_neon_end(); + data += (blocks - rem) * SHA1_BLOCK_SIZE; + blocks = rem; + } while (unlikely(ti && blocks > 0)); for (i = 0; i < SHA1_DIGEST_SIZE / sizeof(__be32); i++) put_unaligned_be32(sctx->state[i], dst++);
The Crypto Extensions based SHA1 implementation uses the NEON register file, and hence runs with preemption disabled. This patch adds a TIF_NEED_RESCHED check to its inner loop so we at least give up the CPU voluntarily when we are running in process context and have been tagged for preemption by the scheduler. Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org> --- arch/arm64/crypto/sha1-ce-core.S | 19 ++++++++-------- arch/arm64/crypto/sha1-ce-glue.c | 49 +++++++++++++++++++++++++++++++--------- 2 files changed, 48 insertions(+), 20 deletions(-)