x86/kprobes: Fix crash when probing CS CALL instructionsgrafted

When using eBPF to probe CS CALL instructions within a function,
a crash can be triggered.

The eBPF tool probes offset 257 of the __hrtimer_run_queues()
function:

<__hrtimer_run_queues+249>:  nopl   0x0(%rax,%rax,1)
<__hrtimer_run_queues+254>:  mov    %r14,%rdi
<__hrtimer_run_queues+257>:  cs call <__x86_indirect_thunk_r12>
<__hrtimer_run_queues+263>:  mov    %eax,%r12d
<__hrtimer_run_queues+266>:  xchg   %ax,%ax
<__hrtimer_run_queues+268>:  mov    %r13,%rdi

Which triggers this crash:

  BUG: unable to handle page fault for address: 00000000000f41c9
  #PF: supervisor write access in kernel mode
  #PF: error_code(0x0002) - not-present page
  PGD 0 P4D 0
  Oops: 0002 [#1] SMP NOPTI
  CPU: 1 PID: 0 Comm: swapper/1 Kdump: loaded Tainted: P
  RIP: 0010:__hrtimer_run_queues+0x106/0x230

Note that __hrtimer_run_queues+0x106 is __hrtimer_run_queues+262, which is
at the 6th byte of the above CS CALL instruction. Since the CS CALL
instruction occupies 6 bytes, the exception occurred in the middle of that
call instruction.

The root cause is that when using eBPF tools to probe in the middle of a
function, a kprobe with INT3 is used as the underlying implementation.

During single-step emulation of the original CALL instruction,
int3_emulate_call() assumes that the probed CALL instruction is 5 bytes
long. However, the actual CS-prefixed CALL instruction occupies 6 bytes,
so it constructs an incorrect exception return address. When the CPU
returns from the kprobe handler, the next instruction to be executed is at
the address of the last byte of that CS CALL instruction. Coincidentally,
starting from that address, the CPU fetches and decodes a completely
different instruction, which ultimately triggers a kernel crash.

Fix the issue by using the actual instruction length obtained from
the instruction decoder when constructing the exception return
address, rather than relying on the hardcoded CALL_INSN_SIZE macro.

[ mingo: Refined the changelog ]

Fixes: 6256e668b7af ("x86/kprobes: Use int3 instead of debug trap for single-step")
Suggested-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Jinke Han <jinkehan@didiglobal.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Borislav Petkov <bp@alien8.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://patch.msgid.link/20260908073742.GA10517@didi-ThinkCentre-M920t-N000

14 files changed, 3926 insertions, 0 deletions

diff --git a/arch/powerpc/crypto/.gitignore b/arch/powerpc/crypto/.gitignore
new file mode 100644
index 000000000..e9fe73aac
--- /dev/null
+++ b/arch/powerpc/crypto/.gitignore
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0-only
+aesp10-ppc.S
+aesp8-ppc.S
+ghashp10-ppc.S
+ghashp8-ppc.S
diff --git a/arch/powerpc/crypto/Kconfig b/arch/powerpc/crypto/Kconfig
new file mode 100644
index 000000000..b247f7ed9
--- /dev/null
+++ b/arch/powerpc/crypto/Kconfig
@@ -0,0 +1,64 @@
+# SPDX-License-Identifier: GPL-2.0
+
+menu "Accelerated Cryptographic Algorithms for CPU (powerpc)"
+
+config CRYPTO_AES_PPC_SPE
+	tristate "Ciphers: AES, modes: ECB/CBC/CTR/XTS (SPE)"
+	depends on SPE
+	select CRYPTO_LIB_AES
+	select CRYPTO_SKCIPHER
+	help
+	  Length-preserving ciphers: AES with ECB, CBC, CTR, and XTS modes
+
+	  Architecture: powerpc using:
+	  - SPE (Signal Processing Engine) extensions
+
+	  SPE is available for:
+	  - Processor Type: Freescale 8500
+	  - CPU selection: e500 (8540)
+
+	  This module should only be used for low power (router) devices
+	  without hardware AES acceleration (e.g. caam crypto). It reduces the
+	  size of the AES tables from 16KB to 8KB + 256 bytes and mitigates
+	  timining attacks. Nevertheless it might be not as secure as other
+	  architecture specific assembler implementations that work on 1KB
+	  tables or 256 bytes S-boxes.
+
+config CRYPTO_AES_GCM_P10
+	tristate "Stitched AES/GCM acceleration support on P10 or later CPU (PPC)"
+	depends on PPC64 && CPU_LITTLE_ENDIAN && VSX
+	select CRYPTO_LIB_AES
+	select CRYPTO_ALGAPI
+	select CRYPTO_AEAD
+	select CRYPTO_SKCIPHER
+	select CRYPTO_SIMD
+	help
+	  AEAD cipher: AES cipher algorithms (FIPS-197)
+	  GCM (Galois/Counter Mode) authenticated encryption mode (NIST SP800-38D)
+	  Architecture: powerpc64 using:
+	    - little-endian
+	    - Power10 or later features
+
+	  Support for cryptographic acceleration instructions on Power10 or
+	  later CPU. This module supports stitched acceleration for AES/GCM.
+
+config CRYPTO_DEV_VMX
+        bool "Support for VMX cryptographic acceleration instructions"
+        depends on PPC64 && VSX
+        help
+          Support for VMX cryptographic acceleration instructions.
+
+config CRYPTO_DEV_VMX_ENCRYPT
+	tristate "Encryption acceleration support on P8 CPU"
+	depends on CRYPTO_DEV_VMX
+	select CRYPTO_AES
+	select CRYPTO_CBC
+	select CRYPTO_CTR
+	select CRYPTO_XTS
+	default m
+	help
+	  Support for VMX cryptographic acceleration instructions on Power8 CPU.
+	  This module supports acceleration for AES in hardware. If you choose
+	  'M' here, this module will be called vmx-crypto.
+
+endmenu
diff --git a/arch/powerpc/crypto/Makefile b/arch/powerpc/crypto/Makefile
new file mode 100644
index 000000000..a1fe102a9
--- /dev/null
+++ b/arch/powerpc/crypto/Makefile
@@ -0,0 +1,35 @@
+# SPDX-License-Identifier: GPL-2.0
+#
+# powerpc/crypto/Makefile
+#
+# Arch-specific CryptoAPI modules.
+#
+
+obj-$(CONFIG_CRYPTO_AES_PPC_SPE) += aes-ppc-spe.o
+obj-$(CONFIG_CRYPTO_AES_GCM_P10) += aes-gcm-p10-crypto.o
+obj-$(CONFIG_CRYPTO_DEV_VMX_ENCRYPT) += vmx-crypto.o
+
+aes-ppc-spe-y := aes-spe-glue.o
+aes-gcm-p10-crypto-y := aes-gcm-p10-glue.o aes-gcm-p10.o ghashp10-ppc.o aesp10-ppc.o
+vmx-crypto-objs := vmx.o aes_cbc.o aes_ctr.o aes_xts.o
+
+ifeq ($(CONFIG_CPU_LITTLE_ENDIAN),y)
+override flavour := linux-ppc64le
+else
+ifdef CONFIG_PPC64_ELF_ABI_V2
+override flavour := linux-ppc64-elfv2
+else
+override flavour := linux-ppc64
+endif
+endif
+
+quiet_cmd_perl = PERL    $@
+      cmd_perl = $(PERL) $< $(flavour) > $@
+
+targets += aesp10-ppc.S ghashp10-ppc.S
+
+$(obj)/aesp10-ppc.S $(obj)/ghashp10-ppc.S: $(obj)/%.S: $(src)/%.pl FORCE
+	$(call if_changed,perl)
+
+OBJECT_FILES_NON_STANDARD_aesp10-ppc.o := y
+OBJECT_FILES_NON_STANDARD_ghashp10-ppc.o := y
diff --git a/arch/powerpc/crypto/aes-gcm-p10-glue.c b/arch/powerpc/crypto/aes-gcm-p10-glue.c
new file mode 100644
index 000000000..f3417436d
--- /dev/null
+++ b/arch/powerpc/crypto/aes-gcm-p10-glue.c
@@ -0,0 +1,433 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Glue code for accelerated AES-GCM stitched implementation for ppc64le.
+ *
+ * Copyright 2022- IBM Inc. All rights reserved
+ */
+
+#include <linux/unaligned.h>
+#include <asm/simd.h>
+#include <asm/switch_to.h>
+#include <crypto/gcm.h>
+#include <crypto/aes.h>
+#include <crypto/algapi.h>
+#include <crypto/b128ops.h>
+#include <crypto/gf128mul.h>
+#include <crypto/internal/simd.h>
+#include <crypto/internal/aead.h>
+#include <crypto/internal/hash.h>
+#include <crypto/internal/skcipher.h>
+#include <crypto/scatterwalk.h>
+#include <linux/cpufeature.h>
+#include <linux/crypto.h>
+#include <linux/module.h>
+#include <linux/types.h>
+
+#define	PPC_ALIGN		16
+#define GCM_IV_SIZE		12
+#define RFC4106_NONCE_SIZE	4
+
+MODULE_DESCRIPTION("PPC64le AES-GCM with Stitched implementation");
+MODULE_AUTHOR("Danny Tsen <dtsen@linux.ibm.com");
+MODULE_LICENSE("GPL v2");
+MODULE_ALIAS_CRYPTO("aes");
+
+asmlinkage int aes_p10_set_encrypt_key(const u8 *userKey, const int bits,
+				       void *key);
+asmlinkage void aes_p10_encrypt(const u8 *in, u8 *out, const void *key);
+asmlinkage void aes_p10_gcm_encrypt(const u8 *in, u8 *out, size_t len,
+				    void *rkey, u8 *iv, void *Xi);
+asmlinkage void aes_p10_gcm_decrypt(const u8 *in, u8 *out, size_t len,
+				    void *rkey, u8 *iv, void *Xi);
+asmlinkage void gcm_init_htable(unsigned char htable[], unsigned char Xi[]);
+asmlinkage void gcm_ghash_p10(unsigned char *Xi, unsigned char *Htable,
+			      unsigned char *aad, unsigned int alen);
+asmlinkage void gcm_update(u8 *iv, void *Xi);
+
+struct p10_aes_key {
+	u8 key[AES_MAX_KEYLENGTH];
+	u64 rounds;
+};
+
+struct gcm_ctx {
+	u8 iv[16];
+	u8 ivtag[16];
+	u8 aad_hash[16];
+	u64 aadLen;
+	u64 Plen;	/* offset 56 - used in aes_p10_gcm_{en/de}crypt */
+	u8 pblock[16];
+};
+struct Hash_ctx {
+	u8 H[16];	/* subkey */
+	u8 Htable[256];	/* Xi, Hash table(offset 32) */
+};
+
+struct p10_aes_gcm_ctx {
+	struct p10_aes_key enc_key;
+	u8 nonce[RFC4106_NONCE_SIZE];
+};
+
+static void vsx_begin(void)
+{
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+}
+
+static void vsx_end(void)
+{
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+}
+
+static void set_subkey(unsigned char *hash)
+{
+	*(u64 *)&hash[0] = be64_to_cpup((__be64 *)&hash[0]);
+	*(u64 *)&hash[8] = be64_to_cpup((__be64 *)&hash[8]);
+}
+
+/*
+ * Compute aad if any.
+ *   - Hash aad and copy to Xi.
+ */
+static void set_aad(struct gcm_ctx *gctx, struct Hash_ctx *hash,
+		    unsigned char *aad, int alen)
+{
+	int i;
+	u8 nXi[16] = {0, };
+
+	gctx->aadLen = alen;
+	i = alen & ~0xf;
+	if (i) {
+		gcm_ghash_p10(nXi, hash->Htable+32, aad, i);
+		aad += i;
+		alen -= i;
+	}
+	if (alen) {
+		for (i = 0; i < alen; i++)
+			nXi[i] ^= aad[i];
+
+		memset(gctx->aad_hash, 0, 16);
+		gcm_ghash_p10(gctx->aad_hash, hash->Htable+32, nXi, 16);
+	} else {
+		memcpy(gctx->aad_hash, nXi, 16);
+	}
+
+	memcpy(hash->Htable, gctx->aad_hash, 16);
+}
+
+static void gcmp10_init(struct gcm_ctx *gctx, u8 *iv, unsigned char *rdkey,
+			struct Hash_ctx *hash, u8 *assoc, unsigned int assoclen)
+{
+	__be32 counter = cpu_to_be32(1);
+
+	aes_p10_encrypt(hash->H, hash->H, rdkey);
+	set_subkey(hash->H);
+	gcm_init_htable(hash->Htable+32, hash->H);
+
+	*((__be32 *)(iv+12)) = counter;
+
+	gctx->Plen = 0;
+
+	/*
+	 * Encrypt counter vector as iv tag and increment counter.
+	 */
+	aes_p10_encrypt(iv, gctx->ivtag, rdkey);
+
+	counter = cpu_to_be32(2);
+	*((__be32 *)(iv+12)) = counter;
+	memcpy(gctx->iv, iv, 16);
+
+	gctx->aadLen = assoclen;
+	memset(gctx->aad_hash, 0, 16);
+	if (assoclen)
+		set_aad(gctx, hash, assoc, assoclen);
+}
+
+static void finish_tag(struct gcm_ctx *gctx, struct Hash_ctx *hash, int len)
+{
+	int i;
+	unsigned char len_ac[16 + PPC_ALIGN];
+	unsigned char *aclen = PTR_ALIGN((void *)len_ac, PPC_ALIGN);
+	__be64 clen = cpu_to_be64(len << 3);
+	__be64 alen = cpu_to_be64(gctx->aadLen << 3);
+
+	if (len == 0 && gctx->aadLen == 0) {
+		memcpy(hash->Htable, gctx->ivtag, 16);
+		return;
+	}
+
+	/*
+	 * Len is in bits.
+	 */
+	*((__be64 *)(aclen)) = alen;
+	*((__be64 *)(aclen+8)) = clen;
+
+	/*
+	 * hash (AAD len and len)
+	 */
+	gcm_ghash_p10(hash->Htable, hash->Htable+32, aclen, 16);
+
+	for (i = 0; i < 16; i++)
+		hash->Htable[i] ^= gctx->ivtag[i];
+}
+
+static int set_authsize(struct crypto_aead *tfm, unsigned int authsize)
+{
+	switch (authsize) {
+	case 4:
+	case 8:
+	case 12:
+	case 13:
+	case 14:
+	case 15:
+	case 16:
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static int p10_aes_gcm_setkey(struct crypto_aead *aead, const u8 *key,
+			      unsigned int keylen)
+{
+	struct crypto_tfm *tfm = crypto_aead_tfm(aead);
+	struct p10_aes_gcm_ctx *ctx = crypto_tfm_ctx(tfm);
+	int ret;
+
+	vsx_begin();
+	ret = aes_p10_set_encrypt_key(key, keylen * 8, &ctx->enc_key);
+	vsx_end();
+
+	return ret ? -EINVAL : 0;
+}
+
+static int p10_aes_gcm_crypt(struct aead_request *req, u8 *riv,
+			     int assoclen, int enc)
+{
+	struct crypto_tfm *tfm = req->base.tfm;
+	struct p10_aes_gcm_ctx *ctx = crypto_tfm_ctx(tfm);
+	u8 databuf[sizeof(struct gcm_ctx) + PPC_ALIGN];
+	struct gcm_ctx *gctx = PTR_ALIGN((void *)databuf, PPC_ALIGN);
+	u8 hashbuf[sizeof(struct Hash_ctx) + PPC_ALIGN];
+	struct Hash_ctx *hash = PTR_ALIGN((void *)hashbuf, PPC_ALIGN);
+	struct skcipher_walk walk;
+	u8 *assocmem = NULL;
+	u8 *assoc;
+	unsigned int cryptlen = req->cryptlen;
+	unsigned char ivbuf[AES_BLOCK_SIZE+PPC_ALIGN];
+	unsigned char *iv = PTR_ALIGN((void *)ivbuf, PPC_ALIGN);
+	int ret;
+	unsigned long auth_tag_len = crypto_aead_authsize(__crypto_aead_cast(tfm));
+	u8 otag[16];
+	int total_processed = 0;
+	int nbytes;
+
+	memset(databuf, 0, sizeof(databuf));
+	memset(hashbuf, 0, sizeof(hashbuf));
+	memset(ivbuf, 0, sizeof(ivbuf));
+	memcpy(iv, riv, GCM_IV_SIZE);
+
+	/* Linearize assoc, if not already linear */
+	if (req->src->length >= assoclen && req->src->length) {
+		assoc = sg_virt(req->src); /* ppc64 is !HIGHMEM */
+	} else {
+		gfp_t flags = (req->base.flags & CRYPTO_TFM_REQ_MAY_SLEEP) ?
+			      GFP_KERNEL : GFP_ATOMIC;
+
+		/* assoc can be any length, so must be on heap */
+		assocmem = kmalloc(assoclen, flags);
+		if (unlikely(!assocmem))
+			return -ENOMEM;
+		assoc = assocmem;
+
+		scatterwalk_map_and_copy(assoc, req->src, 0, assoclen, 0);
+	}
+
+	vsx_begin();
+	gcmp10_init(gctx, iv, (unsigned char *) &ctx->enc_key, hash, assoc, assoclen);
+	vsx_end();
+
+	kfree(assocmem);
+
+	if (enc)
+		ret = skcipher_walk_aead_encrypt(&walk, req, false);
+	else
+		ret = skcipher_walk_aead_decrypt(&walk, req, false);
+	if (ret)
+		return ret;
+
+	while ((nbytes = walk.nbytes) > 0 && ret == 0) {
+		const u8 *src = walk.src.virt.addr;
+		u8 *dst = walk.dst.virt.addr;
+		u8 buf[AES_BLOCK_SIZE];
+
+		if (unlikely(nbytes > 0 && nbytes < AES_BLOCK_SIZE))
+			src = dst = memcpy(buf, src, nbytes);
+
+		vsx_begin();
+		if (enc)
+			aes_p10_gcm_encrypt(src, dst, nbytes,
+					    &ctx->enc_key, gctx->iv, hash->Htable);
+		else
+			aes_p10_gcm_decrypt(src, dst, nbytes,
+					    &ctx->enc_key, gctx->iv, hash->Htable);
+
+		if (unlikely(nbytes > 0 && nbytes < AES_BLOCK_SIZE))
+			memcpy(walk.dst.virt.addr, buf, nbytes);
+
+		vsx_end();
+
+		total_processed += walk.nbytes;
+		ret = skcipher_walk_done(&walk, 0);
+	}
+
+	if (ret)
+		return ret;
+
+	/* Finalize hash */
+	vsx_begin();
+	gcm_update(gctx->iv, hash->Htable);
+	finish_tag(gctx, hash, total_processed);
+	vsx_end();
+
+	/* copy Xi to end of dst */
+	if (enc)
+		scatterwalk_map_and_copy(hash->Htable, req->dst, req->assoclen + cryptlen,
+					 auth_tag_len, 1);
+	else {
+		scatterwalk_map_and_copy(otag, req->src,
+					 req->assoclen + cryptlen - auth_tag_len,
+					 auth_tag_len, 0);
+
+		if (crypto_memneq(otag, hash->Htable, auth_tag_len)) {
+			memzero_explicit(hash->Htable, 16);
+			return -EBADMSG;
+		}
+	}
+
+	return 0;
+}
+
+static int rfc4106_setkey(struct crypto_aead *tfm, const u8 *inkey,
+			  unsigned int keylen)
+{
+	struct p10_aes_gcm_ctx *ctx = crypto_aead_ctx(tfm);
+	int err;
+
+	keylen -= RFC4106_NONCE_SIZE;
+	err = p10_aes_gcm_setkey(tfm, inkey, keylen);
+	if (err)
+		return err;
+
+	memcpy(ctx->nonce, inkey + keylen, RFC4106_NONCE_SIZE);
+	return 0;
+}
+
+static int rfc4106_setauthsize(struct crypto_aead *tfm, unsigned int authsize)
+{
+	return crypto_rfc4106_check_authsize(authsize);
+}
+
+static int rfc4106_encrypt(struct aead_request *req)
+{
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct p10_aes_gcm_ctx *ctx = crypto_aead_ctx(aead);
+	u8 iv[AES_BLOCK_SIZE];
+
+	memcpy(iv, ctx->nonce, RFC4106_NONCE_SIZE);
+	memcpy(iv + RFC4106_NONCE_SIZE, req->iv, GCM_RFC4106_IV_SIZE);
+
+	return crypto_ipsec_check_assoclen(req->assoclen) ?:
+	       p10_aes_gcm_crypt(req, iv, req->assoclen - GCM_RFC4106_IV_SIZE, 1);
+}
+
+static int rfc4106_decrypt(struct aead_request *req)
+{
+	struct crypto_aead *aead = crypto_aead_reqtfm(req);
+	struct p10_aes_gcm_ctx *ctx = crypto_aead_ctx(aead);
+	u8 iv[AES_BLOCK_SIZE];
+
+	memcpy(iv, ctx->nonce, RFC4106_NONCE_SIZE);
+	memcpy(iv + RFC4106_NONCE_SIZE, req->iv, GCM_RFC4106_IV_SIZE);
+
+	return crypto_ipsec_check_assoclen(req->assoclen) ?:
+	       p10_aes_gcm_crypt(req, iv, req->assoclen - GCM_RFC4106_IV_SIZE, 0);
+}
+
+static int p10_aes_gcm_encrypt(struct aead_request *req)
+{
+	return p10_aes_gcm_crypt(req, req->iv, req->assoclen, 1);
+}
+
+static int p10_aes_gcm_decrypt(struct aead_request *req)
+{
+	return p10_aes_gcm_crypt(req, req->iv, req->assoclen, 0);
+}
+
+static struct aead_alg gcm_aes_algs[] = {{
+	.ivsize			= GCM_IV_SIZE,
+	.maxauthsize		= 16,
+
+	.setauthsize		= set_authsize,
+	.setkey			= p10_aes_gcm_setkey,
+	.encrypt		= p10_aes_gcm_encrypt,
+	.decrypt		= p10_aes_gcm_decrypt,
+
+	.base.cra_name		= "__gcm(aes)",
+	.base.cra_driver_name	= "__aes_gcm_p10",
+	.base.cra_priority	= 2100,
+	.base.cra_blocksize	= 1,
+	.base.cra_ctxsize	= sizeof(struct p10_aes_gcm_ctx)+
+				  4 * sizeof(u64[2]),
+	.base.cra_module	= THIS_MODULE,
+	.base.cra_flags		= CRYPTO_ALG_INTERNAL,
+}, {
+	.ivsize			= GCM_RFC4106_IV_SIZE,
+	.maxauthsize		= 16,
+	.setkey			= rfc4106_setkey,
+	.setauthsize		= rfc4106_setauthsize,
+	.encrypt		= rfc4106_encrypt,
+	.decrypt		= rfc4106_decrypt,
+
+	.base.cra_name		= "__rfc4106(gcm(aes))",
+	.base.cra_driver_name	= "__rfc4106_aes_gcm_p10",
+	.base.cra_priority	= 2100,
+	.base.cra_blocksize	= 1,
+	.base.cra_ctxsize	= sizeof(struct p10_aes_gcm_ctx) +
+				  4 * sizeof(u64[2]),
+	.base.cra_module	= THIS_MODULE,
+	.base.cra_flags		= CRYPTO_ALG_INTERNAL,
+}};
+
+static struct simd_aead_alg *p10_simd_aeads[ARRAY_SIZE(gcm_aes_algs)];
+
+static int __init p10_init(void)
+{
+	int ret;
+
+	if (!cpu_has_feature(CPU_FTR_ARCH_31))
+		return 0;
+
+	ret = simd_register_aeads_compat(gcm_aes_algs,
+					 ARRAY_SIZE(gcm_aes_algs),
+					 p10_simd_aeads);
+	if (ret) {
+		simd_unregister_aeads(gcm_aes_algs, ARRAY_SIZE(gcm_aes_algs),
+				      p10_simd_aeads);
+		return ret;
+	}
+	return 0;
+}
+
+static void __exit p10_exit(void)
+{
+	simd_unregister_aeads(gcm_aes_algs, ARRAY_SIZE(gcm_aes_algs),
+			      p10_simd_aeads);
+}
+
+module_init(p10_init);
+module_exit(p10_exit);
diff --git a/arch/powerpc/crypto/aes-gcm-p10.S b/arch/powerpc/crypto/aes-gcm-p10.S
new file mode 100644
index 000000000..89f50eef3
--- /dev/null
+++ b/arch/powerpc/crypto/aes-gcm-p10.S
@@ -0,0 +1,1236 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+#
+# Accelerated AES-GCM stitched implementation for ppc64le.
+#
+# Copyright 2024- IBM Inc.
+#
+#===================================================================================
+# Written by Danny Tsen <dtsen@us.ibm.com>
+#
+# GHASH is based on the Karatsuba multiplication method.
+#
+#    Xi xor X1
+#
+#    X1 * H^4 + X2 * H^3 + x3 * H^2 + X4 * H =
+#      (X1.h * H4.h + xX.l * H4.l + X1 * H4) +
+#      (X2.h * H3.h + X2.l * H3.l + X2 * H3) +
+#      (X3.h * H2.h + X3.l * H2.l + X3 * H2) +
+#      (X4.h * H.h + X4.l * H.l + X4 * H)
+#
+# Xi = v0
+# H Poly = v2
+# Hash keys = v3 - v14
+#     ( H.l, H, H.h)
+#     ( H^2.l, H^2, H^2.h)
+#     ( H^3.l, H^3, H^3.h)
+#     ( H^4.l, H^4, H^4.h)
+#
+# v30 is IV
+# v31 - counter 1
+#
+# AES used,
+#     vs0 - round key 0
+#     v15, v16, v17, v18, v19, v20, v21, v22 for 8 blocks (encrypted)
+#
+# This implementation uses stitched AES-GCM approach to improve overall performance.
+# AES is implemented with 8x blocks and GHASH is using 2 4x blocks.
+#
+# ===================================================================================
+#
+
+#include <asm/ppc_asm.h>
+#include <linux/linkage.h>
+
+.machine        "any"
+.text
+
+.macro	SAVE_GPR GPR OFFSET FRAME
+	std	\GPR,\OFFSET(\FRAME)
+.endm
+
+.macro	SAVE_VRS VRS OFFSET FRAME
+	stxv	\VRS+32, \OFFSET(\FRAME)
+.endm
+
+.macro	RESTORE_GPR GPR OFFSET FRAME
+	ld	\GPR,\OFFSET(\FRAME)
+.endm
+
+.macro	RESTORE_VRS VRS OFFSET FRAME
+	lxv	\VRS+32, \OFFSET(\FRAME)
+.endm
+
+.macro SAVE_REGS
+	mflr 0
+	std 0, 16(1)
+	stdu 1,-512(1)
+
+	SAVE_GPR 14, 112, 1
+	SAVE_GPR 15, 120, 1
+	SAVE_GPR 16, 128, 1
+	SAVE_GPR 17, 136, 1
+	SAVE_GPR 18, 144, 1
+	SAVE_GPR 19, 152, 1
+	SAVE_GPR 20, 160, 1
+	SAVE_GPR 21, 168, 1
+	SAVE_GPR 22, 176, 1
+	SAVE_GPR 23, 184, 1
+	SAVE_GPR 24, 192, 1
+
+	addi	9, 1, 256
+	SAVE_VRS 20, 0, 9
+	SAVE_VRS 21, 16, 9
+	SAVE_VRS 22, 32, 9
+	SAVE_VRS 23, 48, 9
+	SAVE_VRS 24, 64, 9
+	SAVE_VRS 25, 80, 9
+	SAVE_VRS 26, 96, 9
+	SAVE_VRS 27, 112, 9
+	SAVE_VRS 28, 128, 9
+	SAVE_VRS 29, 144, 9
+	SAVE_VRS 30, 160, 9
+	SAVE_VRS 31, 176, 9
+.endm # SAVE_REGS
+
+.macro RESTORE_REGS
+	addi	9, 1, 256
+	RESTORE_VRS 20, 0, 9
+	RESTORE_VRS 21, 16, 9
+	RESTORE_VRS 22, 32, 9
+	RESTORE_VRS 23, 48, 9
+	RESTORE_VRS 24, 64, 9
+	RESTORE_VRS 25, 80, 9
+	RESTORE_VRS 26, 96, 9
+	RESTORE_VRS 27, 112, 9
+	RESTORE_VRS 28, 128, 9
+	RESTORE_VRS 29, 144, 9
+	RESTORE_VRS 30, 160, 9
+	RESTORE_VRS 31, 176, 9
+
+	RESTORE_GPR 14, 112, 1
+	RESTORE_GPR 15, 120, 1
+	RESTORE_GPR 16, 128, 1
+	RESTORE_GPR 17, 136, 1
+	RESTORE_GPR 18, 144, 1
+	RESTORE_GPR 19, 152, 1
+	RESTORE_GPR 20, 160, 1
+	RESTORE_GPR 21, 168, 1
+	RESTORE_GPR 22, 176, 1
+	RESTORE_GPR 23, 184, 1
+	RESTORE_GPR 24, 192, 1
+
+	addi    1, 1, 512
+	ld 0, 16(1)
+	mtlr 0
+.endm # RESTORE_REGS
+
+# 4x loops
+.macro AES_CIPHER_4x _VCIPHER ST r
+	\_VCIPHER	\ST, \ST, \r
+	\_VCIPHER	\ST+1, \ST+1, \r
+	\_VCIPHER	\ST+2, \ST+2, \r
+	\_VCIPHER	\ST+3, \ST+3, \r
+.endm
+
+# 8x loops
+.macro AES_CIPHER_8x _VCIPHER ST r
+	\_VCIPHER	\ST, \ST, \r
+	\_VCIPHER	\ST+1, \ST+1, \r
+	\_VCIPHER	\ST+2, \ST+2, \r
+	\_VCIPHER	\ST+3, \ST+3, \r
+	\_VCIPHER	\ST+4, \ST+4, \r
+	\_VCIPHER	\ST+5, \ST+5, \r
+	\_VCIPHER	\ST+6, \ST+6, \r
+	\_VCIPHER	\ST+7, \ST+7, \r
+.endm
+
+.macro LOOP_8AES_STATE
+	xxlor	32+23, 1, 1
+	xxlor	32+24, 2, 2
+	xxlor	32+25, 3, 3
+	xxlor	32+26, 4, 4
+	AES_CIPHER_8x vcipher, 15, 23
+	AES_CIPHER_8x vcipher, 15, 24
+	AES_CIPHER_8x vcipher, 15, 25
+	AES_CIPHER_8x vcipher, 15, 26
+	xxlor	32+23, 5, 5
+	xxlor	32+24, 6, 6
+	xxlor	32+25, 7, 7
+	xxlor	32+26, 8, 8
+	AES_CIPHER_8x vcipher, 15, 23
+	AES_CIPHER_8x vcipher, 15, 24
+	AES_CIPHER_8x vcipher, 15, 25
+	AES_CIPHER_8x vcipher, 15, 26
+.endm
+
+#
+# PPC_GHASH4x(H, S1, S2, S3, S4): Compute 4x hash values based on Karatsuba method.
+# H: returning digest
+# S#: states
+#
+# S1 should xor with the previous digest
+#
+# Xi = v0
+# H Poly = v2
+# Hash keys = v3 - v14
+# Scratch: v23 - v29
+#
+.macro PPC_GHASH4x H S1 S2 S3 S4
+
+	vpmsumd	23, 12, \S1		# H4.L * X.L
+	vpmsumd	24, 9, \S2
+	vpmsumd	25, 6, \S3
+	vpmsumd	26, 3, \S4
+
+	vpmsumd	27, 13, \S1		# H4.L * X.H + H4.H * X.L
+	vpmsumd	28, 10, \S2		# H3.L * X1.H + H3.H * X1.L
+
+	vxor	23, 23, 24
+	vxor	23, 23, 25
+	vxor	23, 23, 26		# L
+
+	vxor	24, 27, 28
+	vpmsumd	25, 7, \S3
+	vpmsumd	26, 4, \S4
+
+	vxor	24, 24, 25
+	vxor	24, 24, 26		# M
+
+	# sum hash and reduction with H Poly
+	vpmsumd	28, 23, 2		# reduction
+
+	vxor	1, 1, 1
+	vsldoi	25, 24, 1, 8		# mL
+	vsldoi	1, 1, 24, 8		# mH
+	vxor	23, 23, 25		# mL + L
+
+	# This performs swap and xor like,
+	#   vsldoi	23, 23, 23, 8		# swap
+	#   vxor	23, 23, 28
+	xxlor	32+25, 10, 10
+	vpermxor 23, 23, 28, 25
+
+	vpmsumd	26, 14, \S1		# H4.H * X.H
+	vpmsumd	27, 11, \S2
+	vpmsumd	28, 8, \S3
+	vpmsumd	29, 5, \S4
+
+	vxor	24, 26, 27
+	vxor	24, 24, 28
+	vxor	24, 24, 29
+
+	vxor	24, 24, 1
+
+	# sum hash and reduction with H Poly
+	vsldoi	25, 23, 23, 8		# swap
+	vpmsumd	23, 23, 2
+	vxor	27, 25, 24
+	vxor	\H, 23, 27
+.endm
+
+#
+# Compute update single ghash
+# scratch: v1, v22..v27
+#
+.macro PPC_GHASH1x H S1
+
+	vxor	1, 1, 1
+
+	vpmsumd	22, 3, \S1		# L
+	vpmsumd	23, 4, \S1		# M
+	vpmsumd	24, 5, \S1		# H
+
+	vpmsumd	27, 22, 2		# reduction
+
+	vsldoi	25, 23, 1, 8		# mL
+	vsldoi	26, 1, 23, 8		# mH
+	vxor	22, 22, 25		# LL + LL
+	vxor	24, 24, 26		# HH + HH
+
+	xxlor	32+25, 10, 10
+	vpermxor 22, 22, 27, 25
+
+	vsldoi	23, 22, 22, 8		# swap
+	vpmsumd	22, 22, 2		# reduction
+	vxor	23, 23, 24
+	vxor	\H, 22, 23
+.endm
+
+#
+# LOAD_HASH_TABLE
+# Xi = v0
+# H Poly = v2
+# Hash keys = v3 - v14
+#
+.macro LOAD_HASH_TABLE
+	# Load Xi
+	lxvb16x	32, 0, 8	# load Xi
+
+	# load Hash - h^4, h^3, h^2, h
+	li	10, 32
+	lxvd2x	2+32, 10, 8	# H Poli
+	li	10, 48
+	lxvd2x	3+32, 10, 8	# Hl
+	li	10, 64
+	lxvd2x	4+32, 10, 8	# H
+	li	10, 80
+	lxvd2x	5+32, 10, 8	# Hh
+
+	li	10, 96
+	lxvd2x	6+32, 10, 8	# H^2l
+	li	10, 112
+	lxvd2x	7+32, 10, 8	# H^2
+	li	10, 128
+	lxvd2x	8+32, 10, 8	# H^2h
+
+	li	10, 144
+	lxvd2x	9+32, 10, 8	# H^3l
+	li	10, 160
+	lxvd2x	10+32, 10, 8	# H^3
+	li	10, 176
+	lxvd2x	11+32, 10, 8	# H^3h
+
+	li	10, 192
+	lxvd2x	12+32, 10, 8	# H^4l
+	li	10, 208
+	lxvd2x	13+32, 10, 8	# H^4
+	li	10, 224
+	lxvd2x	14+32, 10, 8	# H^4h
+.endm
+
+################################################################################
+# Compute AES and ghash one block at a time.
+# r23: AES rounds
+# v30: current IV
+# vs0: roundkey 0
+#
+################################################################################
+SYM_FUNC_START_LOCAL(aes_gcm_crypt_1x)
+
+	cmpdi	5, 16
+	bge	__More_1x
+	blr
+__More_1x:
+	li      10, 16
+	divdu   12, 5, 10
+
+	xxlxor	32+15, 32+30, 0
+
+	# Pre-load 8 AES rounds to scratch vectors.
+	xxlor	32+16, 1, 1
+	xxlor	32+17, 2, 2
+	xxlor	32+18, 3, 3
+	xxlor	32+19, 4, 4
+	xxlor	32+20, 5, 5
+	xxlor	32+21, 6, 6
+	xxlor	32+28, 7, 7
+	xxlor	32+29, 8, 8
+	lwz	23, 240(6)	# n rounds
+	addi	22, 23, -9	# remaing AES rounds
+
+	cmpdi	12, 0
+	bgt	__Loop_1x
+	blr
+
+__Loop_1x:
+	mtctr	22
+	addi	10, 6, 144
+	vcipher	15, 15, 16
+	vcipher	15, 15, 17
+	vcipher	15, 15, 18
+	vcipher	15, 15, 19
+	vcipher	15, 15, 20
+	vcipher	15, 15, 21
+	vcipher	15, 15, 28
+	vcipher	15, 15, 29
+
+__Loop_aes_1state:
+	lxv	32+1, 0(10)
+	vcipher	15, 15, 1
+	addi	10, 10, 16
+	bdnz	__Loop_aes_1state
+	lxv	32+1, 0(10)		# last round key
+	lxvb16x 11, 0, 14		# load input block
+	vcipherlast 15, 15, 1
+
+	xxlxor	32+15, 32+15, 11
+	stxvb16x 32+15, 0, 9	# store output
+	addi	14, 14, 16
+	addi	9, 9, 16
+
+	cmpdi	24, 0	# decrypt?
+	bne	__Encrypt_1x
+	xxlor	15+32, 11, 11
+__Encrypt_1x:
+	vxor	15, 15, 0
+	PPC_GHASH1x 0, 15
+
+	addi	5, 5, -16
+	addi	11, 11, 16
+
+	vadduwm 30, 30, 31		# IV + counter
+	xxlxor	32+15, 32+30, 0
+	addi	12, 12, -1
+	cmpdi	12, 0
+	bgt	__Loop_1x
+
+	stxvb16x 32+30, 0, 7		# update IV
+	stxvb16x 32+0, 0, 8		# update Xi
+	blr
+SYM_FUNC_END(aes_gcm_crypt_1x)
+
+################################################################################
+# Process a normal partial block when we come here.
+#  Compute partial mask, Load and store partial block to stack.
+#  Update partial_len and pblock.
+#  pblock is (encrypted ^ AES state) for encrypt
+#        and (input ^ AES state) for decrypt.
+#
+################################################################################
+SYM_FUNC_START_LOCAL(__Process_partial)
+
+	# create partial mask
+	vspltisb 16, -1
+	li	12, 16
+	sub	12, 12, 5
+	sldi	12, 12, 3
+	mtvsrdd	32+17, 0, 12
+	vslo	16, 16, 17		# partial block mask
+
+	lxvb16x 11, 0, 14		# load partial block
+	xxland	11, 11, 32+16
+
+	# AES crypt partial
+	xxlxor	32+15, 32+30, 0
+	lwz	23, 240(6)		# n rounds
+	addi	22, 23, -1		# loop - 1
+	mtctr	22
+	addi	10, 6, 16
+
+__Loop_aes_pstate:
+	lxv	32+1, 0(10)
+	vcipher	15, 15, 1
+	addi	10, 10, 16
+	bdnz	__Loop_aes_pstate
+	lxv	32+1, 0(10)		# last round key
+	vcipherlast 15, 15, 1
+
+	xxlxor	32+15, 32+15, 11
+	vand	15, 15, 16
+
+	# AES crypt output v15
+	# Write partial
+	li	10, 224
+	stxvb16x 15+32, 10, 1		# write v15 to stack
+	addi	10, 1, 223
+	addi	12, 9, -1
+        mtctr	5			# partial block len
+__Write_partial:
+        lbzu	22, 1(10)
+	stbu	22, 1(12)
+        bdnz	__Write_partial
+
+	cmpdi	24, 0			# decrypt?
+	bne	__Encrypt_partial
+	xxlor	32+15, 11, 11		# decrypt using the input block
+__Encrypt_partial:
+	#vxor	15, 15, 0		# ^ previous hash
+	#PPC_GHASH1x 0, 15
+
+	add	14, 14, 5
+	add	9, 9, 5
+	std	5, 56(7)		# update partial
+	sub	11, 11, 5
+	li	5, 0			# done last byte
+
+	#
+	# Don't increase IV since this is the last partial.
+	# It should get updated in gcm_update if no more data blocks.
+	#vadduwm	30, 30, 31		# increase IV
+	stxvb16x 32+30, 0, 7		# update IV
+	li	10, 64
+	stxvb16x 32+0, 0, 8		# Update X1
+	stxvb16x 32+15, 10, 7		# Update pblock
+	blr
+SYM_FUNC_END(__Process_partial)
+
+################################################################################
+# Combine partial blocks and ghash when we come here.
+#
+# The partial block has to be shifted to the right location to encrypt/decrypt
+# and compute ghash if combing the previous partial block is needed.
+# - Compute ghash for a full block. Clear Partial_len and pblock. Update IV.
+#   Write Xi.
+# - Don't compute ghash if not full block.  gcm_update will take care of it
+#   is the last block. Update Partial_len and pblock.
+#
+################################################################################
+SYM_FUNC_START_LOCAL(__Combine_partial)
+
+	ld	12, 56(7)
+	mr	21, 5			# these bytes to be processed
+
+	li	17, 0
+	li	16, 16
+	sub	22, 16, 12		# bytes to complete a block
+	sub	17, 22, 5		# remaining bytes in a block
+	cmpdi	5, 16
+	ble	__Inp_msg_less16
+	li	17, 0
+	mr	21, 22
+	b	__Combine_continue
+__Inp_msg_less16:
+	cmpd	22, 5
+	bgt	__Combine_continue
+	li	17, 0
+	mr	21, 22			# these bytes to be processed
+
+__Combine_continue:
+	# load msg and shift to the proper location and mask
+	vspltisb 16, -1
+	sldi	15, 12, 3
+	mtvsrdd	32+17, 0, 15
+	vslo	16, 16, 17
+	vsro	16, 16, 17
+	sldi	15, 17, 3
+	mtvsrdd	32+17, 0, 15
+	vsro	16, 16, 17
+	vslo	16, 16, 17		# mask
+
+	lxvb16x 32+19, 0, 14		# load partial block
+	sldi	15, 12, 3
+	mtvsrdd	32+17, 0, 15
+	vsro	19, 19, 17		# 0x00..xxxx??..??
+	sldi	15, 17, 3
+	mtvsrdd	32+17, 0, 15
+	vsro	19, 19, 17		# 0x00..xxxx
+	vslo	19, 19, 17		# shift back to form 0x00..xxxx00..00
+
+	# AES crypt partial
+	xxlxor	32+15, 32+30, 0
+	lwz	23, 240(6)	# n rounds
+	addi	22, 23, -1	# loop - 1
+	mtctr	22
+	addi	10, 6, 16
+
+__Loop_aes_cpstate:
+	lxv	32+1, 0(10)
+	vcipher	15, 15, 1
+	addi	10, 10, 16
+	bdnz	__Loop_aes_cpstate
+	lxv	32+1, 0(10)		# last round key
+	vcipherlast 15, 15, 1
+
+	vxor	15, 15, 19
+	vand	15, 15, 16
+
+	# AES crypt output v15
+	# Write partial
+	li	10, 224
+	stxvb16x 15+32, 10, 1		# write v15 to stack
+	addi	10, 1, 223
+	add	10, 10, 12		# add offset
+	addi	15, 9, -1
+        mtctr	21			# partial block len
+__Write_combine_partial:
+        lbzu	22, 1(10)
+	stbu	22, 1(15)
+        bdnz	__Write_combine_partial
+
+	add	14, 14, 21
+	add	11, 11, 21
+	add	9, 9, 21
+	sub	5, 5, 21
+
+	# Encrypt/Decrypt?
+	cmpdi	24, 0			# decrypt?
+	bne	__Encrypt_combine_partial
+	vmr	15, 19		# decrypt using the input block
+
+__Encrypt_combine_partial:
+	#
+	# Update partial flag and combine ghash.
+__Update_partial_ghash:
+	li	10, 64
+	lxvb16x 32+17, 10, 7		# load previous pblock
+	add	12, 12, 21		# combined pprocessed
+	vxor	15, 15, 17		# combined pblock
+
+	cmpdi	12, 16
+	beq	__Clear_partial_flag
+	std	12, 56(7)		# update partial len
+	stxvb16x 32+15, 10, 7		# Update current pblock
+	blr
+
+__Clear_partial_flag:
+	li	12, 0
+	std	12, 56(7)
+	# Update IV and ghash here
+	vadduwm	30, 30, 31		# increase IV
+	stxvb16x 32+30, 0, 7		# update IV
+
+	# v15 either is either (input blockor encrypted)^(AES state)
+	vxor	15, 15, 0
+	PPC_GHASH1x 0, 15
+	stxvb16x 32+0, 10, 7		# update pblock for debug?
+	stxvb16x 32+0, 0, 8		# update Xi
+	blr
+SYM_FUNC_END(__Combine_partial)
+
+################################################################################
+# gcm_update(iv, Xi) - compute last hash
+#
+################################################################################
+SYM_FUNC_START(gcm_update)
+
+	ld	10, 56(3)
+	cmpdi	10, 0
+	beq	__no_update
+
+	lxvb16x	32, 0, 4	# load Xi
+	# load Hash - h^4, h^3, h^2, h
+	li	10, 32
+	lxvd2x	2+32, 10, 4	# H Poli
+	li	10, 48
+	lxvd2x	3+32, 10, 4	# Hl
+	li	10, 64
+	lxvd2x	4+32, 10, 4	# H
+	li	10, 80
+	lxvd2x	5+32, 10, 4	# Hh
+
+	addis	11, 2, permx@toc@ha
+	addi	11, 11, permx@toc@l
+	lxv	10, 0(11)	# vs10: vpermxor vector
+
+	li	9, 64
+	lxvb16x 32+6, 9, 3		# load pblock
+	vxor	6, 6, 0
+
+	vxor	1, 1, 1
+	vpmsumd	12, 3, 6		# L
+	vpmsumd	13, 4, 6		# M
+	vpmsumd	14, 5, 6		# H
+	vpmsumd	17, 12, 2		# reduction
+	vsldoi	15, 13, 1, 8		# mL
+	vsldoi	16, 1, 13, 8		# mH
+	vxor	12, 12, 15		# LL + LL
+	vxor	14, 14, 16		# HH + HH
+	xxlor	32+15, 10, 10
+	vpermxor 12, 12, 17, 15
+	vsldoi	13, 12, 12, 8		# swap
+	vpmsumd	12, 12, 2		# reduction
+	vxor	13, 13, 14
+	vxor	7, 12, 13
+
+	#vxor	0, 0, 0
+	#stxvb16x 32+0, 9, 3
+	li	10, 0
+	std	10, 56(3)
+	stxvb16x 32+7, 0, 4
+
+__no_update:
+	blr
+SYM_FUNC_END(gcm_update)
+
+################################################################################
+# aes_p10_gcm_encrypt (const void *inp, void *out, size_t len,
+#               const char *rk, unsigned char iv[16], void *Xip);
+#
+#    r3 - inp
+#    r4 - out
+#    r5 - len
+#    r6 - AES round keys
+#    r7 - iv and other data
+#    r8 - Xi, HPoli, hash keys
+#
+#    rounds is at offset 240 in rk
+#    Xi is at 0 in gcm_table (Xip).
+#
+################################################################################
+SYM_FUNC_START(aes_p10_gcm_encrypt)
+
+	cmpdi	5, 0
+	ble	__Invalid_msg_len
+
+	SAVE_REGS
+	LOAD_HASH_TABLE
+
+	# initialize ICB: GHASH( IV ), IV - r7
+	lxvb16x	30+32, 0, 7	# load IV  - v30
+
+	mr	14, 3
+	mr	9, 4
+
+	# counter 1
+	vxor	31, 31, 31
+	vspltisb 22, 1
+	vsldoi	31, 31, 22,1	# counter 1
+
+	addis	11, 2, permx@toc@ha
+	addi	11, 11, permx@toc@l
+	lxv	10, 0(11)	# vs10: vpermxor vector
+	li	11, 0
+
+	# load 9 round keys to VSR
+	lxv	0, 0(6)			# round key 0
+	lxv	1, 16(6)		# round key 1
+	lxv	2, 32(6)		# round key 2
+	lxv	3, 48(6)		# round key 3
+	lxv	4, 64(6)		# round key 4
+	lxv	5, 80(6)		# round key 5
+	lxv	6, 96(6)		# round key 6
+	lxv	7, 112(6)		# round key 7
+	lxv	8, 128(6)		# round key 8
+
+	# load rounds - 10 (128), 12 (192), 14 (256)
+	lwz	23, 240(6)		# n rounds
+	li	24, 1			# encrypt
+
+__Process_encrypt:
+	#
+	# Process different blocks
+	#
+	ld	12, 56(7)
+	cmpdi	12, 0
+	bgt	__Do_combine_enc
+	cmpdi	5, 128
+	blt	__Process_more_enc
+
+#
+# Process 8x AES/GCM blocks
+#
+__Process_8x_enc:
+	# 8x blcoks
+	li	10, 128
+	divdu	12, 5, 10	# n 128 bytes-blocks
+
+	addi	12, 12, -1	# loop - 1
+
+	vmr	15, 30		# first state: IV
+	vadduwm	16, 15, 31	# state + counter
+	vadduwm	17, 16, 31
+	vadduwm	18, 17, 31
+	vadduwm	19, 18, 31
+	vadduwm	20, 19, 31
+	vadduwm	21, 20, 31
+	vadduwm	22, 21, 31
+	xxlor	9, 32+22, 32+22	# save last state
+
+	# vxor  state, state, w # addroundkey
+	xxlor	32+29, 0, 0
+        vxor    15, 15, 29      # IV + round key - add round key 0
+	vxor	16, 16, 29
+	vxor	17, 17, 29
+	vxor	18, 18, 29
+	vxor	19, 19, 29
+	vxor	20, 20, 29
+	vxor	21, 21, 29
+	vxor	22, 22, 29
+
+	li	15, 16
+	li	16, 32
+	li	17, 48
+	li	18, 64
+	li	19, 80
+	li	20, 96
+	li	21, 112
+
+	#
+	# Pre-compute first 8 AES state and leave 1/3/5 more rounds
+	# for the loop.
+	#
+	addi	22, 23, -9		# process 8 keys
+	mtctr	22			# AES key loop
+	addi	10, 6, 144
+
+	LOOP_8AES_STATE			# process 8 AES keys
+
+__PreLoop_aes_state:
+	lxv	32+1, 0(10)		# round key
+	AES_CIPHER_8x vcipher 15 1
+	addi	10, 10, 16
+	bdnz	__PreLoop_aes_state
+	lxv	32+1, 0(10)		# last round key (v1)
+
+	cmpdi	12, 0			# Only one loop (8 block)
+	beq	__Finish_ghash
+
+#
+# Loop 8x blocks and compute ghash
+#
+__Loop_8x_block_enc:
+	vcipherlast     15, 15, 1
+	vcipherlast     16, 16, 1
+	vcipherlast     17, 17, 1
+	vcipherlast     18, 18, 1
+	vcipherlast     19, 19, 1
+	vcipherlast     20, 20, 1
+	vcipherlast     21, 21, 1
+	vcipherlast     22, 22, 1
+
+	lxvb16x	32+23, 0, 14	# load block
+	lxvb16x	32+24, 15, 14	# load block
+	lxvb16x	32+25, 16, 14	# load block
+	lxvb16x	32+26, 17, 14	# load block
+	lxvb16x	32+27, 18, 14	# load block
+	lxvb16x	32+28, 19, 14	# load block
+	lxvb16x	32+29, 20, 14	# load block
+	lxvb16x	32+30, 21, 14	# load block
+	addi	14, 14, 128
+
+	vxor	15, 15, 23
+	vxor	16, 16, 24
+	vxor	17, 17, 25
+	vxor	18, 18, 26
+	vxor	19, 19, 27
+	vxor	20, 20, 28
+	vxor	21, 21, 29
+	vxor	22, 22, 30
+
+	stxvb16x 47, 0, 9	# store output
+	stxvb16x 48, 15, 9	# store output
+	stxvb16x 49, 16, 9	# store output
+	stxvb16x 50, 17, 9	# store output
+	stxvb16x 51, 18, 9	# store output
+	stxvb16x 52, 19, 9	# store output
+	stxvb16x 53, 20, 9	# store output
+	stxvb16x 54, 21, 9	# store output
+	addi	9, 9, 128
+
+	# ghash here
+	vxor	15, 15, 0
+	PPC_GHASH4x 0, 15, 16, 17, 18
+
+	vxor	19, 19, 0
+	PPC_GHASH4x 0, 19, 20, 21, 22
+
+	xxlor	32+15, 9, 9		# last state
+	vadduwm 15, 15, 31		# state + counter
+	vadduwm 16, 15, 31
+	vadduwm 17, 16, 31
+	vadduwm 18, 17, 31
+	vadduwm 19, 18, 31
+	vadduwm 20, 19, 31
+	vadduwm 21, 20, 31
+	vadduwm 22, 21, 31
+	xxlor	9, 32+22, 32+22		# save last state
+
+	xxlor	32+27, 0, 0		# restore roundkey 0
+        vxor    15, 15, 27		# IV + round key - add round key 0
+	vxor	16, 16, 27
+	vxor	17, 17, 27
+	vxor	18, 18, 27
+	vxor	19, 19, 27
+	vxor	20, 20, 27
+	vxor	21, 21, 27
+	vxor	22, 22, 27
+
+	addi    5, 5, -128
+	addi    11, 11, 128
+
+	LOOP_8AES_STATE			# process 8 AES keys
+	mtctr	22			# AES key loop
+	addi	10, 6, 144
+__LastLoop_aes_state:
+	lxv	32+1, 0(10)		# round key
+	AES_CIPHER_8x vcipher 15 1
+	addi	10, 10, 16
+	bdnz	__LastLoop_aes_state
+	lxv	32+1, 0(10)		# last round key (v1)
+
+	addi	12, 12, -1
+	cmpdi	12, 0
+	bne	__Loop_8x_block_enc
+
+__Finish_ghash:
+	vcipherlast     15, 15, 1
+	vcipherlast     16, 16, 1
+	vcipherlast     17, 17, 1
+	vcipherlast     18, 18, 1
+	vcipherlast     19, 19, 1
+	vcipherlast     20, 20, 1
+	vcipherlast     21, 21, 1
+	vcipherlast     22, 22, 1
+
+	lxvb16x	32+23, 0, 14	# load block
+	lxvb16x	32+24, 15, 14	# load block
+	lxvb16x	32+25, 16, 14	# load block
+	lxvb16x	32+26, 17, 14	# load block
+	lxvb16x	32+27, 18, 14	# load block
+	lxvb16x	32+28, 19, 14	# load block
+	lxvb16x	32+29, 20, 14	# load block
+	lxvb16x	32+30, 21, 14	# load block
+	addi	14, 14, 128
+
+	vxor	15, 15, 23
+	vxor	16, 16, 24
+	vxor	17, 17, 25
+	vxor	18, 18, 26
+	vxor	19, 19, 27
+	vxor	20, 20, 28
+	vxor	21, 21, 29
+	vxor	22, 22, 30
+
+	stxvb16x 47, 0, 9	# store output
+	stxvb16x 48, 15, 9	# store output
+	stxvb16x 49, 16, 9	# store output
+	stxvb16x 50, 17, 9	# store output
+	stxvb16x 51, 18, 9	# store output
+	stxvb16x 52, 19, 9	# store output
+	stxvb16x 53, 20, 9	# store output
+	stxvb16x 54, 21, 9	# store output
+	addi	9, 9, 128
+
+	vxor	15, 15, 0
+	PPC_GHASH4x 0, 15, 16, 17, 18
+
+	vxor	19, 19, 0
+	PPC_GHASH4x 0, 19, 20, 21, 22
+
+	xxlor	30+32, 9, 9		# last ctr
+	vadduwm	30, 30, 31		# increase ctr
+	stxvb16x 32+30, 0, 7		# update IV
+	stxvb16x 32+0, 0, 8		# update Xi
+
+	addi    5, 5, -128
+	addi    11, 11, 128
+
+	#
+	# Done 8x blocks
+	#
+
+	cmpdi   5, 0
+	beq     aes_gcm_out
+
+__Process_more_enc:
+	li	24, 1			# encrypt
+	bl	aes_gcm_crypt_1x
+	cmpdi   5, 0
+	beq     aes_gcm_out
+
+	bl	__Process_partial
+	cmpdi   5, 0
+	beq     aes_gcm_out
+__Do_combine_enc:
+	bl	__Combine_partial
+	cmpdi	5, 0
+	bgt	__Process_encrypt
+	b	aes_gcm_out
+
+SYM_FUNC_END(aes_p10_gcm_encrypt)
+
+################################################################################
+# aes_p10_gcm_decrypt (const void *inp, void *out, size_t len,
+#               const char *rk, unsigned char iv[16], void *Xip);
+# 8x Decrypt
+#
+################################################################################
+SYM_FUNC_START(aes_p10_gcm_decrypt)
+
+	cmpdi	5, 0
+	ble	__Invalid_msg_len
+
+	SAVE_REGS
+	LOAD_HASH_TABLE
+
+	# initialize ICB: GHASH( IV ), IV - r7
+	lxvb16x	30+32, 0, 7	# load IV  - v30
+
+	mr	14, 3
+	mr	9, 4
+
+	# counter 1
+	vxor	31, 31, 31
+	vspltisb 22, 1
+	vsldoi	31, 31, 22,1	# counter 1
+
+	addis	11, 2, permx@toc@ha
+	addi	11, 11, permx@toc@l
+	lxv	10, 0(11)	# vs10: vpermxor vector
+	li	11, 0
+
+	# load 9 round keys to VSR
+	lxv	0, 0(6)			# round key 0
+	lxv	1, 16(6)		# round key 1
+	lxv	2, 32(6)		# round key 2
+	lxv	3, 48(6)		# round key 3
+	lxv	4, 64(6)		# round key 4
+	lxv	5, 80(6)		# round key 5
+	lxv	6, 96(6)		# round key 6
+	lxv	7, 112(6)		# round key 7
+	lxv	8, 128(6)		# round key 8
+
+	# load rounds - 10 (128), 12 (192), 14 (256)
+	lwz	23, 240(6)		# n rounds
+	li	24, 0			# decrypt
+
+__Process_decrypt:
+	#
+	# Process different blocks
+	#
+	ld	12, 56(7)
+	cmpdi	12, 0
+	bgt	__Do_combine_dec
+	cmpdi	5, 128
+	blt	__Process_more_dec
+
+#
+# Process 8x AES/GCM blocks
+#
+__Process_8x_dec:
+	# 8x blcoks
+	li	10, 128
+	divdu	12, 5, 10	# n 128 bytes-blocks
+
+	addi	12, 12, -1	# loop - 1
+
+	vmr	15, 30		# first state: IV
+	vadduwm	16, 15, 31	# state + counter
+	vadduwm	17, 16, 31
+	vadduwm	18, 17, 31
+	vadduwm	19, 18, 31
+	vadduwm	20, 19, 31
+	vadduwm	21, 20, 31
+	vadduwm	22, 21, 31
+	xxlor	9, 32+22, 32+22	# save last state
+
+	# vxor  state, state, w # addroundkey
+	xxlor	32+29, 0, 0
+        vxor    15, 15, 29      # IV + round key - add round key 0
+	vxor	16, 16, 29
+	vxor	17, 17, 29
+	vxor	18, 18, 29
+	vxor	19, 19, 29
+	vxor	20, 20, 29
+	vxor	21, 21, 29
+	vxor	22, 22, 29
+
+	li	15, 16
+	li	16, 32
+	li	17, 48
+	li	18, 64
+	li	19, 80
+	li	20, 96
+	li	21, 112
+
+	#
+	# Pre-compute first 8 AES state and leave 1/3/5 more rounds
+	# for the loop.
+	#
+	addi	22, 23, -9		# process 8 keys
+	mtctr	22			# AES key loop
+	addi	10, 6, 144
+
+	LOOP_8AES_STATE			# process 8 AES keys
+
+__PreLoop_aes_state_dec:
+	lxv	32+1, 0(10)		# round key
+	AES_CIPHER_8x vcipher 15 1
+	addi	10, 10, 16
+	bdnz	__PreLoop_aes_state_dec
+	lxv	32+1, 0(10)		# last round key (v1)
+
+	cmpdi	12, 0			# Only one loop (8 block)
+	beq	__Finish_ghash_dec
+
+#
+# Loop 8x blocks and compute ghash
+#
+__Loop_8x_block_dec:
+	vcipherlast     15, 15, 1
+	vcipherlast     16, 16, 1
+	vcipherlast     17, 17, 1
+	vcipherlast     18, 18, 1
+	vcipherlast     19, 19, 1
+	vcipherlast     20, 20, 1
+	vcipherlast     21, 21, 1
+	vcipherlast     22, 22, 1
+
+	lxvb16x	32+23, 0, 14	# load block
+	lxvb16x	32+24, 15, 14	# load block
+	lxvb16x	32+25, 16, 14	# load block
+	lxvb16x	32+26, 17, 14	# load block
+	lxvb16x	32+27, 18, 14	# load block
+	lxvb16x	32+28, 19, 14	# load block
+	lxvb16x	32+29, 20, 14	# load block
+	lxvb16x	32+30, 21, 14	# load block
+	addi	14, 14, 128
+
+	vxor	15, 15, 23
+	vxor	16, 16, 24
+	vxor	17, 17, 25
+	vxor	18, 18, 26
+	vxor	19, 19, 27
+	vxor	20, 20, 28
+	vxor	21, 21, 29
+	vxor	22, 22, 30
+
+	stxvb16x 47, 0, 9	# store output
+	stxvb16x 48, 15, 9	# store output
+	stxvb16x 49, 16, 9	# store output
+	stxvb16x 50, 17, 9	# store output
+	stxvb16x 51, 18, 9	# store output
+	stxvb16x 52, 19, 9	# store output
+	stxvb16x 53, 20, 9	# store output
+	stxvb16x 54, 21, 9	# store output
+
+	addi	9, 9, 128
+
+	vmr	15, 23
+	vmr	16, 24
+	vmr	17, 25
+	vmr	18, 26
+	vmr	19, 27
+	vmr	20, 28
+	vmr	21, 29
+	vmr	22, 30
+
+	# ghash here
+	vxor	15, 15, 0
+	PPC_GHASH4x 0, 15, 16, 17, 18
+
+	vxor	19, 19, 0
+	PPC_GHASH4x 0, 19, 20, 21, 22
+
+	xxlor	32+15, 9, 9		# last state
+	vadduwm 15, 15, 31		# state + counter
+	vadduwm 16, 15, 31
+	vadduwm 17, 16, 31
+	vadduwm 18, 17, 31
+	vadduwm 19, 18, 31
+	vadduwm 20, 19, 31
+	vadduwm 21, 20, 31
+	vadduwm 22, 21, 31
+	xxlor	9, 32+22, 32+22		# save last state
+
+	xxlor	32+27, 0, 0		# restore roundkey 0
+        vxor    15, 15, 27		# IV + round key - add round key 0
+	vxor	16, 16, 27
+	vxor	17, 17, 27
+	vxor	18, 18, 27
+	vxor	19, 19, 27
+	vxor	20, 20, 27
+	vxor	21, 21, 27
+	vxor	22, 22, 27
+
+	addi    5, 5, -128
+	addi    11, 11, 128
+
+	LOOP_8AES_STATE			# process 8 AES keys
+	mtctr	22			# AES key loop
+	addi	10, 6, 144
+__LastLoop_aes_state_dec:
+	lxv	32+1, 0(10)		# round key
+	AES_CIPHER_8x vcipher 15 1
+	addi	10, 10, 16
+	bdnz	__LastLoop_aes_state_dec
+	lxv	32+1, 0(10)		# last round key (v1)
+
+	addi	12, 12, -1
+	cmpdi	12, 0
+	bne	__Loop_8x_block_dec
+
+__Finish_ghash_dec:
+	vcipherlast     15, 15, 1
+	vcipherlast     16, 16, 1
+	vcipherlast     17, 17, 1
+	vcipherlast     18, 18, 1
+	vcipherlast     19, 19, 1
+	vcipherlast     20, 20, 1
+	vcipherlast     21, 21, 1
+	vcipherlast     22, 22, 1
+
+	lxvb16x	32+23, 0, 14	# load block
+	lxvb16x	32+24, 15, 14	# load block
+	lxvb16x	32+25, 16, 14	# load block
+	lxvb16x	32+26, 17, 14	# load block
+	lxvb16x	32+27, 18, 14	# load block
+	lxvb16x	32+28, 19, 14	# load block
+	lxvb16x	32+29, 20, 14	# load block
+	lxvb16x	32+30, 21, 14	# load block
+	addi	14, 14, 128
+
+	vxor	15, 15, 23
+	vxor	16, 16, 24
+	vxor	17, 17, 25
+	vxor	18, 18, 26
+	vxor	19, 19, 27
+	vxor	20, 20, 28
+	vxor	21, 21, 29
+	vxor	22, 22, 30
+
+	stxvb16x 47, 0, 9	# store output
+	stxvb16x 48, 15, 9	# store output
+	stxvb16x 49, 16, 9	# store output
+	stxvb16x 50, 17, 9	# store output
+	stxvb16x 51, 18, 9	# store output
+	stxvb16x 52, 19, 9	# store output
+	stxvb16x 53, 20, 9	# store output
+	stxvb16x 54, 21, 9	# store output
+	addi	9, 9, 128
+
+	#vmr	15, 23
+	vxor	15, 23, 0
+	vmr	16, 24
+	vmr	17, 25
+	vmr	18, 26
+	vmr	19, 27
+	vmr	20, 28
+	vmr	21, 29
+	vmr	22, 30
+
+	#vxor	15, 15, 0
+	PPC_GHASH4x 0, 15, 16, 17, 18
+
+	vxor	19, 19, 0
+	PPC_GHASH4x 0, 19, 20, 21, 22
+
+	xxlor	30+32, 9, 9		# last ctr
+	vadduwm	30, 30, 31		# increase ctr
+	stxvb16x 32+30, 0, 7		# update IV
+	stxvb16x 32+0, 0, 8		# update Xi
+
+	addi    5, 5, -128
+	addi    11, 11, 128
+
+	#
+	# Done 8x blocks
+	#
+
+	cmpdi   5, 0
+	beq     aes_gcm_out
+
+__Process_more_dec:
+	li	24, 0			# decrypt
+	bl	aes_gcm_crypt_1x
+	cmpdi   5, 0
+	beq     aes_gcm_out
+
+	bl	__Process_partial
+	cmpdi   5, 0
+	beq     aes_gcm_out
+__Do_combine_dec:
+	bl	__Combine_partial
+	cmpdi	5, 0
+	bgt	__Process_decrypt
+	b	aes_gcm_out
+SYM_FUNC_END(aes_p10_gcm_decrypt)
+
+SYM_FUNC_START_LOCAL(aes_gcm_out)
+
+	mr	3, 11			# return count
+
+	RESTORE_REGS
+	blr
+
+__Invalid_msg_len:
+	li	3, 0
+	blr
+SYM_FUNC_END(aes_gcm_out)
+
+SYM_DATA_START_LOCAL(PERMX)
+.align 4
+# for vector permute and xor
+permx:
+.long 0x4c5d6e7f, 0x08192a3b, 0xc4d5e6f7, 0x8091a2b3
+SYM_DATA_END(permx)
diff --git a/arch/powerpc/crypto/aes-spe-glue.c b/arch/powerpc/crypto/aes-spe-glue.c
new file mode 100644
index 000000000..e03848808
--- /dev/null
+++ b/arch/powerpc/crypto/aes-spe-glue.c
@@ -0,0 +1,445 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Glue code for AES implementation for SPE instructions (PPC)
+ *
+ * Based on generic implementation. The assembler module takes care
+ * about the SPE registers so it can run from interrupt context.
+ *
+ * Copyright (c) 2015 Markus Stockhausen <stockhausen@collogia.de>
+ */
+
+#include <crypto/aes.h>
+#include <linux/minmax.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/crypto.h>
+#include <asm/byteorder.h>
+#include <asm/switch_to.h>
+#include <crypto/algapi.h>
+#include <crypto/internal/skcipher.h>
+#include <crypto/xts.h>
+#include <crypto/gf128mul.h>
+#include <crypto/scatterwalk.h>
+
+/*
+ * MAX_BYTES defines the number of bytes that are allowed to be processed
+ * between preempt_disable() and preempt_enable(). e500 cores can issue two
+ * instructions per clock cycle using one 32/64 bit unit (SU1) and one 32
+ * bit unit (SU2). One of these can be a memory access that is executed via
+ * a single load and store unit (LSU). XTS-AES-256 takes ~780 operations per
+ * 16 byte block or 25 cycles per byte. Thus 768 bytes of input data
+ * will need an estimated maximum of 20,000 cycles. Headroom for cache misses
+ * included. Even with the low end model clocked at 667 MHz this equals to a
+ * critical time window of less than 30us. The value has been chosen to
+ * process a 512 byte disk block in one or a large 1400 bytes IPsec network
+ * packet in two runs.
+ *
+ */
+#define MAX_BYTES 768
+
+struct ppc_aes_ctx {
+	u32 key_enc[AES_MAX_KEYLENGTH_U32];
+	u32 key_dec[AES_MAX_KEYLENGTH_U32];
+	u32 rounds;
+};
+
+struct ppc_xts_ctx {
+	u32 key_enc[AES_MAX_KEYLENGTH_U32];
+	u32 key_dec[AES_MAX_KEYLENGTH_U32];
+	u32 key_twk[AES_MAX_KEYLENGTH_U32];
+	u32 rounds;
+};
+
+static void spe_begin(void)
+{
+	/* disable preemption and save users SPE registers if required */
+	preempt_disable();
+	enable_kernel_spe();
+}
+
+static void spe_end(void)
+{
+	disable_kernel_spe();
+	/* reenable preemption */
+	preempt_enable();
+}
+
+static int ppc_aes_setkey_skcipher(struct crypto_skcipher *tfm,
+				   const u8 *in_key, unsigned int key_len)
+{
+	struct ppc_aes_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+	switch (key_len) {
+	case AES_KEYSIZE_128:
+		ctx->rounds = 4;
+		ppc_expand_key_128(ctx->key_enc, in_key);
+		break;
+	case AES_KEYSIZE_192:
+		ctx->rounds = 5;
+		ppc_expand_key_192(ctx->key_enc, in_key);
+		break;
+	case AES_KEYSIZE_256:
+		ctx->rounds = 6;
+		ppc_expand_key_256(ctx->key_enc, in_key);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len);
+
+	return 0;
+}
+
+static int ppc_xts_setkey(struct crypto_skcipher *tfm, const u8 *in_key,
+		   unsigned int key_len)
+{
+	struct ppc_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int err;
+
+	err = xts_verify_key(tfm, in_key, key_len);
+	if (err)
+		return err;
+
+	key_len >>= 1;
+
+	switch (key_len) {
+	case AES_KEYSIZE_128:
+		ctx->rounds = 4;
+		ppc_expand_key_128(ctx->key_enc, in_key);
+		ppc_expand_key_128(ctx->key_twk, in_key + AES_KEYSIZE_128);
+		break;
+	case AES_KEYSIZE_192:
+		ctx->rounds = 5;
+		ppc_expand_key_192(ctx->key_enc, in_key);
+		ppc_expand_key_192(ctx->key_twk, in_key + AES_KEYSIZE_192);
+		break;
+	case AES_KEYSIZE_256:
+		ctx->rounds = 6;
+		ppc_expand_key_256(ctx->key_enc, in_key);
+		ppc_expand_key_256(ctx->key_twk, in_key + AES_KEYSIZE_256);
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	ppc_generate_decrypt_key(ctx->key_dec, ctx->key_enc, key_len);
+
+	return 0;
+}
+
+static int ppc_ecb_crypt(struct skcipher_request *req, bool enc)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_aes_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int err;
+
+	err = skcipher_walk_virt(&walk, req, false);
+
+	while ((nbytes = walk.nbytes) != 0) {
+		nbytes = min(nbytes, MAX_BYTES);
+		nbytes = round_down(nbytes, AES_BLOCK_SIZE);
+
+		spe_begin();
+		if (enc)
+			ppc_encrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_enc, ctx->rounds, nbytes);
+		else
+			ppc_decrypt_ecb(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_dec, ctx->rounds, nbytes);
+		spe_end();
+
+		err = skcipher_walk_done(&walk, walk.nbytes - nbytes);
+	}
+
+	return err;
+}
+
+static int ppc_ecb_encrypt(struct skcipher_request *req)
+{
+	return ppc_ecb_crypt(req, true);
+}
+
+static int ppc_ecb_decrypt(struct skcipher_request *req)
+{
+	return ppc_ecb_crypt(req, false);
+}
+
+static int ppc_cbc_crypt(struct skcipher_request *req, bool enc)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_aes_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int err;
+
+	err = skcipher_walk_virt(&walk, req, false);
+
+	while ((nbytes = walk.nbytes) != 0) {
+		nbytes = min(nbytes, MAX_BYTES);
+		nbytes = round_down(nbytes, AES_BLOCK_SIZE);
+
+		spe_begin();
+		if (enc)
+			ppc_encrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_enc, ctx->rounds, nbytes,
+					walk.iv);
+		else
+			ppc_decrypt_cbc(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_dec, ctx->rounds, nbytes,
+					walk.iv);
+		spe_end();
+
+		err = skcipher_walk_done(&walk, walk.nbytes - nbytes);
+	}
+
+	return err;
+}
+
+static int ppc_cbc_encrypt(struct skcipher_request *req)
+{
+	return ppc_cbc_crypt(req, true);
+}
+
+static int ppc_cbc_decrypt(struct skcipher_request *req)
+{
+	return ppc_cbc_crypt(req, false);
+}
+
+static int ppc_ctr_crypt(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_aes_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int err;
+
+	err = skcipher_walk_virt(&walk, req, false);
+
+	while ((nbytes = walk.nbytes) != 0) {
+		nbytes = min(nbytes, MAX_BYTES);
+		if (nbytes < walk.total)
+			nbytes = round_down(nbytes, AES_BLOCK_SIZE);
+
+		spe_begin();
+		ppc_crypt_ctr(walk.dst.virt.addr, walk.src.virt.addr,
+			      ctx->key_enc, ctx->rounds, nbytes, walk.iv);
+		spe_end();
+
+		err = skcipher_walk_done(&walk, walk.nbytes - nbytes);
+	}
+
+	return err;
+}
+
+static int ppc_xts_crypt(struct skcipher_request *req, bool enc)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int err;
+	u32 *twk;
+
+	err = skcipher_walk_virt(&walk, req, false);
+	twk = ctx->key_twk;
+
+	while ((nbytes = walk.nbytes) != 0) {
+		nbytes = min(nbytes, MAX_BYTES);
+		nbytes = round_down(nbytes, AES_BLOCK_SIZE);
+
+		spe_begin();
+		if (enc)
+			ppc_encrypt_xts(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_enc, ctx->rounds, nbytes,
+					walk.iv, twk);
+		else
+			ppc_decrypt_xts(walk.dst.virt.addr, walk.src.virt.addr,
+					ctx->key_dec, ctx->rounds, nbytes,
+					walk.iv, twk);
+		spe_end();
+
+		twk = NULL;
+		err = skcipher_walk_done(&walk, walk.nbytes - nbytes);
+	}
+
+	return err;
+}
+
+static int ppc_xts_encrypt(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int tail = req->cryptlen % AES_BLOCK_SIZE;
+	int offset = req->cryptlen - tail - AES_BLOCK_SIZE;
+	struct skcipher_request subreq;
+	u8 b[2][AES_BLOCK_SIZE];
+	int err;
+
+	if (req->cryptlen < AES_BLOCK_SIZE)
+		return -EINVAL;
+
+	if (tail) {
+		subreq = *req;
+		skcipher_request_set_crypt(&subreq, req->src, req->dst,
+					   req->cryptlen - tail, req->iv);
+		req = &subreq;
+	}
+
+	err = ppc_xts_crypt(req, true);
+	if (err || !tail)
+		return err;
+
+	scatterwalk_map_and_copy(b[0], req->dst, offset, AES_BLOCK_SIZE, 0);
+	memcpy(b[1], b[0], tail);
+	scatterwalk_map_and_copy(b[0], req->src, offset + AES_BLOCK_SIZE, tail, 0);
+
+	spe_begin();
+	ppc_encrypt_xts(b[0], b[0], ctx->key_enc, ctx->rounds, AES_BLOCK_SIZE,
+			req->iv, NULL);
+	spe_end();
+
+	scatterwalk_map_and_copy(b[0], req->dst, offset, AES_BLOCK_SIZE + tail, 1);
+
+	return 0;
+}
+
+static int ppc_xts_decrypt(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	struct ppc_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int tail = req->cryptlen % AES_BLOCK_SIZE;
+	int offset = req->cryptlen - tail - AES_BLOCK_SIZE;
+	struct skcipher_request subreq;
+	u8 b[3][AES_BLOCK_SIZE];
+	le128 twk;
+	int err;
+
+	if (req->cryptlen < AES_BLOCK_SIZE)
+		return -EINVAL;
+
+	if (tail) {
+		subreq = *req;
+		skcipher_request_set_crypt(&subreq, req->src, req->dst,
+					   offset, req->iv);
+		req = &subreq;
+	}
+
+	err = ppc_xts_crypt(req, false);
+	if (err || !tail)
+		return err;
+
+	scatterwalk_map_and_copy(b[1], req->src, offset, AES_BLOCK_SIZE + tail, 0);
+
+	spe_begin();
+	if (!offset)
+		ppc_encrypt_ecb(req->iv, req->iv, ctx->key_twk, ctx->rounds,
+				AES_BLOCK_SIZE);
+
+	gf128mul_x_ble(&twk, (le128 *)req->iv);
+
+	ppc_decrypt_xts(b[1], b[1], ctx->key_dec, ctx->rounds, AES_BLOCK_SIZE,
+			(u8 *)&twk, NULL);
+	memcpy(b[0], b[2], tail);
+	memcpy(b[0] + tail, b[1] + tail, AES_BLOCK_SIZE - tail);
+	ppc_decrypt_xts(b[0], b[0], ctx->key_dec, ctx->rounds, AES_BLOCK_SIZE,
+			req->iv, NULL);
+	spe_end();
+
+	scatterwalk_map_and_copy(b[0], req->dst, offset, AES_BLOCK_SIZE + tail, 1);
+
+	return 0;
+}
+
+/*
+ * Algorithm definitions. Disabling alignment (cra_alignmask=0) was chosen
+ * because the e500 platform can handle unaligned reads/writes very efficiently.
+ * This improves IPsec thoughput by another few percent. Additionally we assume
+ * that AES context is always aligned to at least 8 bytes because it is created
+ * with kmalloc() in the crypto infrastructure
+ */
+
+static struct skcipher_alg aes_skcipher_algs[] = {
+	{
+		.base.cra_name		=	"ecb(aes)",
+		.base.cra_driver_name	=	"ecb-ppc-spe",
+		.base.cra_priority	=	300,
+		.base.cra_blocksize	=	AES_BLOCK_SIZE,
+		.base.cra_ctxsize	=	sizeof(struct ppc_aes_ctx),
+		.base.cra_module	=	THIS_MODULE,
+		.min_keysize		=	AES_MIN_KEY_SIZE,
+		.max_keysize		=	AES_MAX_KEY_SIZE,
+		.setkey			=	ppc_aes_setkey_skcipher,
+		.encrypt		=	ppc_ecb_encrypt,
+		.decrypt		=	ppc_ecb_decrypt,
+	}, {
+		.base.cra_name		=	"cbc(aes)",
+		.base.cra_driver_name	=	"cbc-ppc-spe",
+		.base.cra_priority	=	300,
+		.base.cra_blocksize	=	AES_BLOCK_SIZE,
+		.base.cra_ctxsize	=	sizeof(struct ppc_aes_ctx),
+		.base.cra_module	=	THIS_MODULE,
+		.min_keysize		=	AES_MIN_KEY_SIZE,
+		.max_keysize		=	AES_MAX_KEY_SIZE,
+		.ivsize			=	AES_BLOCK_SIZE,
+		.setkey			=	ppc_aes_setkey_skcipher,
+		.encrypt		=	ppc_cbc_encrypt,
+		.decrypt		=	ppc_cbc_decrypt,
+	}, {
+		.base.cra_name		=	"ctr(aes)",
+		.base.cra_driver_name	=	"ctr-ppc-spe",
+		.base.cra_priority	=	300,
+		.base.cra_blocksize	=	1,
+		.base.cra_ctxsize	=	sizeof(struct ppc_aes_ctx),
+		.base.cra_module	=	THIS_MODULE,
+		.min_keysize		=	AES_MIN_KEY_SIZE,
+		.max_keysize		=	AES_MAX_KEY_SIZE,
+		.ivsize			=	AES_BLOCK_SIZE,
+		.setkey			=	ppc_aes_setkey_skcipher,
+		.encrypt		=	ppc_ctr_crypt,
+		.decrypt		=	ppc_ctr_crypt,
+		.chunksize		=	AES_BLOCK_SIZE,
+	}, {
+		.base.cra_name		=	"xts(aes)",
+		.base.cra_driver_name	=	"xts-ppc-spe",
+		.base.cra_priority	=	300,
+		.base.cra_blocksize	=	AES_BLOCK_SIZE,
+		.base.cra_ctxsize	=	sizeof(struct ppc_xts_ctx),
+		.base.cra_module	=	THIS_MODULE,
+		.min_keysize		=	AES_MIN_KEY_SIZE * 2,
+		.max_keysize		=	AES_MAX_KEY_SIZE * 2,
+		.ivsize			=	AES_BLOCK_SIZE,
+		.setkey			=	ppc_xts_setkey,
+		.encrypt		=	ppc_xts_encrypt,
+		.decrypt		=	ppc_xts_decrypt,
+	}
+};
+
+static int __init ppc_aes_mod_init(void)
+{
+	return crypto_register_skciphers(aes_skcipher_algs,
+					 ARRAY_SIZE(aes_skcipher_algs));
+}
+
+static void __exit ppc_aes_mod_fini(void)
+{
+	crypto_unregister_skciphers(aes_skcipher_algs,
+				    ARRAY_SIZE(aes_skcipher_algs));
+}
+
+module_init(ppc_aes_mod_init);
+module_exit(ppc_aes_mod_fini);
+
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("AES-ECB/CBC/CTR/XTS, SPE optimized");
+
+MODULE_ALIAS_CRYPTO("aes");
+MODULE_ALIAS_CRYPTO("ecb(aes)");
+MODULE_ALIAS_CRYPTO("cbc(aes)");
+MODULE_ALIAS_CRYPTO("ctr(aes)");
+MODULE_ALIAS_CRYPTO("xts(aes)");
+MODULE_ALIAS_CRYPTO("aes-ppc-spe");
diff --git a/arch/powerpc/crypto/aes_cbc.c b/arch/powerpc/crypto/aes_cbc.c
new file mode 100644
index 000000000..9c271b464
--- /dev/null
+++ b/arch/powerpc/crypto/aes_cbc.c
@@ -0,0 +1,137 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * AES CBC routines supporting VMX instructions on the Power 8
+ *
+ * Copyright (C) 2015 International Business Machines Inc.
+ *
+ * Author: Marcelo Henrique Cerri <mhcerri@br.ibm.com>
+ */
+
+#include <asm/simd.h>
+#include <asm/switch_to.h>
+#include <crypto/aes.h>
+#include <crypto/internal/simd.h>
+#include <crypto/internal/skcipher.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/uaccess.h>
+
+#include "aesp8-ppc.h"
+
+struct p8_aes_cbc_ctx {
+	struct crypto_skcipher *fallback;
+	struct p8_aes_key enc_key;
+	struct p8_aes_key dec_key;
+};
+
+static int p8_aes_cbc_init(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_cbc_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct crypto_skcipher *fallback;
+
+	fallback = crypto_alloc_skcipher("cbc(aes)", 0,
+					 CRYPTO_ALG_NEED_FALLBACK |
+					 CRYPTO_ALG_ASYNC);
+	if (IS_ERR(fallback)) {
+		pr_err("Failed to allocate cbc(aes) fallback: %ld\n",
+		       PTR_ERR(fallback));
+		return PTR_ERR(fallback);
+	}
+
+	crypto_skcipher_set_reqsize(tfm, sizeof(struct skcipher_request) +
+				    crypto_skcipher_reqsize(fallback));
+	ctx->fallback = fallback;
+	return 0;
+}
+
+static void p8_aes_cbc_exit(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_cbc_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+	crypto_free_skcipher(ctx->fallback);
+}
+
+static int p8_aes_cbc_setkey(struct crypto_skcipher *tfm, const u8 *key,
+			     unsigned int keylen)
+{
+	struct p8_aes_cbc_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int ret;
+
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+	ret = aes_p8_set_encrypt_key(key, keylen * 8, &ctx->enc_key);
+	ret |= aes_p8_set_decrypt_key(key, keylen * 8, &ctx->dec_key);
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+
+	ret |= crypto_skcipher_setkey(ctx->fallback, key, keylen);
+
+	return ret ? -EINVAL : 0;
+}
+
+static int p8_aes_cbc_crypt(struct skcipher_request *req, bool enc)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	const struct p8_aes_cbc_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int ret;
+
+	if (!crypto_simd_usable()) {
+		struct skcipher_request *subreq = skcipher_request_ctx(req);
+
+		*subreq = *req;
+		skcipher_request_set_tfm(subreq, ctx->fallback);
+		return enc ? crypto_skcipher_encrypt(subreq) :
+			     crypto_skcipher_decrypt(subreq);
+	}
+
+	ret = skcipher_walk_virt(&walk, req, false);
+	while ((nbytes = walk.nbytes) != 0) {
+		preempt_disable();
+		pagefault_disable();
+		enable_kernel_vsx();
+		aes_p8_cbc_encrypt(walk.src.virt.addr,
+				   walk.dst.virt.addr,
+				   round_down(nbytes, AES_BLOCK_SIZE),
+				   enc ? &ctx->enc_key : &ctx->dec_key,
+				   walk.iv, enc);
+		disable_kernel_vsx();
+		pagefault_enable();
+		preempt_enable();
+
+		ret = skcipher_walk_done(&walk, nbytes % AES_BLOCK_SIZE);
+	}
+	return ret;
+}
+
+static int p8_aes_cbc_encrypt(struct skcipher_request *req)
+{
+	return p8_aes_cbc_crypt(req, true);
+}
+
+static int p8_aes_cbc_decrypt(struct skcipher_request *req)
+{
+	return p8_aes_cbc_crypt(req, false);
+}
+
+struct skcipher_alg p8_aes_cbc_alg = {
+	.base.cra_name = "cbc(aes)",
+	.base.cra_driver_name = "p8_aes_cbc",
+	.base.cra_module = THIS_MODULE,
+	.base.cra_priority = 2000,
+	.base.cra_flags = CRYPTO_ALG_NEED_FALLBACK,
+	.base.cra_blocksize = AES_BLOCK_SIZE,
+	.base.cra_ctxsize = sizeof(struct p8_aes_cbc_ctx),
+	.setkey = p8_aes_cbc_setkey,
+	.encrypt = p8_aes_cbc_encrypt,
+	.decrypt = p8_aes_cbc_decrypt,
+	.init = p8_aes_cbc_init,
+	.exit = p8_aes_cbc_exit,
+	.min_keysize = AES_MIN_KEY_SIZE,
+	.max_keysize = AES_MAX_KEY_SIZE,
+	.ivsize = AES_BLOCK_SIZE,
+};
diff --git a/arch/powerpc/crypto/aes_ctr.c b/arch/powerpc/crypto/aes_ctr.c
new file mode 100644
index 000000000..7dbd06f44
--- /dev/null
+++ b/arch/powerpc/crypto/aes_ctr.c
@@ -0,0 +1,153 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * AES CTR routines supporting VMX instructions on the Power 8
+ *
+ * Copyright (C) 2015 International Business Machines Inc.
+ *
+ * Author: Marcelo Henrique Cerri <mhcerri@br.ibm.com>
+ */
+
+#include <asm/simd.h>
+#include <asm/switch_to.h>
+#include <crypto/aes.h>
+#include <crypto/internal/simd.h>
+#include <crypto/internal/skcipher.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/uaccess.h>
+
+#include "aesp8-ppc.h"
+
+struct p8_aes_ctr_ctx {
+	struct crypto_skcipher *fallback;
+	struct p8_aes_key enc_key;
+};
+
+static int p8_aes_ctr_init(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_ctr_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct crypto_skcipher *fallback;
+
+	fallback = crypto_alloc_skcipher("ctr(aes)", 0,
+					 CRYPTO_ALG_NEED_FALLBACK |
+					 CRYPTO_ALG_ASYNC);
+	if (IS_ERR(fallback)) {
+		pr_err("Failed to allocate ctr(aes) fallback: %ld\n",
+		       PTR_ERR(fallback));
+		return PTR_ERR(fallback);
+	}
+
+	crypto_skcipher_set_reqsize(tfm, sizeof(struct skcipher_request) +
+				    crypto_skcipher_reqsize(fallback));
+	ctx->fallback = fallback;
+	return 0;
+}
+
+static void p8_aes_ctr_exit(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_ctr_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+	crypto_free_skcipher(ctx->fallback);
+}
+
+static int p8_aes_ctr_setkey(struct crypto_skcipher *tfm, const u8 *key,
+			     unsigned int keylen)
+{
+	struct p8_aes_ctr_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int ret;
+
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+	ret = aes_p8_set_encrypt_key(key, keylen * 8, &ctx->enc_key);
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+
+	ret |= crypto_skcipher_setkey(ctx->fallback, key, keylen);
+
+	return ret ? -EINVAL : 0;
+}
+
+static void p8_aes_ctr_final(const struct p8_aes_ctr_ctx *ctx,
+			     struct skcipher_walk *walk)
+{
+	const u8 *src = walk->src.virt.addr;
+	u8 *ctrblk = walk->iv;
+	u8 keystream[AES_BLOCK_SIZE];
+	u8 *dst = walk->dst.virt.addr;
+	unsigned int nbytes = walk->nbytes;
+
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+	aes_p8_encrypt(ctrblk, keystream, &ctx->enc_key);
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+
+	crypto_xor_cpy(dst, keystream, src, nbytes);
+	crypto_inc(ctrblk, AES_BLOCK_SIZE);
+}
+
+static int p8_aes_ctr_crypt(struct skcipher_request *req)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	const struct p8_aes_ctr_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	int ret;
+
+	if (!crypto_simd_usable()) {
+		struct skcipher_request *subreq = skcipher_request_ctx(req);
+
+		*subreq = *req;
+		skcipher_request_set_tfm(subreq, ctx->fallback);
+		return crypto_skcipher_encrypt(subreq);
+	}
+
+	ret = skcipher_walk_virt(&walk, req, false);
+	while ((nbytes = walk.nbytes) >= AES_BLOCK_SIZE) {
+		preempt_disable();
+		pagefault_disable();
+		enable_kernel_vsx();
+		aes_p8_ctr32_encrypt_blocks(walk.src.virt.addr,
+					    walk.dst.virt.addr,
+					    nbytes / AES_BLOCK_SIZE,
+					    &ctx->enc_key, walk.iv);
+		disable_kernel_vsx();
+		pagefault_enable();
+		preempt_enable();
+
+		do {
+			crypto_inc(walk.iv, AES_BLOCK_SIZE);
+		} while ((nbytes -= AES_BLOCK_SIZE) >= AES_BLOCK_SIZE);
+
+		ret = skcipher_walk_done(&walk, nbytes);
+	}
+	if (nbytes) {
+		p8_aes_ctr_final(ctx, &walk);
+		ret = skcipher_walk_done(&walk, 0);
+	}
+	return ret;
+}
+
+struct skcipher_alg p8_aes_ctr_alg = {
+	.base.cra_name = "ctr(aes)",
+	.base.cra_driver_name = "p8_aes_ctr",
+	.base.cra_module = THIS_MODULE,
+	.base.cra_priority = 2000,
+	.base.cra_flags = CRYPTO_ALG_NEED_FALLBACK,
+	.base.cra_blocksize = 1,
+	.base.cra_ctxsize = sizeof(struct p8_aes_ctr_ctx),
+	.setkey = p8_aes_ctr_setkey,
+	.encrypt = p8_aes_ctr_crypt,
+	.decrypt = p8_aes_ctr_crypt,
+	.init = p8_aes_ctr_init,
+	.exit = p8_aes_ctr_exit,
+	.min_keysize = AES_MIN_KEY_SIZE,
+	.max_keysize = AES_MAX_KEY_SIZE,
+	.ivsize = AES_BLOCK_SIZE,
+	.chunksize = AES_BLOCK_SIZE,
+};
diff --git a/arch/powerpc/crypto/aes_xts.c b/arch/powerpc/crypto/aes_xts.c
new file mode 100644
index 000000000..b4c760e46
--- /dev/null
+++ b/arch/powerpc/crypto/aes_xts.c
@@ -0,0 +1,166 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * AES XTS routines supporting VMX In-core instructions on Power 8
+ *
+ * Copyright (C) 2015 International Business Machines Inc.
+ *
+ * Author: Leonidas S. Barbosa <leosilva@linux.vnet.ibm.com>
+ */
+
+#include <asm/simd.h>
+#include <asm/switch_to.h>
+#include <crypto/aes.h>
+#include <crypto/internal/simd.h>
+#include <crypto/internal/skcipher.h>
+#include <crypto/xts.h>
+#include <linux/err.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/uaccess.h>
+
+#include "aesp8-ppc.h"
+
+struct p8_aes_xts_ctx {
+	struct crypto_skcipher *fallback;
+	struct p8_aes_key enc_key;
+	struct p8_aes_key dec_key;
+	struct p8_aes_key tweak_key;
+};
+
+static int p8_aes_xts_init(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct crypto_skcipher *fallback;
+
+	fallback = crypto_alloc_skcipher("xts(aes)", 0,
+					 CRYPTO_ALG_NEED_FALLBACK |
+					 CRYPTO_ALG_ASYNC);
+	if (IS_ERR(fallback)) {
+		pr_err("Failed to allocate xts(aes) fallback: %ld\n",
+		       PTR_ERR(fallback));
+		return PTR_ERR(fallback);
+	}
+
+	crypto_skcipher_set_reqsize(tfm, sizeof(struct skcipher_request) +
+				    crypto_skcipher_reqsize(fallback));
+	ctx->fallback = fallback;
+	return 0;
+}
+
+static void p8_aes_xts_exit(struct crypto_skcipher *tfm)
+{
+	struct p8_aes_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+
+	crypto_free_skcipher(ctx->fallback);
+}
+
+static int p8_aes_xts_setkey(struct crypto_skcipher *tfm, const u8 *key,
+			     unsigned int keylen)
+{
+	struct p8_aes_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	int ret;
+
+	ret = xts_verify_key(tfm, key, keylen);
+	if (ret)
+		return ret;
+
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+	ret = aes_p8_set_encrypt_key(key + keylen/2, (keylen/2) * 8, &ctx->tweak_key);
+	ret |= aes_p8_set_encrypt_key(key, (keylen/2) * 8, &ctx->enc_key);
+	ret |= aes_p8_set_decrypt_key(key, (keylen/2) * 8, &ctx->dec_key);
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+
+	ret |= crypto_skcipher_setkey(ctx->fallback, key, keylen);
+
+	return ret ? -EINVAL : 0;
+}
+
+static int p8_aes_xts_crypt(struct skcipher_request *req, int enc)
+{
+	struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req);
+	const struct p8_aes_xts_ctx *ctx = crypto_skcipher_ctx(tfm);
+	struct skcipher_walk walk;
+	unsigned int nbytes;
+	u8 tweak[AES_BLOCK_SIZE];
+	int ret;
+
+	if (req->cryptlen < AES_BLOCK_SIZE)
+		return -EINVAL;
+
+	if (!crypto_simd_usable() || (req->cryptlen % XTS_BLOCK_SIZE) != 0) {
+		struct skcipher_request *subreq = skcipher_request_ctx(req);
+
+		*subreq = *req;
+		skcipher_request_set_tfm(subreq, ctx->fallback);
+		return enc ? crypto_skcipher_encrypt(subreq) :
+			     crypto_skcipher_decrypt(subreq);
+	}
+
+	ret = skcipher_walk_virt(&walk, req, false);
+	if (ret)
+		return ret;
+
+	preempt_disable();
+	pagefault_disable();
+	enable_kernel_vsx();
+
+	aes_p8_encrypt(walk.iv, tweak, &ctx->tweak_key);
+
+	disable_kernel_vsx();
+	pagefault_enable();
+	preempt_enable();
+
+	while ((nbytes = walk.nbytes) != 0) {
+		preempt_disable();
+		pagefault_disable();
+		enable_kernel_vsx();
+		if (enc)
+			aes_p8_xts_encrypt(walk.src.virt.addr,
+					   walk.dst.virt.addr,
+					   round_down(nbytes, AES_BLOCK_SIZE),
+					   &ctx->enc_key, NULL, tweak);
+		else
+			aes_p8_xts_decrypt(walk.src.virt.addr,
+					   walk.dst.virt.addr,
+					   round_down(nbytes, AES_BLOCK_SIZE),
+					   &ctx->dec_key, NULL, tweak);
+		disable_kernel_vsx();
+		pagefault_enable();
+		preempt_enable();
+
+		ret = skcipher_walk_done(&walk, nbytes % AES_BLOCK_SIZE);
+	}
+	return ret;
+}
+
+static int p8_aes_xts_encrypt(struct skcipher_request *req)
+{
+	return p8_aes_xts_crypt(req, 1);
+}
+
+static int p8_aes_xts_decrypt(struct skcipher_request *req)
+{
+	return p8_aes_xts_crypt(req, 0);
+}
+
+struct skcipher_alg p8_aes_xts_alg = {
+	.base.cra_name = "xts(aes)",
+	.base.cra_driver_name = "p8_aes_xts",
+	.base.cra_module = THIS_MODULE,
+	.base.cra_priority = 2000,
+	.base.cra_flags = CRYPTO_ALG_NEED_FALLBACK,
+	.base.cra_blocksize = AES_BLOCK_SIZE,
+	.base.cra_ctxsize = sizeof(struct p8_aes_xts_ctx),
+	.setkey = p8_aes_xts_setkey,
+	.encrypt = p8_aes_xts_encrypt,
+	.decrypt = p8_aes_xts_decrypt,
+	.init = p8_aes_xts_init,
+	.exit = p8_aes_xts_exit,
+	.min_keysize = 2 * AES_MIN_KEY_SIZE,
+	.max_keysize = 2 * AES_MAX_KEY_SIZE,
+	.ivsize = AES_BLOCK_SIZE,
+};
diff --git a/arch/powerpc/crypto/aesp10-ppc.pl b/arch/powerpc/crypto/aesp10-ppc.pl
new file mode 100644
index 000000000..2c06ce2a2
--- /dev/null
+++ b/arch/powerpc/crypto/aesp10-ppc.pl
@@ -0,0 +1,585 @@
+#! /usr/bin/env perl
+# SPDX-License-Identifier: GPL-2.0
+
+# This code is taken from CRYPTOGAMs[1] and is included here using the option
+# in the license to distribute the code under the GPL. Therefore this program
+# is free software; you can redistribute it and/or modify it under the terms of
+# the GNU General Public License version 2 as published by the Free Software
+# Foundation.
+#
+# [1] https://www.openssl.org/~appro/cryptogams/
+
+# Copyright (c) 2006-2017, CRYPTOGAMS by <appro@openssl.org>
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions
+# are met:
+#
+#       * Redistributions of source code must retain copyright notices,
+#         this list of conditions and the following disclaimer.
+#
+#       * Redistributions in binary form must reproduce the above
+#         copyright notice, this list of conditions and the following
+#         disclaimer in the documentation and/or other materials
+#         provided with the distribution.
+#
+#       * Neither the name of the CRYPTOGAMS nor the names of its
+#         copyright holder and contributors may be used to endorse or
+#         promote products derived from this software without specific
+#         prior written permission.
+#
+# ALTERNATIVELY, provided that this notice is retained in full, this
+# product may be distributed under the terms of the GNU General Public
+# License (GPL), in which case the provisions of the GPL apply INSTEAD OF
+# those given above.
+#
+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER AND CONTRIBUTORS
+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see https://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# This module implements support for AES instructions as per PowerISA
+# specification version 2.07, first implemented by POWER8 processor.
+# The module is endian-agnostic in sense that it supports both big-
+# and little-endian cases. Data alignment in parallelizable modes is
+# handled with VSX loads and stores, which implies MSR.VSX flag being
+# set. It should also be noted that ISA specification doesn't prohibit
+# alignment exceptions for these instructions on page boundaries.
+# Initially alignment was handled in pure AltiVec/VMX way [when data
+# is aligned programmatically, which in turn guarantees exception-
+# free execution], but it turned to hamper performance when vcipher
+# instructions are interleaved. It's reckoned that eventual
+# misalignment penalties at page boundaries are in average lower
+# than additional overhead in pure AltiVec approach.
+#
+# May 2016
+#
+# Add XTS subroutine, 9x on little- and 12x improvement on big-endian
+# systems were measured.
+#
+######################################################################
+# Current large-block performance in cycles per byte processed with
+# 128-bit key (less is better).
+#
+#		CBC en-/decrypt	CTR	XTS
+# POWER8[le]	3.96/0.72	0.74	1.1
+# POWER8[be]	3.75/0.65	0.66	1.0
+
+$flavour = shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T	=8;
+	$LRSAVE	=2*$SIZE_T;
+	$STU	="stdu";
+	$POP	="ld";
+	$PUSH	="std";
+	$UCMP	="cmpld";
+	$SHL	="sldi";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T	=4;
+	$LRSAVE	=$SIZE_T;
+	$STU	="stwu";
+	$POP	="lwz";
+	$PUSH	="stw";
+	$UCMP	="cmplw";
+	$SHL	="slwi";
+} else { die "nonsense $flavour"; }
+
+$LITTLE_ENDIAN = ($flavour=~/le$/) ? $SIZE_T : 0;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour ".shift || die "can't call $xlate: $!";
+
+$FRAME=8*$SIZE_T;
+$prefix="aes_p10";
+
+$sp="r1";
+$vrsave="r12";
+
+#########################################################################
+{{{	# Key setup procedures						#
+my ($inp,$bits,$out,$ptr,$cnt,$rounds)=map("r$_",(3..8));
+my ($zero,$in0,$in1,$key,$rcon,$mask,$tmp)=map("v$_",(0..6));
+my ($stage,$outperm,$outmask,$outhead,$outtail)=map("v$_",(7..11));
+
+$code.=<<___;
+.machine	"any"
+
+.text
+
+.align	7
+rcon:
+.long	0x01000000, 0x01000000, 0x01000000, 0x01000000	?rev
+.long	0x1b000000, 0x1b000000, 0x1b000000, 0x1b000000	?rev
+.long	0x0d0e0f0c, 0x0d0e0f0c, 0x0d0e0f0c, 0x0d0e0f0c	?rev
+.long	0,0,0,0						?asis
+Lconsts:
+	mflr	r0
+	bcl	20,31,\$+4
+	mflr	$ptr	 #vvvvv "distance between . and rcon
+	addi	$ptr,$ptr,-0x48
+	mtlr	r0
+	blr
+	.long	0
+	.byte	0,12,0x14,0,0,0,0,0
+.asciz	"AES for PowerISA 2.07, CRYPTOGAMS by <appro\@openssl.org>"
+
+.globl	.${prefix}_set_encrypt_key
+Lset_encrypt_key:
+	mflr		r11
+	$PUSH		r11,$LRSAVE($sp)
+
+	li		$ptr,-1
+	${UCMP}i	$inp,0
+	beq-		Lenc_key_abort		# if ($inp==0) return -1;
+	${UCMP}i	$out,0
+	beq-		Lenc_key_abort		# if ($out==0) return -1;
+	li		$ptr,-2
+	cmpwi		$bits,128
+	blt-		Lenc_key_abort
+	cmpwi		$bits,256
+	bgt-		Lenc_key_abort
+	andi.		r0,$bits,0x3f
+	bne-		Lenc_key_abort
+
+	lis		r0,0xfff0
+	mfspr		$vrsave,256
+	mtspr		256,r0
+
+	bl		Lconsts
+	mtlr		r11
+
+	neg		r9,$inp
+	lvx		$in0,0,$inp
+	addi		$inp,$inp,15		# 15 is not typo
+	lvsr		$key,0,r9		# borrow $key
+	li		r8,0x20
+	cmpwi		$bits,192
+	lvx		$in1,0,$inp
+	le?vspltisb	$mask,0x0f		# borrow $mask
+	lvx		$rcon,0,$ptr
+	le?vxor		$key,$key,$mask		# adjust for byte swap
+	lvx		$mask,r8,$ptr
+	addi		$ptr,$ptr,0x10
+	vperm		$in0,$in0,$in1,$key	# align [and byte swap in LE]
+	li		$cnt,8
+	vxor		$zero,$zero,$zero
+	mtctr		$cnt
+
+	?lvsr		$outperm,0,$out
+	vspltisb	$outmask,-1
+	lvx		$outhead,0,$out
+	?vperm		$outmask,$zero,$outmask,$outperm
+
+	blt		Loop128
+	addi		$inp,$inp,8
+	beq		L192
+	addi		$inp,$inp,8
+	b		L256
+
+.align	4
+Loop128:
+	vperm		$key,$in0,$in0,$mask	# rotate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vcipherlast	$key,$key,$rcon
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	 vadduwm	$rcon,$rcon,$rcon
+	vxor		$in0,$in0,$key
+	bdnz		Loop128
+
+	lvx		$rcon,0,$ptr		# last two round keys
+
+	vperm		$key,$in0,$in0,$mask	# rotate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vcipherlast	$key,$key,$rcon
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	 vadduwm	$rcon,$rcon,$rcon
+	vxor		$in0,$in0,$key
+
+	vperm		$key,$in0,$in0,$mask	# rotate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vcipherlast	$key,$key,$rcon
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vxor		$in0,$in0,$key
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	 stvx		$stage,0,$out
+
+	addi		$inp,$out,15		# 15 is not typo
+	addi		$out,$out,0x50
+
+	li		$rounds,10
+	b		Ldone
+
+.align	4
+L192:
+	lvx		$tmp,0,$inp
+	li		$cnt,4
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+	vperm		$in1,$in1,$tmp,$key	# align [and byte swap in LE]
+	vspltisb	$key,8			# borrow $key
+	mtctr		$cnt
+	vsububm		$mask,$mask,$key	# adjust the mask
+
+Loop192:
+	vperm		$key,$in1,$in1,$mask	# roate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	vcipherlast	$key,$key,$rcon
+
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+
+	 vsldoi		$stage,$zero,$in1,8
+	vspltw		$tmp,$in0,3
+	vxor		$tmp,$tmp,$in1
+	vsldoi		$in1,$zero,$in1,12	# >>32
+	 vadduwm	$rcon,$rcon,$rcon
+	vxor		$in1,$in1,$tmp
+	vxor		$in0,$in0,$key
+	vxor		$in1,$in1,$key
+	 vsldoi		$stage,$stage,$in0,8
+
+	vperm		$key,$in1,$in1,$mask	# rotate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	 vperm		$outtail,$stage,$stage,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vcipherlast	$key,$key,$rcon
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	 vsldoi		$stage,$in0,$in1,8
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	 vperm		$outtail,$stage,$stage,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	vspltw		$tmp,$in0,3
+	vxor		$tmp,$tmp,$in1
+	vsldoi		$in1,$zero,$in1,12	# >>32
+	 vadduwm	$rcon,$rcon,$rcon
+	vxor		$in1,$in1,$tmp
+	vxor		$in0,$in0,$key
+	vxor		$in1,$in1,$key
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	 stvx		$stage,0,$out
+	 addi		$inp,$out,15		# 15 is not typo
+	 addi		$out,$out,16
+	bdnz		Loop192
+
+	li		$rounds,12
+	addi		$out,$out,0x20
+	b		Ldone
+
+.align	4
+L256:
+	lvx		$tmp,0,$inp
+	li		$cnt,7
+	li		$rounds,14
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+	vperm		$in1,$in1,$tmp,$key	# align [and byte swap in LE]
+	mtctr		$cnt
+
+Loop256:
+	vperm		$key,$in1,$in1,$mask	# rotate-n-splat
+	vsldoi		$tmp,$zero,$in0,12	# >>32
+	 vperm		$outtail,$in1,$in1,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	vcipherlast	$key,$key,$rcon
+	 stvx		$stage,0,$out
+	 addi		$out,$out,16
+
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in0,$in0,$tmp
+	 vadduwm	$rcon,$rcon,$rcon
+	vxor		$in0,$in0,$key
+	 vperm		$outtail,$in0,$in0,$outperm	# rotate
+	 vsel		$stage,$outhead,$outtail,$outmask
+	 vmr		$outhead,$outtail
+	 stvx		$stage,0,$out
+	 addi		$inp,$out,15		# 15 is not typo
+	 addi		$out,$out,16
+	bdz		Ldone
+
+	vspltw		$key,$in0,3		# just splat
+	vsldoi		$tmp,$zero,$in1,12	# >>32
+	vsbox		$key,$key
+
+	vxor		$in1,$in1,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in1,$in1,$tmp
+	vsldoi		$tmp,$zero,$tmp,12	# >>32
+	vxor		$in1,$in1,$tmp
+
+	vxor		$in1,$in1,$key
+	b		Loop256
+
+.align	4
+Ldone:
+	lvx		$in1,0,$inp		# redundant in aligned case
+	vsel		$in1,$outhead,$in1,$outmask
+	stvx		$in1,0,$inp
+	li		$ptr,0
+	mtspr		256,$vrsave
+	stw		$rounds,0($out)
+
+Lenc_key_abort:
+	mr		r3,$ptr
+	blr
+	.long		0
+	.byte		0,12,0x14,1,0,0,3,0
+	.long		0
+.size	.${prefix}_set_encrypt_key,.-.${prefix}_set_encrypt_key
+
+.globl	.${prefix}_set_decrypt_key
+	$STU		$sp,-$FRAME($sp)
+	mflr		r10
+	$PUSH		r10,$FRAME+$LRSAVE($sp)
+	bl		Lset_encrypt_key
+	mtlr		r10
+
+	cmpwi		r3,0
+	bne-		Ldec_key_abort
+
+	slwi		$cnt,$rounds,4
+	subi		$inp,$out,240		# first round key
+	srwi		$rounds,$rounds,1
+	add		$out,$inp,$cnt		# last round key
+	mtctr		$rounds
+
+Ldeckey:
+	lwz		r0, 0($inp)
+	lwz		r6, 4($inp)
+	lwz		r7, 8($inp)
+	lwz		r8, 12($inp)
+	addi		$inp,$inp,16
+	lwz		r9, 0($out)
+	lwz		r10,4($out)
+	lwz		r11,8($out)
+	lwz		r12,12($out)
+	stw		r0, 0($out)
+	stw		r6, 4($out)
+	stw		r7, 8($out)
+	stw		r8, 12($out)
+	subi		$out,$out,16
+	stw		r9, -16($inp)
+	stw		r10,-12($inp)
+	stw		r11,-8($inp)
+	stw		r12,-4($inp)
+	bdnz		Ldeckey
+
+	xor		r3,r3,r3		# return value
+Ldec_key_abort:
+	addi		$sp,$sp,$FRAME
+	blr
+	.long		0
+	.byte		0,12,4,1,0x80,0,3,0
+	.long		0
+.size	.${prefix}_set_decrypt_key,.-.${prefix}_set_decrypt_key
+___
+}}}
+#########################################################################
+{{{	# Single block en- and decrypt procedures			#
+sub gen_block () {
+my $dir = shift;
+my $n   = $dir eq "de" ? "n" : "";
+my ($inp,$out,$key,$rounds,$idx)=map("r$_",(3..7));
+
+$code.=<<___;
+.globl	.${prefix}_${dir}crypt
+	lwz		$rounds,240($key)
+	lis		r0,0xfc00
+	mfspr		$vrsave,256
+	li		$idx,15			# 15 is not typo
+	mtspr		256,r0
+
+	lvx		v0,0,$inp
+	neg		r11,$out
+	lvx		v1,$idx,$inp
+	lvsl		v2,0,$inp		# inpperm
+	le?vspltisb	v4,0x0f
+	?lvsl		v3,0,r11		# outperm
+	le?vxor		v2,v2,v4
+	li		$idx,16
+	vperm		v0,v0,v1,v2		# align [and byte swap in LE]
+	lvx		v1,0,$key
+	?lvsl		v5,0,$key		# keyperm
+	srwi		$rounds,$rounds,1
+	lvx		v2,$idx,$key
+	addi		$idx,$idx,16
+	subi		$rounds,$rounds,1
+	?vperm		v1,v1,v2,v5		# align round key
+
+	vxor		v0,v0,v1
+	lvx		v1,$idx,$key
+	addi		$idx,$idx,16
+	mtctr		$rounds
+
+Loop_${dir}c:
+	?vperm		v2,v2,v1,v5
+	v${n}cipher	v0,v0,v2
+	lvx		v2,$idx,$key
+	addi		$idx,$idx,16
+	?vperm		v1,v1,v2,v5
+	v${n}cipher	v0,v0,v1
+	lvx		v1,$idx,$key
+	addi		$idx,$idx,16
+	bdnz		Loop_${dir}c
+
+	?vperm		v2,v2,v1,v5
+	v${n}cipher	v0,v0,v2
+	lvx		v2,$idx,$key
+	?vperm		v1,v1,v2,v5
+	v${n}cipherlast	v0,v0,v1
+
+	vspltisb	v2,-1
+	vxor		v1,v1,v1
+	li		$idx,15			# 15 is not typo
+	?vperm		v2,v1,v2,v3		# outmask
+	le?vxor		v3,v3,v4
+	lvx		v1,0,$out		# outhead
+	vperm		v0,v0,v0,v3		# rotate [and byte swap in LE]
+	vsel		v1,v1,v0,v2
+	lvx		v4,$idx,$out
+	stvx		v1,0,$out
+	vsel		v0,v0,v4,v2
+	stvx		v0,$idx,$out
+
+	mtspr		256,$vrsave
+	blr
+	.long		0
+	.byte		0,12,0x14,0,0,0,3,0
+	.long		0
+.size	.${prefix}_${dir}crypt,.-.${prefix}_${dir}crypt
+___
+}
+&gen_block("en");
+&gen_block("de");
+}}}
+
+my $consts=1;
+foreach(split("\n",$code)) {
+        s/\`([^\`]*)\`/eval($1)/geo;
+
+	# constants table endian-specific conversion
+	if ($consts && m/\.(long|byte)\s+(.+)\s+(\?[a-z]*)$/o) {
+	    my $conv=$3;
+	    my @bytes=();
+
+	    # convert to endian-agnostic format
+	    if ($1 eq "long") {
+	      foreach (split(/,\s*/,$2)) {
+		my $l = /^0/?oct:int;
+		push @bytes,($l>>24)&0xff,($l>>16)&0xff,($l>>8)&0xff,$l&0xff;
+	      }
+	    } else {
+		@bytes = map(/^0/?oct:int,split(/,\s*/,$2));
+	    }
+
+	    # little-endian conversion
+	    if ($flavour =~ /le$/o) {
+		SWITCH: for($conv)  {
+		    /\?inv/ && do   { @bytes=map($_^0xf,@bytes); last; };
+		    /\?rev/ && do   { @bytes=reverse(@bytes);    last; };
+		}
+	    }
+
+	    #emit
+	    print ".byte\t",join(',',map (sprintf("0x%02x",$_),@bytes)),"\n";
+	    next;
+	}
+	$consts=0 if (m/Lconsts:/o);	# end of table
+
+	# instructions prefixed with '?' are endian-specific and need
+	# to be adjusted accordingly...
+	if ($flavour =~ /le$/o) {	# little-endian
+	    s/le\?//o		or
+	    s/be\?/#be#/o	or
+	    s/\?lvsr/lvsl/o	or
+	    s/\?lvsl/lvsr/o	or
+	    s/\?(vperm\s+v[0-9]+,\s*)(v[0-9]+,\s*)(v[0-9]+,\s*)(v[0-9]+)/$1$3$2$4/o or
+	    s/\?(vsldoi\s+v[0-9]+,\s*)(v[0-9]+,)\s*(v[0-9]+,\s*)([0-9]+)/$1$3$2 16-$4/o or
+	    s/\?(vspltw\s+v[0-9]+,\s*)(v[0-9]+,)\s*([0-9])/$1$2 3-$3/o;
+	} else {			# big-endian
+	    s/le\?/#le#/o	or
+	    s/be\?//o		or
+	    s/\?([a-z]+)/$1/o;
+	}
+
+        print $_,"\n";
+}
+
+close STDOUT;
diff --git a/arch/powerpc/crypto/aesp8-ppc.h b/arch/powerpc/crypto/aesp8-ppc.h
new file mode 100644
index 000000000..c68f5b696
--- /dev/null
+++ b/arch/powerpc/crypto/aesp8-ppc.h
@@ -0,0 +1,7 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <linux/types.h>
+#include <crypto/aes.h>
+
+extern struct skcipher_alg p8_aes_cbc_alg;
+extern struct skcipher_alg p8_aes_ctr_alg;
+extern struct skcipher_alg p8_aes_xts_alg;
diff --git a/arch/powerpc/crypto/ghashp10-ppc.pl b/arch/powerpc/crypto/ghashp10-ppc.pl
new file mode 100644
index 000000000..27a6b0bec
--- /dev/null
+++ b/arch/powerpc/crypto/ghashp10-ppc.pl
@@ -0,0 +1,370 @@
+#!/usr/bin/env perl
+# SPDX-License-Identifier: GPL-2.0
+
+# This code is taken from the OpenSSL project but the author (Andy Polyakov)
+# has relicensed it under the GPLv2. Therefore this program is free software;
+# you can redistribute it and/or modify it under the terms of the GNU General
+# Public License version 2 as published by the Free Software Foundation.
+#
+# The original headers, including the original license headers, are
+# included below for completeness.
+
+# ====================================================================
+# Written by Andy Polyakov <appro@openssl.org> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see https://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# GHASH for PowerISA v2.07.
+#
+# July 2014
+#
+# Accurate performance measurements are problematic, because it's
+# always virtualized setup with possibly throttled processor.
+# Relative comparison is therefore more informative. This initial
+# version is ~2.1x slower than hardware-assisted AES-128-CTR, ~12x
+# faster than "4-bit" integer-only compiler-generated 64-bit code.
+# "Initial version" means that there is room for futher improvement.
+
+$flavour=shift;
+$output =shift;
+
+if ($flavour =~ /64/) {
+	$SIZE_T=8;
+	$LRSAVE=2*$SIZE_T;
+	$STU="stdu";
+	$POP="ld";
+	$PUSH="std";
+} elsif ($flavour =~ /32/) {
+	$SIZE_T=4;
+	$LRSAVE=$SIZE_T;
+	$STU="stwu";
+	$POP="lwz";
+	$PUSH="stw";
+} else { die "nonsense $flavour"; }
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}ppc-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/ppc-xlate.pl" and -f $xlate) or
+die "can't locate ppc-xlate.pl";
+
+open STDOUT,"| $^X $xlate $flavour $output" || die "can't call $xlate: $!";
+
+my ($Xip,$Htbl,$inp,$len)=map("r$_",(3..6));	# argument block
+
+my ($Xl,$Xm,$Xh,$IN)=map("v$_",(0..3));
+my ($zero,$t0,$t1,$t2,$xC2,$H,$Hh,$Hl,$lemask)=map("v$_",(4..12));
+my ($Xl1,$Xm1,$Xh1,$IN1,$H2,$H2h,$H2l)=map("v$_",(13..19));
+my $vrsave="r12";
+my ($t4,$t5,$t6) = ($Hl,$H,$Hh);
+
+$code=<<___;
+.machine	"any"
+
+.text
+
+.globl	.gcm_init_p10
+	lis		r0,0xfff0
+	li		r8,0x10
+	mfspr		$vrsave,256
+	li		r9,0x20
+	mtspr		256,r0
+	li		r10,0x30
+	lvx_u		$H,0,r4			# load H
+	le?xor		r7,r7,r7
+	le?addi		r7,r7,0x8		# need a vperm start with 08
+	le?lvsr		5,0,r7
+	le?vspltisb	6,0x0f
+	le?vxor		5,5,6			# set a b-endian mask
+	le?vperm	$H,$H,$H,5
+
+	vspltisb	$xC2,-16		# 0xf0
+	vspltisb	$t0,1			# one
+	vaddubm		$xC2,$xC2,$xC2		# 0xe0
+	vxor		$zero,$zero,$zero
+	vor		$xC2,$xC2,$t0		# 0xe1
+	vsldoi		$xC2,$xC2,$zero,15	# 0xe1...
+	vsldoi		$t1,$zero,$t0,1		# ...1
+	vaddubm		$xC2,$xC2,$xC2		# 0xc2...
+	vspltisb	$t2,7
+	vor		$xC2,$xC2,$t1		# 0xc2....01
+	vspltb		$t1,$H,0		# most significant byte
+	vsl		$H,$H,$t0		# H<<=1
+	vsrab		$t1,$t1,$t2		# broadcast carry bit
+	vand		$t1,$t1,$xC2
+	vxor		$H,$H,$t1		# twisted H
+
+	vsldoi		$H,$H,$H,8		# twist even more ...
+	vsldoi		$xC2,$zero,$xC2,8	# 0xc2.0
+	vsldoi		$Hl,$zero,$H,8		# ... and split
+	vsldoi		$Hh,$H,$zero,8
+
+	stvx_u		$xC2,0,r3		# save pre-computed table
+	stvx_u		$Hl,r8,r3
+	stvx_u		$H, r9,r3
+	stvx_u		$Hh,r10,r3
+
+	mtspr		256,$vrsave
+	blr
+	.long		0
+	.byte		0,12,0x14,0,0,0,2,0
+	.long		0
+.size	.gcm_init_p10,.-.gcm_init_p10
+
+.globl	.gcm_init_htable
+	lis		r0,0xfff0
+	li		r8,0x10
+	mfspr		$vrsave,256
+	li		r9,0x20
+	mtspr		256,r0
+	li		r10,0x30
+	lvx_u		$H,0,r4			# load H
+
+	vspltisb	$xC2,-16		# 0xf0
+	vspltisb	$t0,1			# one
+	vaddubm		$xC2,$xC2,$xC2		# 0xe0
+	vxor		$zero,$zero,$zero
+	vor		$xC2,$xC2,$t0		# 0xe1
+	vsldoi		$xC2,$xC2,$zero,15	# 0xe1...
+	vsldoi		$t1,$zero,$t0,1		# ...1
+	vaddubm		$xC2,$xC2,$xC2		# 0xc2...
+	vspltisb	$t2,7
+	vor		$xC2,$xC2,$t1		# 0xc2....01
+	vspltb		$t1,$H,0		# most significant byte
+	vsl		$H,$H,$t0		# H<<=1
+	vsrab		$t1,$t1,$t2		# broadcast carry bit
+	vand		$t1,$t1,$xC2
+	vxor		$IN,$H,$t1		# twisted H
+
+	vsldoi		$H,$IN,$IN,8		# twist even more ...
+	vsldoi		$xC2,$zero,$xC2,8	# 0xc2.0
+	vsldoi		$Hl,$zero,$H,8		# ... and split
+	vsldoi		$Hh,$H,$zero,8
+
+	stvx_u		$xC2,0,r3		# save pre-computed table
+	stvx_u		$Hl,r8,r3
+	li		r8,0x40
+	stvx_u		$H, r9,r3
+	li		r9,0x50
+	stvx_u		$Hh,r10,r3
+	li		r10,0x60
+
+	vpmsumd		$Xl,$IN,$Hl		# H.lo·H.lo
+	vpmsumd		$Xm,$IN,$H		# H.hi·H.lo+H.lo·H.hi
+	vpmsumd		$Xh,$IN,$Hh		# H.hi·H.hi
+
+	vpmsumd		$t2,$Xl,$xC2		# 1st reduction phase
+
+	vsldoi		$t0,$Xm,$zero,8
+	vsldoi		$t1,$zero,$Xm,8
+	vxor		$Xl,$Xl,$t0
+	vxor		$Xh,$Xh,$t1
+
+	vsldoi		$Xl,$Xl,$Xl,8
+	vxor		$Xl,$Xl,$t2
+
+	vsldoi		$t1,$Xl,$Xl,8		# 2nd reduction phase
+	vpmsumd		$Xl,$Xl,$xC2
+	vxor		$t1,$t1,$Xh
+	vxor		$IN1,$Xl,$t1
+
+	vsldoi		$H2,$IN1,$IN1,8
+	vsldoi		$H2l,$zero,$H2,8
+	vsldoi		$H2h,$H2,$zero,8
+
+	stvx_u		$H2l,r8,r3		# save H^2
+	li		r8,0x70
+	stvx_u		$H2,r9,r3
+	li		r9,0x80
+	stvx_u		$H2h,r10,r3
+	li		r10,0x90
+
+	vpmsumd		$Xl,$IN,$H2l		# H.lo·H^2.lo
+	 vpmsumd	$Xl1,$IN1,$H2l		# H^2.lo·H^2.lo
+	vpmsumd		$Xm,$IN,$H2		# H.hi·H^2.lo+H.lo·H^2.hi
+	 vpmsumd	$Xm1,$IN1,$H2		# H^2.hi·H^2.lo+H^2.lo·H^2.hi
+	vpmsumd		$Xh,$IN,$H2h		# H.hi·H^2.hi
+	 vpmsumd	$Xh1,$IN1,$H2h		# H^2.hi·H^2.hi
+
+	vpmsumd		$t2,$Xl,$xC2		# 1st reduction phase
+	 vpmsumd	$t6,$Xl1,$xC2		# 1st reduction phase
+
+	vsldoi		$t0,$Xm,$zero,8
+	vsldoi		$t1,$zero,$Xm,8
+	 vsldoi		$t4,$Xm1,$zero,8
+	 vsldoi		$t5,$zero,$Xm1,8
+	vxor		$Xl,$Xl,$t0
+	vxor		$Xh,$Xh,$t1
+	 vxor		$Xl1,$Xl1,$t4
+	 vxor		$Xh1,$Xh1,$t5
+
+	vsldoi		$Xl,$Xl,$Xl,8
+	 vsldoi		$Xl1,$Xl1,$Xl1,8
+	vxor		$Xl,$Xl,$t2
+	 vxor		$Xl1,$Xl1,$t6
+
+	vsldoi		$t1,$Xl,$Xl,8		# 2nd reduction phase
+	 vsldoi		$t5,$Xl1,$Xl1,8		# 2nd reduction phase
+	vpmsumd		$Xl,$Xl,$xC2
+	 vpmsumd	$Xl1,$Xl1,$xC2
+	vxor		$t1,$t1,$Xh
+	 vxor		$t5,$t5,$Xh1
+	vxor		$Xl,$Xl,$t1
+	 vxor		$Xl1,$Xl1,$t5
+
+	vsldoi		$H,$Xl,$Xl,8
+	 vsldoi		$H2,$Xl1,$Xl1,8
+	vsldoi		$Hl,$zero,$H,8
+	vsldoi		$Hh,$H,$zero,8
+	 vsldoi		$H2l,$zero,$H2,8
+	 vsldoi		$H2h,$H2,$zero,8
+
+	stvx_u		$Hl,r8,r3		# save H^3
+	li		r8,0xa0
+	stvx_u		$H,r9,r3
+	li		r9,0xb0
+	stvx_u		$Hh,r10,r3
+	li		r10,0xc0
+	 stvx_u		$H2l,r8,r3		# save H^4
+	 stvx_u		$H2,r9,r3
+	 stvx_u		$H2h,r10,r3
+
+	mtspr		256,$vrsave
+	blr
+	.long		0
+	.byte		0,12,0x14,0,0,0,2,0
+	.long		0
+.size	.gcm_init_htable,.-.gcm_init_htable
+
+.globl	.gcm_gmult_p10
+	lis		r0,0xfff8
+	li		r8,0x10
+	mfspr		$vrsave,256
+	li		r9,0x20
+	mtspr		256,r0
+	li		r10,0x30
+	lvx_u		$IN,0,$Xip		# load Xi
+
+	lvx_u		$Hl,r8,$Htbl		# load pre-computed table
+	 le?lvsl	$lemask,r0,r0
+	lvx_u		$H, r9,$Htbl
+	 le?vspltisb	$t0,0x07
+	lvx_u		$Hh,r10,$Htbl
+	 le?vxor	$lemask,$lemask,$t0
+	lvx_u		$xC2,0,$Htbl
+	 le?vperm	$IN,$IN,$IN,$lemask
+	vxor		$zero,$zero,$zero
+
+	vpmsumd		$Xl,$IN,$Hl		# H.lo·Xi.lo
+	vpmsumd		$Xm,$IN,$H		# H.hi·Xi.lo+H.lo·Xi.hi
+	vpmsumd		$Xh,$IN,$Hh		# H.hi·Xi.hi
+
+	vpmsumd		$t2,$Xl,$xC2		# 1st phase
+
+	vsldoi		$t0,$Xm,$zero,8
+	vsldoi		$t1,$zero,$Xm,8
+	vxor		$Xl,$Xl,$t0
+	vxor		$Xh,$Xh,$t1
+
+	vsldoi		$Xl,$Xl,$Xl,8
+	vxor		$Xl,$Xl,$t2
+
+	vsldoi		$t1,$Xl,$Xl,8		# 2nd phase
+	vpmsumd		$Xl,$Xl,$xC2
+	vxor		$t1,$t1,$Xh
+	vxor		$Xl,$Xl,$t1
+
+	le?vperm	$Xl,$Xl,$Xl,$lemask
+	stvx_u		$Xl,0,$Xip		# write out Xi
+
+	mtspr		256,$vrsave
+	blr
+	.long		0
+	.byte		0,12,0x14,0,0,0,2,0
+	.long		0
+.size	.gcm_gmult_p10,.-.gcm_gmult_p10
+
+.globl	.gcm_ghash_p10
+	lis		r0,0xfff8
+	li		r8,0x10
+	mfspr		$vrsave,256
+	li		r9,0x20
+	mtspr		256,r0
+	li		r10,0x30
+	lvx_u		$Xl,0,$Xip		# load Xi
+
+	lvx_u		$Hl,r8,$Htbl		# load pre-computed table
+	 le?lvsl	$lemask,r0,r0
+	lvx_u		$H, r9,$Htbl
+	 le?vspltisb	$t0,0x07
+	lvx_u		$Hh,r10,$Htbl
+	 le?vxor	$lemask,$lemask,$t0
+	lvx_u		$xC2,0,$Htbl
+	 le?vperm	$Xl,$Xl,$Xl,$lemask
+	vxor		$zero,$zero,$zero
+
+	lvx_u		$IN,0,$inp
+	addi		$inp,$inp,16
+	subi		$len,$len,16
+	 le?vperm	$IN,$IN,$IN,$lemask
+	vxor		$IN,$IN,$Xl
+	b		Loop
+
+.align	5
+Loop:
+	 subic		$len,$len,16
+	vpmsumd		$Xl,$IN,$Hl		# H.lo·Xi.lo
+	 subfe.		r0,r0,r0		# borrow?-1:0
+	vpmsumd		$Xm,$IN,$H		# H.hi·Xi.lo+H.lo·Xi.hi
+	 and		r0,r0,$len
+	vpmsumd		$Xh,$IN,$Hh		# H.hi·Xi.hi
+	 add		$inp,$inp,r0
+
+	vpmsumd		$t2,$Xl,$xC2		# 1st phase
+
+	vsldoi		$t0,$Xm,$zero,8
+	vsldoi		$t1,$zero,$Xm,8
+	vxor		$Xl,$Xl,$t0
+	vxor		$Xh,$Xh,$t1
+
+	vsldoi		$Xl,$Xl,$Xl,8
+	vxor		$Xl,$Xl,$t2
+	 lvx_u		$IN,0,$inp
+	 addi		$inp,$inp,16
+
+	vsldoi		$t1,$Xl,$Xl,8		# 2nd phase
+	vpmsumd		$Xl,$Xl,$xC2
+	 le?vperm	$IN,$IN,$IN,$lemask
+	vxor		$t1,$t1,$Xh
+	vxor		$IN,$IN,$t1
+	vxor		$IN,$IN,$Xl
+	beq		Loop			# did $len-=16 borrow?
+
+	vxor		$Xl,$Xl,$t1
+	le?vperm	$Xl,$Xl,$Xl,$lemask
+	stvx_u		$Xl,0,$Xip		# write out Xi
+
+	mtspr		256,$vrsave
+	blr
+	.long		0
+	.byte		0,12,0x14,0,0,0,4,0
+	.long		0
+.size	.gcm_ghash_p10,.-.gcm_ghash_p10
+
+.asciz  "GHASH for PowerISA 2.07, CRYPTOGAMS by <appro\@openssl.org>"
+.align  2
+___
+
+foreach (split("\n",$code)) {
+	if ($flavour =~ /le$/o) {	# little-endian
+	    s/le\?//o		or
+	    s/be\?/#be#/o;
+	} else {
+	    s/le\?/#le#/o	or
+	    s/be\?//o;
+	}
+	print $_,"\n";
+}
+
+close STDOUT; # enforce flush
diff --git a/arch/powerpc/crypto/ppc-xlate.pl b/arch/powerpc/crypto/ppc-xlate.pl
new file mode 100644
index 000000000..23cca703c
--- /dev/null
+++ b/arch/powerpc/crypto/ppc-xlate.pl
@@ -0,0 +1,229 @@
+#!/usr/bin/env perl
+# SPDX-License-Identifier: GPL-2.0
+
+# PowerPC assembler distiller by <appro>.
+
+my $flavour = shift;
+my $output = shift;
+open STDOUT,">$output" || die "can't open $output: $!";
+
+my %GLOBALS;
+my $dotinlocallabels=($flavour=~/linux/)?1:0;
+
+################################################################
+# directives which need special treatment on different platforms
+################################################################
+my $globl = sub {
+    my $junk = shift;
+    my $name = shift;
+    my $global = \$GLOBALS{$name};
+    my $ret;
+
+    $name =~ s|^[\.\_]||;
+
+    SWITCH: for ($flavour) {
+	/aix/		&& do { $name = ".$name";
+				last;
+			      };
+	/osx/		&& do { $name = "_$name";
+				last;
+			      };
+	/linux/
+			&& do {	$ret = "_GLOBAL($name)";
+				last;
+			      };
+    }
+
+    $ret = ".globl	$name\nalign 5\n$name:" if (!$ret);
+    $$global = $name;
+    $ret;
+};
+my $text = sub {
+    my $ret = ($flavour =~ /aix/) ? ".csect\t.text[PR],7" : ".text";
+    $ret = ".abiversion	2\n".$ret	if ($flavour =~ /linux.*64le/);
+    $ret;
+};
+my $machine = sub {
+    my $junk = shift;
+    my $arch = shift;
+    if ($flavour =~ /osx/)
+    {	$arch =~ s/\"//g;
+	$arch = ($flavour=~/64/) ? "ppc970-64" : "ppc970" if ($arch eq "any");
+    }
+    ".machine	$arch";
+};
+my $size = sub {
+    if ($flavour =~ /linux/)
+    {	shift;
+	my $name = shift; $name =~ s|^[\.\_]||;
+	my $ret  = ".size	$name,.-".($flavour=~/64$/?".":"").$name;
+	$ret .= "\n.size	.$name,.-.$name" if ($flavour=~/64$/);
+	$ret;
+    }
+    else
+    {	"";	}
+};
+my $asciz = sub {
+    shift;
+    my $line = join(",",@_);
+    if ($line =~ /^"(.*)"$/)
+    {	".byte	" . join(",",unpack("C*",$1),0) . "\n.align	2";	}
+    else
+    {	"";	}
+};
+my $quad = sub {
+    shift;
+    my @ret;
+    my ($hi,$lo);
+    for (@_) {
+	if (/^0x([0-9a-f]*?)([0-9a-f]{1,8})$/io)
+	{  $hi=$1?"0x$1":"0"; $lo="0x$2";  }
+	elsif (/^([0-9]+)$/o)
+	{  $hi=$1>>32; $lo=$1&0xffffffff;  } # error-prone with 32-bit perl
+	else
+	{  $hi=undef; $lo=$_; }
+
+	if (defined($hi))
+	{  push(@ret,$flavour=~/le$/o?".long\t$lo,$hi":".long\t$hi,$lo");  }
+	else
+	{  push(@ret,".quad	$lo");  }
+    }
+    join("\n",@ret);
+};
+
+################################################################
+# simplified mnemonics not handled by at least one assembler
+################################################################
+my $cmplw = sub {
+    my $f = shift;
+    my $cr = 0; $cr = shift if ($#_>1);
+    # Some out-of-date 32-bit GNU assembler just can't handle cmplw...
+    ($flavour =~ /linux.*32/) ?
+	"	.long	".sprintf "0x%x",31<<26|$cr<<23|$_[0]<<16|$_[1]<<11|64 :
+	"	cmplw	".join(',',$cr,@_);
+};
+my $bdnz = sub {
+    my $f = shift;
+    my $bo = $f=~/[\+\-]/ ? 16+9 : 16;	# optional "to be taken" hint
+    "	bc	$bo,0,".shift;
+} if ($flavour!~/linux/);
+my $bltlr = sub {
+    my $f = shift;
+    my $bo = $f=~/\-/ ? 12+2 : 12;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%x",19<<26|$bo<<21|16<<1 :
+	"	bclr	$bo,0";
+};
+my $bnelr = sub {
+    my $f = shift;
+    my $bo = $f=~/\-/ ? 4+2 : 4;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%x",19<<26|$bo<<21|2<<16|16<<1 :
+	"	bclr	$bo,2";
+};
+my $beqlr = sub {
+    my $f = shift;
+    my $bo = $f=~/-/ ? 12+2 : 12;	# optional "not to be taken" hint
+    ($flavour =~ /linux/) ?		# GNU as doesn't allow most recent hints
+	"	.long	".sprintf "0x%X",19<<26|$bo<<21|2<<16|16<<1 :
+	"	bclr	$bo,2";
+};
+# GNU assembler can't handle extrdi rA,rS,16,48, or when sum of last two
+# arguments is 64, with "operand out of range" error.
+my $extrdi = sub {
+    my ($f,$ra,$rs,$n,$b) = @_;
+    $b = ($b+$n)&63; $n = 64-$n;
+    "	rldicl	$ra,$rs,$b,$n";
+};
+my $vmr = sub {
+    my ($f,$vx,$vy) = @_;
+    "	vor	$vx,$vy,$vy";
+};
+
+# Some ABIs specify vrsave, special-purpose register #256, as reserved
+# for system use.
+my $no_vrsave = ($flavour =~ /linux-ppc64le/);
+my $mtspr = sub {
+    my ($f,$idx,$ra) = @_;
+    if ($idx == 256 && $no_vrsave) {
+	"	or	$ra,$ra,$ra";
+    } else {
+	"	mtspr	$idx,$ra";
+    }
+};
+my $mfspr = sub {
+    my ($f,$rd,$idx) = @_;
+    if ($idx == 256 && $no_vrsave) {
+	"	li	$rd,-1";
+    } else {
+	"	mfspr	$rd,$idx";
+    }
+};
+
+# PowerISA 2.06 stuff
+sub vsxmem_op {
+    my ($f, $vrt, $ra, $rb, $op) = @_;
+    "	.long	".sprintf "0x%X",(31<<26)|($vrt<<21)|($ra<<16)|($rb<<11)|($op*2+1);
+}
+# made-up unaligned memory reference AltiVec/VMX instructions
+my $lvx_u	= sub {	vsxmem_op(@_, 844); };	# lxvd2x
+my $stvx_u	= sub {	vsxmem_op(@_, 972); };	# stxvd2x
+my $lvdx_u	= sub {	vsxmem_op(@_, 588); };	# lxsdx
+my $stvdx_u	= sub {	vsxmem_op(@_, 716); };	# stxsdx
+my $lvx_4w	= sub { vsxmem_op(@_, 780); };	# lxvw4x
+my $stvx_4w	= sub { vsxmem_op(@_, 908); };	# stxvw4x
+
+# PowerISA 2.07 stuff
+sub vcrypto_op {
+    my ($f, $vrt, $vra, $vrb, $op) = @_;
+    "	.long	".sprintf "0x%X",(4<<26)|($vrt<<21)|($vra<<16)|($vrb<<11)|$op;
+}
+my $vcipher	= sub { vcrypto_op(@_, 1288); };
+my $vcipherlast	= sub { vcrypto_op(@_, 1289); };
+my $vncipher	= sub { vcrypto_op(@_, 1352); };
+my $vncipherlast= sub { vcrypto_op(@_, 1353); };
+my $vsbox	= sub { vcrypto_op(@_, 0, 1480); };
+my $vshasigmad	= sub { my ($st,$six)=splice(@_,-2); vcrypto_op(@_, $st<<4|$six, 1730); };
+my $vshasigmaw	= sub { my ($st,$six)=splice(@_,-2); vcrypto_op(@_, $st<<4|$six, 1666); };
+my $vpmsumb	= sub { vcrypto_op(@_, 1032); };
+my $vpmsumd	= sub { vcrypto_op(@_, 1224); };
+my $vpmsubh	= sub { vcrypto_op(@_, 1096); };
+my $vpmsumw	= sub { vcrypto_op(@_, 1160); };
+my $vaddudm	= sub { vcrypto_op(@_, 192);  };
+my $vadduqm	= sub { vcrypto_op(@_, 256);  };
+
+my $mtsle	= sub {
+    my ($f, $arg) = @_;
+    "	.long	".sprintf "0x%X",(31<<26)|($arg<<21)|(147*2);
+};
+
+print "#include <asm/ppc_asm.h>\n" if $flavour =~ /linux/;
+
+while($line=<>) {
+
+    $line =~ s|[#!;].*$||;	# get rid of asm-style comments...
+    $line =~ s|/\*.*\*/||;	# ... and C-style comments...
+    $line =~ s|^\s+||;		# ... and skip white spaces in beginning...
+    $line =~ s|\s+$||;		# ... and at the end
+
+    {
+	$line =~ s|\b\.L(\w+)|L$1|g;	# common denominator for Locallabel
+	$line =~ s|\bL(\w+)|\.L$1|g	if ($dotinlocallabels);
+    }
+
+    {
+	$line =~ s|^\s*(\.?)(\w+)([\.\+\-]?)\s*||;
+	my $c = $1; $c = "\t" if ($c eq "");
+	my $mnemonic = $2;
+	my $f = $3;
+	my $opcode = eval("\$$mnemonic");
+	$line =~ s/\b(c?[rf]|v|vs)([0-9]+)\b/$2/g if ($c ne "." and $flavour !~ /osx/);
+	if (ref($opcode) eq 'CODE') { $line = &$opcode($f,split(',',$line)); }
+	elsif ($mnemonic)           { $line = $c.$mnemonic.$f."\t".$line; }
+    }
+
+    print $line if ($line);
+    print "\n";
+}
+
+close STDOUT;
diff --git a/arch/powerpc/crypto/vmx.c b/arch/powerpc/crypto/vmx.c
new file mode 100644
index 000000000..08da5311d
--- /dev/null
+++ b/arch/powerpc/crypto/vmx.c
@@ -0,0 +1,61 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Routines supporting VMX instructions on the Power 8
+ *
+ * Copyright (C) 2015 International Business Machines Inc.
+ *
+ * Author: Marcelo Henrique Cerri <mhcerri@br.ibm.com>
+ */
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/types.h>
+#include <linux/err.h>
+#include <linux/cpufeature.h>
+#include <linux/crypto.h>
+#include <asm/cputable.h>
+#include <crypto/internal/skcipher.h>
+
+#include "aesp8-ppc.h"
+
+static int __init p8_init(void)
+{
+	int ret;
+
+	ret = crypto_register_skcipher(&p8_aes_cbc_alg);
+	if (ret)
+		goto err;
+
+	ret = crypto_register_skcipher(&p8_aes_ctr_alg);
+	if (ret)
+		goto err_unregister_aes_cbc;
+
+	ret = crypto_register_skcipher(&p8_aes_xts_alg);
+	if (ret)
+		goto err_unregister_aes_ctr;
+
+	return 0;
+
+err_unregister_aes_ctr:
+	crypto_unregister_skcipher(&p8_aes_ctr_alg);
+err_unregister_aes_cbc:
+	crypto_unregister_skcipher(&p8_aes_cbc_alg);
+err:
+	return ret;
+}
+
+static void __exit p8_exit(void)
+{
+	crypto_unregister_skcipher(&p8_aes_xts_alg);
+	crypto_unregister_skcipher(&p8_aes_ctr_alg);
+	crypto_unregister_skcipher(&p8_aes_cbc_alg);
+}
+
+module_cpu_feature_match(PPC_MODULE_FEATURE_VEC_CRYPTO, p8_init);
+module_exit(p8_exit);
+
+MODULE_AUTHOR("Marcelo Cerri<mhcerri@br.ibm.com>");
+MODULE_DESCRIPTION("IBM VMX cryptographic acceleration instructions "
+		   "support on Power 8");
+MODULE_LICENSE("GPL");
+MODULE_VERSION("1.0.0");