Hi, On Thu, Feb 02, 2023 at 12:32:33PM +0100, Jiri Olsa wrote: > On Wed, Feb 01, 2023 at 02:21:16PM -0700, Nathan Chancellor wrote: > > SNIP > > > > > Should we do this and the hostprogs migration as a follow up? There > > > > isn't that much use of hostprogs in tools, but I like that your change > > > > will show HOSTCC rather than CC during compilation. If we use > > > > hostprogs can we just avoid the overrides altogether? > > > > > > right, I think so.. we can now remove it for the BINARY target and > > > then for the rest of 'prepare' once they are converted to hostprogs.. > > > not sure how 'hostprogs' will work for library, will need to check > > > > > > I'll try to send the format patch with the fix below tomorrow > > > > > > Nathan, any chance you could test it? > > > > > > Sure thing, would you happen to have a single patch file of what I > > should test? I am a little confused from reading the thread (unless what > > you want tested has not been sent yet). > > I meant the change below, but I posted the formal patch now: > https://lore.kernel.org/bpf/20230202112839.1131892-1-jolsa@xxxxxxxxxx/T/#u I've been chasing the links posted by Ian, but now I can't comment on patch "tools/resolve_btfids: Compile resolve_btfids as host program", because lkml wasn't CCed, and I'm not subscribed to the bpf ML. So I'm commenting here. Can someone please apply the patch linked above? Compiling with CONFIG_DEBUG_INFO_BTF=y is also broken on net-next. Maintainers can add my Tested-by: Vladimir Oltean <vladimir.oltean@xxxxxxx> on that patch if it helps speed up the process.
