https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120182
Richard Biener <rguenth at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Ever confirmed|0 |1 Last reconfirmed| |2025-05-09 Component|tree-optimization |rtl-optimization Status|UNCONFIRMED |NEW --- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> --- It looks like an rtl optimization issue: call mmap@PLT cmpq $2654208, %rax jne .L2 movb $0, 2654208 movq 2654208, %rax testq %rax, %rax je .L3 movq $256, (%rax) .L3: movq $256, 2654208 movb $0, 2654208 movq 2654208, %rax testq %rax, %rax je .L4 movq $256, (%rax) .L4: movq $256, 2654208 movb $0, 2654208 movb $0, 2654209 movb $0, 2654210 so all the zeroing besides the first byte is sunk somehow.