Merge tag 'pull-tcg-20230620' of https://gitlab.com/rth7680/qemu into…

… staging tcg: Define _CALL_AIX for clang on ppc64 accel/tcg: Build fix for macos catalina accel/tcg: Handle MO_ATOM_WITHIN16 in do_st16_leN accel/tcg: Restrict SavedIOTLB to system emulation accel/tcg: Use generic 'helper-proto-common.h' header plugins: Remove unused 'exec/helper-proto.h' header *: Check for CONFIG_USER_ONLY instead of CONFIG_SOFTMMU # -----BEGIN PGP SIGNATURE----- # # iQFRBAABCgA7FiEEekgeeIaLTbaoWgXAZN846K9+IV8FAmSRYmIdHHJpY2hhcmQu # aGVuZGVyc29uQGxpbmFyby5vcmcACgkQZN846K9+IV8zbAgAlX4GcShS1OU1BDRe # b0HHHj1fFBB/9yk8f/5WuQb2snYS+pcZCez9XeT175ugovXSOz+shvmFrbRPvpfj # q8C88CIKCJRsXnhWqKWOKDqgTttu2WNXOvCe0eCZbUoGQ9K1seMvUBq6T50fNv2H # fXeHtLSu/+jiHIN3+woJqdgrkp0cko2rrpnwIpjuIsY1iz/J/VKEHmnv7Ah+GsRs # OTYnR7iN6uhBXVll14r3UCylbgdEz58sSSEi3dYYfaTRuijDwOzM0evhk6+5XzHP # DYwGdbtDE5HJOrCLiKegk80Gh6v1XVZQWnn9PdiN1eJcQsWNT9mYV9/4IsCVrsF4 # 8r5KUg== # =JmjK # -----END PGP SIGNATURE----- # gpg: Signature made Tue 20 Jun 2023 10:25:06 AM CEST # gpg: using RSA key 7A481E78868B4DB6A85A05C064DF38E8AF7E215F # gpg: issuer "richard.henderson@linaro.org" # gpg: Good signature from "Richard Henderson <richard.henderson@linaro.org>" [ultimate] * tag 'pull-tcg-20230620' of https://gitlab.com/rth7680/qemu: cputlb: Restrict SavedIOTLB to system emulation exec/cpu-defs: Check for SOFTMMU instead of !USER_ONLY accel/tcg/cpu-exec: Use generic 'helper-proto-common.h' header plugins: Remove unused 'exec/helper-proto.h' header meson: Replace softmmu_ss -> system_ss meson: Replace CONFIG_SOFTMMU -> CONFIG_SYSTEM_ONLY meson: Alias CONFIG_SOFTMMU -> CONFIG_SYSTEM_ONLY accel/tcg: Check for USER_ONLY definition instead of SOFTMMU one hw/core/cpu: Check for USER_ONLY definition instead of SOFTMMU one target/ppc: Check for USER_ONLY definition instead of SOFTMMU one target/m68k: Check for USER_ONLY definition instead of SOFTMMU one target/tricore: Remove pointless CONFIG_SOFTMMU guard target/i386: Simplify i386_tr_init_disas_context() tcg/ppc: Define _CALL_AIX for clang on ppc64(be) accel/tcg: Handle MO_ATOM_WITHIN16 in do_st16_leN host/include/x86_64: Use __m128i for "x" constraints Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
qemu · Jun 20, 2023 · cab35c7 · cab35c7
2 parents 48ab886 + d7ee93e
commit cab35c7
Show file tree

Hide file tree

Showing 125 changed files with 872 additions and 862 deletions.
diff --git a/accel/meson.build b/accel/meson.build
@@ -1,5 +1,5 @@
 specific_ss.add(files('accel-common.c', 'accel-blocker.c'))
-softmmu_ss.add(files('accel-softmmu.c'))
+system_ss.add(files('accel-softmmu.c'))
 user_ss.add(files('accel-user.c'))
 
 subdir('tcg')
@@ -12,4 +12,4 @@ if have_system
 endif
 
 # qtest
-softmmu_ss.add(files('dummy-cpus.c'))
+system_ss.add(files('dummy-cpus.c'))
diff --git a/accel/qtest/meson.build b/accel/qtest/meson.build
@@ -1 +1 @@
-qtest_module_ss.add(when: ['CONFIG_SOFTMMU'], if_true: files('qtest.c'))
+qtest_module_ss.add(when: ['CONFIG_SYSTEM_ONLY'], if_true: files('qtest.c'))
diff --git a/accel/stubs/meson.build b/accel/stubs/meson.build
@@ -4,4 +4,4 @@ sysemu_stubs_ss.add(when: 'CONFIG_XEN', if_false: files('xen-stub.c'))
 sysemu_stubs_ss.add(when: 'CONFIG_KVM', if_false: files('kvm-stub.c'))
 sysemu_stubs_ss.add(when: 'CONFIG_TCG', if_false: files('tcg-stub.c'))
 
-specific_ss.add_all(when: ['CONFIG_SOFTMMU'], if_true: sysemu_stubs_ss)
+specific_ss.add_all(when: ['CONFIG_SYSTEM_ONLY'], if_true: sysemu_stubs_ss)
diff --git a/accel/tcg/cpu-exec.c b/accel/tcg/cpu-exec.c
@@ -38,7 +38,7 @@
 #include "sysemu/cpu-timers.h"
 #include "exec/replay-core.h"
 #include "sysemu/tcg.h"
-#include "exec/helper-proto.h"
+#include "exec/helper-proto-common.h"
 #include "tb-jmp-cache.h"
 #include "tb-hash.h"
 #include "tb-context.h"
@@ -566,7 +566,7 @@ void cpu_exec_step_atomic(CPUState *cpu)
         cpu_tb_exec(cpu, tb, &tb_exit);
         cpu_exec_exit(cpu);
     } else {
-#ifndef CONFIG_SOFTMMU
+#ifdef CONFIG_USER_ONLY
         clear_helper_retaddr();
         if (have_mmap_lock()) {
             mmap_unlock();
@@ -1023,7 +1023,7 @@ static int cpu_exec_setjmp(CPUState *cpu, SyncClocks *sc)
         /* Non-buggy compilers preserve this; assert the correct value. */
         g_assert(cpu == current_cpu);
 
-#ifndef CONFIG_SOFTMMU
+#ifdef CONFIG_USER_ONLY
         clear_helper_retaddr();
         if (have_mmap_lock()) {
             mmap_unlock();

diff --git a/accel/tcg/cputlb.c b/accel/tcg/cputlb.c
@@ -2727,6 +2727,7 @@ static uint64_t do_st16_leN(CPUArchState *env, MMULookupPageData *p,
          * and so neither is atomic.
          */
     case MO_ATOM_IFALIGN:
+    case MO_ATOM_WITHIN16:
     case MO_ATOM_NONE:
         stq_le_p(p->haddr, int128_getlo(val_le));
         return store_bytes_leN(p->haddr + 8, p->size - 8,

diff --git a/accel/tcg/internal.h b/accel/tcg/internal.h
@@ -17,10 +17,10 @@
  * memory related structures are protected with mmap_lock.
  * In !user-mode we use per-page locks.
  */
-#ifdef CONFIG_SOFTMMU
-#define assert_memory_lock()
-#else
+#ifdef CONFIG_USER_ONLY
 #define assert_memory_lock() tcg_debug_assert(have_mmap_lock())
+#else
+#define assert_memory_lock()
 #endif
 
 #if defined(CONFIG_SOFTMMU) && defined(CONFIG_DEBUG_TCG)

diff --git a/accel/tcg/meson.build b/accel/tcg/meson.build
@@ -10,18 +10,18 @@ tcg_ss.add(files(
   'translator.c',
 ))
 tcg_ss.add(when: 'CONFIG_USER_ONLY', if_true: files('user-exec.c'))
-tcg_ss.add(when: 'CONFIG_SOFTMMU', if_false: files('user-exec-stub.c'))
+tcg_ss.add(when: 'CONFIG_SYSTEM_ONLY', if_false: files('user-exec-stub.c'))
 tcg_ss.add(when: 'CONFIG_PLUGIN', if_true: [files('plugin-gen.c')])
 tcg_ss.add(when: libdw, if_true: files('debuginfo.c'))
 tcg_ss.add(when: 'CONFIG_LINUX', if_true: files('perf.c'))
 specific_ss.add_all(when: 'CONFIG_TCG', if_true: tcg_ss)
 
-specific_ss.add(when: ['CONFIG_SOFTMMU', 'CONFIG_TCG'], if_true: files(
+specific_ss.add(when: ['CONFIG_SYSTEM_ONLY', 'CONFIG_TCG'], if_true: files(
   'cputlb.c',
   'monitor.c',
 ))
 
-tcg_module_ss.add(when: ['CONFIG_SOFTMMU', 'CONFIG_TCG'], if_true: files(
+tcg_module_ss.add(when: ['CONFIG_SYSTEM_ONLY', 'CONFIG_TCG'], if_true: files(
   'tcg-accel-ops.c',
   'tcg-accel-ops-mttcg.c',
   'tcg-accel-ops-icount.c',

diff --git a/audio/meson.build b/audio/meson.build
@@ -1,5 +1,5 @@
-softmmu_ss.add([spice_headers, files('audio.c')])
-softmmu_ss.add(files(
+system_ss.add([spice_headers, files('audio.c')])
+system_ss.add(files(
   'audio-hmp-cmds.c',
   'audio_legacy.c',
   'mixeng.c',
@@ -8,8 +8,8 @@ softmmu_ss.add(files(
   'wavcapture.c',
 ))
 
-softmmu_ss.add(when: coreaudio, if_true: files('coreaudio.m'))
-softmmu_ss.add(when: dsound, if_true: files('dsoundaudio.c', 'audio_win_int.c'))
+system_ss.add(when: coreaudio, if_true: files('coreaudio.m'))
+system_ss.add(when: dsound, if_true: files('dsoundaudio.c', 'audio_win_int.c'))
 
 audio_modules = {}
 foreach m : [

diff --git a/backends/meson.build b/backends/meson.build
@@ -1,4 +1,4 @@
-softmmu_ss.add([files(
+system_ss.add([files(
   'cryptodev-builtin.c',
   'cryptodev-hmp-cmds.c',
   'cryptodev.c',
@@ -10,20 +10,20 @@ softmmu_ss.add([files(
   'confidential-guest-support.c',
 ), numa])
 
-softmmu_ss.add(when: 'CONFIG_POSIX', if_true: files('rng-random.c'))
-softmmu_ss.add(when: 'CONFIG_POSIX', if_true: files('hostmem-file.c'))
-softmmu_ss.add(when: 'CONFIG_LINUX', if_true: files('hostmem-memfd.c'))
+system_ss.add(when: 'CONFIG_POSIX', if_true: files('rng-random.c'))
+system_ss.add(when: 'CONFIG_POSIX', if_true: files('hostmem-file.c'))
+system_ss.add(when: 'CONFIG_LINUX', if_true: files('hostmem-memfd.c'))
 if keyutils.found()
-    softmmu_ss.add(keyutils, files('cryptodev-lkcf.c'))
+    system_ss.add(keyutils, files('cryptodev-lkcf.c'))
 endif
 if have_vhost_user
-  softmmu_ss.add(when: 'CONFIG_VIRTIO', if_true: files('vhost-user.c'))
+  system_ss.add(when: 'CONFIG_VIRTIO', if_true: files('vhost-user.c'))
 endif
-softmmu_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost.c'))
+system_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost.c'))
 if have_vhost_user_crypto
-  softmmu_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost-user.c'))
+  system_ss.add(when: 'CONFIG_VIRTIO_CRYPTO', if_true: files('cryptodev-vhost-user.c'))
 endif
-softmmu_ss.add(when: gio, if_true: files('dbus-vmstate.c'))
-softmmu_ss.add(when: 'CONFIG_SGX', if_true: files('hostmem-epc.c'))
+system_ss.add(when: gio, if_true: files('dbus-vmstate.c'))
+system_ss.add(when: 'CONFIG_SGX', if_true: files('hostmem-epc.c'))
 
 subdir('tpm')
diff --git a/backends/tpm/meson.build b/backends/tpm/meson.build
@@ -1,6 +1,6 @@
 if have_tpm
-  softmmu_ss.add(files('tpm_backend.c'))
-  softmmu_ss.add(files('tpm_util.c'))
-  softmmu_ss.add(when: 'CONFIG_TPM_PASSTHROUGH', if_true: files('tpm_passthrough.c'))
-  softmmu_ss.add(when: 'CONFIG_TPM_EMULATOR', if_true: files('tpm_emulator.c'))
+  system_ss.add(files('tpm_backend.c'))
+  system_ss.add(files('tpm_util.c'))
+  system_ss.add(when: 'CONFIG_TPM_PASSTHROUGH', if_true: files('tpm_passthrough.c'))
+  system_ss.add(when: 'CONFIG_TPM_EMULATOR', if_true: files('tpm_emulator.c'))
 endif
diff --git a/block/meson.build b/block/meson.build
@@ -42,8 +42,8 @@ block_ss.add(files(
   'write-threshold.c',
 ), zstd, zlib, gnutls)
 
-softmmu_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
-softmmu_ss.add(files('block-ram-registrar.c'))
+system_ss.add(when: 'CONFIG_TCG', if_true: files('blkreplay.c'))
+system_ss.add(files('block-ram-registrar.c'))
 
 if get_option('qcow1').allowed()
   block_ss.add(files('qcow.c'))
@@ -159,7 +159,7 @@ block_ss.add(block_gen_c)
 
 block_ss.add(files('stream.c'))
 
-softmmu_ss.add(files('qapi-sysemu.c'))
+system_ss.add(files('qapi-sysemu.c'))
 
 subdir('export')
 subdir('monitor')

diff --git a/block/monitor/meson.build b/block/monitor/meson.build
@@ -1,2 +1,2 @@
-softmmu_ss.add(files('block-hmp-cmds.c'))
+system_ss.add(files('block-hmp-cmds.c'))
 block_ss.add(files('bitmap-qmp-cmds.c'))
diff --git a/chardev/meson.build b/chardev/meson.build
@@ -28,7 +28,7 @@ chardev_ss.add(when: 'CONFIG_WIN32', if_true: files(
 
 chardev_ss = chardev_ss.apply(config_host, strict: false)
 
-softmmu_ss.add(files(
+system_ss.add(files(
     'char-hmp-cmds.c',
     'msmouse.c',
     'wctablet.c',

diff --git a/disas/meson.build b/disas/meson.build
@@ -13,5 +13,5 @@ common_ss.add(when: 'CONFIG_XTENSA_DIS', if_true: files('xtensa.c'))
 common_ss.add(when: capstone, if_true: [files('capstone.c'), capstone])
 common_ss.add(files('disas.c'))
 
-softmmu_ss.add(files('disas-mon.c'))
+system_ss.add(files('disas-mon.c'))
 specific_ss.add(capstone)
diff --git a/docs/devel/build-system.rst b/docs/devel/build-system.rst
@@ -183,9 +183,9 @@ Target-independent emulator sourcesets:
   This includes error handling infrastructure, standard data structures,
   platform portability wrapper functions, etc.
 
-  Target-independent code lives in the ``common_ss``, ``softmmu_ss`` and
+  Target-independent code lives in the ``common_ss``, ``system_ss`` and
   ``user_ss`` sourcesets.  ``common_ss`` is linked into all emulators,
-  ``softmmu_ss`` only in system emulators, ``user_ss`` only in user-mode
+  ``system_ss`` only in system emulators, ``user_ss`` only in user-mode
   emulators.
 
   Target-independent sourcesets must exercise particular care when using
@@ -197,11 +197,11 @@ Target-independent emulator sourcesets:
   symbol::
 
     # Some targets have CONFIG_ACPI, some don't, so this is not enough
-    softmmu_ss.add(when: 'CONFIG_ACPI', if_true: files('acpi.c'),
+    system_ss.add(when: 'CONFIG_ACPI', if_true: files('acpi.c'),
                                         if_false: files('acpi-stub.c'))
 
     # This is required as well:
-    softmmu_ss.add(when: 'CONFIG_ALL', if_true: files('acpi-stub.c'))
+    system_ss.add(when: 'CONFIG_ALL', if_true: files('acpi-stub.c'))
 
 Target-dependent emulator sourcesets:
   In the target-dependent set lives CPU emulation, some device emulation and
@@ -229,16 +229,16 @@ Target-dependent emulator sourcesets:
   for all emulators and for system emulators only.  For example::
 
     arm_ss = ss.source_set()
-    arm_softmmu_ss = ss.source_set()
+    arm_system_ss = ss.source_set()
     ...
     target_arch += {'arm': arm_ss}
-    target_softmmu_arch += {'arm': arm_softmmu_ss}
+    target_softmmu_arch += {'arm': arm_system_ss}
 
 Module sourcesets:
   There are two dictionaries for modules: ``modules`` is used for
   target-independent modules and ``target_modules`` is used for
   target-dependent modules.  When modules are disabled the ``module``
-  source sets are added to ``softmmu_ss`` and the ``target_modules``
+  source sets are added to ``system_ss`` and the ``target_modules``
   source sets are added to ``specific_ss``.
 
   Both dictionaries are nested.  One dictionary is created per

diff --git a/dump/meson.build b/dump/meson.build
@@ -1,2 +1,2 @@
-softmmu_ss.add([files('dump.c', 'dump-hmp-cmds.c'), snappy, lzo])
-specific_ss.add(when: 'CONFIG_SOFTMMU', if_true: files('win_dump.c'))
+system_ss.add([files('dump.c', 'dump-hmp-cmds.c'), snappy, lzo])
+specific_ss.add(when: 'CONFIG_SYSTEM_ONLY', if_true: files('win_dump.c'))
diff --git a/ebpf/meson.build b/ebpf/meson.build
@@ -1 +1 @@
-softmmu_ss.add(when: libbpf, if_true: files('ebpf_rss.c'), if_false: files('ebpf_rss-stub.c'))
+system_ss.add(when: libbpf, if_true: files('ebpf_rss.c'), if_false: files('ebpf_rss-stub.c'))
diff --git a/fsdev/meson.build b/fsdev/meson.build
@@ -6,8 +6,8 @@ fsdev_ss.add(when: ['CONFIG_FSDEV_9P'], if_true: files(
   '9p-marshal.c',
   'qemu-fsdev.c',
 ), if_false: files('qemu-fsdev-dummy.c'))
-softmmu_ss.add_all(when: 'CONFIG_LINUX', if_true: fsdev_ss)
-softmmu_ss.add_all(when: 'CONFIG_DARWIN', if_true: fsdev_ss)
+system_ss.add_all(when: 'CONFIG_LINUX', if_true: fsdev_ss)
+system_ss.add_all(when: 'CONFIG_DARWIN', if_true: fsdev_ss)
 
 if have_virtfs_proxy_helper
   executable('virtfs-proxy-helper',

diff --git a/gdbstub/meson.build b/gdbstub/meson.build
@@ -8,14 +8,14 @@
 # cflags so:
 
 gdb_user_ss = ss.source_set()
-gdb_softmmu_ss = ss.source_set()
+gdb_system_ss = ss.source_set()
 
 # We build two versions of gdbstub, one for each mode
 gdb_user_ss.add(files('gdbstub.c', 'user.c'))
-gdb_softmmu_ss.add(files('gdbstub.c', 'softmmu.c'))
+gdb_system_ss.add(files('gdbstub.c', 'softmmu.c'))
 
 gdb_user_ss = gdb_user_ss.apply(config_host, strict: false)
-gdb_softmmu_ss = gdb_softmmu_ss.apply(config_host, strict: false)
+gdb_system_ss = gdb_system_ss.apply(config_host, strict: false)
 
 libgdb_user = static_library('gdb_user',
                              gdb_user_ss.sources() + genh,
@@ -24,14 +24,14 @@ libgdb_user = static_library('gdb_user',
                              build_by_default: have_user)
 
 libgdb_softmmu = static_library('gdb_softmmu',
-                                gdb_softmmu_ss.sources() + genh,
+                                gdb_system_ss.sources() + genh,
                                 name_suffix: 'fa',
                                 build_by_default: have_system)
 
 gdb_user = declare_dependency(link_whole: libgdb_user)
 user_ss.add(gdb_user)
 gdb_softmmu = declare_dependency(link_whole: libgdb_softmmu)
-softmmu_ss.add(gdb_softmmu)
+system_ss.add(gdb_softmmu)
 
 common_ss.add(files('syscalls.c'))
 

diff --git a/host/include/x86_64/host/atomic128-ldst.h b/host/include/x86_64/host/atomic128-ldst.h
@@ -8,12 +8,19 @@
  * atomic primitive is meant to provide.
  */
 
-#ifndef AARCH64_ATOMIC128_LDST_H
-#define AARCH64_ATOMIC128_LDST_H
+#ifndef X86_64_ATOMIC128_LDST_H
+#define X86_64_ATOMIC128_LDST_H
 
 #ifdef CONFIG_INT128_TYPE
 #include "host/cpuinfo.h"
 #include "tcg/debug-assert.h"
+#include <immintrin.h>
+
+typedef union {
+    __m128i v;
+    __int128_t i;
+    Int128 s;
+} X86Int128Union;
 
 /*
  * Through clang 16, with -mcx16, __atomic_load_n is incorrectly
@@ -25,21 +32,21 @@
 
 static inline Int128 atomic16_read_ro(const Int128 *ptr)
 {
-    Int128Alias r;
+    X86Int128Union r;
 
     tcg_debug_assert(HAVE_ATOMIC128_RO);
-    asm("vmovdqa %1, %0" : "=x" (r.i) : "m" (*ptr));
+    asm("vmovdqa %1, %0" : "=x" (r.v) : "m" (*ptr));
 
     return r.s;
 }
 
 static inline Int128 atomic16_read_rw(Int128 *ptr)
 {
     __int128_t *ptr_align = __builtin_assume_aligned(ptr, 16);
-    Int128Alias r;
+    X86Int128Union r;
 
     if (HAVE_ATOMIC128_RO) {
-        asm("vmovdqa %1, %0" : "=x" (r.i) : "m" (*ptr_align));
+        asm("vmovdqa %1, %0" : "=x" (r.v) : "m" (*ptr_align));
     } else {
         r.i = __sync_val_compare_and_swap_16(ptr_align, 0, 0);
     }
@@ -49,10 +56,10 @@ static inline Int128 atomic16_read_rw(Int128 *ptr)
 static inline void atomic16_set(Int128 *ptr, Int128 val)
 {
     __int128_t *ptr_align = __builtin_assume_aligned(ptr, 16);
-    Int128Alias new = { .s = val };
+    X86Int128Union new = { .s = val };
 
     if (HAVE_ATOMIC128_RO) {
-        asm("vmovdqa %1, %0" : "=m"(*ptr_align) : "x" (new.i));
+        asm("vmovdqa %1, %0" : "=m"(*ptr_align) : "x" (new.v));
     } else {
         __int128_t old;
         do {
@@ -65,4 +72,4 @@ static inline void atomic16_set(Int128 *ptr, Int128 val)
 #include "host/include/generic/host/atomic128-ldst.h"
 #endif
 
-#endif /* AARCH64_ATOMIC128_LDST_H */
+#endif /* X86_64_ATOMIC128_LDST_H */
diff --git a/host/include/x86_64/host/load-extract-al16-al8.h b/host/include/x86_64/host/load-extract-al16-al8.h
@@ -9,7 +9,7 @@
 #define X86_64_LOAD_EXTRACT_AL16_AL8_H
 
 #ifdef CONFIG_INT128_TYPE
-#include "host/cpuinfo.h"
+#include "host/atomic128-ldst.h"
 
 /**
  * load_atom_extract_al16_or_al8:
@@ -26,7 +26,7 @@ load_atom_extract_al16_or_al8(void *pv, int s)
     uintptr_t pi = (uintptr_t)pv;
     __int128_t *ptr_align = (__int128_t *)(pi & ~7);
     int shr = (pi & 7) * 8;
-    Int128Alias r;
+    X86Int128Union r;
 
     /*
      * ptr_align % 16 is now only 0 or 8.
@@ -35,9 +35,9 @@ load_atom_extract_al16_or_al8(void *pv, int s)
      * when ptr_align % 16 == 0 for 16-byte atomicity.
      */
     if ((cpuinfo & CPUINFO_ATOMIC_VMOVDQU) || (pi & 8)) {
-        asm("vmovdqu %1, %0" : "=x" (r.i) : "m" (*ptr_align));
+        asm("vmovdqu %1, %0" : "=x" (r.v) : "m" (*ptr_align));
     } else {
-        asm("vmovdqa %1, %0" : "=x" (r.i) : "m" (*ptr_align));
+        asm("vmovdqa %1, %0" : "=x" (r.v) : "m" (*ptr_align));
     }
     return int128_getlo(int128_urshift(r.s, shr));
 }