* [PATCH-for-10.2 0/2] tcg: Fix TCI build on macOS @ 2025-12-02 18:46 Philippe Mathieu-Daudé 2025-12-02 18:46 ` [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning Philippe Mathieu-Daudé 2025-12-02 18:46 ` [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support Philippe Mathieu-Daudé 0 siblings, 2 replies; 6+ messages in thread From: Philippe Mathieu-Daudé @ 2025-12-02 18:46 UTC (permalink / raw) To: qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau, Richard Henderson, Philippe Mathieu-Daudé We don't test TCG+TCI on macOS host and it is broken. Luckyly the fix is easy. Philippe Mathieu-Daudé (2): tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning tcg/tci: Disable Int128 support meson.build | 2 +- include/tcg/helper-info.h | 3 +++ 2 files changed, 4 insertions(+), 1 deletion(-) -- 2.51.0 ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning 2025-12-02 18:46 [PATCH-for-10.2 0/2] tcg: Fix TCI build on macOS Philippe Mathieu-Daudé @ 2025-12-02 18:46 ` Philippe Mathieu-Daudé 2025-12-03 0:34 ` Richard Henderson 2025-12-02 18:46 ` [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support Philippe Mathieu-Daudé 1 sibling, 1 reply; 6+ messages in thread From: Philippe Mathieu-Daudé @ 2025-12-02 18:46 UTC (permalink / raw) To: qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau, Richard Henderson, Philippe Mathieu-Daudé Since we build TCI with FFI (commit 22f15579fa1 "tcg: Build ffi data structures for helpers") we get on Darwin: In file included from ../../tcg/tci.c:22: In file included from include/tcg/helper-info.h:13: /Library/Developer/CommandLineTools/SDKs/MacOSX15.sdk/usr/include/ffi/ffi.h:483:5: warning: 'FFI_GO_CLOSURES' is not defined, evaluates to 0 [-Wundef] 483 | #if FFI_GO_CLOSURES | ^ 1 warning generated. Since this libffi change was committed more than 10 years ago (see https://github.com/libffi/libffi/commit/e951d64c0852), just define the missing definition on QEMU to silence the warning. Reported-by: Stefan Weil <sw@weilnetz.de> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> --- include/tcg/helper-info.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/include/tcg/helper-info.h b/include/tcg/helper-info.h index 909fe73afa3..36bb7f8187f 100644 --- a/include/tcg/helper-info.h +++ b/include/tcg/helper-info.h @@ -10,6 +10,9 @@ #define TCG_HELPER_INFO_H #ifdef CONFIG_TCG_INTERPRETER +#ifndef FFI_GO_CLOSURES +#define FFI_GO_CLOSURES 0 +#endif #include <ffi.h> #endif #include "tcg-target-reg-bits.h" -- 2.51.0 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning 2025-12-02 18:46 ` [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning Philippe Mathieu-Daudé @ 2025-12-03 0:34 ` Richard Henderson 0 siblings, 0 replies; 6+ messages in thread From: Richard Henderson @ 2025-12-03 0:34 UTC (permalink / raw) To: Philippe Mathieu-Daudé, qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau On 12/2/25 10:46, Philippe Mathieu-Daudé wrote: > Since we build TCI with FFI (commit 22f15579fa1 "tcg: Build ffi data > structures for helpers") we get on Darwin: > > In file included from ../../tcg/tci.c:22: > In file included from include/tcg/helper-info.h:13: > /Library/Developer/CommandLineTools/SDKs/MacOSX15.sdk/usr/include/ffi/ffi.h:483:5: warning: 'FFI_GO_CLOSURES' is not defined, evaluates to 0 [-Wundef] > 483 | #if FFI_GO_CLOSURES > | ^ > 1 warning generated. > > Since this libffi change was committed more than 10 years ago (see > https://github.com/libffi/libffi/commit/e951d64c0852), just define > the missing definition on QEMU to silence the warning. See also c23e9a1c from 2023, which fixed that to use an ifdef. I suppose the inclusion or exclusion of the go interface is neither here nor there, so long as it compiles. r~ ^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support 2025-12-02 18:46 [PATCH-for-10.2 0/2] tcg: Fix TCI build on macOS Philippe Mathieu-Daudé 2025-12-02 18:46 ` [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning Philippe Mathieu-Daudé @ 2025-12-02 18:46 ` Philippe Mathieu-Daudé 2025-12-02 22:30 ` Richard Henderson 1 sibling, 1 reply; 6+ messages in thread From: Philippe Mathieu-Daudé @ 2025-12-02 18:46 UTC (permalink / raw) To: qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau, Richard Henderson, Philippe Mathieu-Daudé Disable Int128 on TCI, otherwise build fails: In file included from ../../tcg/optimize.c:29: In file included from include/tcg/tcg-op-common.h:12: In file included from include/exec/helper-proto-common.h:10: In file included from include/qemu/atomic128.h:62: In file included from host/include/aarch64/host/atomic128-cas.h.inc:16: host/include/generic/host/atomic128-cas.h.inc:37:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') 37 | Int128 old = *ptr_align; | ^ ~~~~~~~~~~ host/include/generic/host/atomic128-cas.h.inc:39:52: warning: incompatible pointer types passing 'Int128 *' (aka 'struct Int128 *') to parameter of type '__int128_t *' (aka '__int128 *') [-Wincompatible-pointer-types] 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, | ^~~~ host/include/generic/host/atomic128-cas.h.inc:39:58: error: passing 'Int128' (aka 'struct Int128') to parameter of incompatible type '__int128_t' (aka '__int128') 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, | ^~~ host/include/generic/host/atomic128-cas.h.inc:50:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') 50 | Int128 old = *ptr_align; | ^ ~~~~~~~~~~ host/include/generic/host/atomic128-cas.h.inc:52:62: error: invalid operands to binary expression ('Int128' (aka 'struct Int128') and 'Int128') 52 | while (!__atomic_compare_exchange_n(ptr_align, &old, old & val, true, | ~~~ ^ ~~~ host/include/generic/host/atomic128-cas.h.inc:63:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') 63 | Int128 old = *ptr_align; | ^ ~~~~~~~~~~ host/include/generic/host/atomic128-cas.h.inc:65:62: error: invalid operands to binary expression ('Int128' (aka 'struct Int128') and 'Int128') 65 | while (!__atomic_compare_exchange_n(ptr_align, &old, old | val, true, | ~~~ ^ ~~~ 1 warning and 6 errors generated. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> --- meson.build | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/meson.build b/meson.build index d9293294d8e..692a62e6c18 100644 --- a/meson.build +++ b/meson.build @@ -2996,7 +2996,7 @@ config_host_data.set('CONFIG_ATOMIC64', cc.links(''' # has_int128_type is set to false on Emscripten to avoid errors by libffi # during runtime. -has_int128_type = host_os != 'emscripten' and cc.compiles(''' +has_int128_type = host_os != 'emscripten' and tcg_arch != 'tci' and cc.compiles(''' __int128_t a; __uint128_t b; int main(void) { b = a; }''') -- 2.51.0 ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support 2025-12-02 18:46 ` [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support Philippe Mathieu-Daudé @ 2025-12-02 22:30 ` Richard Henderson 2025-12-03 13:18 ` Richard Henderson 0 siblings, 1 reply; 6+ messages in thread From: Richard Henderson @ 2025-12-02 22:30 UTC (permalink / raw) To: Philippe Mathieu-Daudé, qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau On 12/2/25 10:46, Philippe Mathieu-Daudé wrote: > Disable Int128 on TCI, otherwise build fails: > > In file included from ../../tcg/optimize.c:29: > In file included from include/tcg/tcg-op-common.h:12: > In file included from include/exec/helper-proto-common.h:10: > In file included from include/qemu/atomic128.h:62: > In file included from host/include/aarch64/host/atomic128-cas.h.inc:16: > host/include/generic/host/atomic128-cas.h.inc:37:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') > 37 | Int128 old = *ptr_align; > | ^ ~~~~~~~~~~ > host/include/generic/host/atomic128-cas.h.inc:39:52: warning: incompatible pointer types passing 'Int128 *' (aka 'struct Int128 *') to parameter of type '__int128_t *' (aka '__int128 *') [-Wincompatible-pointer-types] > 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, > | ^~~~ > host/include/generic/host/atomic128-cas.h.inc:39:58: error: passing 'Int128' (aka 'struct Int128') to parameter of incompatible type '__int128_t' (aka '__int128') > 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, > | ^~~ > host/include/generic/host/atomic128-cas.h.inc:50:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') > 50 | Int128 old = *ptr_align; > | ^ ~~~~~~~~~~ > host/include/generic/host/atomic128-cas.h.inc:52:62: error: invalid operands to binary expression ('Int128' (aka 'struct Int128') and 'Int128') > 52 | while (!__atomic_compare_exchange_n(ptr_align, &old, old & val, true, > | ~~~ ^ ~~~ > host/include/generic/host/atomic128-cas.h.inc:63:12: error: initializing 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka '__int128') > 63 | Int128 old = *ptr_align; > | ^ ~~~~~~~~~~ > host/include/generic/host/atomic128-cas.h.inc:65:62: error: invalid operands to binary expression ('Int128' (aka 'struct Int128') and 'Int128') > 65 | while (!__atomic_compare_exchange_n(ptr_align, &old, old | val, true, > | ~~~ ^ ~~~ > 1 warning and 6 errors generated. > > Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> And what of the CONFIG_TCG_INTERPRETER test that already exists in int128.h? There's got to be more to this than that. r~ > --- > meson.build | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/meson.build b/meson.build > index d9293294d8e..692a62e6c18 100644 > --- a/meson.build > +++ b/meson.build > @@ -2996,7 +2996,7 @@ config_host_data.set('CONFIG_ATOMIC64', cc.links(''' > > # has_int128_type is set to false on Emscripten to avoid errors by libffi > # during runtime. > -has_int128_type = host_os != 'emscripten' and cc.compiles(''' > +has_int128_type = host_os != 'emscripten' and tcg_arch != 'tci' and cc.compiles(''' > __int128_t a; > __uint128_t b; > int main(void) { b = a; }''') ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support 2025-12-02 22:30 ` Richard Henderson @ 2025-12-03 13:18 ` Richard Henderson 0 siblings, 0 replies; 6+ messages in thread From: Richard Henderson @ 2025-12-03 13:18 UTC (permalink / raw) To: Philippe Mathieu-Daudé, qemu-devel Cc: Daniel P. Berrangé, Paolo Bonzini, Stefan Weil, Marc-André Lureau On 12/2/25 14:30, Richard Henderson wrote: > On 12/2/25 10:46, Philippe Mathieu-Daudé wrote: >> Disable Int128 on TCI, otherwise build fails: >> >> In file included from ../../tcg/optimize.c:29: >> In file included from include/tcg/tcg-op-common.h:12: >> In file included from include/exec/helper-proto-common.h:10: >> In file included from include/qemu/atomic128.h:62: >> In file included from host/include/aarch64/host/atomic128-cas.h.inc:16: >> host/include/generic/host/atomic128-cas.h.inc:37:12: error: initializing >> 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka >> '__int128') >> 37 | Int128 old = *ptr_align; >> | ^ ~~~~~~~~~~ >> host/include/generic/host/atomic128-cas.h.inc:39:52: warning: incompatible pointer >> types passing 'Int128 *' (aka 'struct Int128 *') to parameter of type '__int128_t >> *' (aka '__int128 *') [-Wincompatible-pointer-types] >> 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, >> | ^~~~ >> host/include/generic/host/atomic128-cas.h.inc:39:58: error: passing 'Int128' (aka >> 'struct Int128') to parameter of incompatible type '__int128_t' (aka '__int128') >> 39 | while (!__atomic_compare_exchange_n(ptr_align, &old, new, true, >> | ^~~ >> host/include/generic/host/atomic128-cas.h.inc:50:12: error: initializing >> 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka >> '__int128') >> 50 | Int128 old = *ptr_align; >> | ^ ~~~~~~~~~~ >> host/include/generic/host/atomic128-cas.h.inc:52:62: error: invalid operands to >> binary expression ('Int128' (aka 'struct Int128') and 'Int128') >> 52 | while (!__atomic_compare_exchange_n(ptr_align, &old, old & val, true, >> | ~~~ ^ ~~~ >> host/include/generic/host/atomic128-cas.h.inc:63:12: error: initializing >> 'Int128' (aka 'struct Int128') with an expression of incompatible type '__int128_t' (aka >> '__int128') >> 63 | Int128 old = *ptr_align; >> | ^ ~~~~~~~~~~ >> host/include/generic/host/atomic128-cas.h.inc:65:62: error: invalid operands to >> binary expression ('Int128' (aka 'struct Int128') and 'Int128') >> 65 | while (!__atomic_compare_exchange_n(ptr_align, &old, old | val, true, >> | ~~~ ^ ~~~ >> 1 warning and 6 errors generated. >> >> Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> > > And what of the CONFIG_TCG_INTERPRETER test that already exists in int128.h? > > There's got to be more to this than that. If you disable Int128 entirely, then we can't use host/include/generic/host/atomic128-cas.h.inc at all. The bug is within that header, were we need to use Int128Alias more, as we already do for atomic16_cmpxchg. I'll post a fix. r~ ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2025-12-03 13:18 UTC | newest] Thread overview: 6+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2025-12-02 18:46 [PATCH-for-10.2 0/2] tcg: Fix TCI build on macOS Philippe Mathieu-Daudé 2025-12-02 18:46 ` [PATCH-for-10.2? 1/2] tcg/tci: Disable -Wundef FFI_GO_CLOSURES warning Philippe Mathieu-Daudé 2025-12-03 0:34 ` Richard Henderson 2025-12-02 18:46 ` [PATCH-for-10.2 2/2] tcg/tci: Disable Int128 support Philippe Mathieu-Daudé 2025-12-02 22:30 ` Richard Henderson 2025-12-03 13:18 ` Richard Henderson
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).