aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--xen/Rules.mk1
-rw-r--r--xen/arch/x86/acpi/power.c8
-rw-r--r--xen/arch/x86/tboot.c235
-rw-r--r--xen/crypto/Makefile2
-rw-r--r--xen/crypto/rijndael.c1269
-rw-r--r--xen/crypto/vmac.c1220
-rw-r--r--xen/include/asm-x86/tboot.h12
-rw-r--r--xen/include/crypto/rijndael.h58
-rw-r--r--xen/include/crypto/vmac.h178
9 files changed, 2956 insertions, 27 deletions
diff --git a/xen/Rules.mk b/xen/Rules.mk
index 3ee2f08593..5601e45191 100644
--- a/xen/Rules.mk
+++ b/xen/Rules.mk
@@ -38,6 +38,7 @@ ALL_OBJS-y += $(BASEDIR)/common/built_in.o
ALL_OBJS-y += $(BASEDIR)/drivers/built_in.o
ALL_OBJS-y += $(BASEDIR)/xsm/built_in.o
ALL_OBJS-y += $(BASEDIR)/arch/$(TARGET_ARCH)/built_in.o
+ALL_OBJS-$(x86) += $(BASEDIR)/crypto/built_in.o
CFLAGS-y += -g -D__XEN__
CFLAGS-$(XSM_ENABLE) += -DXSM_ENABLE
diff --git a/xen/arch/x86/acpi/power.c b/xen/arch/x86/acpi/power.c
index 4c6f76d61e..1e0311c326 100644
--- a/xen/arch/x86/acpi/power.c
+++ b/xen/arch/x86/acpi/power.c
@@ -190,6 +190,7 @@ static int enter_state(u32 state)
case ACPI_STATE_S3:
do_suspend_lowlevel();
system_reset_counter++;
+ error = tboot_s3_resume();
break;
case ACPI_STATE_S5:
acpi_enter_sleep_state(ACPI_STATE_S5);
@@ -206,7 +207,10 @@ static int enter_state(u32 state)
device_power_up();
- printk(XENLOG_INFO "Finishing wakeup from ACPI S%d state.", state);
+ printk(XENLOG_INFO "Finishing wakeup from ACPI S%d state.\n", state);
+
+ if ( (state == ACPI_STATE_S3) && error )
+ panic("Memory integrity was lost on resume (%d)\n", error);
done:
spin_debug_enable();
@@ -318,7 +322,7 @@ static void tboot_sleep(u8 sleep_state)
tboot_shutdown(shutdown_type);
}
-
+
/* System is really put into sleep state by this stub */
acpi_status asmlinkage acpi_enter_sleep_state(u8 sleep_state)
{
diff --git a/xen/arch/x86/tboot.c b/xen/arch/x86/tboot.c
index 92f509ae95..ed20e9951f 100644
--- a/xen/arch/x86/tboot.c
+++ b/xen/arch/x86/tboot.c
@@ -3,11 +3,14 @@
#include <xen/types.h>
#include <xen/lib.h>
#include <xen/sched.h>
+#include <xen/domain_page.h>
+#include <xen/iommu.h>
#include <asm/fixmap.h>
#include <asm/page.h>
#include <asm/processor.h>
#include <asm/e820.h>
#include <asm/tboot.h>
+#include <crypto/vmac.h>
/* tboot=<physical address of shared page> */
static char opt_tboot[20] = "";
@@ -16,6 +19,10 @@ string_param("tboot", opt_tboot);
/* Global pointer to shared data; NULL means no measured launch. */
tboot_shared_t *g_tboot_shared;
+static vmac_t domain_mac; /* MAC for all domains during S3 */
+static vmac_t xenheap_mac; /* MAC for xen heap during S3 */
+static vmac_t frametable_mac; /* MAC for frame table during S3 */
+
static const uuid_t tboot_shared_uuid = TBOOT_SHARED_UUID;
/* used by tboot_protect_mem_regions() and/or tboot_parse_dmar_table() */
@@ -40,6 +47,7 @@ static uint64_t sinit_base, sinit_size;
#define TXTCR_HEAP_SIZE 0x0308
extern char __init_begin[], __per_cpu_start[], __per_cpu_end[], __bss_start[];
+extern unsigned long allocator_bitmap_end;
#define SHA1_SIZE 20
typedef uint8_t sha1_hash_t[SHA1_SIZE];
@@ -82,8 +90,9 @@ void __init tboot_probe(void)
if ( memcmp(&tboot_shared_uuid, (uuid_t *)tboot_shared, sizeof(uuid_t)) )
return;
- /* new tboot_shared (w/ GAS support) is not backwards compatible */
- if ( tboot_shared->version < 3 ) {
+ /* new tboot_shared (w/ GAS support, integrity, etc.) is not backwards
+ compatible */
+ if ( tboot_shared->version < 4 ) {
printk("unsupported version of tboot (%u)\n", tboot_shared->version);
return;
}
@@ -121,6 +130,143 @@ void __init tboot_probe(void)
(unsigned long)__va((map_base + map_size) << PAGE_SHIFT));
}
+/* definitions from xen/drivers/passthrough/vtd/iommu.h
+ * used to walk through vtd page tables */
+#define LEVEL_STRIDE (9)
+#define PTE_NUM (1<<LEVEL_STRIDE)
+#define dma_pte_present(p) (((p).val & 3) != 0)
+#define dma_pte_addr(p) ((p).val & PAGE_MASK_4K)
+#define agaw_to_level(val) ((val)+2)
+struct dma_pte {
+ u64 val;
+};
+
+static void update_iommu_mac(vmac_ctx_t *ctx, uint64_t pt_maddr, int level)
+{
+ int i;
+ struct dma_pte *pt_vaddr, *pte;
+ int next_level = level - 1;
+
+ if ( pt_maddr == 0 )
+ return;
+
+ pt_vaddr = (struct dma_pte *)map_domain_page(pt_maddr >> PAGE_SHIFT_4K);
+ vmac_update((void *)pt_vaddr, PAGE_SIZE, ctx);
+
+ for ( i = 0; i < PTE_NUM; i++ )
+ {
+ pte = &pt_vaddr[i];
+ if ( !dma_pte_present(*pte) )
+ continue;
+
+ if ( next_level >= 1 )
+ update_iommu_mac(ctx, dma_pte_addr(*pte), next_level);
+ }
+
+ unmap_domain_page(pt_vaddr);
+}
+
+#define is_page_in_use(page) \
+ ((page->count_info & PGC_count_mask) != 0 || page->count_info == 0)
+
+static void update_pagetable_mac(vmac_ctx_t *ctx)
+{
+ unsigned long mfn;
+
+ for ( mfn = 0; mfn < max_page; mfn++ )
+ {
+ struct page_info *page = mfn_to_page(mfn);
+ if ( is_page_in_use(page) && !is_xen_heap_page(page) ) {
+ if ( page->count_info & PGC_page_table ) {
+ void *pg = map_domain_page(mfn);
+ vmac_update(pg, PAGE_SIZE, ctx);
+ unmap_domain_page(pg);
+ }
+ }
+ }
+}
+
+static void tboot_gen_domain_integrity(const uint8_t key[TB_KEY_SIZE],
+ vmac_t *mac)
+{
+ struct domain *d;
+ struct page_info *page;
+ uint8_t nonce[16] = {};
+ vmac_ctx_t ctx;
+
+ vmac_set_key((uint8_t *)key, &ctx);
+ for_each_domain( d )
+ {
+ if ( !d->arch.s3_integrity )
+ continue;
+ printk("MACing Domain %u\n", d->domain_id);
+
+ page_list_for_each(page, &d->page_list)
+ {
+ void *pg;
+ pg = map_domain_page(page_to_mfn(page));
+ vmac_update(pg, PAGE_SIZE, &ctx);
+ unmap_domain_page(pg);
+ }
+
+ if ( !is_idle_domain(d) )
+ {
+ struct hvm_iommu *hd = domain_hvm_iommu(d);
+ update_iommu_mac(&ctx, hd->pgd_maddr, agaw_to_level(hd->agaw));
+ }
+ }
+
+ /* MAC all shadow page tables */
+ update_pagetable_mac(&ctx);
+
+ *mac = vmac(NULL, 0, nonce, NULL, &ctx);
+
+ printk("MAC for domains is: 0x%08"PRIx64"\n", *mac);
+
+ /* wipe ctx to ensure key is not left in memory */
+ memset(&ctx, 0, sizeof(ctx));
+}
+
+static void tboot_gen_xenheap_integrity(const uint8_t key[TB_KEY_SIZE],
+ vmac_t *mac)
+{
+ unsigned long mfn;
+ uint8_t nonce[16] = {};
+ vmac_ctx_t ctx;
+
+ vmac_set_key((uint8_t *)key, &ctx);
+ for ( mfn = 0; mfn < max_page; mfn++ )
+ {
+ struct page_info *page = __mfn_to_page(mfn);
+ if ( is_page_in_use(page) && is_xen_heap_page(page) ) {
+ void *pg = mfn_to_virt(mfn);
+ vmac_update((uint8_t *)pg, PAGE_SIZE, &ctx);
+ }
+ }
+ *mac = vmac(NULL, 0, nonce, NULL, &ctx);
+
+ printk("MAC for xenheap is: 0x%08"PRIx64"\n", *mac);
+
+ /* wipe ctx to ensure key is not left in memory */
+ memset(&ctx, 0, sizeof(ctx));
+}
+
+static void tboot_gen_frametable_integrity(const uint8_t key[TB_KEY_SIZE],
+ vmac_t *mac)
+{
+ uint8_t nonce[16] = {};
+ vmac_ctx_t ctx;
+
+ vmac_set_key((uint8_t *)key, &ctx);
+ *mac = vmac((uint8_t *)frame_table,
+ PFN_UP(max_page * sizeof(*frame_table)), nonce, NULL, &ctx);
+
+ printk("MAC for frametable is: 0x%08"PRIx64"\n", *mac);
+
+ /* wipe ctx to ensure key is not left in memory */
+ memset(&ctx, 0, sizeof(ctx));
+}
+
void tboot_shutdown(uint32_t shutdown_type)
{
uint32_t map_base, map_size;
@@ -130,38 +276,60 @@ void tboot_shutdown(uint32_t shutdown_type)
local_irq_disable();
- /* if this is S3 then set regions to MAC */
- if ( shutdown_type == TB_SHUTDOWN_S3 ) {
- g_tboot_shared->num_mac_regions = 4;
- /* S3 resume code (and other real mode trampoline code) */
- g_tboot_shared->mac_regions[0].start =
- (uint64_t)bootsym_phys(trampoline_start);
- g_tboot_shared->mac_regions[0].end =
- (uint64_t)bootsym_phys(trampoline_end);
- /* hypervisor code + data */
- g_tboot_shared->mac_regions[1].start = (uint64_t)__pa(&_stext);
- g_tboot_shared->mac_regions[1].end = (uint64_t)__pa(&__init_begin);
- /* per-cpu data */
- g_tboot_shared->mac_regions[2].start = (uint64_t)__pa(&__per_cpu_start);
- g_tboot_shared->mac_regions[2].end = (uint64_t)__pa(&__per_cpu_end);
- /* bss */
- g_tboot_shared->mac_regions[3].start = (uint64_t)__pa(&__bss_start);
- g_tboot_shared->mac_regions[3].end = (uint64_t)__pa(&_end);
- }
+ /* we may be called from an interrupt context, so to prevent */
+ /* 'ASSERT(!in_irq());' in alloc_domheap_pages(), decrease count */
+ while ( in_irq() )
+ irq_exit();
/* Create identity map for tboot shutdown code. */
+ /* do before S3 integrity because mapping tboot may change xenheap */
map_base = PFN_DOWN(g_tboot_shared->tboot_base);
map_size = PFN_UP(g_tboot_shared->tboot_size);
err = map_pages_to_xen(map_base << PAGE_SHIFT, map_base, map_size,
__PAGE_HYPERVISOR);
- if ( err != 0 )
- {
+ if ( err != 0 ) {
printk("error (0x%x) mapping tboot pages (mfns) @ 0x%x, 0x%x\n", err,
map_base, map_size);
return;
}
+ /* if this is S3 then set regions to MAC */
+ if ( shutdown_type == TB_SHUTDOWN_S3 ) {
+ /*
+ * Xen regions for tboot to MAC
+ */
+ g_tboot_shared->num_mac_regions = 5;
+ /* S3 resume code (and other real mode trampoline code) */
+ g_tboot_shared->mac_regions[0].start = bootsym_phys(trampoline_start);
+ g_tboot_shared->mac_regions[0].size = bootsym_phys(trampoline_end) -
+ bootsym_phys(trampoline_start);
+ /* hypervisor code + data */
+ g_tboot_shared->mac_regions[1].start = (uint64_t)__pa(&_stext);
+ g_tboot_shared->mac_regions[1].size = __pa(&__init_begin) -
+ __pa(&_stext);
+ /* per-cpu data */
+ g_tboot_shared->mac_regions[2].start = (uint64_t)__pa(&__per_cpu_start);
+ g_tboot_shared->mac_regions[2].size = __pa(&__per_cpu_end) -
+ __pa(&__per_cpu_start);
+ /* bss */
+ g_tboot_shared->mac_regions[3].start = (uint64_t)__pa(&__bss_start);
+ g_tboot_shared->mac_regions[3].size = __pa(&_end) - __pa(&__bss_start);
+ /* boot allocator bitmap */
+ g_tboot_shared->mac_regions[4].start = (uint64_t)__pa(&_end);
+ g_tboot_shared->mac_regions[4].size = allocator_bitmap_end -
+ __pa(&_end);
+
+ /*
+ * MAC domains and other Xen memory
+ */
+ /* Xen has no better entropy source for MAC key than tboot's */
+ /* MAC domains first in case it perturbs xenheap */
+ tboot_gen_domain_integrity(g_tboot_shared->s3_key, &domain_mac);
+ tboot_gen_frametable_integrity(g_tboot_shared->s3_key, &frametable_mac);
+ tboot_gen_xenheap_integrity(g_tboot_shared->s3_key, &xenheap_mac);
+ }
+
write_ptbase(idle_vcpu[0]);
((void(*)(void))(unsigned long)g_tboot_shared->shutdown_entry)();
@@ -264,6 +432,29 @@ int __init tboot_parse_dmar_table(acpi_table_handler dmar_handler)
return rc;
}
+int tboot_s3_resume(void)
+{
+ vmac_t mac;
+
+ if ( !tboot_in_measured_env() )
+ return 0;
+
+ /* need to do these in reverse order of shutdown */
+ tboot_gen_xenheap_integrity(g_tboot_shared->s3_key, &mac);
+ if ( mac != xenheap_mac )
+ return -1;
+
+ tboot_gen_frametable_integrity(g_tboot_shared->s3_key, &mac);
+ if ( mac != frametable_mac )
+ return -2;
+
+ tboot_gen_domain_integrity(g_tboot_shared->s3_key, &mac);
+ if ( mac != domain_mac )
+ return 0; /* -3 */
+
+ return 0;
+}
+
/*
* Local variables:
* mode: C
diff --git a/xen/crypto/Makefile b/xen/crypto/Makefile
new file mode 100644
index 0000000000..db29655333
--- /dev/null
+++ b/xen/crypto/Makefile
@@ -0,0 +1,2 @@
+obj-y += rijndael.o
+obj-y += vmac.o
diff --git a/xen/crypto/rijndael.c b/xen/crypto/rijndael.c
new file mode 100644
index 0000000000..f7496187d3
--- /dev/null
+++ b/xen/crypto/rijndael.c
@@ -0,0 +1,1269 @@
+/* $OpenBSD: rijndael.c,v 1.19 2008/06/09 07:49:45 djm Exp $ */
+
+/**
+ * rijndael-alg-fst.c
+ *
+ * @version 3.0 (December 2000)
+ *
+ * Optimised ANSI C code for the Rijndael cipher (now AES)
+ *
+ * @author Vincent Rijmen <vincent.rijmen@esat.kuleuven.ac.be>
+ * @author Antoon Bosselaers <antoon.bosselaers@esat.kuleuven.ac.be>
+ * @author Paulo Barreto <paulo.barreto@terra.com.br>
+ *
+ * This code is hereby placed in the public domain.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''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 AUTHORS 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.
+ */
+
+/* start for Xen */
+#include <xen/config.h>
+#include <xen/init.h>
+#include <xen/types.h>
+#include <xen/lib.h>
+#include <crypto/rijndael.h>
+/* end for Xen */
+
+#undef FULL_UNROLL
+
+/*
+Te0[x] = S [x].[02, 01, 01, 03];
+Te1[x] = S [x].[03, 02, 01, 01];
+Te2[x] = S [x].[01, 03, 02, 01];
+Te3[x] = S [x].[01, 01, 03, 02];
+Te4[x] = S [x].[01, 01, 01, 01];
+
+Td0[x] = Si[x].[0e, 09, 0d, 0b];
+Td1[x] = Si[x].[0b, 0e, 09, 0d];
+Td2[x] = Si[x].[0d, 0b, 0e, 09];
+Td3[x] = Si[x].[09, 0d, 0b, 0e];
+Td4[x] = Si[x].[01, 01, 01, 01];
+*/
+
+static const u32 Te0[256] = {
+ 0xc66363a5U, 0xf87c7c84U, 0xee777799U, 0xf67b7b8dU,
+ 0xfff2f20dU, 0xd66b6bbdU, 0xde6f6fb1U, 0x91c5c554U,
+ 0x60303050U, 0x02010103U, 0xce6767a9U, 0x562b2b7dU,
+ 0xe7fefe19U, 0xb5d7d762U, 0x4dababe6U, 0xec76769aU,
+ 0x8fcaca45U, 0x1f82829dU, 0x89c9c940U, 0xfa7d7d87U,
+ 0xeffafa15U, 0xb25959ebU, 0x8e4747c9U, 0xfbf0f00bU,
+ 0x41adadecU, 0xb3d4d467U, 0x5fa2a2fdU, 0x45afafeaU,
+ 0x239c9cbfU, 0x53a4a4f7U, 0xe4727296U, 0x9bc0c05bU,
+ 0x75b7b7c2U, 0xe1fdfd1cU, 0x3d9393aeU, 0x4c26266aU,
+ 0x6c36365aU, 0x7e3f3f41U, 0xf5f7f702U, 0x83cccc4fU,
+ 0x6834345cU, 0x51a5a5f4U, 0xd1e5e534U, 0xf9f1f108U,
+ 0xe2717193U, 0xabd8d873U, 0x62313153U, 0x2a15153fU,
+ 0x0804040cU, 0x95c7c752U, 0x46232365U, 0x9dc3c35eU,
+ 0x30181828U, 0x379696a1U, 0x0a05050fU, 0x2f9a9ab5U,
+ 0x0e070709U, 0x24121236U, 0x1b80809bU, 0xdfe2e23dU,
+ 0xcdebeb26U, 0x4e272769U, 0x7fb2b2cdU, 0xea75759fU,
+ 0x1209091bU, 0x1d83839eU, 0x582c2c74U, 0x341a1a2eU,
+ 0x361b1b2dU, 0xdc6e6eb2U, 0xb45a5aeeU, 0x5ba0a0fbU,
+ 0xa45252f6U, 0x763b3b4dU, 0xb7d6d661U, 0x7db3b3ceU,
+ 0x5229297bU, 0xdde3e33eU, 0x5e2f2f71U, 0x13848497U,
+ 0xa65353f5U, 0xb9d1d168U, 0x00000000U, 0xc1eded2cU,
+ 0x40202060U, 0xe3fcfc1fU, 0x79b1b1c8U, 0xb65b5bedU,
+ 0xd46a6abeU, 0x8dcbcb46U, 0x67bebed9U, 0x7239394bU,
+ 0x944a4adeU, 0x984c4cd4U, 0xb05858e8U, 0x85cfcf4aU,
+ 0xbbd0d06bU, 0xc5efef2aU, 0x4faaaae5U, 0xedfbfb16U,
+ 0x864343c5U, 0x9a4d4dd7U, 0x66333355U, 0x11858594U,
+ 0x8a4545cfU, 0xe9f9f910U, 0x04020206U, 0xfe7f7f81U,
+ 0xa05050f0U, 0x783c3c44U, 0x259f9fbaU, 0x4ba8a8e3U,
+ 0xa25151f3U, 0x5da3a3feU, 0x804040c0U, 0x058f8f8aU,
+ 0x3f9292adU, 0x219d9dbcU, 0x70383848U, 0xf1f5f504U,
+ 0x63bcbcdfU, 0x77b6b6c1U, 0xafdada75U, 0x42212163U,
+ 0x20101030U, 0xe5ffff1aU, 0xfdf3f30eU, 0xbfd2d26dU,
+ 0x81cdcd4cU, 0x180c0c14U, 0x26131335U, 0xc3ecec2fU,
+ 0xbe5f5fe1U, 0x359797a2U, 0x884444ccU, 0x2e171739U,
+ 0x93c4c457U, 0x55a7a7f2U, 0xfc7e7e82U, 0x7a3d3d47U,
+ 0xc86464acU, 0xba5d5de7U, 0x3219192bU, 0xe6737395U,
+ 0xc06060a0U, 0x19818198U, 0x9e4f4fd1U, 0xa3dcdc7fU,
+ 0x44222266U, 0x542a2a7eU, 0x3b9090abU, 0x0b888883U,
+ 0x8c4646caU, 0xc7eeee29U, 0x6bb8b8d3U, 0x2814143cU,
+ 0xa7dede79U, 0xbc5e5ee2U, 0x160b0b1dU, 0xaddbdb76U,
+ 0xdbe0e03bU, 0x64323256U, 0x743a3a4eU, 0x140a0a1eU,
+ 0x924949dbU, 0x0c06060aU, 0x4824246cU, 0xb85c5ce4U,
+ 0x9fc2c25dU, 0xbdd3d36eU, 0x43acacefU, 0xc46262a6U,
+ 0x399191a8U, 0x319595a4U, 0xd3e4e437U, 0xf279798bU,
+ 0xd5e7e732U, 0x8bc8c843U, 0x6e373759U, 0xda6d6db7U,
+ 0x018d8d8cU, 0xb1d5d564U, 0x9c4e4ed2U, 0x49a9a9e0U,
+ 0xd86c6cb4U, 0xac5656faU, 0xf3f4f407U, 0xcfeaea25U,
+ 0xca6565afU, 0xf47a7a8eU, 0x47aeaee9U, 0x10080818U,
+ 0x6fbabad5U, 0xf0787888U, 0x4a25256fU, 0x5c2e2e72U,
+ 0x381c1c24U, 0x57a6a6f1U, 0x73b4b4c7U, 0x97c6c651U,
+ 0xcbe8e823U, 0xa1dddd7cU, 0xe874749cU, 0x3e1f1f21U,
+ 0x964b4bddU, 0x61bdbddcU, 0x0d8b8b86U, 0x0f8a8a85U,
+ 0xe0707090U, 0x7c3e3e42U, 0x71b5b5c4U, 0xcc6666aaU,
+ 0x904848d8U, 0x06030305U, 0xf7f6f601U, 0x1c0e0e12U,
+ 0xc26161a3U, 0x6a35355fU, 0xae5757f9U, 0x69b9b9d0U,
+ 0x17868691U, 0x99c1c158U, 0x3a1d1d27U, 0x279e9eb9U,
+ 0xd9e1e138U, 0xebf8f813U, 0x2b9898b3U, 0x22111133U,
+ 0xd26969bbU, 0xa9d9d970U, 0x078e8e89U, 0x339494a7U,
+ 0x2d9b9bb6U, 0x3c1e1e22U, 0x15878792U, 0xc9e9e920U,
+ 0x87cece49U, 0xaa5555ffU, 0x50282878U, 0xa5dfdf7aU,
+ 0x038c8c8fU, 0x59a1a1f8U, 0x09898980U, 0x1a0d0d17U,
+ 0x65bfbfdaU, 0xd7e6e631U, 0x844242c6U, 0xd06868b8U,
+ 0x824141c3U, 0x299999b0U, 0x5a2d2d77U, 0x1e0f0f11U,
+ 0x7bb0b0cbU, 0xa85454fcU, 0x6dbbbbd6U, 0x2c16163aU,
+};
+static const u32 Te1[256] = {
+ 0xa5c66363U, 0x84f87c7cU, 0x99ee7777U, 0x8df67b7bU,
+ 0x0dfff2f2U, 0xbdd66b6bU, 0xb1de6f6fU, 0x5491c5c5U,
+ 0x50603030U, 0x03020101U, 0xa9ce6767U, 0x7d562b2bU,
+ 0x19e7fefeU, 0x62b5d7d7U, 0xe64dababU, 0x9aec7676U,
+ 0x458fcacaU, 0x9d1f8282U, 0x4089c9c9U, 0x87fa7d7dU,
+ 0x15effafaU, 0xebb25959U, 0xc98e4747U, 0x0bfbf0f0U,
+ 0xec41adadU, 0x67b3d4d4U, 0xfd5fa2a2U, 0xea45afafU,
+ 0xbf239c9cU, 0xf753a4a4U, 0x96e47272U, 0x5b9bc0c0U,
+ 0xc275b7b7U, 0x1ce1fdfdU, 0xae3d9393U, 0x6a4c2626U,
+ 0x5a6c3636U, 0x417e3f3fU, 0x02f5f7f7U, 0x4f83ccccU,
+ 0x5c683434U, 0xf451a5a5U, 0x34d1e5e5U, 0x08f9f1f1U,
+ 0x93e27171U, 0x73abd8d8U, 0x53623131U, 0x3f2a1515U,
+ 0x0c080404U, 0x5295c7c7U, 0x65462323U, 0x5e9dc3c3U,
+ 0x28301818U, 0xa1379696U, 0x0f0a0505U, 0xb52f9a9aU,
+ 0x090e0707U, 0x36241212U, 0x9b1b8080U, 0x3ddfe2e2U,
+ 0x26cdebebU, 0x694e2727U, 0xcd7fb2b2U, 0x9fea7575U,
+ 0x1b120909U, 0x9e1d8383U, 0x74582c2cU, 0x2e341a1aU,
+ 0x2d361b1bU, 0xb2dc6e6eU, 0xeeb45a5aU, 0xfb5ba0a0U,
+ 0xf6a45252U, 0x4d763b3bU, 0x61b7d6d6U, 0xce7db3b3U,
+ 0x7b522929U, 0x3edde3e3U, 0x715e2f2fU, 0x97138484U,
+ 0xf5a65353U, 0x68b9d1d1U, 0x00000000U, 0x2cc1ededU,
+ 0x60402020U, 0x1fe3fcfcU, 0xc879b1b1U, 0xedb65b5bU,
+ 0xbed46a6aU, 0x468dcbcbU, 0xd967bebeU, 0x4b723939U,
+ 0xde944a4aU, 0xd4984c4cU, 0xe8b05858U, 0x4a85cfcfU,
+ 0x6bbbd0d0U, 0x2ac5efefU, 0xe54faaaaU, 0x16edfbfbU,
+ 0xc5864343U, 0xd79a4d4dU, 0x55663333U, 0x94118585U,
+ 0xcf8a4545U, 0x10e9f9f9U, 0x06040202U, 0x81fe7f7fU,
+ 0xf0a05050U, 0x44783c3cU, 0xba259f9fU, 0xe34ba8a8U,
+ 0xf3a25151U, 0xfe5da3a3U, 0xc0804040U, 0x8a058f8fU,
+ 0xad3f9292U, 0xbc219d9dU, 0x48703838U, 0x04f1f5f5U,
+ 0xdf63bcbcU, 0xc177b6b6U, 0x75afdadaU, 0x63422121U,
+ 0x30201010U, 0x1ae5ffffU, 0x0efdf3f3U, 0x6dbfd2d2U,
+ 0x4c81cdcdU, 0x14180c0cU, 0x35261313U, 0x2fc3ececU,
+ 0xe1be5f5fU, 0xa2359797U, 0xcc884444U, 0x392e1717U,
+ 0x5793c4c4U, 0xf255a7a7U, 0x82fc7e7eU, 0x477a3d3dU,
+ 0xacc86464U, 0xe7ba5d5dU, 0x2b321919U, 0x95e67373U,
+ 0xa0c06060U, 0x98198181U, 0xd19e4f4fU, 0x7fa3dcdcU,
+ 0x66442222U, 0x7e542a2aU, 0xab3b9090U, 0x830b8888U,
+ 0xca8c4646U, 0x29c7eeeeU, 0xd36bb8b8U, 0x3c281414U,
+ 0x79a7dedeU, 0xe2bc5e5eU, 0x1d160b0bU, 0x76addbdbU,
+ 0x3bdbe0e0U, 0x56643232U, 0x4e743a3aU, 0x1e140a0aU,
+ 0xdb924949U, 0x0a0c0606U, 0x6c482424U, 0xe4b85c5cU,
+ 0x5d9fc2c2U, 0x6ebdd3d3U, 0xef43acacU, 0xa6c46262U,
+ 0xa8399191U, 0xa4319595U, 0x37d3e4e4U, 0x8bf27979U,
+ 0x32d5e7e7U, 0x438bc8c8U, 0x596e3737U, 0xb7da6d6dU,
+ 0x8c018d8dU, 0x64b1d5d5U, 0xd29c4e4eU, 0xe049a9a9U,
+ 0xb4d86c6cU, 0xfaac5656U, 0x07f3f4f4U, 0x25cfeaeaU,
+ 0xafca6565U, 0x8ef47a7aU, 0xe947aeaeU, 0x18100808U,
+ 0xd56fbabaU, 0x88f07878U, 0x6f4a2525U, 0x725c2e2eU,
+ 0x24381c1cU, 0xf157a6a6U, 0xc773b4b4U, 0x5197c6c6U,
+ 0x23cbe8e8U, 0x7ca1ddddU, 0x9ce87474U, 0x213e1f1fU,
+ 0xdd964b4bU, 0xdc61bdbdU, 0x860d8b8bU, 0x850f8a8aU,
+ 0x90e07070U, 0x427c3e3eU, 0xc471b5b5U, 0xaacc6666U,
+ 0xd8904848U, 0x05060303U, 0x01f7f6f6U, 0x121c0e0eU,
+ 0xa3c26161U, 0x5f6a3535U, 0xf9ae5757U, 0xd069b9b9U,
+ 0x91178686U, 0x5899c1c1U, 0x273a1d1dU, 0xb9279e9eU,
+ 0x38d9e1e1U, 0x13ebf8f8U, 0xb32b9898U, 0x33221111U,
+ 0xbbd26969U, 0x70a9d9d9U, 0x89078e8eU, 0xa7339494U,
+ 0xb62d9b9bU, 0x223c1e1eU, 0x92158787U, 0x20c9e9e9U,
+ 0x4987ceceU, 0xffaa5555U, 0x78502828U, 0x7aa5dfdfU,
+ 0x8f038c8cU, 0xf859a1a1U, 0x80098989U, 0x171a0d0dU,
+ 0xda65bfbfU, 0x31d7e6e6U, 0xc6844242U, 0xb8d06868U,
+ 0xc3824141U, 0xb0299999U, 0x775a2d2dU, 0x111e0f0fU,
+ 0xcb7bb0b0U, 0xfca85454U, 0xd66dbbbbU, 0x3a2c1616U,
+};
+static const u32 Te2[256] = {
+ 0x63a5c663U, 0x7c84f87cU, 0x7799ee77U, 0x7b8df67bU,
+ 0xf20dfff2U, 0x6bbdd66bU, 0x6fb1de6fU, 0xc55491c5U,
+ 0x30506030U, 0x01030201U, 0x67a9ce67U, 0x2b7d562bU,
+ 0xfe19e7feU, 0xd762b5d7U, 0xabe64dabU, 0x769aec76U,
+ 0xca458fcaU, 0x829d1f82U, 0xc94089c9U, 0x7d87fa7dU,
+ 0xfa15effaU, 0x59ebb259U, 0x47c98e47U, 0xf00bfbf0U,
+ 0xadec41adU, 0xd467b3d4U, 0xa2fd5fa2U, 0xafea45afU,
+ 0x9cbf239cU, 0xa4f753a4U, 0x7296e472U, 0xc05b9bc0U,
+ 0xb7c275b7U, 0xfd1ce1fdU, 0x93ae3d93U, 0x266a4c26U,
+ 0x365a6c36U, 0x3f417e3fU, 0xf702f5f7U, 0xcc4f83ccU,
+ 0x345c6834U, 0xa5f451a5U, 0xe534d1e5U, 0xf108f9f1U,
+ 0x7193e271U, 0xd873abd8U, 0x31536231U, 0x153f2a15U,
+ 0x040c0804U, 0xc75295c7U, 0x23654623U, 0xc35e9dc3U,
+ 0x18283018U, 0x96a13796U, 0x050f0a05U, 0x9ab52f9aU,
+ 0x07090e07U, 0x12362412U, 0x809b1b80U, 0xe23ddfe2U,
+ 0xeb26cdebU, 0x27694e27U, 0xb2cd7fb2U, 0x759fea75U,
+ 0x091b1209U, 0x839e1d83U, 0x2c74582cU, 0x1a2e341aU,
+ 0x1b2d361bU, 0x6eb2dc6eU, 0x5aeeb45aU, 0xa0fb5ba0U,
+ 0x52f6a452U, 0x3b4d763bU, 0xd661b7d6U, 0xb3ce7db3U,
+ 0x297b5229U, 0xe33edde3U, 0x2f715e2fU, 0x84971384U,
+ 0x53f5a653U, 0xd168b9d1U, 0x00000000U, 0xed2cc1edU,
+ 0x20604020U, 0xfc1fe3fcU, 0xb1c879b1U, 0x5bedb65bU,
+ 0x6abed46aU, 0xcb468dcbU, 0xbed967beU, 0x394b7239U,
+ 0x4ade944aU, 0x4cd4984cU, 0x58e8b058U, 0xcf4a85cfU,
+ 0xd06bbbd0U, 0xef2ac5efU, 0xaae54faaU, 0xfb16edfbU,
+ 0x43c58643U, 0x4dd79a4dU, 0x33556633U, 0x85941185U,
+ 0x45cf8a45U, 0xf910e9f9U, 0x02060402U, 0x7f81fe7fU,
+ 0x50f0a050U, 0x3c44783cU, 0x9fba259fU, 0xa8e34ba8U,
+ 0x51f3a251U, 0xa3fe5da3U, 0x40c08040U, 0x8f8a058fU,
+ 0x92ad3f92U, 0x9dbc219dU, 0x38487038U, 0xf504f1f5U,
+ 0xbcdf63bcU, 0xb6c177b6U, 0xda75afdaU, 0x21634221U,
+ 0x10302010U, 0xff1ae5ffU, 0xf30efdf3U, 0xd26dbfd2U,
+ 0xcd4c81cdU, 0x0c14180cU, 0x13352613U, 0xec2fc3ecU,
+ 0x5fe1be5fU, 0x97a23597U, 0x44cc8844U, 0x17392e17U,
+ 0xc45793c4U, 0xa7f255a7U, 0x7e82fc7eU, 0x3d477a3dU,
+ 0x64acc864U, 0x5de7ba5dU, 0x192b3219U, 0x7395e673U,
+ 0x60a0c060U, 0x81981981U, 0x4fd19e4fU, 0xdc7fa3dcU,
+ 0x22664422U, 0x2a7e542aU, 0x90ab3b90U, 0x88830b88U,
+ 0x46ca8c46U, 0xee29c7eeU, 0xb8d36bb8U, 0x143c2814U,
+ 0xde79a7deU, 0x5ee2bc5eU, 0x0b1d160bU, 0xdb76addbU,
+ 0xe03bdbe0U, 0x32566432U, 0x3a4e743aU, 0x0a1e140aU,
+ 0x49db9249U, 0x060a0c06U, 0x246c4824U, 0x5ce4b85cU,
+ 0xc25d9fc2U, 0xd36ebdd3U, 0xacef43acU, 0x62a6c462U,
+ 0x91a83991U, 0x95a43195U, 0xe437d3e4U, 0x798bf279U,
+ 0xe732d5e7U, 0xc8438bc8U, 0x37596e37U, 0x6db7da6dU,
+ 0x8d8c018dU, 0xd564b1d5U, 0x4ed29c4eU, 0xa9e049a9U,
+ 0x6cb4d86cU, 0x56faac56U, 0xf407f3f4U, 0xea25cfeaU,
+ 0x65afca65U, 0x7a8ef47aU, 0xaee947aeU, 0x08181008U,
+ 0xbad56fbaU, 0x7888f078U, 0x256f4a25U, 0x2e725c2eU,
+ 0x1c24381cU, 0xa6f157a6U, 0xb4c773b4U, 0xc65197c6U,
+ 0xe823cbe8U, 0xdd7ca1ddU, 0x749ce874U, 0x1f213e1fU,
+ 0x4bdd964bU, 0xbddc61bdU, 0x8b860d8bU, 0x8a850f8aU,
+ 0x7090e070U, 0x3e427c3eU, 0xb5c471b5U, 0x66aacc66U,
+ 0x48d89048U, 0x03050603U, 0xf601f7f6U, 0x0e121c0eU,
+ 0x61a3c261U, 0x355f6a35U, 0x57f9ae57U, 0xb9d069b9U,
+ 0x86911786U, 0xc15899c1U, 0x1d273a1dU, 0x9eb9279eU,
+ 0xe138d9e1U, 0xf813ebf8U, 0x98b32b98U, 0x11332211U,
+ 0x69bbd269U, 0xd970a9d9U, 0x8e89078eU, 0x94a73394U,
+ 0x9bb62d9bU, 0x1e223c1eU, 0x87921587U, 0xe920c9e9U,
+ 0xce4987ceU, 0x55ffaa55U, 0x28785028U, 0xdf7aa5dfU,
+ 0x8c8f038cU, 0xa1f859a1U, 0x89800989U, 0x0d171a0dU,
+ 0xbfda65bfU, 0xe631d7e6U, 0x42c68442U, 0x68b8d068U,
+ 0x41c38241U, 0x99b02999U, 0x2d775a2dU, 0x0f111e0fU,
+ 0xb0cb7bb0U, 0x54fca854U, 0xbbd66dbbU, 0x163a2c16U,
+};
+static const u32 Te3[256] = {
+ 0x6363a5c6U, 0x7c7c84f8U, 0x777799eeU, 0x7b7b8df6U,
+ 0xf2f20dffU, 0x6b6bbdd6U, 0x6f6fb1deU, 0xc5c55491U,
+ 0x30305060U, 0x01010302U, 0x6767a9ceU, 0x2b2b7d56U,
+ 0xfefe19e7U, 0xd7d762b5U, 0xababe64dU, 0x76769aecU,
+ 0xcaca458fU, 0x82829d1fU, 0xc9c94089U, 0x7d7d87faU,
+ 0xfafa15efU, 0x5959ebb2U, 0x4747c98eU, 0xf0f00bfbU,
+ 0xadadec41U, 0xd4d467b3U, 0xa2a2fd5fU, 0xafafea45U,
+ 0x9c9cbf23U, 0xa4a4f753U, 0x727296e4U, 0xc0c05b9bU,
+ 0xb7b7c275U, 0xfdfd1ce1U, 0x9393ae3dU, 0x26266a4cU,
+ 0x36365a6cU, 0x3f3f417eU, 0xf7f702f5U, 0xcccc4f83U,
+ 0x34345c68U, 0xa5a5f451U, 0xe5e534d1U, 0xf1f108f9U,
+ 0x717193e2U, 0xd8d873abU, 0x31315362U, 0x15153f2aU,
+ 0x04040c08U, 0xc7c75295U, 0x23236546U, 0xc3c35e9dU,
+ 0x18182830U, 0x9696a137U, 0x05050f0aU, 0x9a9ab52fU,
+ 0x0707090eU, 0x12123624U, 0x80809b1bU, 0xe2e23ddfU,
+ 0xebeb26cdU, 0x2727694eU, 0xb2b2cd7fU, 0x75759feaU,
+ 0x09091b12U, 0x83839e1dU, 0x2c2c7458U, 0x1a1a2e34U,
+ 0x1b1b2d36U, 0x6e6eb2dcU, 0x5a5aeeb4U, 0xa0a0fb5bU,
+ 0x5252f6a4U, 0x3b3b4d76U, 0xd6d661b7U, 0xb3b3ce7dU,
+ 0x29297b52U, 0xe3e33eddU, 0x2f2f715eU, 0x84849713U,
+ 0x5353f5a6U, 0xd1d168b9U, 0x00000000U, 0xeded2cc1U,
+ 0x20206040U, 0xfcfc1fe3U, 0xb1b1c879U, 0x5b5bedb6U,
+ 0x6a6abed4U, 0xcbcb468dU, 0xbebed967U, 0x39394b72U,
+ 0x4a4ade94U, 0x4c4cd498U, 0x5858e8b0U, 0xcfcf4a85U,
+ 0xd0d06bbbU, 0xefef2ac5U, 0xaaaae54fU, 0xfbfb16edU,
+ 0x4343c586U, 0x4d4dd79aU, 0x33335566U, 0x85859411U,
+ 0x4545cf8aU, 0xf9f910e9U, 0x02020604U, 0x7f7f81feU,
+ 0x5050f0a0U, 0x3c3c4478U, 0x9f9fba25U, 0xa8a8e34bU,
+ 0x5151f3a2U, 0xa3a3fe5dU, 0x4040c080U, 0x8f8f8a05U,
+ 0x9292ad3fU, 0x9d9dbc21U, 0x38384870U, 0xf5f504f1U,
+ 0xbcbcdf63U, 0xb6b6c177U, 0xdada75afU, 0x21216342U,
+ 0x10103020U, 0xffff1ae5U, 0xf3f30efdU, 0xd2d26dbfU,
+ 0xcdcd4c81U, 0x0c0c1418U, 0x13133526U, 0xecec2fc3U,
+ 0x5f5fe1beU, 0x9797a235U, 0x4444cc88U, 0x1717392eU,
+ 0xc4c45793U, 0xa7a7f255U, 0x7e7e82fcU, 0x3d3d477aU,
+ 0x6464acc8U, 0x5d5de7baU, 0x19192b32U, 0x737395e6U,
+ 0x6060a0c0U, 0x81819819U, 0x4f4fd19eU, 0xdcdc7fa3U,
+ 0x22226644U, 0x2a2a7e54U, 0x9090ab3bU, 0x8888830bU,
+ 0x4646ca8cU, 0xeeee29c7U, 0xb8b8d36bU, 0x14143c28U,
+ 0xdede79a7U, 0x5e5ee2bcU, 0x0b0b1d16U, 0xdbdb76adU,
+ 0xe0e03bdbU, 0x32325664U, 0x3a3a4e74U, 0x0a0a1e14U,
+ 0x4949db92U, 0x06060a0cU, 0x24246c48U, 0x5c5ce4b8U,
+ 0xc2c25d9fU, 0xd3d36ebdU, 0xacacef43U, 0x6262a6c4U,
+ 0x9191a839U, 0x9595a431U, 0xe4e437d3U, 0x79798bf2U,
+ 0xe7e732d5U, 0xc8c8438bU, 0x3737596eU, 0x6d6db7daU,
+ 0x8d8d8c01U, 0xd5d564b1U, 0x4e4ed29cU, 0xa9a9e049U,
+ 0x6c6cb4d8U, 0x5656faacU, 0xf4f407f3U, 0xeaea25cfU,
+ 0x6565afcaU, 0x7a7a8ef4U, 0xaeaee947U, 0x08081810U,
+ 0xbabad56fU, 0x787888f0U, 0x25256f4aU, 0x2e2e725cU,
+ 0x1c1c2438U, 0xa6a6f157U, 0xb4b4c773U, 0xc6c65197U,
+ 0xe8e823cbU, 0xdddd7ca1U, 0x74749ce8U, 0x1f1f213eU,
+ 0x4b4bdd96U, 0xbdbddc61U, 0x8b8b860dU, 0x8a8a850fU,
+ 0x707090e0U, 0x3e3e427cU, 0xb5b5c471U, 0x6666aaccU,
+ 0x4848d890U, 0x03030506U, 0xf6f601f7U, 0x0e0e121cU,
+ 0x6161a3c2U, 0x35355f6aU, 0x5757f9aeU, 0xb9b9d069U,
+ 0x86869117U, 0xc1c15899U, 0x1d1d273aU, 0x9e9eb927U,
+ 0xe1e138d9U, 0xf8f813ebU, 0x9898b32bU, 0x11113322U,
+ 0x6969bbd2U, 0xd9d970a9U, 0x8e8e8907U, 0x9494a733U,
+ 0x9b9bb62dU, 0x1e1e223cU, 0x87879215U, 0xe9e920c9U,
+ 0xcece4987U, 0x5555ffaaU, 0x28287850U, 0xdfdf7aa5U,
+ 0x8c8c8f03U, 0xa1a1f859U, 0x89898009U, 0x0d0d171aU,
+ 0xbfbfda65U, 0xe6e631d7U, 0x4242c684U, 0x6868b8d0U,
+ 0x4141c382U, 0x9999b029U, 0x2d2d775aU, 0x0f0f111eU,
+ 0xb0b0cb7bU, 0x5454fca8U, 0xbbbbd66dU, 0x16163a2cU,
+};
+static const u32 Te4[256] = {
+ 0x63636363U, 0x7c7c7c7cU, 0x77777777U, 0x7b7b7b7bU,
+ 0xf2f2f2f2U, 0x6b6b6b6bU, 0x6f6f6f6fU, 0xc5c5c5c5U,
+ 0x30303030U, 0x01010101U, 0x67676767U, 0x2b2b2b2bU,
+ 0xfefefefeU, 0xd7d7d7d7U, 0xababababU, 0x76767676U,
+ 0xcacacacaU, 0x82828282U, 0xc9c9c9c9U, 0x7d7d7d7dU,
+ 0xfafafafaU, 0x59595959U, 0x47474747U, 0xf0f0f0f0U,
+ 0xadadadadU, 0xd4d4d4d4U, 0xa2a2a2a2U, 0xafafafafU,
+ 0x9c9c9c9cU, 0xa4a4a4a4U, 0x72727272U, 0xc0c0c0c0U,
+ 0xb7b7b7b7U, 0xfdfdfdfdU, 0x93939393U, 0x26262626U,
+ 0x36363636U, 0x3f3f3f3fU, 0xf7f7f7f7U, 0xccccccccU,
+ 0x34343434U, 0xa5a5a5a5U, 0xe5e5e5e5U, 0xf1f1f1f1U,
+ 0x71717171U, 0xd8d8d8d8U, 0x31313131U, 0x15151515U,
+ 0x04040404U, 0xc7c7c7c7U, 0x23232323U, 0xc3c3c3c3U,
+ 0x18181818U, 0x96969696U, 0x05050505U, 0x9a9a9a9aU,
+ 0x07070707U, 0x12121212U, 0x80808080U, 0xe2e2e2e2U,
+ 0xebebebebU, 0x27272727U, 0xb2b2b2b2U, 0x75757575U,
+ 0x09090909U, 0x83838383U, 0x2c2c2c2cU, 0x1a1a1a1aU,
+ 0x1b1b1b1bU, 0x6e6e6e6eU, 0x5a5a5a5aU, 0xa0a0a0a0U,
+ 0x52525252U, 0x3b3b3b3bU, 0xd6d6d6d6U, 0xb3b3b3b3U,
+ 0x29292929U, 0xe3e3e3e3U, 0x2f2f2f2fU, 0x84848484U,
+ 0x53535353U, 0xd1d1d1d1U, 0x00000000U, 0xededededU,
+ 0x20202020U, 0xfcfcfcfcU, 0xb1b1b1b1U, 0x5b5b5b5bU,
+ 0x6a6a6a6aU, 0xcbcbcbcbU, 0xbebebebeU, 0x39393939U,
+ 0x4a4a4a4aU, 0x4c4c4c4cU, 0x58585858U, 0xcfcfcfcfU,
+ 0xd0d0d0d0U, 0xefefefefU, 0xaaaaaaaaU, 0xfbfbfbfbU,
+ 0x43434343U, 0x4d4d4d4dU, 0x33333333U, 0x85858585U,
+ 0x45454545U, 0xf9f9f9f9U, 0x02020202U, 0x7f7f7f7fU,
+ 0x50505050U, 0x3c3c3c3cU, 0x9f9f9f9fU, 0xa8a8a8a8U,
+ 0x51515151U, 0xa3a3a3a3U, 0x40404040U, 0x8f8f8f8fU,
+ 0x92929292U, 0x9d9d9d9dU, 0x38383838U, 0xf5f5f5f5U,
+ 0xbcbcbcbcU, 0xb6b6b6b6U, 0xdadadadaU, 0x21212121U,
+ 0x10101010U, 0xffffffffU, 0xf3f3f3f3U, 0xd2d2d2d2U,
+ 0xcdcdcdcdU, 0x0c0c0c0cU, 0x13131313U, 0xececececU,
+ 0x5f5f5f5fU, 0x97979797U, 0x44444444U, 0x17171717U,
+ 0xc4c4c4c4U, 0xa7a7a7a7U, 0x7e7e7e7eU, 0x3d3d3d3dU,
+ 0x64646464U, 0x5d5d5d5dU, 0x19191919U, 0x73737373U,
+ 0x60606060U, 0x81818181U, 0x4f4f4f4fU, 0xdcdcdcdcU,
+ 0x22222222U, 0x2a2a2a2aU, 0x90909090U, 0x88888888U,
+ 0x46464646U, 0xeeeeeeeeU, 0xb8b8b8b8U, 0x14141414U,
+ 0xdedededeU, 0x5e5e5e5eU, 0x0b0b0b0bU, 0xdbdbdbdbU,
+ 0xe0e0e0e0U, 0x32323232U, 0x3a3a3a3aU, 0x0a0a0a0aU,
+ 0x49494949U, 0x06060606U, 0x24242424U, 0x5c5c5c5cU,
+ 0xc2c2c2c2U, 0xd3d3d3d3U, 0xacacacacU, 0x62626262U,
+ 0x91919191U, 0x95959595U, 0xe4e4e4e4U, 0x79797979U,
+ 0xe7e7e7e7U, 0xc8c8c8c8U, 0x37373737U, 0x6d6d6d6dU,
+ 0x8d8d8d8dU, 0xd5d5d5d5U, 0x4e4e4e4eU, 0xa9a9a9a9U,
+ 0x6c6c6c6cU, 0x56565656U, 0xf4f4f4f4U, 0xeaeaeaeaU,
+ 0x65656565U, 0x7a7a7a7aU, 0xaeaeaeaeU, 0x08080808U,
+ 0xbabababaU, 0x78787878U, 0x25252525U, 0x2e2e2e2eU,
+ 0x1c1c1c1cU, 0xa6a6a6a6U, 0xb4b4b4b4U, 0xc6c6c6c6U,
+ 0xe8e8e8e8U, 0xddddddddU, 0x74747474U, 0x1f1f1f1fU,
+ 0x4b4b4b4bU, 0xbdbdbdbdU, 0x8b8b8b8bU, 0x8a8a8a8aU,
+ 0x70707070U, 0x3e3e3e3eU, 0xb5b5b5b5U, 0x66666666U,
+ 0x48484848U, 0x03030303U, 0xf6f6f6f6U, 0x0e0e0e0eU,
+ 0x61616161U, 0x35353535U, 0x57575757U, 0xb9b9b9b9U,
+ 0x86868686U, 0xc1c1c1c1U, 0x1d1d1d1dU, 0x9e9e9e9eU,
+ 0xe1e1e1e1U, 0xf8f8f8f8U, 0x98989898U, 0x11111111U,
+ 0x69696969U, 0xd9d9d9d9U, 0x8e8e8e8eU, 0x94949494U,
+ 0x9b9b9b9bU, 0x1e1e1e1eU, 0x87878787U, 0xe9e9e9e9U,
+ 0xcecececeU, 0x55555555U, 0x28282828U, 0xdfdfdfdfU,
+ 0x8c8c8c8cU, 0xa1a1a1a1U, 0x89898989U, 0x0d0d0d0dU,
+ 0xbfbfbfbfU, 0xe6e6e6e6U, 0x42424242U, 0x68686868U,
+ 0x41414141U, 0x99999999U, 0x2d2d2d2dU, 0x0f0f0f0fU,
+ 0xb0b0b0b0U, 0x54545454U, 0xbbbbbbbbU, 0x16161616U,
+};
+static const u32 Td0[256] = {
+ 0x51f4a750U, 0x7e416553U, 0x1a17a4c3U, 0x3a275e96U,
+ 0x3bab6bcbU, 0x1f9d45f1U, 0xacfa58abU, 0x4be30393U,
+ 0x2030fa55U, 0xad766df6U, 0x88cc7691U, 0xf5024c25U,
+ 0x4fe5d7fcU, 0xc52acbd7U, 0x26354480U, 0xb562a38fU,
+ 0xdeb15a49U, 0x25ba1b67U, 0x45ea0e98U, 0x5dfec0e1U,
+ 0xc32f7502U, 0x814cf012U, 0x8d4697a3U, 0x6bd3f9c6U,
+ 0x038f5fe7U, 0x15929c95U, 0xbf6d7aebU, 0x955259daU,
+ 0xd4be832dU, 0x587421d3U, 0x49e06929U, 0x8ec9c844U,
+ 0x75c2896aU, 0xf48e7978U, 0x99583e6bU, 0x27b971ddU,
+ 0xbee14fb6U, 0xf088ad17U, 0xc920ac66U, 0x7dce3ab4U,
+ 0x63df4a18U, 0xe51a3182U, 0x97513360U, 0x62537f45U,
+ 0xb16477e0U, 0xbb6bae84U, 0xfe81a01cU, 0xf9082b94U,
+ 0x70486858U, 0x8f45fd19U, 0x94de6c87U, 0x527bf8b7U,
+ 0xab73d323U, 0x724b02e2U, 0xe31f8f57U, 0x6655ab2aU,
+ 0xb2eb2807U, 0x2fb5c203U, 0x86c57b9aU, 0xd33708a5U,
+ 0x302887f2U, 0x23bfa5b2U, 0x02036abaU, 0xed16825cU,
+ 0x8acf1c2bU, 0xa779b492U, 0xf307f2f0U, 0x4e69e2a1U,
+ 0x65daf4cdU, 0x0605bed5U, 0xd134621fU, 0xc4a6fe8aU,
+ 0x342e539dU, 0xa2f355a0U, 0x058ae132U, 0xa4f6eb75U,
+ 0x0b83ec39U, 0x4060efaaU, 0x5e719f06U, 0xbd6e1051U,
+ 0x3e218af9U, 0x96dd063dU, 0xdd3e05aeU, 0x4de6bd46U,
+ 0x91548db5U, 0x71c45d05U, 0x0406d46fU, 0x605015ffU,
+ 0x1998fb24U, 0xd6bde997U, 0x894043ccU, 0x67d99e77U,
+ 0xb0e842bdU, 0x07898b88U, 0xe7195b38U, 0x79c8eedbU,
+ 0xa17c0a47U, 0x7c420fe9U, 0xf8841ec9U, 0x00000000U,
+ 0x09808683U, 0x322bed48U, 0x1e1170acU, 0x6c5a724eU,
+ 0xfd0efffbU, 0x0f853856U, 0x3daed51eU, 0x362d3927U,
+ 0x0a0fd964U, 0x685ca621U, 0x9b5b54d1U, 0x24362e3aU,
+ 0x0c0a67b1U, 0x9357e70fU, 0xb4ee96d2U, 0x1b9b919eU,
+ 0x80c0c54fU, 0x61dc20a2U, 0x5a774b69U, 0x1c121a16U,
+ 0xe293ba0aU, 0xc0a02ae5U, 0x3c22e043U, 0x121b171dU,
+ 0x0e090d0bU, 0xf28bc7adU, 0x2db6a8b9U, 0x141ea9c8U,
+ 0x57f11985U, 0xaf75074cU, 0xee99ddbbU, 0xa37f60fdU,
+ 0xf701269fU, 0x5c72f5bcU, 0x44663bc5U, 0x5bfb7e34U,
+ 0x8b432976U, 0xcb23c6dcU, 0xb6edfc68U, 0xb8e4f163U,
+ 0xd731dccaU, 0x42638510U, 0x13972240U, 0x84c61120U,
+ 0x854a247dU, 0xd2bb3df8U, 0xaef93211U, 0xc729a16dU,
+ 0x1d9e2f4bU, 0xdcb230f3U, 0x0d8652ecU, 0x77c1e3d0U,
+ 0x2bb3166cU, 0xa970b999U, 0x119448faU, 0x47e96422U,
+ 0xa8fc8cc4U, 0xa0f03f1aU, 0x567d2cd8U, 0x223390efU,
+ 0x87494ec7U, 0xd938d1c1U, 0x8ccaa2feU, 0x98d40b36U,
+ 0xa6f581cfU, 0xa57ade28U, 0xdab78e26U, 0x3fadbfa4U,
+ 0x2c3a9de4U, 0x5078920dU, 0x6a5fcc9bU, 0x547e4662U,
+ 0xf68d13c2U, 0x90d8b8e8U, 0x2e39f75eU, 0x82c3aff5U,
+ 0x9f5d80beU, 0x69d0937cU, 0x6fd52da9U, 0xcf2512b3U,
+ 0xc8ac993bU, 0x10187da7U, 0xe89c636eU, 0xdb3bbb7bU,
+ 0xcd267809U, 0x6e5918f4U, 0xec9ab701U, 0x834f9aa8U,
+ 0xe6956e65U, 0xaaffe67eU, 0x21bccf08U, 0xef15e8e6U,
+ 0xbae79bd9U, 0x4a6f36ceU, 0xea9f09d4U, 0x29b07cd6U,
+ 0x31a4b2afU, 0x2a3f2331U, 0xc6a59430U, 0x35a266c0U,
+ 0x744ebc37U, 0xfc82caa6U, 0xe090d0b0U, 0x33a7d815U,
+ 0xf104984aU, 0x41ecdaf7U, 0x7fcd500eU, 0x1791f62fU,
+ 0x764dd68dU, 0x43efb04dU, 0xccaa4d54U, 0xe49604dfU,
+ 0x9ed1b5e3U, 0x4c6a881bU, 0xc12c1fb8U, 0x4665517fU,
+ 0x9d5eea04U, 0x018c355dU, 0xfa877473U, 0xfb0b412eU,
+ 0xb3671d5aU, 0x92dbd252U, 0xe9105633U, 0x6dd64713U,
+ 0x9ad7618cU, 0x37a10c7aU, 0x59f8148eU, 0xeb133c89U,
+ 0xcea927eeU, 0xb761c935U, 0xe11ce5edU, 0x7a47b13cU,
+ 0x9cd2df59U, 0x55f2733fU, 0x1814ce79U, 0x73c737bfU,
+ 0x53f7cdeaU, 0x5ffdaa5bU, 0xdf3d6f14U, 0x7844db86U,
+ 0xcaaff381U, 0xb968c43eU, 0x3824342cU, 0xc2a3405fU,
+ 0x161dc372U, 0xbce2250cU, 0x283c498bU, 0xff0d9541U,
+ 0x39a80171U, 0x080cb3deU, 0xd8b4e49cU, 0x6456c190U,
+ 0x7bcb8461U, 0xd532b670U, 0x486c5c74U, 0xd0b85742U,
+};
+static const u32 Td1[256] = {
+ 0x5051f4a7U, 0x537e4165U, 0xc31a17a4U, 0x963a275eU,
+ 0xcb3bab6bU, 0xf11f9d45U, 0xabacfa58U, 0x934be303U,
+ 0x552030faU, 0xf6ad766dU, 0x9188cc76U, 0x25f5024cU,
+ 0xfc4fe5d7U, 0xd7c52acbU, 0x80263544U, 0x8fb562a3U,
+ 0x49deb15aU, 0x6725ba1bU, 0x9845ea0eU, 0xe15dfec0U,
+ 0x02c32f75U, 0x12814cf0U, 0xa38d4697U, 0xc66bd3f9U,
+ 0xe7038f5fU, 0x9515929cU, 0xebbf6d7aU, 0xda955259U,
+ 0x2dd4be83U, 0xd3587421U, 0x2949e069U, 0x448ec9c8U,
+ 0x6a75c289U, 0x78f48e79U, 0x6b99583eU, 0xdd27b971U,
+ 0xb6bee14fU, 0x17f088adU, 0x66c920acU, 0xb47dce3aU,
+ 0x1863df4aU, 0x82e51a31U, 0x60975133U, 0x4562537fU,
+ 0xe0b16477U, 0x84bb6baeU, 0x1cfe81a0U, 0x94f9082bU,
+ 0x58704868U, 0x198f45fdU, 0x8794de6cU, 0xb7527bf8U,
+ 0x23ab73d3U, 0xe2724b02U, 0x57e31f8fU, 0x2a6655abU,
+ 0x07b2eb28U, 0x032fb5c2U, 0x9a86c57bU, 0xa5d33708U,
+ 0xf2302887U, 0xb223bfa5U, 0xba02036aU, 0x5ced1682U,
+ 0x2b8acf1cU, 0x92a779b4U, 0xf0f307f2U, 0xa14e69e2U,
+ 0xcd65daf4U, 0xd50605beU, 0x1fd13462U, 0x8ac4a6feU,
+ 0x9d342e53U, 0xa0a2f355U, 0x32058ae1U, 0x75a4f6ebU,
+ 0x390b83ecU, 0xaa4060efU, 0x065e719fU, 0x51bd6e10U,
+ 0xf93e218aU, 0x3d96dd06U, 0xaedd3e05U, 0x464de6bdU,
+ 0xb591548dU, 0x0571c45dU, 0x6f0406d4U, 0xff605015U,
+ 0x241998fbU, 0x97d6bde9U, 0xcc894043U, 0x7767d99eU,
+ 0xbdb0e842U, 0x8807898bU, 0x38e7195bU, 0xdb79c8eeU,
+ 0x47a17c0aU, 0xe97c420fU, 0xc9f8841eU, 0x00000000U,
+ 0x83098086U, 0x48322bedU, 0xac1e1170U, 0x4e6c5a72U,
+ 0xfbfd0effU, 0x560f8538U, 0x1e3daed5U, 0x27362d39U,
+ 0x640a0fd9U, 0x21685ca6U, 0xd19b5b54U, 0x3a24362eU,
+ 0xb10c0a67U, 0x0f9357e7U, 0xd2b4ee96U, 0x9e1b9b91U,
+ 0x4f80c0c5U, 0xa261dc20U, 0x695a774bU, 0x161c121aU,
+ 0x0ae293baU, 0xe5c0a02aU, 0x433c22e0U, 0x1d121b17U,
+ 0x0b0e090dU, 0xadf28bc7U, 0xb92db6a8U, 0xc8141ea9U,
+ 0x8557f119U, 0x4caf7507U, 0xbbee99ddU, 0xfda37f60U,
+ 0x9ff70126U, 0xbc5c72f5U, 0xc544663bU, 0x345bfb7eU,
+ 0x768b4329U, 0xdccb23c6U, 0x68b6edfcU, 0x63b8e4f1U,
+ 0xcad731dcU, 0x10426385U, 0x40139722U, 0x2084c611U,
+ 0x7d854a24U, 0xf8d2bb3dU, 0x11aef932U, 0x6dc729a1U,
+ 0x4b1d9e2fU, 0xf3dcb230U, 0xec0d8652U, 0xd077c1e3U,
+ 0x6c2bb316U, 0x99a970b9U, 0xfa119448U, 0x2247e964U,
+ 0xc4a8fc8cU, 0x1aa0f03fU, 0xd8567d2cU, 0xef223390U,
+ 0xc787494eU, 0xc1d938d1U, 0xfe8ccaa2U, 0x3698d40bU,
+ 0xcfa6f581U, 0x28a57adeU, 0x26dab78eU, 0xa43fadbfU,
+ 0xe42c3a9dU, 0x0d507892U, 0x9b6a5fccU, 0x62547e46U,
+ 0xc2f68d13U, 0xe890d8b8U, 0x5e2e39f7U, 0xf582c3afU,
+ 0xbe9f5d80U, 0x7c69d093U, 0xa96fd52dU, 0xb3cf2512U,
+ 0x3bc8ac99U, 0xa710187dU, 0x6ee89c63U, 0x7bdb3bbbU,
+ 0x09cd2678U, 0xf46e5918U, 0x01ec9ab7U, 0xa8834f9aU,
+ 0x65e6956eU, 0x7eaaffe6U, 0x0821bccfU, 0xe6ef15e8U,
+ 0xd9bae79bU, 0xce4a6f36U, 0xd4ea9f09U, 0xd629b07cU,
+ 0xaf31a4b2U, 0x312a3f23U, 0x30c6a594U, 0xc035a266U,
+ 0x37744ebcU, 0xa6fc82caU, 0xb0e090d0U, 0x1533a7d8U,
+ 0x4af10498U, 0xf741ecdaU, 0x0e7fcd50U, 0x2f1791f6U,
+ 0x8d764dd6U, 0x4d43efb0U, 0x54ccaa4dU, 0xdfe49604U,
+ 0xe39ed1b5U, 0x1b4c6a88U, 0xb8c12c1fU, 0x7f466551U,
+ 0x049d5eeaU, 0x5d018c35U, 0x73fa8774U, 0x2efb0b41U,
+ 0x5ab3671dU, 0x5292dbd2U, 0x33e91056U, 0x136dd647U,
+ 0x8c9ad761U, 0x7a37a10cU, 0x8e59f814U, 0x89eb133cU,
+ 0xeecea927U, 0x35b761c9U, 0xede11ce5U, 0x3c7a47b1U,
+ 0x599cd2dfU, 0x3f55f273U, 0x791814ceU, 0xbf73c737U,
+ 0xea53f7cdU, 0x5b5ffdaaU, 0x14df3d6fU, 0x867844dbU,
+ 0x81caaff3U, 0x3eb968c4U, 0x2c382434U, 0x5fc2a340U,
+ 0x72161dc3U, 0x0cbce225U, 0x8b283c49U, 0x41ff0d95U,
+ 0x7139a801U, 0xde080cb3U, 0x9cd8b4e4U, 0x906456c1U,
+ 0x617bcb84U, 0x70d532b6U, 0x74486c5cU, 0x42d0b857U,
+};
+static const u32 Td2[256] = {
+ 0xa75051f4U, 0x65537e41U, 0xa4c31a17U, 0x5e963a27U,
+ 0x6bcb3babU, 0x45f11f9dU, 0x58abacfaU, 0x03934be3U,
+ 0xfa552030U, 0x6df6ad76U, 0x769188ccU, 0x4c25f502U,
+ 0xd7fc4fe5U, 0xcbd7c52aU, 0x44802635U, 0xa38fb562U,
+ 0x5a49deb1U, 0x1b6725baU, 0x0e9845eaU, 0xc0e15dfeU,
+ 0x7502c32fU, 0xf012814cU, 0x97a38d46U, 0xf9c66bd3U,
+ 0x5fe7038fU, 0x9c951592U, 0x7aebbf6dU, 0x59da9552U,
+ 0x832dd4beU, 0x21d35874U, 0x692949e0U, 0xc8448ec9U,
+ 0x896a75c2U, 0x7978f48eU, 0x3e6b9958U, 0x71dd27b9U,
+ 0x4fb6bee1U, 0xad17f088U, 0xac66c920U, 0x3ab47dceU,
+ 0x4a1863dfU, 0x3182e51aU, 0x33609751U, 0x7f456253U,
+ 0x77e0b164U, 0xae84bb6bU, 0xa01cfe81U, 0x2b94f908U,
+ 0x68587048U, 0xfd198f45U, 0x6c8794deU, 0xf8b7527bU,
+ 0xd323ab73U, 0x02e2724bU, 0x8f57e31fU, 0xab2a6655U,
+ 0x2807b2ebU, 0xc2032fb5U, 0x7b9a86c5U, 0x08a5d337U,
+ 0x87f23028U, 0xa5b223bfU, 0x6aba0203U, 0x825ced16U,
+ 0x1c2b8acfU, 0xb492a779U, 0xf2f0f307U, 0xe2a14e69U,
+ 0xf4cd65daU, 0xbed50605U, 0x621fd134U, 0xfe8ac4a6U,
+ 0x539d342eU, 0x55a0a2f3U, 0xe132058aU, 0xeb75a4f6U,
+ 0xec390b83U, 0xefaa4060U, 0x9f065e71U, 0x1051bd6eU,
+ 0x8af93e21U, 0x063d96ddU, 0x05aedd3eU, 0xbd464de6U,
+ 0x8db59154U, 0x5d0571c4U, 0xd46f0406U, 0x15ff6050U,
+ 0xfb241998U, 0xe997d6bdU, 0x43cc8940U, 0x9e7767d9U,
+ 0x42bdb0e8U, 0x8b880789U, 0x5b38e719U, 0xeedb79c8U,
+ 0x0a47a17cU, 0x0fe97c42U, 0x1ec9f884U, 0x00000000U,
+ 0x86830980U, 0xed48322bU, 0x70ac1e11U, 0x724e6c5aU,
+ 0xfffbfd0eU, 0x38560f85U, 0xd51e3daeU, 0x3927362dU,
+ 0xd9640a0fU, 0xa621685cU, 0x54d19b5bU, 0x2e3a2436U,
+ 0x67b10c0aU, 0xe70f9357U, 0x96d2b4eeU, 0x919e1b9bU,
+ 0xc54f80c0U, 0x20a261dcU, 0x4b695a77U, 0x1a161c12U,
+ 0xba0ae293U, 0x2ae5c0a0U, 0xe0433c22U, 0x171d121bU,
+ 0x0d0b0e09U, 0xc7adf28bU, 0xa8b92db6U, 0xa9c8141eU,
+ 0x198557f1U, 0x074caf75U, 0xddbbee99U, 0x60fda37fU,
+ 0x269ff701U, 0xf5bc5c72U, 0x3bc54466U, 0x7e345bfbU,
+ 0x29768b43U, 0xc6dccb23U, 0xfc68b6edU, 0xf163b8e4U,
+ 0xdccad731U, 0x85104263U, 0x22401397U, 0x112084c6U,
+ 0x247d854aU, 0x3df8d2bbU, 0x3211aef9U, 0xa16dc729U,
+ 0x2f4b1d9eU, 0x30f3dcb2U, 0x52ec0d86U, 0xe3d077c1U,
+ 0x166c2bb3U, 0xb999a970U, 0x48fa1194U, 0x642247e9U,
+ 0x8cc4a8fcU, 0x3f1aa0f0U, 0x2cd8567dU, 0x90ef2233U,
+ 0x4ec78749U, 0xd1c1d938U, 0xa2fe8ccaU, 0x0b3698d4U,
+ 0x81cfa6f5U, 0xde28a57aU, 0x8e26dab7U, 0xbfa43fadU,
+ 0x9de42c3aU, 0x920d5078U, 0xcc9b6a5fU, 0x4662547eU,
+ 0x13c2f68dU, 0xb8e890d8U, 0xf75e2e39U, 0xaff582c3U,
+ 0x80be9f5dU, 0x937c69d0U, 0x2da96fd5U, 0x12b3cf25U,
+ 0x993bc8acU, 0x7da71018U, 0x636ee89cU, 0xbb7bdb3bU,
+ 0x7809cd26U, 0x18f46e59U, 0xb701ec9aU, 0x9aa8834fU,
+ 0x6e65e695U, 0xe67eaaffU, 0xcf0821bcU, 0xe8e6ef15U,
+ 0x9bd9bae7U, 0x36ce4a6fU, 0x09d4ea9fU, 0x7cd629b0U,
+ 0xb2af31a4U, 0x23312a3fU, 0x9430c6a5U, 0x66c035a2U,
+ 0xbc37744eU, 0xcaa6fc82U, 0xd0b0e090U, 0xd81533a7U,
+ 0x984af104U, 0xdaf741ecU, 0x500e7fcdU, 0xf62f1791U,
+ 0xd68d764dU, 0xb04d43efU, 0x4d54ccaaU, 0x04dfe496U,
+ 0xb5e39ed1U, 0x881b4c6aU, 0x1fb8c12cU, 0x517f4665U,
+ 0xea049d5eU, 0x355d018cU, 0x7473fa87U, 0x412efb0bU,
+ 0x1d5ab367U, 0xd25292dbU, 0x5633e910U, 0x47136dd6U,
+ 0x618c9ad7U, 0x0c7a37a1U, 0x148e59f8U, 0x3c89eb13U,
+ 0x27eecea9U, 0xc935b761U, 0xe5ede11cU, 0xb13c7a47U,
+ 0xdf599cd2U, 0x733f55f2U, 0xce791814U, 0x37bf73c7U,
+ 0xcdea53f7U, 0xaa5b5ffdU, 0x6f14df3dU, 0xdb867844U,
+ 0xf381caafU, 0xc43eb968U, 0x342c3824U, 0x405fc2a3U,
+ 0xc372161dU, 0x250cbce2U, 0x498b283cU, 0x9541ff0dU,
+ 0x017139a8U, 0xb3de080cU, 0xe49cd8b4U, 0xc1906456U,
+ 0x84617bcbU, 0xb670d532U, 0x5c74486cU, 0x5742d0b8U,
+};
+static const u32 Td3[256] = {
+ 0xf4a75051U, 0x4165537eU, 0x17a4c31aU, 0x275e963aU,
+ 0xab6bcb3bU, 0x9d45f11fU, 0xfa58abacU, 0xe303934bU,
+ 0x30fa5520U, 0x766df6adU, 0xcc769188U, 0x024c25f5U,
+ 0xe5d7fc4fU, 0x2acbd7c5U, 0x35448026U, 0x62a38fb5U,
+ 0xb15a49deU, 0xba1b6725U, 0xea0e9845U, 0xfec0e15dU,
+ 0x2f7502c3U, 0x4cf01281U, 0x4697a38dU, 0xd3f9c66bU,
+ 0x8f5fe703U, 0x929c9515U, 0x6d7aebbfU, 0x5259da95U,
+ 0xbe832dd4U, 0x7421d358U, 0xe0692949U, 0xc9c8448eU,
+ 0xc2896a75U, 0x8e7978f4U, 0x583e6b99U, 0xb971dd27U,
+ 0xe14fb6beU, 0x88ad17f0U, 0x20ac66c9U, 0xce3ab47dU,
+ 0xdf4a1863U, 0x1a3182e5U, 0x51336097U, 0x537f4562U,
+ 0x6477e0b1U, 0x6bae84bbU, 0x81a01cfeU, 0x082b94f9U,
+ 0x48685870U, 0x45fd198fU, 0xde6c8794U, 0x7bf8b752U,
+ 0x73d323abU, 0x4b02e272U, 0x1f8f57e3U, 0x55ab2a66U,
+ 0xeb2807b2U, 0xb5c2032fU, 0xc57b9a86U, 0x3708a5d3U,
+ 0x2887f230U, 0xbfa5b223U, 0x036aba02U, 0x16825cedU,
+ 0xcf1c2b8aU, 0x79b492a7U, 0x07f2f0f3U, 0x69e2a14eU,
+ 0xdaf4cd65U, 0x05bed506U, 0x34621fd1U, 0xa6fe8ac4U,
+ 0x2e539d34U, 0xf355a0a2U, 0x8ae13205U, 0xf6eb75a4U,
+ 0x83ec390bU, 0x60efaa40U, 0x719f065eU, 0x6e1051bdU,
+ 0x218af93eU, 0xdd063d96U, 0x3e05aeddU, 0xe6bd464dU,
+ 0x548db591U, 0xc45d0571U, 0x06d46f04U, 0x5015ff60U,
+ 0x98fb2419U, 0xbde997d6U, 0x4043cc89U, 0xd99e7767U,
+ 0xe842bdb0U, 0x898b8807U, 0x195b38e7U, 0xc8eedb79U,
+ 0x7c0a47a1U, 0x420fe97cU, 0x841ec9f8U, 0x00000000U,
+ 0x80868309U, 0x2bed4832U, 0x1170ac1eU, 0x5a724e6cU,
+ 0x0efffbfdU, 0x8538560fU, 0xaed51e3dU, 0x2d392736U,
+ 0x0fd9640aU, 0x5ca62168U, 0x5b54d19bU, 0x362e3a24U,
+ 0x0a67b10cU, 0x57e70f93U, 0xee96d2b4U, 0x9b919e1bU,
+ 0xc0c54f80U, 0xdc20a261U, 0x774b695aU, 0x121a161cU,
+ 0x93ba0ae2U, 0xa02ae5c0U, 0x22e0433cU, 0x1b171d12U,
+ 0x090d0b0eU, 0x8bc7adf2U, 0xb6a8b92dU, 0x1ea9c814U,
+ 0xf1198557U, 0x75074cafU, 0x99ddbbeeU, 0x7f60fda3U,
+ 0x01269ff7U, 0x72f5bc5cU, 0x663bc544U, 0xfb7e345bU,
+ 0x4329768bU, 0x23c6dccbU, 0xedfc68b6U, 0xe4f163b8U,
+ 0x31dccad7U, 0x63851042U, 0x97224013U, 0xc6112084U,
+ 0x4a247d85U, 0xbb3df8d2U, 0xf93211aeU, 0x29a16dc7U,
+ 0x9e2f4b1dU, 0xb230f3dcU, 0x8652ec0dU, 0xc1e3d077U,
+ 0xb3166c2bU, 0x70b999a9U, 0x9448fa11U, 0xe9642247U,
+ 0xfc8cc4a8U, 0xf03f1aa0U, 0x7d2cd856U, 0x3390ef22U,
+ 0x494ec787U, 0x38d1c1d9U, 0xcaa2fe8cU, 0xd40b3698U,
+ 0xf581cfa6U, 0x7ade28a5U, 0xb78e26daU, 0xadbfa43fU,
+ 0x3a9de42cU, 0x78920d50U, 0x5fcc9b6aU, 0x7e466254U,
+ 0x8d13c2f6U, 0xd8b8e890U, 0x39f75e2eU, 0xc3aff582U,
+ 0x5d80be9fU, 0xd0937c69U, 0xd52da96fU, 0x2512b3cfU,
+ 0xac993bc8U, 0x187da710U, 0x9c636ee8U, 0x3bbb7bdbU,
+ 0x267809cdU, 0x5918f46eU, 0x9ab701ecU, 0x4f9aa883U,
+ 0x956e65e6U, 0xffe67eaaU, 0xbccf0821U, 0x15e8e6efU,
+ 0xe79bd9baU, 0x6f36ce4aU, 0x9f09d4eaU, 0xb07cd629U,
+ 0xa4b2af31U, 0x3f23312aU, 0xa59430c6U, 0xa266c035U,
+ 0x4ebc3774U, 0x82caa6fcU, 0x90d0b0e0U, 0xa7d81533U,
+ 0x04984af1U, 0xecdaf741U, 0xcd500e7fU, 0x91f62f17U,
+ 0x4dd68d76U, 0xefb04d43U, 0xaa4d54ccU, 0x9604dfe4U,
+ 0xd1b5e39eU, 0x6a881b4cU, 0x2c1fb8c1U, 0x65517f46U,
+ 0x5eea049dU, 0x8c355d01U, 0x877473faU, 0x0b412efbU,
+ 0x671d5ab3U, 0xdbd25292U, 0x105633e9U, 0xd647136dU,
+ 0xd7618c9aU, 0xa10c7a37U, 0xf8148e59U, 0x133c89ebU,
+ 0xa927eeceU, 0x61c935b7U, 0x1ce5ede1U, 0x47b13c7aU,
+ 0xd2df599cU, 0xf2733f55U, 0x14ce7918U, 0xc737bf73U,
+ 0xf7cdea53U, 0xfdaa5b5fU, 0x3d6f14dfU, 0x44db8678U,
+ 0xaff381caU, 0x68c43eb9U, 0x24342c38U, 0xa3405fc2U,
+ 0x1dc37216U, 0xe2250cbcU, 0x3c498b28U, 0x0d9541ffU,
+ 0xa8017139U, 0x0cb3de08U, 0xb4e49cd8U, 0x56c19064U,
+ 0xcb84617bU, 0x32b670d5U, 0x6c5c7448U, 0xb85742d0U,
+};
+static const u32 Td4[256] = {
+ 0x52525252U, 0x09090909U, 0x6a6a6a6aU, 0xd5d5d5d5U,
+ 0x30303030U, 0x36363636U, 0xa5a5a5a5U, 0x38383838U,
+ 0xbfbfbfbfU, 0x40404040U, 0xa3a3a3a3U, 0x9e9e9e9eU,
+ 0x81818181U, 0xf3f3f3f3U, 0xd7d7d7d7U, 0xfbfbfbfbU,
+ 0x7c7c7c7cU, 0xe3e3e3e3U, 0x39393939U, 0x82828282U,
+ 0x9b9b9b9bU, 0x2f2f2f2fU, 0xffffffffU, 0x87878787U,
+ 0x34343434U, 0x8e8e8e8eU, 0x43434343U, 0x44444444U,
+ 0xc4c4c4c4U, 0xdedededeU, 0xe9e9e9e9U, 0xcbcbcbcbU,
+ 0x54545454U, 0x7b7b7b7bU, 0x94949494U, 0x32323232U,
+ 0xa6a6a6a6U, 0xc2c2c2c2U, 0x23232323U, 0x3d3d3d3dU,
+ 0xeeeeeeeeU, 0x4c4c4c4cU, 0x95959595U, 0x0b0b0b0bU,
+ 0x42424242U, 0xfafafafaU, 0xc3c3c3c3U, 0x4e4e4e4eU,
+ 0x08080808U, 0x2e2e2e2eU, 0xa1a1a1a1U, 0x66666666U,
+ 0x28282828U, 0xd9d9d9d9U, 0x24242424U, 0xb2b2b2b2U,
+ 0x76767676U, 0x5b5b5b5bU, 0xa2a2a2a2U, 0x49494949U,
+ 0x6d6d6d6dU, 0x8b8b8b8bU, 0xd1d1d1d1U, 0x25252525U,
+ 0x72727272U, 0xf8f8f8f8U, 0xf6f6f6f6U, 0x64646464U,
+ 0x86868686U, 0x68686868U, 0x98989898U, 0x16161616U,
+ 0xd4d4d4d4U, 0xa4a4a4a4U, 0x5c5c5c5cU, 0xccccccccU,
+ 0x5d5d5d5dU, 0x65656565U, 0xb6b6b6b6U, 0x92929292U,
+ 0x6c6c6c6cU, 0x70707070U, 0x48484848U, 0x50505050U,
+ 0xfdfdfdfdU, 0xededededU, 0xb9b9b9b9U, 0xdadadadaU,
+ 0x5e5e5e5eU, 0x15151515U, 0x46464646U, 0x57575757U,
+ 0xa7a7a7a7U, 0x8d8d8d8dU, 0x9d9d9d9dU, 0x84848484U,
+ 0x90909090U, 0xd8d8d8d8U, 0xababababU, 0x00000000U,
+ 0x8c8c8c8cU, 0xbcbcbcbcU, 0xd3d3d3d3U, 0x0a0a0a0aU,
+ 0xf7f7f7f7U, 0xe4e4e4e4U, 0x58585858U, 0x05050505U,
+ 0xb8b8b8b8U, 0xb3b3b3b3U, 0x45454545U, 0x06060606U,
+ 0xd0d0d0d0U, 0x2c2c2c2cU, 0x1e1e1e1eU, 0x8f8f8f8fU,
+ 0xcacacacaU, 0x3f3f3f3fU, 0x0f0f0f0fU, 0x02020202U,
+ 0xc1c1c1c1U, 0xafafafafU, 0xbdbdbdbdU, 0x03030303U,
+ 0x01010101U, 0x13131313U, 0x8a8a8a8aU, 0x6b6b6b6bU,
+ 0x3a3a3a3aU, 0x91919191U, 0x11111111U, 0x41414141U,
+ 0x4f4f4f4fU, 0x67676767U, 0xdcdcdcdcU, 0xeaeaeaeaU,
+ 0x97979797U, 0xf2f2f2f2U, 0xcfcfcfcfU, 0xcecececeU,
+ 0xf0f0f0f0U, 0xb4b4b4b4U, 0xe6e6e6e6U, 0x73737373U,
+ 0x96969696U, 0xacacacacU, 0x74747474U, 0x22222222U,
+ 0xe7e7e7e7U, 0xadadadadU, 0x35353535U, 0x85858585U,
+ 0xe2e2e2e2U, 0xf9f9f9f9U, 0x37373737U, 0xe8e8e8e8U,
+ 0x1c1c1c1cU, 0x75757575U, 0xdfdfdfdfU, 0x6e6e6e6eU,
+ 0x47474747U, 0xf1f1f1f1U, 0x1a1a1a1aU, 0x71717171U,
+ 0x1d1d1d1dU, 0x29292929U, 0xc5c5c5c5U, 0x89898989U,
+ 0x6f6f6f6fU, 0xb7b7b7b7U, 0x62626262U, 0x0e0e0e0eU,
+ 0xaaaaaaaaU, 0x18181818U, 0xbebebebeU, 0x1b1b1b1bU,
+ 0xfcfcfcfcU, 0x56565656U, 0x3e3e3e3eU, 0x4b4b4b4bU,
+ 0xc6c6c6c6U, 0xd2d2d2d2U, 0x79797979U, 0x20202020U,
+ 0x9a9a9a9aU, 0xdbdbdbdbU, 0xc0c0c0c0U, 0xfefefefeU,
+ 0x78787878U, 0xcdcdcdcdU, 0x5a5a5a5aU, 0xf4f4f4f4U,
+ 0x1f1f1f1fU, 0xddddddddU, 0xa8a8a8a8U, 0x33333333U,
+ 0x88888888U, 0x07070707U, 0xc7c7c7c7U, 0x31313131U,
+ 0xb1b1b1b1U, 0x12121212U, 0x10101010U, 0x59595959U,
+ 0x27272727U, 0x80808080U, 0xececececU, 0x5f5f5f5fU,
+ 0x60606060U, 0x51515151U, 0x7f7f7f7fU, 0xa9a9a9a9U,
+ 0x19191919U, 0xb5b5b5b5U, 0x4a4a4a4aU, 0x0d0d0d0dU,
+ 0x2d2d2d2dU, 0xe5e5e5e5U, 0x7a7a7a7aU, 0x9f9f9f9fU,
+ 0x93939393U, 0xc9c9c9c9U, 0x9c9c9c9cU, 0xefefefefU,
+ 0xa0a0a0a0U, 0xe0e0e0e0U, 0x3b3b3b3bU, 0x4d4d4d4dU,
+ 0xaeaeaeaeU, 0x2a2a2a2aU, 0xf5f5f5f5U, 0xb0b0b0b0U,
+ 0xc8c8c8c8U, 0xebebebebU, 0xbbbbbbbbU, 0x3c3c3c3cU,
+ 0x83838383U, 0x53535353U, 0x99999999U, 0x61616161U,
+ 0x17171717U, 0x2b2b2b2bU, 0x04040404U, 0x7e7e7e7eU,
+ 0xbabababaU, 0x77777777U, 0xd6d6d6d6U, 0x26262626U,
+ 0xe1e1e1e1U, 0x69696969U, 0x14141414U, 0x63636363U,
+ 0x55555555U, 0x21212121U, 0x0c0c0c0cU, 0x7d7d7d7dU,
+};
+static const u32 rcon[] = {
+ 0x01000000, 0x02000000, 0x04000000, 0x08000000,
+ 0x10000000, 0x20000000, 0x40000000, 0x80000000,
+ 0x1B000000, 0x36000000, /* for 128-bit blocks, Rijndael never uses more than 10 rcon values */
+};
+
+#define GETU32(pt) (((u32)(pt)[0] << 24) ^ ((u32)(pt)[1] << 16) ^ ((u32)(pt)[2] << 8) ^ ((u32)(pt)[3]))
+#define PUTU32(ct, st) { (ct)[0] = (u8)((st) >> 24); (ct)[1] = (u8)((st) >> 16); (ct)[2] = (u8)((st) >> 8); (ct)[3] = (u8)(st); }
+
+/**
+ * Expand the cipher key into the encryption key schedule.
+ *
+ * @return the number of rounds for the given cipher key size.
+ */
+int
+rijndaelKeySetupEnc(u32 rk[/*4*(Nr + 1)*/], const u8 cipherKey[], int keyBits)
+{
+ int i = 0;
+ u32 temp;
+
+ rk[0] = GETU32(cipherKey );
+ rk[1] = GETU32(cipherKey + 4);
+ rk[2] = GETU32(cipherKey + 8);
+ rk[3] = GETU32(cipherKey + 12);
+ if (keyBits == 128) {
+ for (;;) {
+ temp = rk[3];
+ rk[4] = rk[0] ^
+ (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
+ (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
+ (Te4[(temp ) & 0xff] & 0x0000ff00) ^
+ (Te4[(temp >> 24) ] & 0x000000ff) ^
+ rcon[i];
+ rk[5] = rk[1] ^ rk[4];
+ rk[6] = rk[2] ^ rk[5];
+ rk[7] = rk[3] ^ rk[6];
+ if (++i == 10) {
+ return 10;
+ }
+ rk += 4;
+ }
+ }
+ rk[4] = GETU32(cipherKey + 16);
+ rk[5] = GETU32(cipherKey + 20);
+ if (keyBits == 192) {
+ for (;;) {
+ temp = rk[ 5];
+ rk[ 6] = rk[ 0] ^
+ (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
+ (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
+ (Te4[(temp ) & 0xff] & 0x0000ff00) ^
+ (Te4[(temp >> 24) ] & 0x000000ff) ^
+ rcon[i];
+ rk[ 7] = rk[ 1] ^ rk[ 6];
+ rk[ 8] = rk[ 2] ^ rk[ 7];
+ rk[ 9] = rk[ 3] ^ rk[ 8];
+ if (++i == 8) {
+ return 12;
+ }
+ rk[10] = rk[ 4] ^ rk[ 9];
+ rk[11] = rk[ 5] ^ rk[10];
+ rk += 6;
+ }
+ }
+ rk[6] = GETU32(cipherKey + 24);
+ rk[7] = GETU32(cipherKey + 28);
+ if (keyBits == 256) {
+ for (;;) {
+ temp = rk[ 7];
+ rk[ 8] = rk[ 0] ^
+ (Te4[(temp >> 16) & 0xff] & 0xff000000) ^
+ (Te4[(temp >> 8) & 0xff] & 0x00ff0000) ^
+ (Te4[(temp ) & 0xff] & 0x0000ff00) ^
+ (Te4[(temp >> 24) ] & 0x000000ff) ^
+ rcon[i];
+ rk[ 9] = rk[ 1] ^ rk[ 8];
+ rk[10] = rk[ 2] ^ rk[ 9];
+ rk[11] = rk[ 3] ^ rk[10];
+ if (++i == 7) {
+ return 14;
+ }
+ temp = rk[11];
+ rk[12] = rk[ 4] ^
+ (Te4[(temp >> 24) ] & 0xff000000) ^
+ (Te4[(temp >> 16) & 0xff] & 0x00ff0000) ^
+ (Te4[(temp >> 8) & 0xff] & 0x0000ff00) ^
+ (Te4[(temp ) & 0xff] & 0x000000ff);
+ rk[13] = rk[ 5] ^ rk[12];
+ rk[14] = rk[ 6] ^ rk[13];
+ rk[15] = rk[ 7] ^ rk[14];
+ rk += 8;
+ }
+ }
+ return 0;
+}
+
+/**
+ * Expand the cipher key into the decryption key schedule.
+ *
+ * @return the number of rounds for the given cipher key size.
+ */
+int
+rijndaelKeySetupDec(u32 rk[/*4*(Nr + 1)*/], const u8 cipherKey[], int keyBits)
+{
+ int Nr, i, j;
+ u32 temp;
+
+ /* expand the cipher key: */
+ Nr = rijndaelKeySetupEnc(rk, cipherKey, keyBits);
+
+ /* invert the order of the round keys: */
+ for (i = 0, j = 4*Nr; i < j; i += 4, j -= 4) {
+ temp = rk[i ]; rk[i ] = rk[j ]; rk[j ] = temp;
+ temp = rk[i + 1]; rk[i + 1] = rk[j + 1]; rk[j + 1] = temp;
+ temp = rk[i + 2]; rk[i + 2] = rk[j + 2]; rk[j + 2] = temp;
+ temp = rk[i + 3]; rk[i + 3] = rk[j + 3]; rk[j + 3] = temp;
+ }
+ /* apply the inverse MixColumn transform to all round keys but the first and the last: */
+ for (i = 1; i < Nr; i++) {
+ rk += 4;
+ rk[0] =
+ Td0[Te4[(rk[0] >> 24) ] & 0xff] ^
+ Td1[Te4[(rk[0] >> 16) & 0xff] & 0xff] ^
+ Td2[Te4[(rk[0] >> 8) & 0xff] & 0xff] ^
+ Td3[Te4[(rk[0] ) & 0xff] & 0xff];
+ rk[1] =
+ Td0[Te4[(rk[1] >> 24) ] & 0xff] ^
+ Td1[Te4[(rk[1] >> 16) & 0xff] & 0xff] ^
+ Td2[Te4[(rk[1] >> 8) & 0xff] & 0xff] ^
+ Td3[Te4[(rk[1] ) & 0xff] & 0xff];
+ rk[2] =
+ Td0[Te4[(rk[2] >> 24) ] & 0xff] ^
+ Td1[Te4[(rk[2] >> 16) & 0xff] & 0xff] ^
+ Td2[Te4[(rk[2] >> 8) & 0xff] & 0xff] ^
+ Td3[Te4[(rk[2] ) & 0xff] & 0xff];
+ rk[3] =
+ Td0[Te4[(rk[3] >> 24) ] & 0xff] ^
+ Td1[Te4[(rk[3] >> 16) & 0xff] & 0xff] ^
+ Td2[Te4[(rk[3] >> 8) & 0xff] & 0xff] ^
+ Td3[Te4[(rk[3] ) & 0xff] & 0xff];
+ }
+ return Nr;
+}
+
+void
+rijndaelEncrypt(const u32 rk[/*4*(Nr + 1)*/], int Nr, const u8 pt[16],
+ u8 ct[16])
+{
+ u32 s0, s1, s2, s3, t0, t1, t2, t3;
+#ifndef FULL_UNROLL
+ int r;
+#endif /* ?FULL_UNROLL */
+
+ /*
+ * map byte array block to cipher state
+ * and add initial round key:
+ */
+ s0 = GETU32(pt ) ^ rk[0];
+ s1 = GETU32(pt + 4) ^ rk[1];
+ s2 = GETU32(pt + 8) ^ rk[2];
+ s3 = GETU32(pt + 12) ^ rk[3];
+#ifdef FULL_UNROLL
+ /* round 1: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[ 4];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[ 5];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[ 6];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[ 7];
+ /* round 2: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[ 8];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[ 9];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[10];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[11];
+ /* round 3: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[12];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[13];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[14];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[15];
+ /* round 4: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[16];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[17];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[18];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[19];
+ /* round 5: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[20];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[21];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[22];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[23];
+ /* round 6: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[24];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[25];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[26];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[27];
+ /* round 7: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[28];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[29];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[30];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[31];
+ /* round 8: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[32];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[33];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[34];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[35];
+ /* round 9: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[36];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[37];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[38];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[39];
+ if (Nr > 10) {
+ /* round 10: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[40];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[41];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[42];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[43];
+ /* round 11: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[44];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[45];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[46];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[47];
+ if (Nr > 12) {
+ /* round 12: */
+ s0 = Te0[t0 >> 24] ^ Te1[(t1 >> 16) & 0xff] ^ Te2[(t2 >> 8) & 0xff] ^ Te3[t3 & 0xff] ^ rk[48];
+ s1 = Te0[t1 >> 24] ^ Te1[(t2 >> 16) & 0xff] ^ Te2[(t3 >> 8) & 0xff] ^ Te3[t0 & 0xff] ^ rk[49];
+ s2 = Te0[t2 >> 24] ^ Te1[(t3 >> 16) & 0xff] ^ Te2[(t0 >> 8) & 0xff] ^ Te3[t1 & 0xff] ^ rk[50];
+ s3 = Te0[t3 >> 24] ^ Te1[(t0 >> 16) & 0xff] ^ Te2[(t1 >> 8) & 0xff] ^ Te3[t2 & 0xff] ^ rk[51];
+ /* round 13: */
+ t0 = Te0[s0 >> 24] ^ Te1[(s1 >> 16) & 0xff] ^ Te2[(s2 >> 8) & 0xff] ^ Te3[s3 & 0xff] ^ rk[52];
+ t1 = Te0[s1 >> 24] ^ Te1[(s2 >> 16) & 0xff] ^ Te2[(s3 >> 8) & 0xff] ^ Te3[s0 & 0xff] ^ rk[53];
+ t2 = Te0[s2 >> 24] ^ Te1[(s3 >> 16) & 0xff] ^ Te2[(s0 >> 8) & 0xff] ^ Te3[s1 & 0xff] ^ rk[54];
+ t3 = Te0[s3 >> 24] ^ Te1[(s0 >> 16) & 0xff] ^ Te2[(s1 >> 8) & 0xff] ^ Te3[s2 & 0xff] ^ rk[55];
+ }
+ }
+ rk += Nr << 2;
+#else /* !FULL_UNROLL */
+ /*
+ * Nr - 1 full rounds:
+ */
+ r = Nr >> 1;
+ for (;;) {
+ t0 =
+ Te0[(s0 >> 24) ] ^
+ Te1[(s1 >> 16) & 0xff] ^
+ Te2[(s2 >> 8) & 0xff] ^
+ Te3[(s3 ) & 0xff] ^
+ rk[4];
+ t1 =
+ Te0[(s1 >> 24) ] ^
+ Te1[(s2 >> 16) & 0xff] ^
+ Te2[(s3 >> 8) & 0xff] ^
+ Te3[(s0 ) & 0xff] ^
+ rk[5];
+ t2 =
+ Te0[(s2 >> 24) ] ^
+ Te1[(s3 >> 16) & 0xff] ^
+ Te2[(s0 >> 8) & 0xff] ^
+ Te3[(s1 ) & 0xff] ^
+ rk[6];
+ t3 =
+ Te0[(s3 >> 24) ] ^
+ Te1[(s0 >> 16) & 0xff] ^
+ Te2[(s1 >> 8) & 0xff] ^
+ Te3[(s2 ) & 0xff] ^
+ rk[7];
+
+ rk += 8;
+ if (--r == 0) {
+ break;
+ }
+
+ s0 =
+ Te0[(t0 >> 24) ] ^
+ Te1[(t1 >> 16) & 0xff] ^
+ Te2[(t2 >> 8) & 0xff] ^
+ Te3[(t3 ) & 0xff] ^
+ rk[0];
+ s1 =
+ Te0[(t1 >> 24) ] ^
+ Te1[(t2 >> 16) & 0xff] ^
+ Te2[(t3 >> 8) & 0xff] ^
+ Te3[(t0 ) & 0xff] ^
+ rk[1];
+ s2 =
+ Te0[(t2 >> 24) ] ^
+ Te1[(t3 >> 16) & 0xff] ^
+ Te2[(t0 >> 8) & 0xff] ^
+ Te3[(t1 ) & 0xff] ^
+ rk[2];
+ s3 =
+ Te0[(t3 >> 24) ] ^
+ Te1[(t0 >> 16) & 0xff] ^
+ Te2[(t1 >> 8) & 0xff] ^
+ Te3[(t2 ) & 0xff] ^
+ rk[3];
+ }
+#endif /* ?FULL_UNROLL */
+ /*
+ * apply last round and
+ * map cipher state to byte array block:
+ */
+ s0 =
+ (Te4[(t0 >> 24) ] & 0xff000000) ^
+ (Te4[(t1 >> 16) & 0xff] & 0x00ff0000) ^
+ (Te4[(t2 >> 8) & 0xff] & 0x0000ff00) ^
+ (Te4[(t3 ) & 0xff] & 0x000000ff) ^
+ rk[0];
+ PUTU32(ct , s0);
+ s1 =
+ (Te4[(t1 >> 24) ] & 0xff000000) ^
+ (Te4[(t2 >> 16) & 0xff] & 0x00ff0000) ^
+ (Te4[(t3 >> 8) & 0xff] & 0x0000ff00) ^
+ (Te4[(t0 ) & 0xff] & 0x000000ff) ^
+ rk[1];
+ PUTU32(ct + 4, s1);
+ s2 =
+ (Te4[(t2 >> 24) ] & 0xff000000) ^
+ (Te4[(t3 >> 16) & 0xff] & 0x00ff0000) ^
+ (Te4[(t0 >> 8) & 0xff] & 0x0000ff00) ^
+ (Te4[(t1 ) & 0xff] & 0x000000ff) ^
+ rk[2];
+ PUTU32(ct + 8, s2);
+ s3 =
+ (Te4[(t3 >> 24) ] & 0xff000000) ^
+ (Te4[(t0 >> 16) & 0xff] & 0x00ff0000) ^
+ (Te4[(t1 >> 8) & 0xff] & 0x0000ff00) ^
+ (Te4[(t2 ) & 0xff] & 0x000000ff) ^
+ rk[3];
+ PUTU32(ct + 12, s3);
+}
+
+static void
+rijndaelDecrypt(const u32 rk[/*4*(Nr + 1)*/], int Nr, const u8 ct[16],
+ u8 pt[16])
+{
+ u32 s0, s1, s2, s3, t0, t1, t2, t3;
+#ifndef FULL_UNROLL
+ int r;
+#endif /* ?FULL_UNROLL */
+
+ /*
+ * map byte array block to cipher state
+ * and add initial round key:
+ */
+ s0 = GETU32(ct ) ^ rk[0];
+ s1 = GETU32(ct + 4) ^ rk[1];
+ s2 = GETU32(ct + 8) ^ rk[2];
+ s3 = GETU32(ct + 12) ^ rk[3];
+#ifdef FULL_UNROLL
+ /* round 1: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[ 4];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[ 5];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[ 6];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[ 7];
+ /* round 2: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[ 8];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[ 9];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[10];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[11];
+ /* round 3: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[12];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[13];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[14];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[15];
+ /* round 4: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[16];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[17];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[18];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[19];
+ /* round 5: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[20];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[21];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[22];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[23];
+ /* round 6: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[24];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[25];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[26];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[27];
+ /* round 7: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[28];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[29];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[30];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[31];
+ /* round 8: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[32];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[33];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[34];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[35];
+ /* round 9: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[36];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[37];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[38];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[39];
+ if (Nr > 10) {
+ /* round 10: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[40];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[41];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[42];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[43];
+ /* round 11: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[44];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[45];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[46];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[47];
+ if (Nr > 12) {
+ /* round 12: */
+ s0 = Td0[t0 >> 24] ^ Td1[(t3 >> 16) & 0xff] ^ Td2[(t2 >> 8) & 0xff] ^ Td3[t1 & 0xff] ^ rk[48];
+ s1 = Td0[t1 >> 24] ^ Td1[(t0 >> 16) & 0xff] ^ Td2[(t3 >> 8) & 0xff] ^ Td3[t2 & 0xff] ^ rk[49];
+ s2 = Td0[t2 >> 24] ^ Td1[(t1 >> 16) & 0xff] ^ Td2[(t0 >> 8) & 0xff] ^ Td3[t3 & 0xff] ^ rk[50];
+ s3 = Td0[t3 >> 24] ^ Td1[(t2 >> 16) & 0xff] ^ Td2[(t1 >> 8) & 0xff] ^ Td3[t0 & 0xff] ^ rk[51];
+ /* round 13: */
+ t0 = Td0[s0 >> 24] ^ Td1[(s3 >> 16) & 0xff] ^ Td2[(s2 >> 8) & 0xff] ^ Td3[s1 & 0xff] ^ rk[52];
+ t1 = Td0[s1 >> 24] ^ Td1[(s0 >> 16) & 0xff] ^ Td2[(s3 >> 8) & 0xff] ^ Td3[s2 & 0xff] ^ rk[53];
+ t2 = Td0[s2 >> 24] ^ Td1[(s1 >> 16) & 0xff] ^ Td2[(s0 >> 8) & 0xff] ^ Td3[s3 & 0xff] ^ rk[54];
+ t3 = Td0[s3 >> 24] ^ Td1[(s2 >> 16) & 0xff] ^ Td2[(s1 >> 8) & 0xff] ^ Td3[s0 & 0xff] ^ rk[55];
+ }
+ }
+ rk += Nr << 2;
+#else /* !FULL_UNROLL */
+ /*
+ * Nr - 1 full rounds:
+ */
+ r = Nr >> 1;
+ for (;;) {
+ t0 =
+ Td0[(s0 >> 24) ] ^
+ Td1[(s3 >> 16) & 0xff] ^
+ Td2[(s2 >> 8) & 0xff] ^
+ Td3[(s1 ) & 0xff] ^
+ rk[4];
+ t1 =
+ Td0[(s1 >> 24) ] ^
+ Td1[(s0 >> 16) & 0xff] ^
+ Td2[(s3 >> 8) & 0xff] ^
+ Td3[(s2 ) & 0xff] ^
+ rk[5];
+ t2 =
+ Td0[(s2 >> 24) ] ^
+ Td1[(s1 >> 16) & 0xff] ^
+ Td2[(s0 >> 8) & 0xff] ^
+ Td3[(s3 ) & 0xff] ^
+ rk[6];
+ t3 =
+ Td0[(s3 >> 24) ] ^
+ Td1[(s2 >> 16) & 0xff] ^
+ Td2[(s1 >> 8) & 0xff] ^
+ Td3[(s0 ) & 0xff] ^
+ rk[7];
+
+ rk += 8;
+ if (--r == 0) {
+ break;
+ }
+
+ s0 =
+ Td0[(t0 >> 24) ] ^
+ Td1[(t3 >> 16) & 0xff] ^
+ Td2[(t2 >> 8) & 0xff] ^
+ Td3[(t1 ) & 0xff] ^
+ rk[0];
+ s1 =
+ Td0[(t1 >> 24) ] ^
+ Td1[(t0 >> 16) & 0xff] ^
+ Td2[(t3 >> 8) & 0xff] ^
+ Td3[(t2 ) & 0xff] ^
+ rk[1];
+ s2 =
+ Td0[(t2 >> 24) ] ^
+ Td1[(t1 >> 16) & 0xff] ^
+ Td2[(t0 >> 8) & 0xff] ^
+ Td3[(t3 ) & 0xff] ^
+ rk[2];
+ s3 =
+ Td0[(t3 >> 24) ] ^
+ Td1[(t2 >> 16) & 0xff] ^
+ Td2[(t1 >> 8) & 0xff] ^
+ Td3[(t0 ) & 0xff] ^
+ rk[3];
+ }
+#endif /* ?FULL_UNROLL */
+ /*
+ * apply last round and
+ * map cipher state to byte array block:
+ */
+ s0 =
+ (Td4[(t0 >> 24) ] & 0xff000000) ^
+ (Td4[(t3 >> 16) & 0xff] & 0x00ff0000) ^
+ (Td4[(t2 >> 8) & 0xff] & 0x0000ff00) ^
+ (Td4[(t1 ) & 0xff] & 0x000000ff) ^
+ rk[0];
+ PUTU32(pt , s0);
+ s1 =
+ (Td4[(t1 >> 24) ] & 0xff000000) ^
+ (Td4[(t0 >> 16) & 0xff] & 0x00ff0000) ^
+ (Td4[(t3 >> 8) & 0xff] & 0x0000ff00) ^
+ (Td4[(t2 ) & 0xff] & 0x000000ff) ^
+ rk[1];
+ PUTU32(pt + 4, s1);
+ s2 =
+ (Td4[(t2 >> 24) ] & 0xff000000) ^
+ (Td4[(t1 >> 16) & 0xff] & 0x00ff0000) ^
+ (Td4[(t0 >> 8) & 0xff] & 0x0000ff00) ^
+ (Td4[(t3 ) & 0xff] & 0x000000ff) ^
+ rk[2];
+ PUTU32(pt + 8, s2);
+ s3 =
+ (Td4[(t3 >> 24) ] & 0xff000000) ^
+ (Td4[(t2 >> 16) & 0xff] & 0x00ff0000) ^
+ (Td4[(t1 >> 8) & 0xff] & 0x0000ff00) ^
+ (Td4[(t0 ) & 0xff] & 0x000000ff) ^
+ rk[3];
+ PUTU32(pt + 12, s3);
+}
+
+/* setup key context for encryption only */
+int
+rijndael_set_key_enc_only(rijndael_ctx *ctx, const u_char *key, int bits)
+{
+ int rounds;
+
+ rounds = rijndaelKeySetupEnc(ctx->ek, key, bits);
+ if (rounds == 0)
+ return -1;
+
+ ctx->Nr = rounds;
+ ctx->enc_only = 1;
+
+ return 0;
+}
+
+/* setup key context for both encryption and decryption */
+int
+rijndael_set_key(rijndael_ctx *ctx, const u_char *key, int bits)
+{
+ int rounds;
+
+ rounds = rijndaelKeySetupEnc(ctx->ek, key, bits);
+ if (rounds == 0)
+ return -1;
+ if (rijndaelKeySetupDec(ctx->dk, key, bits) != rounds)
+ return -1;
+
+ ctx->Nr = rounds;
+ ctx->enc_only = 0;
+
+ return 0;
+}
+
+void
+rijndael_decrypt(rijndael_ctx *ctx, const u_char *src, u_char *dst)
+{
+ rijndaelDecrypt(ctx->dk, ctx->Nr, src, dst);
+}
+
+void
+rijndael_encrypt(rijndael_ctx *ctx, const u_char *src, u_char *dst)
+{
+ rijndaelEncrypt(ctx->ek, ctx->Nr, src, dst);
+}
diff --git a/xen/crypto/vmac.c b/xen/crypto/vmac.c
new file mode 100644
index 0000000000..e5558ba782
--- /dev/null
+++ b/xen/crypto/vmac.c
@@ -0,0 +1,1220 @@
+/* --------------------------------------------------------------------------
+ * VMAC and VHASH Implementation by Ted Krovetz (tdk@acm.org) and Wei Dai.
+ * This implementation is herby placed in the public domain.
+ * The authors offers no warranty. Use at your own risk.
+ * Please send bug reports to the authors.
+ * Last modified: 17 APR 08, 1700 PDT
+ * ----------------------------------------------------------------------- */
+
+/* start for Xen */
+#include <xen/config.h>
+#include <xen/init.h>
+#include <xen/types.h>
+#include <xen/lib.h>
+#include <crypto/vmac.h>
+#define UINT64_C(x) x##ULL
+/* end for Xen */
+
+/* Enable code tuned for 64-bit registers; otherwise tuned for 32-bit */
+#ifndef VMAC_ARCH_64
+#define VMAC_ARCH_64 (__x86_64__ || __ppc64__ || _M_X64)
+#endif
+
+/* Enable code tuned for Intel SSE2 instruction set */
+#if ((__SSE2__ || (_M_IX86_FP >= 2)) && ( ! VMAC_ARCH_64))
+#define VMAC_USE_SSE2 1
+#include <emmintrin.h>
+#endif
+
+/* Native word reads. Update (or define via compiler) if incorrect */
+#ifndef VMAC_ARCH_BIG_ENDIAN /* Assume big-endian unless on the list */
+#define VMAC_ARCH_BIG_ENDIAN \
+ (!(__x86_64__ || __i386__ || _M_IX86 || \
+ _M_X64 || __ARMEL__ || __MIPSEL__))
+#endif
+
+/* ----------------------------------------------------------------------- */
+/* Constants and masks */
+
+const uint64_t p64 = UINT64_C(0xfffffffffffffeff); /* 2^64 - 257 prime */
+const uint64_t m62 = UINT64_C(0x3fffffffffffffff); /* 62-bit mask */
+const uint64_t m63 = UINT64_C(0x7fffffffffffffff); /* 63-bit mask */
+const uint64_t m64 = UINT64_C(0xffffffffffffffff); /* 64-bit mask */
+const uint64_t mpoly = UINT64_C(0x1fffffff1fffffff); /* Poly key mask */
+
+/* ----------------------------------------------------------------------- *
+ * The following routines are used in this implementation. They are
+ * written via macros to simulate zero-overhead call-by-reference.
+ * All have default implemantations for when they are not defined in an
+ * architecture-specific manner.
+ *
+ * MUL64: 64x64->128-bit multiplication
+ * PMUL64: assumes top bits cleared on inputs
+ * ADD128: 128x128->128-bit addition
+ * GET_REVERSED_64: load and byte-reverse 64-bit word
+ * ----------------------------------------------------------------------- */
+
+/* ----------------------------------------------------------------------- */
+#if (__GNUC__ && (__x86_64__ || __amd64__))
+/* ----------------------------------------------------------------------- */
+
+#define ADD128(rh,rl,ih,il) \
+ asm ("addq %3, %1 \n\t" \
+ "adcq %2, %0" \
+ : "+r"(rh),"+r"(rl) \
+ : "r"(ih),"r"(il) : "cc");
+
+#define MUL64(rh,rl,i1,i2) \
+ asm ("mulq %3" : "=a"(rl), "=d"(rh) : "a"(i1), "r"(i2) : "cc")
+
+#define PMUL64 MUL64
+
+#define GET_REVERSED_64(p) \
+ ({uint64_t x; \
+ asm ("bswapq %0" : "=r" (x) : "0"(*(uint64_t *)(p))); x;})
+
+/* ----------------------------------------------------------------------- */
+#elif (__GNUC__ && __i386__)
+/* ----------------------------------------------------------------------- */
+
+#define GET_REVERSED_64(p) \
+ ({ uint64_t x; \
+ uint32_t *tp = (uint32_t *)(p); \
+ asm ("bswap %%edx\n\t" \
+ "bswap %%eax" \
+ : "=A"(x) \
+ : "a"(tp[1]), "d"(tp[0])); \
+ x; })
+
+/* ----------------------------------------------------------------------- */
+#elif (__GNUC__ && __ppc64__)
+/* ----------------------------------------------------------------------- */
+
+#define ADD128(rh,rl,ih,il) \
+ asm volatile ( "addc %1, %1, %3 \n\t" \
+ "adde %0, %0, %2" \
+ : "+r"(rh),"+r"(rl) \
+ : "r"(ih),"r"(il));
+
+#define MUL64(rh,rl,i1,i2) \
+{ uint64_t _i1 = (i1), _i2 = (i2); \
+ rl = _i1 * _i2; \
+ asm volatile ("mulhdu %0, %1, %2" : "=r" (rh) : "r" (_i1), "r" (_i2));\
+}
+
+#define PMUL64 MUL64
+
+#define GET_REVERSED_64(p) \
+ ({ uint32_t hi, lo, *_p = (uint32_t *)(p); \
+ asm volatile ("lwbrx %0, %1, %2" : "=r"(lo) : "b%"(0), "r"(_p) ); \
+ asm volatile ("lwbrx %0, %1, %2" : "=r"(hi) : "b%"(4), "r"(_p) ); \
+ ((uint64_t)hi << 32) | (uint64_t)lo; } )
+
+/* ----------------------------------------------------------------------- */
+#elif (__GNUC__ && (__ppc__ || __PPC__))
+/* ----------------------------------------------------------------------- */
+
+#define GET_REVERSED_64(p) \
+ ({ uint32_t hi, lo, *_p = (uint32_t *)(p); \
+ asm volatile ("lwbrx %0, %1, %2" : "=r"(lo) : "b%"(0), "r"(_p) ); \
+ asm volatile ("lwbrx %0, %1, %2" : "=r"(hi) : "b%"(4), "r"(_p) ); \
+ ((uint64_t)hi << 32) | (uint64_t)lo; } )
+
+/* ----------------------------------------------------------------------- */
+#elif (__GNUC__ && (__ARMEL__ || __ARM__))
+/* ----------------------------------------------------------------------- */
+
+#define bswap32(v) \
+({ uint32_t tmp,out; \
+ asm volatile( \
+ "eor %1, %2, %2, ror #16\n" \
+ "bic %1, %1, #0x00ff0000\n" \
+ "mov %0, %2, ror #8\n" \
+ "eor %0, %0, %1, lsr #8" \
+ : "=r" (out), "=&r" (tmp) \
+ : "r" (v)); \
+ out;})
+
+/* ----------------------------------------------------------------------- */
+#elif _MSC_VER
+/* ----------------------------------------------------------------------- */
+
+#include <intrin.h>
+
+#if (_M_IA64 || _M_X64) && \
+ (!defined(__INTEL_COMPILER) || __INTEL_COMPILER >= 1000)
+#define MUL64(rh,rl,i1,i2) (rl) = _umul128(i1,i2,&(rh));
+#pragma intrinsic(_umul128)
+#define PMUL64 MUL64
+#endif
+
+/* MSVC uses add, adc in this version */
+#define ADD128(rh,rl,ih,il) \
+ { uint64_t _il = (il); \
+ (rl) += (_il); \
+ (rh) += (ih) + ((rl) < (_il)); \
+ }
+
+#if _MSC_VER >= 1300
+#define GET_REVERSED_64(p) _byteswap_uint64(*(uint64_t *)(p))
+#pragma intrinsic(_byteswap_uint64)
+#endif
+
+#if _MSC_VER >= 1400 && \
+ (!defined(__INTEL_COMPILER) || __INTEL_COMPILER >= 1000)
+#define MUL32(i1,i2) (__emulu((uint32_t)(i1),(uint32_t)(i2)))
+#pragma intrinsic(__emulu)
+#endif
+
+/* ----------------------------------------------------------------------- */
+#endif
+/* ----------------------------------------------------------------------- */
+
+#if __GNUC__
+#define ALIGN(n) __attribute__ ((aligned(n)))
+#define NOINLINE __attribute__ ((noinline))
+#define FASTCALL
+#elif _MSC_VER
+#define ALIGN(n) __declspec(align(n))
+#define NOINLINE __declspec(noinline)
+#define FASTCALL __fastcall
+#else
+#define ALIGN(n)
+#define NOINLINE
+#define FASTCALL
+#endif
+
+/* ----------------------------------------------------------------------- */
+/* Default implementations, if not defined above */
+/* ----------------------------------------------------------------------- */
+
+#ifndef ADD128
+#define ADD128(rh,rl,ih,il) \
+ { uint64_t _il = (il); \
+ (rl) += (_il); \
+ if ((rl) < (_il)) (rh)++; \
+ (rh) += (ih); \
+ }
+#endif
+
+#ifndef MUL32
+#define MUL32(i1,i2) ((uint64_t)(uint32_t)(i1)*(uint32_t)(i2))
+#endif
+
+#ifndef PMUL64 /* rh may not be same as i1 or i2 */
+#define PMUL64(rh,rl,i1,i2) /* Assumes m doesn't overflow */ \
+ { uint64_t _i1 = (i1), _i2 = (i2); \
+ uint64_t m = MUL32(_i1,_i2>>32) + MUL32(_i1>>32,_i2); \
+ rh = MUL32(_i1>>32,_i2>>32); \
+ rl = MUL32(_i1,_i2); \
+ ADD128(rh,rl,(m >> 32),(m << 32)); \
+ }
+#endif
+
+#ifndef MUL64
+#define MUL64(rh,rl,i1,i2) \
+ { uint64_t _i1 = (i1), _i2 = (i2); \
+ uint64_t m1= MUL32(_i1,_i2>>32); \
+ uint64_t m2= MUL32(_i1>>32,_i2); \
+ rh = MUL32(_i1>>32,_i2>>32); \
+ rl = MUL32(_i1,_i2); \
+ ADD128(rh,rl,(m1 >> 32),(m1 << 32)); \
+ ADD128(rh,rl,(m2 >> 32),(m2 << 32)); \
+ }
+#endif
+
+#ifndef GET_REVERSED_64
+#ifndef bswap64
+#ifndef bswap32
+#define bswap32(x) \
+ ({ uint32_t bsx = (x); \
+ ((((bsx) & 0xff000000u) >> 24) | (((bsx) & 0x00ff0000u) >> 8) | \
+ (((bsx) & 0x0000ff00u) << 8) | (((bsx) & 0x000000ffu) << 24)); })
+#endif
+#define bswap64(x) \
+ ({ union { uint64_t ll; uint32_t l[2]; } w, r; \
+ w.ll = (x); \
+ r.l[0] = bswap32 (w.l[1]); \
+ r.l[1] = bswap32 (w.l[0]); \
+ r.ll; })
+#endif
+#define GET_REVERSED_64(p) bswap64(*(uint64_t *)(p))
+#endif
+
+/* ----------------------------------------------------------------------- */
+
+#if (VMAC_PREFER_BIG_ENDIAN)
+# define get64PE get64BE
+#else
+# define get64PE get64LE
+#endif
+
+#if (VMAC_ARCH_BIG_ENDIAN)
+# define get64BE(ptr) (*(uint64_t *)(ptr))
+# define get64LE(ptr) GET_REVERSED_64(ptr)
+#else /* assume little-endian */
+# define get64BE(ptr) GET_REVERSED_64(ptr)
+# define get64LE(ptr) (*(uint64_t *)(ptr))
+#endif
+
+
+/* --------------------------------------------------------------------- *
+ * For highest performance the L1 NH and L2 polynomial hashes should be
+ * carefully implemented to take advantage of one's target architechture.
+ * Here these two hash functions are defined multiple time; once for
+ * 64-bit architectures, once for 32-bit SSE2 architectures, and once
+ * for the rest (32-bit) architectures.
+ * For each, nh_16 *must* be defined (works on multiples of 16 bytes).
+ * Optionally, nh_vmac_nhbytes can be defined (for multiples of
+ * VMAC_NHBYTES), and nh_16_2 and nh_vmac_nhbytes_2 (versions that do two
+ * NH computations at once).
+ * --------------------------------------------------------------------- */
+
+/* ----------------------------------------------------------------------- */
+#if VMAC_ARCH_64
+/* ----------------------------------------------------------------------- */
+
+#define nh_16(mp, kp, nw, rh, rl) \
+{ int i; uint64_t th, tl; \
+ rh = rl = 0; \
+ for (i = 0; i < nw; i+= 2) { \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i ],get64PE((mp)+i+1)+(kp)[i+1]);\
+ ADD128(rh,rl,th,tl); \
+ } \
+}
+#define nh_16_2(mp, kp, nw, rh, rl, rh1, rl1) \
+{ int i; uint64_t th, tl; \
+ rh1 = rl1 = rh = rl = 0; \
+ for (i = 0; i < nw; i+= 2) { \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i ],get64PE((mp)+i+1)+(kp)[i+1]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i+2],get64PE((mp)+i+1)+(kp)[i+3]);\
+ ADD128(rh1,rl1,th,tl); \
+ } \
+}
+
+#if (VMAC_NHBYTES >= 64) /* These versions do 64-bytes of message at a time */
+#define nh_vmac_nhbytes(mp, kp, nw, rh, rl) \
+{ int i; uint64_t th, tl; \
+ rh = rl = 0; \
+ for (i = 0; i < nw; i+= 8) { \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i ],get64PE((mp)+i+1)+(kp)[i+1]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+2)+(kp)[i+2],get64PE((mp)+i+3)+(kp)[i+3]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+4)+(kp)[i+4],get64PE((mp)+i+5)+(kp)[i+5]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+6)+(kp)[i+6],get64PE((mp)+i+7)+(kp)[i+7]);\
+ ADD128(rh,rl,th,tl); \
+ } \
+}
+#define nh_vmac_nhbytes_2(mp, kp, nw, rh, rl, rh1, rl1) \
+{ int i; uint64_t th, tl; \
+ rh1 = rl1 = rh = rl = 0; \
+ for (i = 0; i < nw; i+= 8) { \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i ],get64PE((mp)+i+1)+(kp)[i+1]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i )+(kp)[i+2],get64PE((mp)+i+1)+(kp)[i+3]);\
+ ADD128(rh1,rl1,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+2)+(kp)[i+2],get64PE((mp)+i+3)+(kp)[i+3]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+2)+(kp)[i+4],get64PE((mp)+i+3)+(kp)[i+5]);\
+ ADD128(rh1,rl1,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+4)+(kp)[i+4],get64PE((mp)+i+5)+(kp)[i+5]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+4)+(kp)[i+6],get64PE((mp)+i+5)+(kp)[i+7]);\
+ ADD128(rh1,rl1,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+6)+(kp)[i+6],get64PE((mp)+i+7)+(kp)[i+7]);\
+ ADD128(rh,rl,th,tl); \
+ MUL64(th,tl,get64PE((mp)+i+6)+(kp)[i+8],get64PE((mp)+i+7)+(kp)[i+9]);\
+ ADD128(rh1,rl1,th,tl); \
+ } \
+}
+#endif
+
+#define poly_step(ah, al, kh, kl, mh, ml) \
+{ uint64_t t1h, t1l, t2h, t2l, t3h, t3l, z=0; \
+ /* compute ab*cd, put bd into result registers */ \
+ PMUL64(t3h,t3l,al,kh); \
+ PMUL64(t2h,t2l,ah,kl); \
+ PMUL64(t1h,t1l,ah,2*kh); \
+ PMUL64(ah,al,al,kl); \
+ /* add 2 * ac to result */ \
+ ADD128(ah,al,t1h,t1l); \
+ /* add together ad + bc */ \
+ ADD128(t2h,t2l,t3h,t3l); \
+ /* now (ah,al), (t2l,2*t2h) need summing */ \
+ /* first add the high registers, carrying into t2h */ \
+ ADD128(t2h,ah,z,t2l); \
+ /* double t2h and add top bit of ah */ \
+ t2h = 2 * t2h + (ah >> 63); \
+ ah &= m63; \
+ /* now add the low registers */ \
+ ADD128(ah,al,mh,ml); \
+ ADD128(ah,al,z,t2h); \
+}
+
+/* ----------------------------------------------------------------------- */
+#elif VMAC_USE_SSE2
+/* ----------------------------------------------------------------------- */
+
+// macros from Crypto++ for sharing inline assembly code between MSVC and GNU C
+#if defined(__GNUC__)
+ // define these in two steps to allow arguments to be expanded
+ #define GNU_AS2(x, y) #x ", " #y ";"
+ #define GNU_AS3(x, y, z) #x ", " #y ", " #z ";"
+ #define GNU_ASL(x) "\n" #x ":"
+ #define GNU_ASJ(x, y, z) #x " " #y #z ";"
+ #define AS2(x, y) GNU_AS2(x, y)
+ #define AS3(x, y, z) GNU_AS3(x, y, z)
+ #define ASS(x, y, a, b, c, d) #x ", " #y ", " #a "*64+" #b "*16+" #c "*4+" #d ";"
+ #define ASL(x) GNU_ASL(x)
+ #define ASJ(x, y, z) GNU_ASJ(x, y, z)
+#else
+ #define AS2(x, y) __asm {x, y}
+ #define AS3(x, y, z) __asm {x, y, z}
+ #define ASS(x, y, a, b, c, d) __asm {x, y, _MM_SHUFFLE(a, b, c, d)}
+ #define ASL(x) __asm {label##x:}
+ #define ASJ(x, y, z) __asm {x label##y}
+#endif
+
+static void NOINLINE nh_16_func(const uint64_t *mp, const uint64_t *kp, size_t nw, uint64_t *rh, uint64_t *rl)
+{
+ // This assembly version, using MMX registers, is just as fast as the
+ // intrinsics version (which uses XMM registers) on the Intel Core 2,
+ // but is much faster on the Pentium 4. In order to schedule multiplies
+ // as early as possible, the loop interleaves operations for the current
+ // block and the next block. To mask out high 32-bits, we use "movd"
+ // to move the lower 32-bits to the stack and then back. Surprisingly,
+ // this is faster than any other method.
+#ifdef __GNUC__
+ __asm__ __volatile__
+ (
+ ".intel_syntax noprefix;"
+#else
+ AS2( mov esi, mp)
+ AS2( mov edi, kp)
+ AS2( mov ecx, nw)
+ AS2( mov eax, rl)
+ AS2( mov edx, rh)
+#endif
+ AS2( sub esp, 12)
+ AS2( movq mm6, [esi])
+ AS2( paddq mm6, [edi])
+ AS2( movq mm5, [esi+8])
+ AS2( paddq mm5, [edi+8])
+ AS2( add esi, 16)
+ AS2( add edi, 16)
+ AS2( movq mm4, mm6)
+ ASS( pshufw mm2, mm6, 1, 0, 3, 2)
+ AS2( pmuludq mm6, mm5)
+ ASS( pshufw mm3, mm5, 1, 0, 3, 2)
+ AS2( pmuludq mm5, mm2)
+ AS2( pmuludq mm2, mm3)
+ AS2( pmuludq mm3, mm4)
+ AS2( pxor mm7, mm7)
+ AS2( movd [esp], mm6)
+ AS2( psrlq mm6, 32)
+ AS2( movd [esp+4], mm5)
+ AS2( psrlq mm5, 32)
+ AS2( sub ecx, 2)
+ ASJ( jz, 1, f)
+ ASL(0)
+ AS2( movq mm0, [esi])
+ AS2( paddq mm0, [edi])
+ AS2( movq mm1, [esi+8])
+ AS2( paddq mm1, [edi+8])
+ AS2( add esi, 16)
+ AS2( add edi, 16)
+ AS2( movq mm4, mm0)
+ AS2( paddq mm5, mm2)
+ ASS( pshufw mm2, mm0, 1, 0, 3, 2)
+ AS2( pmuludq mm0, mm1)
+ AS2( movd [esp+8], mm3)
+ AS2( psrlq mm3, 32)
+ AS2( paddq mm5, mm3)
+ ASS( pshufw mm3, mm1, 1, 0, 3, 2)
+ AS2( pmuludq mm1, mm2)
+ AS2( pmuludq mm2, mm3)
+ AS2( pmuludq mm3, mm4)
+ AS2( movd mm4, [esp])
+ AS2( paddq mm7, mm4)
+ AS2( movd mm4, [esp+4])
+ AS2( paddq mm6, mm4)
+ AS2( movd mm4, [esp+8])
+ AS2( paddq mm6, mm4)
+ AS2( movd [esp], mm0)
+ AS2( psrlq mm0, 32)
+ AS2( paddq mm6, mm0)
+ AS2( movd [esp+4], mm1)
+ AS2( psrlq mm1, 32)
+ AS2( paddq mm5, mm1)
+ AS2( sub ecx, 2)
+ ASJ( jnz, 0, b)
+ ASL(1)
+ AS2( paddq mm5, mm2)
+ AS2( movd [esp+8], mm3)
+ AS2( psrlq mm3, 32)
+ AS2( paddq mm5, mm3)
+ AS2( movd mm4, [esp])
+ AS2( paddq mm7, mm4)
+ AS2( movd mm4, [esp+4])
+ AS2( paddq mm6, mm4)
+ AS2( movd mm4, [esp+8])
+ AS2( paddq mm6, mm4)
+
+ ASS( pshufw mm0, mm7, 3, 2, 1, 0)
+ AS2( psrlq mm7, 32)
+ AS2( paddq mm6, mm7)
+ AS2( punpckldq mm0, mm6)
+ AS2( psrlq mm6, 32)
+ AS2( paddq mm5, mm6)
+ AS2( movq [eax], mm0)
+ AS2( movq [edx], mm5)
+ AS2( add esp, 12)
+#ifdef __GNUC__
+ ".att_syntax prefix;"
+ :
+ : "S" (mp), "D" (kp), "c" (nw), "a" (rl), "d" (rh)
+ : "memory", "cc"
+ );
+#endif
+}
+#define nh_16(mp, kp, nw, rh, rl) nh_16_func(mp, kp, nw, &(rh), &(rl));
+
+static void poly_step_func(uint64_t *ahi, uint64_t *alo, const uint64_t *kh,
+ const uint64_t *kl, const uint64_t *mh, const uint64_t *ml)
+{
+ // This code tries to schedule the multiplies as early as possible to overcome
+ // the long latencies on the Pentium 4. It also minimizes "movq" instructions
+ // which are very expensive on the P4.
+
+#define a0 [eax+0]
+#define a1 [eax+4]
+#define a2 [ebx+0]
+#define a3 [ebx+4]
+#define k0 [ecx+0]
+#define k1 [ecx+4]
+#define k2 [edx+0]
+#define k3 [edx+4]
+
+#ifdef __GNUC__
+ uint32_t temp;
+ __asm__ __volatile__
+ (
+ "mov %%ebx, %0;"
+ "mov %1, %%ebx;"
+ ".intel_syntax noprefix;"
+#else
+ AS2( mov ebx, ahi)
+ AS2( mov edx, kh)
+ AS2( mov eax, alo)
+ AS2( mov ecx, kl)
+ AS2( mov esi, mh)
+ AS2( mov edi, ml)
+#endif
+
+ AS2( movd mm0, a3)
+ AS2( movq mm4, mm0)
+ AS2( pmuludq mm0, k3) // a3*k3
+ AS2( movd mm1, a0)
+ AS2( pmuludq mm1, k2) // a0*k2
+ AS2( movd mm2, a1)
+ AS2( movd mm6, k1)
+ AS2( pmuludq mm2, mm6) // a1*k1
+ AS2( movd mm3, a2)
+ AS2( movq mm5, mm3)
+ AS2( movd mm7, k0)
+ AS2( pmuludq mm3, mm7) // a2*k0
+ AS2( pmuludq mm4, mm7) // a3*k0
+ AS2( pmuludq mm5, mm6) // a2*k1
+ AS2( psllq mm0, 1)
+ AS2( paddq mm0, [esi])
+ AS2( paddq mm0, mm1)
+ AS2( movd mm1, a1)
+ AS2( paddq mm4, mm5)
+ AS2( movq mm5, mm1)
+ AS2( pmuludq mm1, k2) // a1*k2
+ AS2( paddq mm0, mm2)
+ AS2( movd mm2, a0)
+ AS2( paddq mm0, mm3)
+ AS2( movq mm3, mm2)
+ AS2( pmuludq mm2, k3) // a0*k3
+ AS2( pmuludq mm3, mm7) // a0*k0
+ AS2( movd esi, mm0)
+ AS2( psrlq mm0, 32)
+ AS2( pmuludq mm7, mm5) // a1*k0
+ AS2( pmuludq mm5, k3) // a1*k3
+ AS2( paddq mm0, mm1)
+ AS2( movd mm1, a2)
+ AS2( pmuludq mm1, k2) // a2*k2
+ AS2( paddq mm0, mm2)
+ AS2( paddq mm0, mm4)
+ AS2( movq mm4, mm0)
+ AS2( movd mm2, a3)
+ AS2( pmuludq mm2, mm6) // a3*k1
+ AS2( pmuludq mm6, a0) // a0*k1
+ AS2( psrlq mm0, 31)
+ AS2( paddq mm0, mm3)
+ AS2( movd mm3, [edi])
+ AS2( paddq mm0, mm3)
+ AS2( movd mm3, a2)
+ AS2( pmuludq mm3, k3) // a2*k3
+ AS2( paddq mm5, mm1)
+ AS2( movd mm1, a3)
+ AS2( pmuludq mm1, k2) // a3*k2
+ AS2( paddq mm5, mm2)
+ AS2( movd mm2, [edi+4])
+ AS2( psllq mm5, 1)
+ AS2( paddq mm0, mm5)
+ AS2( movq mm5, mm0)
+ AS2( psllq mm4, 33)
+ AS2( psrlq mm0, 32)
+ AS2( paddq mm6, mm7)
+ AS2( movd mm7, esi)
+ AS2( paddq mm0, mm6)
+ AS2( paddq mm0, mm2)
+ AS2( paddq mm3, mm1)
+ AS2( psllq mm3, 1)
+ AS2( paddq mm0, mm3)
+ AS2( psrlq mm4, 1)
+ AS2( punpckldq mm5, mm0)
+ AS2( psrlq mm0, 32)
+ AS2( por mm4, mm7)
+ AS2( paddq mm0, mm4)
+ AS2( movq a0, mm5)
+ AS2( movq a2, mm0)
+#ifdef __GNUC__
+ ".att_syntax prefix;"
+ "mov %0, %%ebx;"
+ : "=m" (temp)
+ : "m" (ahi), "D" (ml), "d" (kh), "a" (alo), "S" (mh), "c" (kl)
+ : "memory", "cc"
+ );
+#endif
+
+
+#undef a0
+#undef a1
+#undef a2
+#undef a3
+#undef k0
+#undef k1
+#undef k2
+#undef k3
+}
+
+#define poly_step(ah, al, kh, kl, mh, ml) \
+ poly_step_func(&(ah), &(al), &(kh), &(kl), &(mh), &(ml))
+
+/* ----------------------------------------------------------------------- */
+#else /* not VMAC_ARCH_64 and not SSE2 */
+/* ----------------------------------------------------------------------- */
+
+#ifndef nh_16
+#define nh_16(mp, kp, nw, rh, rl) \
+{ uint64_t t1,t2,m1,m2,t; \
+ int i; \
+ rh = rl = t = 0; \
+ for (i = 0; i < nw; i+=2) { \
+ t1 = get64PE(mp+i) + kp[i]; \
+ t2 = get64PE(mp+i+1) + kp[i+1]; \
+ m2 = MUL32(t1 >> 32, t2); \
+ m1 = MUL32(t1, t2 >> 32); \
+ ADD128(rh,rl,MUL32(t1 >> 32,t2 >> 32),MUL32(t1,t2)); \
+ rh += (uint64_t)(uint32_t)(m1 >> 32) + (uint32_t)(m2 >> 32); \
+ t += (uint64_t)(uint32_t)m1 + (uint32_t)m2; \
+ } \
+ ADD128(rh,rl,(t >> 32),(t << 32)); \
+}
+#endif
+
+static void poly_step_func(uint64_t *ahi, uint64_t *alo, const uint64_t *kh,
+ const uint64_t *kl, const uint64_t *mh, const uint64_t *ml)
+{
+
+#if VMAC_ARCH_BIG_ENDIAN
+#define INDEX_HIGH 0
+#define INDEX_LOW 1
+#else
+#define INDEX_HIGH 1
+#define INDEX_LOW 0
+#endif
+
+#define a0 *(((uint32_t*)alo)+INDEX_LOW)
+#define a1 *(((uint32_t*)alo)+INDEX_HIGH)
+#define a2 *(((uint32_t*)ahi)+INDEX_LOW)
+#define a3 *(((uint32_t*)ahi)+INDEX_HIGH)
+#define k0 *(((uint32_t*)kl)+INDEX_LOW)
+#define k1 *(((uint32_t*)kl)+INDEX_HIGH)
+#define k2 *(((uint32_t*)kh)+INDEX_LOW)
+#define k3 *(((uint32_t*)kh)+INDEX_HIGH)
+
+ uint64_t p, q, t;
+ uint32_t t2;
+
+ p = MUL32(a3, k3);
+ p += p;
+ p += *(uint64_t *)mh;
+ p += MUL32(a0, k2);
+ p += MUL32(a1, k1);
+ p += MUL32(a2, k0);
+ t = (uint32_t)(p);
+ p >>= 32;
+ p += MUL32(a0, k3);
+ p += MUL32(a1, k2);
+ p += MUL32(a2, k1);
+ p += MUL32(a3, k0);
+ t |= ((uint64_t)((uint32_t)p & 0x7fffffff)) << 32;
+ p >>= 31;
+ p += (uint64_t)(((uint32_t*)ml)[INDEX_LOW]);
+ p += MUL32(a0, k0);
+ q = MUL32(a1, k3);
+ q += MUL32(a2, k2);
+ q += MUL32(a3, k1);
+ q += q;
+ p += q;
+ t2 = (uint32_t)(p);
+ p >>= 32;
+ p += (uint64_t)(((uint32_t*)ml)[INDEX_HIGH]);
+ p += MUL32(a0, k1);
+ p += MUL32(a1, k0);
+ q = MUL32(a2, k3);
+ q += MUL32(a3, k2);
+ q += q;
+ p += q;
+ *(uint64_t *)(alo) = (p << 32) | t2;
+ p >>= 32;
+ *(uint64_t *)(ahi) = p + t;
+
+#undef a0
+#undef a1
+#undef a2
+#undef a3
+#undef k0
+#undef k1
+#undef k2
+#undef k3
+}
+
+#define poly_step(ah, al, kh, kl, mh, ml) \
+ poly_step_func(&(ah), &(al), &(kh), &(kl), &(mh), &(ml))
+
+/* ----------------------------------------------------------------------- */
+#endif /* end of specialized NH and poly definitions */
+/* ----------------------------------------------------------------------- */
+
+/* At least nh_16 is defined. Defined others as needed here */
+#ifndef nh_16_2
+#define nh_16_2(mp, kp, nw, rh, rl, rh2, rl2) \
+ nh_16(mp, kp, nw, rh, rl); \
+ nh_16(mp, ((kp)+2), nw, rh2, rl2);
+#endif
+#ifndef nh_vmac_nhbytes
+#define nh_vmac_nhbytes(mp, kp, nw, rh, rl) \
+ nh_16(mp, kp, nw, rh, rl)
+#endif
+#ifndef nh_vmac_nhbytes_2
+#define nh_vmac_nhbytes_2(mp, kp, nw, rh, rl, rh2, rl2) \
+ nh_vmac_nhbytes(mp, kp, nw, rh, rl); \
+ nh_vmac_nhbytes(mp, ((kp)+2), nw, rh2, rl2);
+#endif
+
+/* ----------------------------------------------------------------------- */
+
+void vhash_abort(vmac_ctx_t *ctx)
+{
+ ctx->polytmp[0] = ctx->polykey[0] ;
+ ctx->polytmp[1] = ctx->polykey[1] ;
+ #if (VMAC_TAG_LEN == 128)
+ ctx->polytmp[2] = ctx->polykey[2] ;
+ ctx->polytmp[3] = ctx->polykey[3] ;
+ #endif
+ ctx->first_block_processed = 0;
+}
+
+/* ----------------------------------------------------------------------- */
+static uint64_t l3hash(uint64_t p1, uint64_t p2,
+ uint64_t k1, uint64_t k2, uint64_t len)
+{
+ uint64_t rh, rl, t, z=0;
+
+ /* fully reduce (p1,p2)+(len,0) mod p127 */
+ t = p1 >> 63;
+ p1 &= m63;
+ ADD128(p1, p2, len, t);
+ /* At this point, (p1,p2) is at most 2^127+(len<<64) */
+ t = (p1 > m63) + ((p1 == m63) && (p2 == m64));
+ ADD128(p1, p2, z, t);
+ p1 &= m63;
+
+ /* compute (p1,p2)/(2^64-2^32) and (p1,p2)%(2^64-2^32) */
+ t = p1 + (p2 >> 32);
+ t += (t >> 32);
+ t += (uint32_t)t > 0xfffffffeu;
+ p1 += (t >> 32);
+ p2 += (p1 << 32);
+
+ /* compute (p1+k1)%p64 and (p2+k2)%p64 */
+ p1 += k1;
+ p1 += (0 - (p1 < k1)) & 257;
+ p2 += k2;
+ p2 += (0 - (p2 < k2)) & 257;
+
+ /* compute (p1+k1)*(p2+k2)%p64 */
+ MUL64(rh, rl, p1, p2);
+ t = rh >> 56;
+ ADD128(t, rl, z, rh);
+ rh <<= 8;
+ ADD128(t, rl, z, rh);
+ t += t << 8;
+ rl += t;
+ rl += (0 - (rl < t)) & 257;
+ rl += (0 - (rl > p64-1)) & 257;
+ return rl;
+}
+
+/* ----------------------------------------------------------------------- */
+
+void vhash_update(unsigned char *m,
+ unsigned int mbytes, /* Pos multiple of VMAC_NHBYTES */
+ vmac_ctx_t *ctx)
+{
+ uint64_t rh, rl, *mptr;
+ const uint64_t *kptr = (uint64_t *)ctx->nhkey;
+ int i;
+ uint64_t ch, cl;
+ uint64_t pkh = ctx->polykey[0];
+ uint64_t pkl = ctx->polykey[1];
+ #if (VMAC_TAG_LEN == 128)
+ uint64_t ch2, cl2, rh2, rl2;
+ uint64_t pkh2 = ctx->polykey[2];
+ uint64_t pkl2 = ctx->polykey[3];
+ #endif
+
+ mptr = (uint64_t *)m;
+ i = mbytes / VMAC_NHBYTES; /* Must be non-zero */
+
+ ch = ctx->polytmp[0];
+ cl = ctx->polytmp[1];
+ #if (VMAC_TAG_LEN == 128)
+ ch2 = ctx->polytmp[2];
+ cl2 = ctx->polytmp[3];
+ #endif
+
+ if ( ! ctx->first_block_processed) {
+ ctx->first_block_processed = 1;
+ #if (VMAC_TAG_LEN == 64)
+ nh_vmac_nhbytes(mptr,kptr,VMAC_NHBYTES/8,rh,rl);
+ #else
+ nh_vmac_nhbytes_2(mptr,kptr,VMAC_NHBYTES/8,rh,rl,rh2,rl2);
+ rh2 &= m62;
+ ADD128(ch2,cl2,rh2,rl2);
+ #endif
+ rh &= m62;
+ ADD128(ch,cl,rh,rl);
+ mptr += (VMAC_NHBYTES/sizeof(uint64_t));
+ i--;
+ }
+
+ while (i--) {
+ #if (VMAC_TAG_LEN == 64)
+ nh_vmac_nhbytes(mptr,kptr,VMAC_NHBYTES/8,rh,rl);
+ #else
+ nh_vmac_nhbytes_2(mptr,kptr,VMAC_NHBYTES/8,rh,rl,rh2,rl2);
+ rh2 &= m62;
+ poly_step(ch2,cl2,pkh2,pkl2,rh2,rl2);
+ #endif
+ rh &= m62;
+ poly_step(ch,cl,pkh,pkl,rh,rl);
+ mptr += (VMAC_NHBYTES/sizeof(uint64_t));
+ }
+
+ ctx->polytmp[0] = ch;
+ ctx->polytmp[1] = cl;
+ #if (VMAC_TAG_LEN == 128)
+ ctx->polytmp[2] = ch2;
+ ctx->polytmp[3] = cl2;
+ #endif
+ #if VMAC_USE_SSE2
+ _mm_empty(); /* SSE2 version of poly_step uses mmx instructions */
+ #endif
+}
+
+/* ----------------------------------------------------------------------- */
+
+uint64_t xvhash(unsigned char m[],
+ unsigned int mbytes,
+ uint64_t *tagl,
+ vmac_ctx_t *ctx)
+{
+ uint64_t ch, cl, rh, rl, *mptr;
+ #if (VMAC_TAG_LEN == 128)
+ uint64_t ch2, cl2, rh2, rl2;
+ #endif
+ const uint64_t *kptr = (uint64_t *)ctx->nhkey;
+ int i, remaining;
+
+ remaining = mbytes % VMAC_NHBYTES;
+ i = mbytes-remaining;
+ mptr = (uint64_t *)(m+i);
+ if (i) vhash_update(m,i,ctx);
+
+ ch = ctx->polytmp[0];
+ cl = ctx->polytmp[1];
+ #if (VMAC_TAG_LEN == 128)
+ ch2 = ctx->polytmp[2];
+ cl2 = ctx->polytmp[3];
+ #endif
+
+ if (remaining) {
+ #if (VMAC_TAG_LEN == 128)
+ nh_16_2(mptr,kptr,2*((remaining+15)/16),rh,rl,rh2,rl2);
+ rh2 &= m62;
+ #else
+ nh_16(mptr,kptr,2*((remaining+15)/16),rh,rl);
+ #endif
+ rh &= m62;
+ if (i) {
+ poly_step(ch,cl,ctx->polykey[0],ctx->polykey[1],rh,rl);
+ #if (VMAC_TAG_LEN == 128)
+ poly_step(ch2,cl2,ctx->polykey[2],ctx->polykey[3],rh2,rl2);
+ #endif
+ } else {
+ ADD128(ch,cl,rh,rl);
+ #if (VMAC_TAG_LEN == 128)
+ ADD128(ch2,cl2,rh2,rl2);
+ #endif
+ }
+ }
+
+ #if VMAC_USE_SSE2
+ _mm_empty(); /* SSE2 version of poly_step uses mmx instructions */
+ #endif
+ vhash_abort(ctx);
+ remaining *= 8;
+#if (VMAC_TAG_LEN == 128)
+ *tagl = l3hash(ch2, cl2, ctx->l3key[2], ctx->l3key[3],remaining);
+#endif
+ return l3hash(ch, cl, ctx->l3key[0], ctx->l3key[1],remaining);
+}
+
+uint64_t vhash(unsigned char m[],
+ unsigned int mbytes,
+ uint64_t *tagl,
+ vmac_ctx_t *ctx)
+{
+ uint64_t rh, rl, *mptr;
+ const uint64_t *kptr = (uint64_t *)ctx->nhkey;
+ int i, remaining;
+ uint64_t ch, cl;
+ uint64_t pkh = ctx->polykey[0];
+ uint64_t pkl = ctx->polykey[1];
+ #if (VMAC_TAG_LEN == 128)
+ uint64_t ch2, cl2, rh2, rl2;
+ uint64_t pkh2 = ctx->polykey[2];
+ uint64_t pkl2 = ctx->polykey[3];
+ #endif
+
+ mptr = (uint64_t *)m;
+ i = mbytes / VMAC_NHBYTES;
+ remaining = mbytes % VMAC_NHBYTES;
+
+ if (ctx->first_block_processed)
+ {
+ ch = ctx->polytmp[0];
+ cl = ctx->polytmp[1];
+ #if (VMAC_TAG_LEN == 128)
+ ch2 = ctx->polytmp[2];
+ cl2 = ctx->polytmp[3];
+ #endif
+ }
+ else if (i)
+ {
+ #if (VMAC_TAG_LEN == 64)
+ nh_vmac_nhbytes(mptr,kptr,VMAC_NHBYTES/8,ch,cl);
+ #else
+ nh_vmac_nhbytes_2(mptr,kptr,VMAC_NHBYTES/8,ch,cl,ch2,cl2);
+ ch2 &= m62;
+ ADD128(ch2,cl2,pkh2,pkl2);
+ #endif
+ ch &= m62;
+ ADD128(ch,cl,pkh,pkl);
+ mptr += (VMAC_NHBYTES/sizeof(uint64_t));
+ i--;
+ }
+ else if (remaining)
+ {
+ #if (VMAC_TAG_LEN == 64)
+ nh_16(mptr,kptr,2*((remaining+15)/16),ch,cl);
+ #else
+ nh_16_2(mptr,kptr,2*((remaining+15)/16),ch,cl,ch2,cl2);
+ ch2 &= m62;
+ ADD128(ch2,cl2,pkh2,pkl2);
+ #endif
+ ch &= m62;
+ ADD128(ch,cl,pkh,pkl);
+ mptr += (VMAC_NHBYTES/sizeof(uint64_t));
+ goto do_l3;
+ }
+ else /* Empty String */
+ {
+ ch = pkh; cl = pkl;
+ #if (VMAC_TAG_LEN == 128)
+ ch2 = pkh2; cl2 = pkl2;
+ #endif
+ goto do_l3;
+ }
+
+ while (i--) {
+ #if (VMAC_TAG_LEN == 64)
+ nh_vmac_nhbytes(mptr,kptr,VMAC_NHBYTES/8,rh,rl);
+ #else
+ nh_vmac_nhbytes_2(mptr,kptr,VMAC_NHBYTES/8,rh,rl,rh2,rl2);
+ rh2 &= m62;
+ poly_step(ch2,cl2,pkh2,pkl2,rh2,rl2);
+ #endif
+ rh &= m62;
+ poly_step(ch,cl,pkh,pkl,rh,rl);
+ mptr += (VMAC_NHBYTES/sizeof(uint64_t));
+ }
+ if (remaining) {
+ #if (VMAC_TAG_LEN == 64)
+ nh_16(mptr,kptr,2*((remaining+15)/16),rh,rl);
+ #else
+ nh_16_2(mptr,kptr,2*((remaining+15)/16),rh,rl,rh2,rl2);
+ rh2 &= m62;
+ poly_step(ch2,cl2,pkh2,pkl2,rh2,rl2);
+ #endif
+ rh &= m62;
+ poly_step(ch,cl,pkh,pkl,rh,rl);
+ }
+
+do_l3:
+ #if VMAC_USE_SSE2
+ _mm_empty(); /* SSE2 version of poly_step uses mmx instructions */
+ #endif
+ vhash_abort(ctx);
+ remaining *= 8;
+#if (VMAC_TAG_LEN == 128)
+ *tagl = l3hash(ch2, cl2, ctx->l3key[2], ctx->l3key[3],remaining);
+#endif
+ return l3hash(ch, cl, ctx->l3key[0], ctx->l3key[1],remaining);
+}
+
+/* ----------------------------------------------------------------------- */
+
+uint64_t vmac(unsigned char m[],
+ unsigned int mbytes,
+ unsigned char n[16],
+ uint64_t *tagl,
+ vmac_ctx_t *ctx)
+{
+#if (VMAC_TAG_LEN == 64)
+ uint64_t *in_n, *out_p;
+ uint64_t p, h;
+ int i;
+
+ #if VMAC_CACHE_NONCES
+ in_n = ctx->cached_nonce;
+ out_p = ctx->cached_aes;
+ #else
+ uint64_t tmp[2];
+ in_n = out_p = tmp;
+ #endif
+
+ i = n[15] & 1;
+ #if VMAC_CACHE_NONCES
+ if ((*(uint64_t *)(n+8) != in_n[1]) ||
+ (*(uint64_t *)(n ) != in_n[0])) {
+ #endif
+
+ in_n[0] = *(uint64_t *)(n );
+ in_n[1] = *(uint64_t *)(n+8);
+ ((unsigned char *)in_n)[15] &= 0xFE;
+ aes_encryption(in_n, out_p, &ctx->cipher_key);
+
+ #if VMAC_CACHE_NONCES
+ ((unsigned char *)in_n)[15] |= (unsigned char)(1-i);
+ }
+ #endif
+ p = get64BE(out_p + i);
+ h = vhash(m, mbytes, (uint64_t *)0, ctx);
+ return p + h;
+#else
+ uint64_t tmp[2];
+ uint64_t th,tl;
+ aes_encryption(n, (unsigned char *)tmp, &ctx->cipher_key);
+ th = vhash(m, mbytes, &tl, ctx);
+ th += get64BE(tmp);
+ *tagl = tl + get64BE(tmp+1);
+ return th;
+#endif
+}
+
+/* ----------------------------------------------------------------------- */
+
+void vmac_set_key(unsigned char user_key[], vmac_ctx_t *ctx)
+{
+ uint64_t in[2] = {0}, out[2];
+ unsigned i;
+ aes_key_setup(user_key, &ctx->cipher_key);
+
+ /* Fill nh key */
+ ((unsigned char *)in)[0] = 0x80;
+ for (i = 0; i < sizeof(ctx->nhkey)/8; i+=2) {
+ aes_encryption((unsigned char *)in, (unsigned char *)out,
+ &ctx->cipher_key);
+ ctx->nhkey[i ] = get64BE(out);
+ ctx->nhkey[i+1] = get64BE(out+1);
+ ((unsigned char *)in)[15] += 1;
+ }
+
+ /* Fill poly key */
+ ((unsigned char *)in)[0] = 0xC0;
+ in[1] = 0;
+ for (i = 0; i < sizeof(ctx->polykey)/8; i+=2) {
+ aes_encryption((unsigned char *)in, (unsigned char *)out,
+ &ctx->cipher_key);
+ ctx->polytmp[i ] = ctx->polykey[i ] = get64BE(out) & mpoly;
+ ctx->polytmp[i+1] = ctx->polykey[i+1] = get64BE(out+1) & mpoly;
+ ((unsigned char *)in)[15] += 1;
+ }
+
+ /* Fill ip key */
+ ((unsigned char *)in)[0] = 0xE0;
+ in[1] = 0;
+ for (i = 0; i < sizeof(ctx->l3key)/8; i+=2) {
+ do {
+ aes_encryption((unsigned char *)in, (unsigned char *)out,
+ &ctx->cipher_key);
+ ctx->l3key[i ] = get64BE(out);
+ ctx->l3key[i+1] = get64BE(out+1);
+ ((unsigned char *)in)[15] += 1;
+ } while (ctx->l3key[i] >= p64 || ctx->l3key[i+1] >= p64);
+ }
+
+ /* Invalidate nonce/aes cache and reset other elements */
+ #if (VMAC_TAG_LEN == 64) && (VMAC_CACHE_NONCES)
+ ctx->cached_nonce[0] = (uint64_t)-1; /* Ensure illegal nonce */
+ ctx->cached_nonce[1] = (uint64_t)0; /* Ensure illegal nonce */
+ #endif
+ ctx->first_block_processed = 0;
+}
+
+/* ----------------------------------------------------------------------- */
+
+
+#if VMAC_RUN_TESTS
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <time.h>
+#include <string.h>
+
+unsigned prime(void) /* Wake variable speed cpu, get rough speed estimate */
+{
+ volatile uint64_t i;
+ volatile uint64_t j=1;
+ unsigned cnt=0;
+ volatile clock_t ticks = clock();
+ do {
+ for (i = 0; i < 500000; i++) {
+ uint64_t x = get64PE(&j);
+ j = x * x + (uint64_t)ticks;
+ }
+ cnt++;
+ } while (clock() - ticks < (CLOCKS_PER_SEC/2));
+ return cnt; /* cnt is millions of iterations per second */
+}
+
+int main(void)
+{
+ ALIGN(16) vmac_ctx_t ctx, ctx_aio, ctx_inc1, ctx_inc2;
+ uint64_t res, tagl;
+ void *p;
+ unsigned char *m;
+ ALIGN(4) unsigned char key[] = "abcdefghijklmnop";
+ ALIGN(4) unsigned char nonce[] = "\0\0\0\0\0\0\0\0bcdefghi";
+ unsigned int vector_lengths[] = {0,3,48,300,3000000};
+ #if (VMAC_TAG_LEN == 64)
+ ALIGN(4) char *should_be[] = {"2576BE1C56D8B81B","2D376CF5B1813CE5",
+ "E8421F61D573D298","4492DF6C5CAC1BBE",
+ "09BA597DD7601113"};
+ #else
+ ALIGN(4) char *should_be[] = {"472766C70F74ED23481D6D7DE4E80DAC",
+ "4EE815A06A1D71EDD36FC75D51188A42",
+ "09F2C80C8E1007A0C12FAE19FE4504AE",
+ "66438817154850C61D8A412164803BCB",
+ "2B6B02288FFC461B75485DE893C629DC"};
+ #endif
+ unsigned speed_lengths[] = {16, 32, 64, 128, 256, 512, 1024, 2048, 4096};
+ unsigned i, j, *speed_iters;
+ clock_t ticks;
+ double cpb;
+ const unsigned int buf_len = 3 * (1 << 20);
+
+ j = prime();
+ i = sizeof(speed_lengths)/sizeof(speed_lengths[0]);
+ speed_iters = (unsigned *)malloc(i*sizeof(speed_iters[0]));
+ speed_iters[i-1] = j * (1 << 12);
+ while (--i) speed_iters[i-1] = (unsigned)(1.3 * speed_iters[i]);
+
+ /* Initialize context and message buffer, all 16-byte aligned */
+ p = malloc(buf_len + 32);
+ m = (unsigned char *)(((size_t)p + 16) & ~((size_t)15));
+ memset(m, 0, buf_len + 16);
+ vmac_set_key(key, &ctx);
+
+ /* Test incremental and all-in-one interfaces for correctness */
+ vmac_set_key(key, &ctx_aio);
+ vmac_set_key(key, &ctx_inc1);
+ vmac_set_key(key, &ctx_inc2);
+
+
+ /*
+ for (i = 0; i <= 512; i++) {
+ vhash_update(m,(i/VMAC_NHBYTES)*VMAC_NHBYTES,&ctx_inc1);
+ tagh = vmac(m+(i/VMAC_NHBYTES)*VMAC_NHBYTES, i%VMAC_NHBYTES,
+ nonce, &tagl, &ctx);
+ vhash_update(m,(i/VMAC_NHBYTES)*VMAC_NHBYTES,&ctx_inc1);
+ for (j = 0; j < vector_lengths[i]; j++)
+ m[j] = (unsigned char)('a'+j%3);
+
+ }
+ */
+
+ /* Generate vectors */
+ for (i = 0; i < sizeof(vector_lengths)/sizeof(unsigned int); i++) {
+ for (j = 0; j < vector_lengths[i]; j++)
+ m[j] = (unsigned char)('a'+j%3);
+ res = vmac(m, vector_lengths[i], nonce, &tagl, &ctx);
+ #if (VMAC_TAG_LEN == 64)
+ printf("\'abc\' * %7u: %016llX Should be: %s\n",
+ vector_lengths[i]/3,res,should_be[i]);
+ #else
+ printf("\'abc\' * %7u: %016llX%016llX\nShould be : %s\n",
+ vector_lengths[i]/3,res,tagl,should_be[i]);
+ #endif
+ }
+
+ /* Speed test */
+ for (i = 0; i < sizeof(speed_lengths)/sizeof(unsigned int); i++) {
+ ticks = clock();
+ for (j = 0; j < speed_iters[i]; j++) {
+ #if HASH_ONLY
+ res = vhash(m, speed_lengths[i], &tagl, &ctx);
+ #else
+ res = vmac(m, speed_lengths[i], nonce, &tagl, &ctx);
+ nonce[7]++;
+ #endif
+ }
+ ticks = clock() - ticks;
+ cpb = ((ticks*VMAC_HZ)/
+ ((double)CLOCKS_PER_SEC*speed_lengths[i]*speed_iters[i]));
+ printf("%4u bytes, %2.2f cpb\n", speed_lengths[i], cpb);
+ }
+ return 1;
+}
+
+#endif
diff --git a/xen/include/asm-x86/tboot.h b/xen/include/asm-x86/tboot.h
index 471e11f25c..0474da0f7b 100644
--- a/xen/include/asm-x86/tboot.h
+++ b/xen/include/asm-x86/tboot.h
@@ -53,10 +53,12 @@ typedef struct __packed {
/* used to communicate between tboot and the launched kernel (i.e. Xen) */
+#define TB_KEY_SIZE 64 /* 512 bits */
+
#define MAX_TB_MAC_REGIONS 32
typedef struct __packed {
- uint64_t start;
- uint64_t end;
+ uint64_t start; /* must be 64 byte -aligned */
+ uint32_t size; /* must be 64 byte -granular */
} tboot_mac_region_t;
/* GAS - Generic Address Structure (ACPI 2.0+) */
@@ -83,7 +85,7 @@ typedef struct __packed {
typedef struct __packed {
/* version 3+ fields: */
uuid_t uuid; /* {663C8DFF-E8B3-4b82-AABF-19EA4D057A08} */
- uint32_t version; /* Version number; currently supports 0.3 */
+ uint32_t version; /* Version number; currently supports 0.4 */
uint32_t log_addr; /* physical addr of tb_log_t log */
uint32_t shutdown_entry; /* entry point for tboot shutdown */
uint32_t shutdown_type; /* type of shutdown (TB_SHUTDOWN_*) */
@@ -94,6 +96,9 @@ typedef struct __packed {
uint8_t num_mac_regions; /* number mem regions to MAC on S3 */
/* contig regions memory to MAC on S3 */
tboot_mac_region_t mac_regions[MAX_TB_MAC_REGIONS];
+ /* version 4+ fields: */
+ /* populated by tboot; will be encrypted */
+ uint8_t s3_key[TB_KEY_SIZE];
} tboot_shared_t;
#define TB_SHUTDOWN_REBOOT 0
@@ -113,6 +118,7 @@ void tboot_shutdown(uint32_t shutdown_type);
int tboot_in_measured_env(void);
int tboot_protect_mem_regions(void);
int tboot_parse_dmar_table(acpi_table_handler dmar_handler);
+int tboot_s3_resume(void);
#endif /* __TBOOT_H__ */
diff --git a/xen/include/crypto/rijndael.h b/xen/include/crypto/rijndael.h
new file mode 100644
index 0000000000..29746029e8
--- /dev/null
+++ b/xen/include/crypto/rijndael.h
@@ -0,0 +1,58 @@
+/* $OpenBSD: rijndael.h,v 1.13 2008/06/09 07:49:45 djm Exp $ */
+
+/**
+ * rijndael-alg-fst.h
+ *
+ * @version 3.0 (December 2000)
+ *
+ * Optimised ANSI C code for the Rijndael cipher (now AES)
+ *
+ * @author Vincent Rijmen <vincent.rijmen@esat.kuleuven.ac.be>
+ * @author Antoon Bosselaers <antoon.bosselaers@esat.kuleuven.ac.be>
+ * @author Paulo Barreto <paulo.barreto@terra.com.br>
+ *
+ * This code is hereby placed in the public domain.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ''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 AUTHORS 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.
+ */
+#ifndef __RIJNDAEL_H
+#define __RIJNDAEL_H
+
+#define AES_MAXKEYBITS (256)
+#define AES_MAXKEYBYTES (AES_MAXKEYBITS/8)
+/* for 256-bit keys, fewer for less */
+#define AES_MAXROUNDS 14
+
+//typedef unsigned char u8;
+//typedef unsigned short u16;
+//typedef unsigned int u32;
+
+/* The structure for key information */
+typedef struct {
+ int enc_only; /* context contains only encrypt schedule */
+ int Nr; /* key-length-dependent number of rounds */
+ u32 ek[4*(AES_MAXROUNDS + 1)]; /* encrypt key schedule */
+ u32 dk[4*(AES_MAXROUNDS + 1)]; /* decrypt key schedule */
+} rijndael_ctx;
+
+int rijndael_set_key(rijndael_ctx *, const u_char *, int);
+int rijndael_set_key_enc_only(rijndael_ctx *, const u_char *, int);
+void rijndael_decrypt(rijndael_ctx *, const u_char *, u_char *);
+void rijndael_encrypt(rijndael_ctx *, const u_char *, u_char *);
+
+int rijndaelKeySetupEnc(unsigned int [], const unsigned char [], int);
+int rijndaelKeySetupDec(unsigned int [], const unsigned char [], int);
+void rijndaelEncrypt(const unsigned int [], int, const unsigned char [],
+ unsigned char []);
+
+#endif /* __RIJNDAEL_H */
diff --git a/xen/include/crypto/vmac.h b/xen/include/crypto/vmac.h
new file mode 100644
index 0000000000..9e92b3e351
--- /dev/null
+++ b/xen/include/crypto/vmac.h
@@ -0,0 +1,178 @@
+#ifndef HEADER_VMAC_H
+#define HEADER_VMAC_H
+
+/* --------------------------------------------------------------------------
+ * VMAC and VHASH Implementation by Ted Krovetz (tdk@acm.org) and Wei Dai.
+ * This implementation is herby placed in the public domain.
+ * The authors offers no warranty. Use at your own risk.
+ * Please send bug reports to the authors.
+ * Last modified: 17 APR 08, 1700 PDT
+ * ----------------------------------------------------------------------- */
+
+/* --------------------------------------------------------------------------
+ * User definable settings.
+ * ----------------------------------------------------------------------- */
+#define VMAC_TAG_LEN 64 /* Must be 64 or 128 - 64 sufficient for most */
+#define VMAC_KEY_LEN 128 /* Must be 128, 192 or 256 */
+#define VMAC_NHBYTES 128 /* Must 2^i for any 3 < i < 13. Standard = 128 */
+#define VMAC_PREFER_BIG_ENDIAN 0 /* Prefer non-x86 */
+
+#define VMAC_USE_OPENSSL 0 /* Set to non-zero to use OpenSSL's AES */
+#define VMAC_CACHE_NONCES 1 /* Set to non-zero to cause caching */
+ /* of consecutive nonces on 64-bit tags */
+
+#define VMAC_RUN_TESTS 0 /* Set to non-zero to check vectors and speed */
+#define VMAC_HZ (448e6) /* Set to hz of host machine to get speed */
+#define VMAC_HASH_ONLY 0 /* Set to non-zero to time hash only (not-mac) */
+/* Speeds of cpus I have access to
+#define hz (2400e6) glyme Core 2 "Conroe"
+#define hz (2000e6) jupiter G5
+#define hz (1592e6) titan
+#define hz (2793e6) athena/gaia
+#define hz (1250e6) isis G4
+#define hz (2160e6) imac Core 2 "Merom"
+#define hz (266e6) ppc/arm
+#define hz (400e6) mips
+*/
+
+/* --------------------------------------------------------------------------
+ * This implementation uses uint32_t and uint64_t as names for unsigned 32-
+ * and 64-bit integer types. These are defined in C99 stdint.h. The
+ * following may need adaptation if you are not running a C99 or
+ * Microsoft C environment.
+ * ----------------------------------------------------------------------- */
+#define VMAC_USE_STDINT 1 /* Set to zero if system has no stdint.h */
+
+#if VMAC_USE_STDINT && !_MSC_VER /* Try stdint.h if non-Microsoft */
+#ifdef __cplusplus
+#define __STDC_CONSTANT_MACROS
+#endif
+//#include <stdint.h>
+#elif (_MSC_VER) /* Microsoft C does not have stdint.h */
+typedef unsigned __int32 uint32_t;
+typedef unsigned __int64 uint64_t;
+#define UINT64_C(v) v ## UI64
+#else /* Guess sensibly - may need adaptation */
+typedef unsigned int uint32_t;
+typedef unsigned long long uint64_t;
+#define UINT64_C(v) v ## ULL
+#endif
+
+/* --------------------------------------------------------------------------
+ * This implementation supports two free AES implementations: OpenSSL's and
+ * Paulo Barreto's. To use OpenSSL's, you will need to include the OpenSSL
+ * crypto library (eg, gcc -lcrypto foo.c). For Barreto's, you will need
+ * to compile rijndael-alg-fst.c, last seen at http://www.iaik.tu-graz.ac.at/
+ * research/krypto/AES/old/~rijmen/rijndael/rijndael-fst-3.0.zip and
+ * http://homes.esat.kuleuven.be/~rijmen/rijndael/rijndael-fst-3.0.zip.
+ * To use a different implementation, use these definitions as a model.
+ * ----------------------------------------------------------------------- */
+#if VMAC_USE_OPENSSL
+
+#include <openssl/aes.h>
+typedef AES_KEY aes_int_key;
+
+#define aes_encryption(in,out,int_key) \
+ AES_encrypt((unsigned char *)(in),(unsigned char *)(out),(int_key))
+#define aes_key_setup(key,int_key) \
+ AES_set_encrypt_key((key),VMAC_KEY_LEN,(int_key))
+
+#else
+
+//#include "rijndael-alg-fst.h"
+typedef uint64_t vmac_t;
+#include "rijndael.h"
+typedef u32 aes_int_key[4*(VMAC_KEY_LEN/32+7)];
+
+#define aes_encryption(in,out,int_key) \
+ rijndaelEncrypt((u32 *)(int_key), \
+ ((VMAC_KEY_LEN/32)+6), \
+ (u8 *)(in), (u8 *)(out))
+#define aes_key_setup(user_key,int_key) \
+ rijndaelKeySetupEnc((u32 *)(int_key), \
+ (u8 *)(user_key), \
+ VMAC_KEY_LEN)
+#endif
+
+/* --------------------------------------------------------------------- */
+
+typedef struct {
+ uint64_t nhkey [(VMAC_NHBYTES/8)+2*(VMAC_TAG_LEN/64-1)];
+ uint64_t polykey[2*VMAC_TAG_LEN/64];
+ uint64_t l3key [2*VMAC_TAG_LEN/64];
+ uint64_t polytmp[2*VMAC_TAG_LEN/64];
+ aes_int_key cipher_key;
+ #if (VMAC_TAG_LEN == 64) && (VMAC_CACHE_NONCES)
+ uint64_t cached_nonce[2];
+ uint64_t cached_aes[2];
+ #endif
+ int first_block_processed;
+} vmac_ctx_t;
+
+/* --------------------------------------------------------------------- */
+#ifdef __cplusplus
+extern "C" {
+#endif
+/* --------------------------------------------------------------------------
+ * <<<<< USAGE NOTES >>>>>
+ *
+ * Given msg m (mbytes in length) and nonce buffer n
+ * this function returns a tag as its output. The tag is returned as
+ * a number. When VMAC_TAG_LEN == 64, the 'return'ed integer is the tag,
+ * and *tagl is meaningless. When VMAC_TAG_LEN == 128 the tag is the
+ * number y * 2^64 + *tagl where y is the function's return value.
+ * If you want to consider tags to be strings, then you must do so with
+ * an agreed upon endian orientation for interoperability, and convert
+ * the results appropriately. VHASH hashes m without creating any tag.
+ * Consecutive substrings forming a prefix of a message may be passed
+ * to vhash_update, with vhash or vmac being called with the remainder
+ * to produce the output.
+ *
+ * Requirements:
+ * - On 32-bit architectures with SSE2 instructions, ctx and m MUST be
+ * begin on 16-byte memory boundaries.
+ * - m MUST be your message followed by zeroes to the nearest 16-byte
+ * boundary. If m is a length multiple of 16 bytes, then it is already
+ * at a 16-byte boundary and needs no padding. mbytes should be your
+ * message length without any padding.
+ * - The first bit of the nonce buffer n must be 0. An i byte nonce, is made
+ * as the first 16-i bytes of n being zero, and the final i the nonce.
+ * - vhash_update MUST have mbytes be a positive multiple of VMAC_NHBYTES
+ * ----------------------------------------------------------------------- */
+
+#define vmac_update vhash_update
+
+void vhash_update(unsigned char m[],
+ unsigned int mbytes,
+ vmac_ctx_t *ctx);
+
+uint64_t vmac(unsigned char m[],
+ unsigned int mbytes,
+ unsigned char n[16],
+ uint64_t *tagl,
+ vmac_ctx_t *ctx);
+
+uint64_t vhash(unsigned char m[],
+ unsigned int mbytes,
+ uint64_t *tagl,
+ vmac_ctx_t *ctx);
+
+/* --------------------------------------------------------------------------
+ * When passed a VMAC_KEY_LEN bit user_key, this function initialazies ctx.
+ * ----------------------------------------------------------------------- */
+
+void vmac_set_key(unsigned char user_key[], vmac_ctx_t *ctx);
+
+/* --------------------------------------------------------------------------
+ * This function aborts current hash and resets ctx, ready for a new message.
+ * ----------------------------------------------------------------------- */
+
+void vhash_abort(vmac_ctx_t *ctx);
+
+/* --------------------------------------------------------------------- */
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif /* HEADER_AES_H */
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-26 20:56:21 +0000