Skip to contentSkip to conversation

toolchain/clang-native: Set BUILD_LDFLAGS instead of LDFLAGS

PATCH In review Khem Raj 4 participants 15 messages Aug 1, 2025 – Aug 11, 2025 RISC-V

lore
Share & tools
lore

Khem Raj

author
#0 lore raw

PATCH 1/7 toolchain/clang-native: Set BUILD_LDFLAGS instead of LDFLAGS

BUILD_LDFLAGS is used by other recipes to emit native LDFLAGS into their cmake toolchain files especially in QT layers. it will usually be inherited by recipes wanting clang provide BUILD_* variables so this is more appropriate than LDFLAGS to pass these options.

---
 meta/classes/toolchain/clang-native.bbclass | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
meta/classes/toolchain/clang-native.bbclass +1 -1
index 4de491a1cb5..006be9fadd6 100644
@@ -15,4 +15,4 @@ BUILD_READELF = "${BUILD_PREFIX}llvm-readelf"
 
 DEPENDS += "clang-native libcxx-native compiler-rt-native"
 
-LDFLAGS += " --rtlib=libgcc --unwindlib=libgcc"
+BUILD_LDFLAGS += " --rtlib=libgcc --unwindlib=libgcc"
author
#1 lore raw

PATCH 2/7 cargo-c: Update patches to latest versions

getrandom patch is accepted upstream with minor changes parking_lot patch has addressed upstream feedback

---
 ...Use-libc-SYS_futex_time64-on-riscv32.patch | 69 +++++++------------
 ...Use-libc-SYS_futex_time64-on-riscv32.patch | 65 +++++++----------
 2 files changed, 51 insertions(+), 83 deletions(-)
meta/recipes-devtools/rust/cargo-c/0001-getrandom-Use-libc-SYS_futex_time64-on-riscv32.patch +26 -43
index dc7130fb574..638832c467f 100644
@@ -1,53 +1,36 @@
 From 71c356a07fbbf1530cfc87960e975f93bc9007e8 Mon Sep 17 00:00:00 2001
 From: Khem Raj <[email protected]>
 Date: Tue, 22 Jul 2025 09:46:03 -0700
-Subject: [PATCH] Use libc::SYS_futex_time64 on riscv32
+Subject: [PATCH] Use getrandom syscall on riscv32/riscv64 linux
 
-On RISC-V 32-bit (riscv32), the SYS_futex system call is
-often handled indirectly due to the use of a 64-bit time_t
-type. Specifically, while SYS_futex is not directly defined,
-a related syscall like SYS_futex_time64 can be used,
+Minimum kernel needed on RISCV is fairly new (4.15+) so we are sure
+to have getrandom syscall, on glibc there is mimimal ABI kernel to denote
+it but musl does not have any other way to indicate it, so add it
+as a condition here to choose getrandom backend for rv32/rv64 on linux
+when using musl.
 
-Upstream-Status: Submitted [https://github.com/rust-random/getrandom/pull/698]
+Upstream-Status: Backport [https://github.com/rust-random/getrandom/pull/699]
 Signed-off-by: Khem Raj <[email protected]>
 ---
  src/backends/use_file.rs | 17 +++++++++++++++++
  1 file changed, 17 insertions(+)
 
-diff --git a/src/backends/use_file.rs b/src/backends/use_file.rs
-index 7b48d43..baa0c66 100644
---- a/src/backends/use_file.rs
-+++ b/src/backends/use_file.rs
-@@ -158,7 +158,18 @@ mod sync {
-     pub(super) fn wait() {
-         let op = libc::FUTEX_WAIT | libc::FUTEX_PRIVATE_FLAG;
-         let timeout_ptr = core::ptr::null::<libc::timespec>();
-+        #[cfg(not(target_arch = "riscv32"))]
-         let ret = unsafe { libc::syscall(libc::SYS_futex, &FD, op, FD_ONGOING_INIT, timeout_ptr) };
-+        #[cfg(target_arch = "riscv32")]
-+        let ret = unsafe {
-+            libc::syscall(
-+                libc::SYS_futex_time64,
-+                &FD,
-+                op,
-+                FD_ONGOING_INIT,
-+                timeout_ptr,
-+            )
-+        };
-         // FUTEX_WAIT should return either 0 or EAGAIN error
-         debug_assert!({
-             match ret {
-@@ -172,7 +183,13 @@ mod sync {
-     /// Wake up all threads which wait for value of atomic `FD` to change.
-     pub(super) fn wake() {
-         let op = libc::FUTEX_WAKE | libc::FUTEX_PRIVATE_FLAG;
-+
-+        #[cfg(not(target_arch = "riscv32"))]
-         let ret = unsafe { libc::syscall(libc::SYS_futex, &FD, op, libc::INT_MAX) };
-+
-+        #[cfg(target_arch = "riscv32")]
-+        let ret = unsafe { libc::syscall(libc::SYS_futex_time64, &FD, op, libc::INT_MAX) };
-+
-         debug_assert!(ret >= 0);
-     }
- 
+--- a/src/backends.rs
++++ b/src/backends.rs
+@@ -93,7 +93,15 @@ cfg_if! {
+                 // Minimum supported Linux kernel version for MUSL targets
+                 // is not specified explicitly (as of Rust 1.77) and they
+                 // are used in practice to target pre-3.17 kernels.
+-                target_env = "musl",
++                all(
++                    target_env = "musl",
++                    not(
++                        any(
++                            target_arch = "riscv64",
++                            target_arch = "riscv32",
++                        ),
++                    ),
++                ),
+             ),
+         )
+     ))] {
meta/recipes-devtools/rust/cargo-c/0001-parking-lot-Use-libc-SYS_futex_time64-on-riscv32.patch +25 -40

91 lines not on this page · raw message

author
#2 lore raw

PATCH 3/7 libxcrypt: Allow undefined symbols in version scripts with lld linker

In reply to Khem Raj

Unlike GNU ld, LLD defaults to erroring about undefined version symbols add commandline parameter to sush lld here

Fixes

| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'crypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_r' failed: symbol not defined
| x86_64-yoesdk-linux-clang: error: linker command failed with exit code 1 (use -v to see invocation)
---
 meta/recipes-core/libxcrypt/libxcrypt.inc | 2 ++
 1 file changed, 2 insertions(+)
meta/recipes-core/libxcrypt/libxcrypt.inc +2 -0
index 77fec832348..10f6cd921d8 100644
@@ -24,3 +24,5 @@ API = "--disable-obsolete-api"
 EXTRA_OECONF += "${API}"
 
 BBCLASSEXTEND = "native nativesdk"
+
+LDFLAGS:append:toolchain-clang = "${@bb.utils.contains('DISTRO_FEATURES', 'ld-is-lld', ' -Wl,--undefined-version', '', d)}"
#3 lore raw

Re: [OE-core] [PATCH 3/7] libxcrypt: Allow undefined symbols in version scripts with lld linker

Quoted 5 lines · Khem Raj via lists.openembedded.org

Unlike GNU ld, LLD defaults to erroring about undefined version symbols add commandline parameter to sush lld here

Fixes

Quoted 8 lines
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'crypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_r' failed: symbol not defined
| x86_64-yoesdk-linux-clang: error: linker command failed with exit code 1 (use -v to see invocation)

+LDFLAGS:append:toolchain-clang = "${@bb.utils.contains('DISTRO_FEATURES', 'ld-is-lld', ' -Wl,--undefined-version', '', d)}"

Wait. Why is there an undefined symbol error in the first place? Shouldn't we address the error, rather than silence it?

The same question stands for similar patches for libtirpc, gzip and binutils-cross-canadian.

Alex

author
#4 lore raw

Re: [OE-core] [PATCH 3/7] libxcrypt: Allow undefined symbols in version scripts with lld linker

Quoted 15 lines · Alexander Kanavin
Quoted 11 lines · Khem Raj via lists.openembedded.org

Unlike GNU ld, LLD defaults to erroring about undefined version symbols add commandline parameter to sush lld here

Fixes

| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'crypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_gensalt_r' failed: symbol not defined
| x86_64-yoesdk-linux-ld.lld: error: version script assignment of 'XCRYPT_2.0' to symbol 'xcrypt_r' failed: symbol not defined
| x86_64-yoesdk-linux-clang: error: linker command failed with exit code 1 (use -v to see invocation)
Quoted 4 lines
Quoted 1 line

+LDFLAGS:append:toolchain-clang = "${@bb.utils.contains('DISTRO_FEATURES', 'ld-is-lld', ' -Wl,--undefined-version', '', d)}"

Wait. Why is there an undefined symbol error in the first place? Shouldn't we address the error, rather than silence it?

LLD has stricter defaults and strict symbol versioning enforcement. LLD requires that all symbols
referenced in version scripts actually exist in the object files being linked Same errors/behavior can be
seen with GNU linker as well if we add -Wl,--undefined-version to LDFLAGS GNU linker chooses to be a bit lenient with defaults. It is a common pattern with new vs old tools.

libxcrypt uses version scripts to manage symbol versioning, but LLD requires that all symbols referenced
in version scripts actually exist in the object files being linked. The symbols might be:

- Conditionally compiled out based on configuration
- Defined with different names/prefixes
- Missing due to build configuration issues

Having LLD to lower its barrier does not change the output it produces, it's the same as GNU ld,
There is no difference - both .so files work identically at runtime since the symbols are exported into .dynstr sections so old binaries needing these versioned symbols can still
link to them at runtime.

Should the version script be fixed ? maybe, I am not sure where all libxcypt is used in legacy systems
to qualify this. Current assumption is to have these symbols. This patch makes LLD behave like GNU LD
does.

Hope this helps you understand what's going on. Feel free to ask if you have more doubts.

Quoted 3 lines

The same question stands for similar patches for libtirpc, gzip and binutils-cross-canadian.

They all have the same issue underneath.

Quoted 2 lines

Alex

#5 lore raw

Re: [OE-core] [PATCH 3/7] libxcrypt: Allow undefined symbols in version scripts with lld linker

Quoted 8 lines · Khem Raj

Hope this helps you understand what's going on. Feel free to ask if you have more doubts.

Quoted 3 lines

The same question stands for similar patches for libtirpc, gzip and binutils-cross-canadian.

They all have the same issue underneath.

The way you describe it, this at least merits upstream tickets, if not actual fixes. Upstreams can then decide if lld is something they care about.

Alex

author
#6 lore raw

Re: [OE-core] [PATCH 3/7] libxcrypt: Allow undefined symbols in version scripts with lld linker

Quoted 7 lines · Alexander Kanavin
Quoted 4 lines · Khem Raj

Hope this helps you understand what's going on. Feel free to ask if you have more doubts.

Quoted 1 line
Quoted 9 lines
Quoted 4 lines
Quoted 2 lines

The same question stands for similar patches for libtirpc, gzip and binutils-cross-canadian.

They all have the same issue underneath.

The way you describe it, this at least merits upstream tickets, if not actual fixes. Upstreams can then decide if lld is something they care about.

Its in plan.
I have reported it for libxcrypt.

https://github.com/besser82/libxcrypt/issues/213

Others are also in works.

Quoted 1 line

Alex

author
#7 lore raw

PATCH 4/7 libtirpc: Allow undefined symbols in version scripts with lld linker

In reply to Khem Raj

Fixes link errors with lld

aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol '_svcauth_gss' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'authgss_create' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'authgss_create_default' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'authgss_free_private_data' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'authgss_get_private_data' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'authgss_service' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'cbc_crypt' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'ecb_crypt' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'endrpcent' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'getrpcent' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'getrpcbynumber' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'getrpcbyname' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'gss_log_debug' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'gss_log_hexdump' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'gss_log_status' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'rpc_gss_get_error' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'rpc_gss_get_mech_info' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'rpc_gss_get_mechanisms' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'rpc_gss_get_principal_name' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'TIRPC_0.3.0' to symbol 'rpc_gss_get_versions' failed: symbol not defined

---
 meta/recipes-extended/libtirpc/libtirpc_1.3.6.bb | 2 ++
 1 file changed, 2 insertions(+)
meta/recipes-extended/libtirpc/libtirpc_1.3.6.bb +2 -0
index 6ea9a725dbf..98bb1128c00 100644
@@ -27,6 +27,8 @@ PACKAGECONFIG ??= "\
 PACKAGECONFIG[ipv6] = "--enable-ipv6,--disable-ipv6"
 PACKAGECONFIG[gssapi] = "--enable-gssapi,--disable-gssapi,krb5"
 
+LDFLAGS:append:toolchain-clang = "${@bb.utils.contains('DISTRO_FEATURES', 'ld-is-lld', ' -Wl,--undefined-version', '', d)}"
+
 do_install:append() {
 	test -e ${D}${sysconfdir}/netconfig && chown root:root ${D}${sysconfdir}/netconfig
 }
author
#8 lore raw

PATCH 5/7 gzip: Always use GNU ld for linking

In reply to Khem Raj

Linking does not yet work with LLD

Fixes
aarch64-yoesdk-linux-ld.lld: error: undefined symbol: gzopen

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

referenced by minigzip.c
minigzip.o:(main)

---
 meta/recipes-core/zlib/zlib_1.3.1.bb | 1 +
 1 file changed, 1 insertion(+)
meta/recipes-core/zlib/zlib_1.3.1.bb +1 -0
index e62c50c5df7..96b7a5821ff 100644
@@ -20,6 +20,7 @@ SRC_URI[sha256sum] = "9a93b2b7dfdac77ceba5a558a580e74667dd6fede4585b91eefb60f03b
 PREMIRRORS:append = " https://zlib.net/ https://zlib.net/fossils/"
 
 CFLAGS += "-D_REENTRANT -fPIE"
+LDFLAGS:append = " -fuse-ld=bfd"
 
 RDEPENDS:${PN}-ptest += "make"
 
#9 lore raw

Re: [OE-core] [PATCH 5/7] gzip: Always use GNU ld for linking

Quoted 6 lines · Khem Raj via lists.openembedded.org

Linking does not yet work with LLD

Fixes
aarch64-yoesdk-linux-ld.lld: error: undefined symbol: gzopen

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

referenced by minigzip.c
minigzip.o:(main)

Do we need minigzip?
Does upstream know about the bug?
../Randy

Quoted 26 lines
---
  meta/recipes-core/zlib/zlib_1.3.1.bb | 1 +
  1 file changed, 1 insertion(+)
meta/recipes-core/zlib/zlib_1.3.1.bb +1 -1
index e62c50c5df7..96b7a5821ff 100644
@@ -20,6 +20,7 @@ SRC_URI[sha256sum] = "9a93b2b7dfdac77ceba5a558a580e74667dd6fede4585b91eefb60f03b
  PREMIRRORS:append =" https://zlib.net/ https://zlib.net/fossils/"
 
  CFLAGS += "-D_REENTRANT -fPIE"
+LDFLAGS:append = " -fuse-ld=bfd"
 
  RDEPENDS:${PN}-ptest += "make"
 
 
-=-=-=-=-=-=-=-=-=-=-=-

Links: You receive all messages sent to this group.
View/Reply Online (#221217):https://lists.openembedded.org/g/openembedded-core/message/221217 Mute This Topic:https://lists.openembedded.org/mt/114479323/3616765
Group Owner:[email protected] Unsubscribe:https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Signature · 2 lines
# Randy MacLeod
# Wind River Linux
author
#10 lore raw

PATCH 6/7 binutils-cross-canadian: Always use GNU linker

In reply to Khem Raj

LLD fails to link gprofng
aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'dlclose' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'dlopen' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'fclose' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'fdopen' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'fgetpos' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'fopen' failed: symbol not defined aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'fsetpos' failed: symbol not defined

---
 meta/recipes-devtools/binutils/binutils-cross-canadian.inc | 5 +++++
 1 file changed, 5 insertions(+)
meta/recipes-devtools/binutils/binutils-cross-canadian.inc +5 -0
index c1fda1a4658..7d7ffe92ce1 100644
@@ -13,6 +13,11 @@ EXTRA_OECONF += "--with-sysroot=${SDKPATH}/sysroots/${TUNE_PKGARCH}${TARGET_VEND
 # e.g. we switch between different machines with different tunes.
 EXTRA_OECONF[vardepsexclude] = "TUNE_PKGARCH"
 
+# lld complains
+# aarch64-yoesdk-linux-ld.lld: error: version script assignment of 'GLIBC_2.0' to symbol 'dlclose' failed: symbol not defined
+
+LDFLAGS:append:toolchain-clang = " -fuse-ld=bfd"
+
 do_install () {
 	autotools_do_install
 
author
#11 lore raw

PATCH 7/7 fmt: Fix build with clang-21

In reply to Khem Raj

Backport a fix from upstream master branch

---
 ...compilation-on-clang-21-libc-21-4477.patch | 46 +++++++++++++++++++
 meta/recipes-devtools/fmt/fmt_11.2.0.bb       |  4 +-
 2 files changed, 49 insertions(+), 1 deletion(-)
 create mode 100644 meta/recipes-devtools/fmt/fmt/0001-Fix-compilation-on-clang-21-libc-21-4477.patch
meta/recipes-devtools/fmt/fmt/0001-Fix-compilation-on-clang-21-libc-21-4477.patch new file +46 -0
new file mode 100644
index 00000000000..911fa90a40c
@@ -0,0 +1,46 @@
+From f4345467fce7edbc6b36c3fa1cf197a67be617e2 Mon Sep 17 00:00:00 2001
+From: Remy Jette <[email protected]>
+Date: Sat, 21 Jun 2025 07:28:14 -0700
+Subject: [PATCH] Fix compilation on clang-21 / libc++-21 (#4477)
+
+`<cstdlib>` was not being included, so malloc and free were only declared
+via transitive includes. Some includes changed in the latest libc++-21
+build which broke fmt.
+
+Also changed `malloc`/`free` to `std::malloc` and `std::free`, as
+putting those symbols in the global namespace is optional for the
+implementation when including `<cstdlib>`.
+
+Upstream-Status: Backport [https://github.com/fmtlib/fmt/pull/4477]
+Signed-off-by: Khem Raj <[email protected]>
+---
+ include/fmt/format.h | 5 +++--
+ 1 file changed, 3 insertions(+), 2 deletions(-)
+
+Index: fmt-11.2.0/include/fmt/format.h
+===================================================================
+--- fmt-11.2.0.orig/include/fmt/format.h
++++ fmt-11.2.0/include/fmt/format.h
+@@ -44,6 +44,7 @@
+ #  include <cmath>    // std::signbit
+ #  include <cstddef>  // std::byte
+ #  include <cstdint>  // uint32_t
++#  include <cstdlib>  // std::malloc, std::free
+ #  include <cstring>  // std::memcpy
+ #  include <limits>   // std::numeric_limits
+ #  include <new>      // std::bad_alloc
+@@ -744,12 +745,12 @@ template <typename T> struct allocator {
+ 
+   T* allocate(size_t n) {
+     FMT_ASSERT(n <= max_value<size_t>() / sizeof(T), "");
+-    T* p = static_cast<T*>(malloc(n * sizeof(T)));
++    T* p = static_cast<T*>(std::malloc(n * sizeof(T)));
+     if (!p) FMT_THROW(std::bad_alloc());
+     return p;
+   }
+ 
+-  void deallocate(T* p, size_t) { free(p); }
++  void deallocate(T* p, size_t) { std::free(p); }
+ };
+ 
+ }  // namespace detail
meta/recipes-devtools/fmt/fmt_11.2.0.bb +3 -1
index d0d4cea23e2..fd5dc0c4ab7 100644
@@ -5,7 +5,9 @@ LICENSE = "MIT"
 LIC_FILES_CHKSUM = "file://LICENSE;md5=b9257785fc4f3803a4b71b76c1412729"
 
 SRC_URI = "git://github.com/fmtlib/fmt;branch=master;protocol=https;tag=${PV}\
-           file://0001-Workaround-an-ABI-issue-in-spdlog.patch"
+           file://0001-Workaround-an-ABI-issue-in-spdlog.patch \
+           file://0001-Fix-compilation-on-clang-21-libc-21-4477.patch \
+           "
 SRCREV = "40626af88bd7df9a5fb80be7b25ac85b122d6c21"
 
 inherit cmake
#12 lore raw
In reply to Khem Raj
Quoted 8 lines · Khem Raj via lists.openembedded.org

BUILD_LDFLAGS is used by other recipes to emit native LDFLAGS into their cmake toolchain files especially in QT layers. it will usually be inherited by recipes wanting clang provide BUILD_* variables so this is more appropriate than LDFLAGS to pass these options.

---

Hi Khem,

It looks like something in there is breaking the reproducibility:

AssertionError: The following deb packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/deb/./core2-64/clang-dev_20.1.8-r0_amd64.deb The following ipk packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/ipk/./core2-64/clang-dev_20.1.8-r0_core2-64.ipk The following rpm packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/rpm/./core2_64/clang-dev-20.1.8-r0.core2_64.rpm

https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/2205
https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20250801-t697t4fj/packages/diff-html/

Can you have a look at this?

Signature · 3 lines
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
#13 lore raw
Quoted 22 lines · Mathieu Dubois-Briand
Quoted 8 lines · Khem Raj via lists.openembedded.org

BUILD_LDFLAGS is used by other recipes to emit native LDFLAGS into their cmake toolchain files especially in QT layers. it will usually be inherited by recipes wanting clang provide BUILD_* variables so this is more appropriate than LDFLAGS to pass these options.

---

Hi Khem,

It looks like something in there is breaking the reproducibility:

AssertionError: The following deb packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/deb/./core2-64/clang-dev_20.1.8-r0_amd64.deb The following ipk packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/ipk/./core2-64/clang-dev_20.1.8-r0_core2-64.ipk The following rpm packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/rpm/./core2_64/clang-dev-20.1.8-r0.core2_64.rpm

Quoted 4 lines

https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/2205
https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20250801-t697t4fj/packages/diff-html/

Can you have a look at this?

I did a rebuild of this one and is succeed. So either we have an intermittent fail or some cache issue, not really sure...

https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/2210

Signature · 3 lines
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
author
#14 lore raw
Quoted 24 lines · Mathieu Dubois-Briand
Quoted 22 lines · Mathieu Dubois-Briand
Quoted 8 lines · Khem Raj via lists.openembedded.org

BUILD_LDFLAGS is used by other recipes to emit native LDFLAGS into their cmake toolchain files especially in QT layers. it will usually be inherited by recipes wanting clang provide BUILD_* variables so this is more appropriate than LDFLAGS to pass these options.

---

Hi Khem,

It looks like something in there is breaking the reproducibility:

AssertionError: The following deb packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/deb/./core2-64/clang-dev_20.1.8-r0_amd64.deb The following ipk packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/ipk/./core2-64/clang-dev_20.1.8-r0_core2-64.ipk The following rpm packages are different and not in exclusion list:
/srv/pokybuild/yocto-worker/reproducible/build/build-st/reproducibleB-extended/tmp/deploy/rpm/./core2_64/clang-dev-20.1.8-r0.core2_64.rpm

Quoted 8 lines
Quoted 4 lines

https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/2205
https://valkyrie.yocto.io/pub/repro-fail/oe-reproducible-20250801-t697t4fj/packages/diff-html/

Can you have a look at this?

I did a rebuild of this one and is succeed. So either we have an intermittent fail or some cache issue, not really sure...

Possible, I was wondering how build linker flags would insert into target builds, it's quite unlikely
but still a case. I have dumped the binaries ( strings ) and tried to catch absolute build path and
did not catch it so far.

Quoted 8 lines

https://autobuilder.yoctoproject.org/valkyrie/#/builders/37/builds/2210

--
Mathieu Dubois-Briand, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com