On Fri, 1 Aug 2025 at 22:44, Richard Henderson <richard.hender...@linaro.org> wrote: > > On 8/2/25 03:45, Peter Maydell wrote: > > On Wed, 30 Jul 2025 at 01:24, Richard Henderson > > <richard.hender...@linaro.org> wrote: > >> > >> Signed-off-by: Richard Henderson <richard.hender...@linaro.org> > >> --- > >> linux-user/elfload.c | 41 +----------------------- > >> linux-user/x86_64/target_coredump.c.inc | 42 +++++++++++++++++++++++++ > >> 2 files changed, 43 insertions(+), 40 deletions(-) > >> create mode 100644 linux-user/x86_64/target_coredump.c.inc > >> > > > > Maybe we can come back to this at some point to convert the > > .c.inc into a proper .c / .h file. > > Are you thinking of something like ELF_NREG + target_elf_gregset_t in the > header and > elf_core_copy_regs in the c file? > > I suppose there's no reason not to do that now, putting the function in the > new > target/elfload.c, and the one or two declarations in target_elf.h.
I hadn't looked closely enough at the details to have a concrete idea; this was just my preference for avoiding .c.inc files unless we really need them (e.g. we're including the same file multiple times or it's just too painful to split a huge file any other way). If it's easy to put the function in the .c file and the declarations in the .h file then I think that's better than the .c.inc approach. -- PMM
