[xen master] build: force compiler to use atomics when coverage is enabled

[patchbot@xxxxxxx]

commit e9339514744904061ca01046e5ca6b97cb4f0fb2
Author:     Volodymyr Babchuk <volodymyr_babchuk@xxxxxxxx>
AuthorDate: Tue Jul 22 08:51:50 2025 +0200
Commit:     Jan Beulich <jbeulich@xxxxxxxx>
CommitDate: Tue Jul 22 08:51:50 2025 +0200

    build: force compiler to use atomics when coverage is enabled
    
    By default GCC uses "simple" coverage counter update
    mechanism. It is perfectly fine for single-threaded (or single CPU in
    our case) setups, but will cause race conditions on SMP systems.
    
    For example, I observed that counters are going backwards when running
    Xen inside QEMU.
    
    GCC starting from version 7 and LLVM/Clang starting from version 11
    support -fprofile-update=atomic option, which forces coverage counter
    updates to be atomic, which resolves the issue. As Xen runs mostly on
    SMP systems, force use this option if it is supported by a compiler.
    
    Signed-off-by: Volodymyr Babchuk <volodymyr_babchuk@xxxxxxxx>
    Acked-by: Jan Beulich <jbeulich@xxxxxxxx>
---
 xen/Rules.mk | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/xen/Rules.mk b/xen/Rules.mk
index da21850926..24f447b957 100644
--- a/xen/Rules.mk
+++ b/xen/Rules.mk
@@ -141,6 +141,9 @@ else
     cov-cflags-$(CONFIG_CONDITION_COVERAGE) += -fcondition-coverage
 endif
 
+# Ensure that profile/coverage data is updated atomically
+$(call cc-option-add,cov-cflags-$(CONFIG_COVERAGE),CC,-fprofile-update=atomic)
+
 # Reset cov-cflags-y in cases where an objects has another one as prerequisite
 $(nocov-y) $(filter %.init.o, $(obj-y) $(obj-bin-y) $(extra-y)): \
     cov-cflags-y :=
--
generated by git-patchbot for /home/xen/git/xen.git#master