Skip to main content

Browse the archive

https://github.com/torvalds/linux
23 October 2021, 00:29:24 UTC
Revision 928501344fc645f80390afc12708c81b3595745d authored by Andrey Konovalov on 26 February 2021, 01:19:55 UTC, committed by Linus Torvalds on 26 February 2021, 17:41:02 UTC 
kasan, mm: don't save alloc stacks twice
 
Patch series "kasan: optimizations and fixes for HW_TAGS", v4.

This patchset makes the HW_TAGS mode more efficient, mostly by reworking
poisoning approaches and simplifying/inlining some internal helpers.

With this change, the overhead of HW_TAGS annotations excluding setting
and checking memory tags is ~3%.  The performance impact caused by tags
will be unknown until we have hardware that supports MTE.

As a side-effect, this patchset speeds up generic KASAN by ~15%.

This patch (of 13):

Currently KASAN saves allocation stacks in both kasan_slab_alloc() and
kasan_kmalloc() annotations.  This patch changes KASAN to save allocation
stacks for slab objects from kmalloc caches in kasan_kmalloc() only, and
stacks for other slab objects in kasan_slab_alloc() only.

This change requires ____kasan_kmalloc() knowing whether the object
belongs to a kmalloc cache.  This is implemented by adding a flag field to
the kasan_info structure.  That flag is only set for kmalloc caches via a
new kasan_cache_create_kmalloc() annotation.

Link: https://lkml.kernel.org/r/cover.1612546384.git.andreyknvl@google.com
Link: https://lkml.kernel.org/r/7c673ebca8d00f40a7ad6f04ab9a2bddeeae2097.1612546384.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Reviewed-by: Marco Elver <elver@google.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Peter Collingbourne <pcc@google.com>
Cc: Evgenii Stepanov <eugenis@google.com>
Cc: Branislav Rankov <Branislav.Rankov@arm.com>
Cc: Kevin Brodsky <kevin.brodsky@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent d3a61f7
Raw File
Tip revision: 928501344fc645f80390afc12708c81b3595745d authored by Andrey Konovalov on 26 February 2021, 01:19:55 UTC
kasan, mm: don't save alloc stacks twice
Tip revision: 9285013
sha1_base.h 
/* SPDX-License-Identifier: GPL-2.0-only */
/*
 * sha1_base.h - core logic for SHA-1 implementations
 *
 * Copyright (C) 2015 Linaro Ltd <ard.biesheuvel@linaro.org>
 */

#ifndef _CRYPTO_SHA1_BASE_H
#define _CRYPTO_SHA1_BASE_H

#include <crypto/internal/hash.h>
#include <crypto/sha1.h>
#include <linux/crypto.h>
#include <linux/module.h>
#include <linux/string.h>

#include <asm/unaligned.h>

typedef void (sha1_block_fn)(struct sha1_state *sst, u8 const *src, int blocks);

static inline int sha1_base_init(struct shash_desc *desc)
{
	struct sha1_state *sctx = shash_desc_ctx(desc);

	sctx->state[0] = SHA1_H0;
	sctx->state[1] = SHA1_H1;
	sctx->state[2] = SHA1_H2;
	sctx->state[3] = SHA1_H3;
	sctx->state[4] = SHA1_H4;
	sctx->count = 0;

	return 0;
}

static inline int sha1_base_do_update(struct shash_desc *desc,
				      const u8 *data,
				      unsigned int len,
				      sha1_block_fn *block_fn)
{
	struct sha1_state *sctx = shash_desc_ctx(desc);
	unsigned int partial = sctx->count % SHA1_BLOCK_SIZE;

	sctx->count += len;

	if (unlikely((partial + len) >= SHA1_BLOCK_SIZE)) {
		int blocks;

		if (partial) {
			int p = SHA1_BLOCK_SIZE - partial;

			memcpy(sctx->buffer + partial, data, p);
			data += p;
			len -= p;

			block_fn(sctx, sctx->buffer, 1);
		}

		blocks = len / SHA1_BLOCK_SIZE;
		len %= SHA1_BLOCK_SIZE;

		if (blocks) {
			block_fn(sctx, data, blocks);
			data += blocks * SHA1_BLOCK_SIZE;
		}
		partial = 0;
	}
	if (len)
		memcpy(sctx->buffer + partial, data, len);

	return 0;
}

static inline int sha1_base_do_finalize(struct shash_desc *desc,
					sha1_block_fn *block_fn)
{
	const int bit_offset = SHA1_BLOCK_SIZE - sizeof(__be64);
	struct sha1_state *sctx = shash_desc_ctx(desc);
	__be64 *bits = (__be64 *)(sctx->buffer + bit_offset);
	unsigned int partial = sctx->count % SHA1_BLOCK_SIZE;

	sctx->buffer[partial++] = 0x80;
	if (partial > bit_offset) {
		memset(sctx->buffer + partial, 0x0, SHA1_BLOCK_SIZE - partial);
		partial = 0;

		block_fn(sctx, sctx->buffer, 1);
	}

	memset(sctx->buffer + partial, 0x0, bit_offset - partial);
	*bits = cpu_to_be64(sctx->count << 3);
	block_fn(sctx, sctx->buffer, 1);

	return 0;
}

static inline int sha1_base_finish(struct shash_desc *desc, u8 *out)
{
	struct sha1_state *sctx = shash_desc_ctx(desc);
	__be32 *digest = (__be32 *)out;
	int i;

	for (i = 0; i < SHA1_DIGEST_SIZE / sizeof(__be32); i++)
		put_unaligned_be32(sctx->state[i], digest++);

	memzero_explicit(sctx, sizeof(*sctx));
	return 0;
}

#endif /* _CRYPTO_SHA1_BASE_H */
back to top