On 6/24/19 3:43 PM, Jakub Kicinski wrote: > On Mon, 24 Jun 2019 22:16:02 +0000, Andrey Ignatov wrote: >> Jakub Kicinski <jakub.kicin...@netronome.com> [Mon, 2019-06-24 14:51 -0700]: >>> This is a cgroup-specific flag, right? It should be a parameter >>> to cgroup show, not a global flag. Can we please drop this patch >>> from the tree? >> >> Hey Jakub, >> >> I had same thought about cgroup-specific flag while reviewing the patch, >> but then found out that all flags in bpftool are now global, no mater if >> they're sub-command-specific or not. >> >> For example, --mapcompat is used only in prog-subcommand, but the option >> is global; --bpffs is used in prog- and map-subcommands, but the option >> is global as well, etc (there are more examples). > > I don't think these are equivalent. BPF_F_QUERY_EFFECTIVE is a flag > for a syscall corresponding to a subcommand quite clearly. > >> I agree that limiting the scope of an option is a good idea in the long >> term and it'd be great to rework all existing options to be available >> only for corresponding sub-commands, but I don't see how the new `-e` >> options is different from existing options and why it should be dropped. > > Agreed, TBH, but we can't change existing options, people may be using > them. Let's drop the patch and make sure we're not making this mistake > again :)
I don't think this patch should be penalized. I'd rather see we fix them all.
