September 2026 - Wine-Devel

Wine staging 11.18 release
by Alistair Leslie-Hughes Sept. 18, 2026

Sept. 18, 2026

Binary packages for various distributions will be available from: https://www.winehq.org/download Summary since last release * Rebased to current wine 11.18 (273 patches are applied to wine vanilla) Upstreamed (Either directly from staging or fixed with a similar patch). * include: Convert d3d10_1shader.h to idl Removed (No longer required). * None Added: * [59846]: quartz - Convert from YV12 to RGB on render. * [47175] - services - Check the binary before setting WOW64 entry. + [50431] - services - Dont write WOW64 for 64bit binaries. Updated: * vkd3d-latest Where can you help * Run Steam/Battle.net/GOG/UPlay/Epic * Test your favorite game. * Test your favorite applications. * Improve staging patches and get them accepted upstream. * Suggest patches to be included in staging. As always, if you find a bug, please report it via https://bugs.winehq.org Best Regards Alistair.

1 0

Fwd: [PATCH 0/2] x86/fred: Reconstruct fault state for rejected INT instructions
by Paul Gofman Sept. 18, 2026

Sept. 18, 2026

-------- Forwarded Message -------- Subject: [PATCH 0/2] x86/fred: Reconstruct fault state for rejected INT instructions Date: Thu, 17 Sep 2026 16:09:05 -0700 From: Matthew Schwartz <matthew.schwartz(a)linux.dev> To: Xin Li <xin(a)zytor.com>, hpa(a)zytor.com CC: Andy Lutomirski <luto(a)kernel.org>, Thomas Gleixner <tglx(a)kernel.org>, Ingo Molnar <mingo(a)redhat.com>, Borislav Petkov <bp(a)alien8.de>, Dave Hansen <dave.hansen(a)linux.intel.com>, x86(a)kernel.org, linux-kernel(a)vger.kernel.org, linux-kselftest(a)vger.kernel.org, Paul Gofman <pgofman(a)codeweavers.com>, Lionel Landwerlin <lionel.g.landwerlin(a)intel.com>, Shuah Khan <shuah(a)kernel.org>, Matthew Schwartz <matthew.schwartz(a)linux.dev> This series reconstructs the #GP signal context for rejected INT instructions with FRED enabled. The incorrect error code and saved IP break Wine's handling of INT 0x2d, causing Elden Ring to fail with a spurious access violation on Panther Lake. The software event flag handling follows the same reasoning as the existing sigreturn fix in prevent_single_step_upon_eretu(). The second patch adds 32-bit and 64-bit signal-context and ptrace resume coverage. Both selftest variants pass all 29 checks on patched FRED-enabled Panther Lake and on a non-FRED AMD host. With the same binaries on unpatched Panther Lake, 16 signal-context checks and the first ptrace IP check fail. The two dependent ptrace resume checks are not reached. Separate continuation testing with the fix matched an IDT baseline on another system in both bitnesses: unchanged IP retries the INT whether RF is set or clear, and advancing IP resumes at the next instruction. One observed ptrace divergence remains separate: after rewriting IP at the stop from raise(SIGSTOP), PTRACE_SINGLESTEP traps before the target instruction on FRED, whereas IDT executes it. The selftest starts from an INT3 stop to isolate rejected-INT behavior. The syscall-stop case is a follow-up investigation outside this series. Matthew Schwartz (2): x86/fred: Reconstruct the #GP context for rejected INT instructions selftests/x86: Check signal state for rejected software interrupts arch/x86/entry/entry_fred.c | 11 +- tools/testing/selftests/x86/Makefile | 2 +- tools/testing/selftests/x86/int_signal.c | 311 +++++++++++++++++++++++ 3 files changed, 322 insertions(+), 2 deletions(-) create mode 100644 tools/testing/selftests/x86/int_signal.c base-commit: b5a051f6b840d48f159166ef073d3021989bfb50 -- 2.55.0

1 0

Wine on ARM64 macOS
by Brendan Shanks Sept. 9, 2026

Sept. 9, 2026

Hi everyone, There’s been a lot of interest in this over the years, and I’m excited to announce that all the pieces are now present in macOS to natively run ARM64 Wine. Thanks to all the teams at Apple and CodeWeavers who have worked on this effort. For anyone who wants to test this out now, last week CodeWeavers released a CrossOver Preview for macOS that uses ARM64 Wine in combination with FEX for i386/x86_64 (ARM64EC) emulation: <https://www.codeweavers.com/blog/mjohnson/2026/7/31/crossover-preview-the-r…> I’ll be sending MRs upstream soon to enable this, but here’s more details on what’s been added to macOS: **x18** Apple’s ARM64 ABI reserves x18 for use by the platform, and on recent OS versions it’s liable to be cleared at any time. The Windows ARM64 ABI uses x18 for the TEB, which conflicts with Apple’s ABI. In macOS 26.5, Apple has added a special “x18 custom ABI” mode that threads can switch into when running code that uses x18, and then switch out of when calling back into system frameworks. See <usr/include/os/arch/arm64.h> in a current SDK for more details. I’m working on patches to use this API in Wine. Alternately, older versions of macOS (11/12) did not clear x18, and current versions of macOS continue to maintain that behavior for binaries built against those older SDKs. Linking with `-Wl,-platform_version,macos,11.0,11.0` results in x18 not being cleared. This is a workaround that can be used until Wine is using the x18 custom ABI. **4K pages** macOS on ARM64 uses 16K pages, and Windows on all platforms uses 4K pages. Wine has some ability to work around this mismatch, but it’s not possible to fully emulate 4K page semantics. macOS 26.5 adds `posix_spawnattr_set_4k_page_size_np()`, a posix_spawn attribute that will spawn/exec a process with 4K pages. **PAGEZERO** macOS on ARM64 enforces a PAGEZERO of at least 4GB, preventing Wine from mapping USER_SHARED_DATA at 0x7ffe0000 (where Windows maps it on all platforms). Emulating i386/x86_64 Windows binaries also requires using the below-4GB address space. macOS/Xcode 26.4 have kernel and linker changes (the new `-Wl,-x86_64_layout_emulation` flag) which allow the PAGEZERO size to be changed, and also remapped. The best approach we’ve found is to actually increase PAGEZERO’s size to around 0x170000000, this reserves that entire address space for Wine. **TSO** Apple’s CPUs include support for the TSO memory model, to speed-up emulation of x86 code. macOS 26.5 allows this to be enabled per-thread with the `thread_set_x86_64_compat()` call. Note that newly-created threads do not inherit this state. **Entitlements** Unfortunately, restricted entitlements are required to run binaries using any of the above features. These entitlements require an Apple developer account to use: `com.apple.developer.cross-architecture-support` is available to paid developer accounts, `com.apple.developer.cross-architecture-support-unmanaged` is the same except that it’s available to free developer accounts. More detailed documentation will be coming along with Wine patches, but the rough process is: in your developer account, create an App ID for the Wine loader and enable the Cross-Process Architecture entitlement for it. Ensure your Mac’s provisioning UDID is added to your account. Create a provisioning profile for the app ID, and download it. The profile needs to be placed in a .app bundle along with the Wine loader. It is possible to bypass all this by disabling System Integrity Protection and setting kernel boot args to disable entitlement and code signing restrictions. We'll document that too, although SIP is a key part of the macOS security model and ultimately we can’t recommend disabling it. Again, I’ll be submitting Wine MRs soon to take advantage of these new features. Brendan

2 3

Wine staging 11.17 release
by Alistair Leslie-Hughes Sept. 5, 2026

Sept. 5, 2026

Binary packages for various distributions will be available from: https://www.winehq.org/download Summary since last release * Rebased to current wine 11.17 (279 patches are applied to wine vanilla) Upstreamed (Either directly from staging or fixed with a similar patch). * kernel32: Implement SetThreadpoolTimerEx Removed (No longer required). * None Added: * None Updated: * vkd3d-latest Where can you help * Run Steam/Battle.net/GOG/UPlay/Epic * Test your favorite game. * Test your favorite applications. * Improve staging patches and get them accepted upstream. * Suggest patches to be included in staging. As always, if you find a bug, please report it via https://bugs.winehq.org Best Regards Alistair.

1 0

[PATCH] selftests: ntsync: add wait termination edge cases
by Elizabeth Figura Sept. 1, 2026

Sept. 1, 2026

From: Iván Ezequiel Rodriguez <ivanrwcm25(a)gmail.com> The existing tests cover timeouts that expire while blocked, but not a deadline that has already passed, and no test exercises interruption by a signal even though the documented behaviour is EINTR. Also extend wait_args_validation() with an invalid flags case. Add coverage for: - an absolute timeout in the past, which must return ETIMEDOUT without blocking - a wait interrupted by a signal, which must return EINTR - undefined bits in ntsync_wait_args::flags, which must return EINVAL Tested on v7.3-rc1 under QEMU with CONFIG_NTSYNC=y: 15/15 pass, and the two new tests were run 50 times each at 1, 2, 4 and 8 vCPUs with no intermittent failures. Signed-off-by: Iván Ezequiel Rodriguez <ivanrwcm25(a)gmail.com> Reviewed-by: Elizabeth Figura <zfigura(a)codeweavers.com> Signed-off-by: Elizabeth Figura <zfigura(a)codeweavers.com> --- .../testing/selftests/drivers/ntsync/ntsync.c | 96 +++++++++++++++++++ 1 file changed, 96 insertions(+) diff --git a/tools/testing/selftests/drivers/ntsync/ntsync.c b/tools/testing/selftests/drivers/ntsync/ntsync.c index 1f0dc43bb4c0..7d944e2a9094 100644 --- a/tools/testing/selftests/drivers/ntsync/ntsync.c +++ b/tools/testing/selftests/drivers/ntsync/ntsync.c @@ -11,6 +11,7 @@ #include <sys/wait.h> #include <fcntl.h> #include <sched.h> +#include <signal.h> #include <time.h> #include <pthread.h> #include <linux/ntsync.h> @@ -1381,6 +1382,13 @@ TEST(wait_args_validation) EXPECT_EQ(-1, ret); EXPECT_EQ(EINVAL, errno); + wait_args.pad = 0; + wait_args.flags = ~NTSYNC_WAIT_REALTIME; + ret = ioctl(fd, NTSYNC_IOC_WAIT_ANY, &wait_args); + EXPECT_EQ(-1, ret); + EXPECT_EQ(EINVAL, errno); + wait_args.flags = 0; + ret = wait_any(fd2, 1, &sem, 123, &index); EXPECT_EQ(-1, ret); EXPECT_EQ(EINVAL, errno); @@ -1471,4 +1479,92 @@ TEST(wait_any_monotonic_timens) close(fd); } +TEST(wait_timeout_expired) +{ + struct ntsync_wait_args wait_args = {0}; + struct ntsync_sem_args sem_args = {0}; + struct timespec start, end; + __u64 elapsed_ns; + int fd, sem, ret; + + fd = open("/dev/ntsync", O_CLOEXEC | O_RDONLY); + ASSERT_LE(0, fd); + + sem_args.count = 0; + sem_args.max = 1; + sem = ioctl(fd, NTSYNC_IOC_CREATE_SEM, &sem_args); + EXPECT_LE(0, sem); + + wait_args.timeout = 0; + wait_args.objs = (uintptr_t)&sem; + wait_args.count = 1; + wait_args.owner = 123; + wait_args.index = 0xdeadbeef; + + clock_gettime(CLOCK_MONOTONIC, &start); + ret = ioctl(fd, NTSYNC_IOC_WAIT_ANY, &wait_args); + clock_gettime(CLOCK_MONOTONIC, &end); + EXPECT_EQ(-1, ret); + EXPECT_EQ(ETIMEDOUT, errno); + + elapsed_ns = (end.tv_sec - start.tv_sec) * 1000000000ULL + + (end.tv_nsec - start.tv_nsec); + EXPECT_LT(elapsed_ns, 50 * 1000000ULL); + + close(sem); + close(fd); +} + +static void ntsync_test_signal_handler(int sig) +{ +} + +TEST(wait_interrupted_by_signal) +{ + struct ntsync_wait_args wait_args = {0}; + struct ntsync_sem_args sem_args = {0}; + struct sigaction sa = {0}, old_sa; + struct wait_args thread_args; + int fd, sem, ret; + pthread_t thread; + + fd = open("/dev/ntsync", O_CLOEXEC | O_RDONLY); + ASSERT_LE(0, fd); + + sem_args.count = 0; + sem_args.max = 1; + sem = ioctl(fd, NTSYNC_IOC_CREATE_SEM, &sem_args); + EXPECT_LE(0, sem); + + sa.sa_handler = ntsync_test_signal_handler; + ret = sigaction(SIGUSR1, &sa, &old_sa); + EXPECT_EQ(0, ret); + + wait_args.timeout = get_abs_timeout(5000); + wait_args.objs = (uintptr_t)&sem; + wait_args.count = 1; + wait_args.owner = 123; + wait_args.index = 0xdeadbeef; + thread_args.fd = fd; + thread_args.args = &wait_args; + thread_args.request = NTSYNC_IOC_WAIT_ANY; + ret = pthread_create(&thread, NULL, wait_thread, &thread_args); + EXPECT_EQ(0, ret); + + ret = wait_for_thread(thread, 50); + EXPECT_EQ(ETIMEDOUT, ret); + + ret = pthread_kill(thread, SIGUSR1); + EXPECT_EQ(0, ret); + + ret = wait_for_thread(thread, 1000); + EXPECT_EQ(0, ret); + EXPECT_EQ(-1, thread_args.ret); + EXPECT_EQ(EINTR, thread_args.err); + + sigaction(SIGUSR1, &old_sa, NULL); + close(sem); + close(fd); +} + TEST_HARNESS_MAIN base-commit: cee9395acd8043be0644b25c34bfa86623f2b935 -- 2.53.0

1 0