On Mon, 27 Sept 2021 at 15:01, Jakub Sitnicki <jakub@xxxxxxxxxxxxxx> wrote: > > I find exposing stats via system configuration variables a bit > unexpected. Not sure if there is any example today that we're following. > > Maybe an entry under /sys/kernel/debug would be a better fit? > > That way we don't have to commit to a sysctl that might go away if we > start charging JIT allocs against memory cgroup quota. I had a look around, there are no other obvious places in debugfs or proc where we already have bpf info exposed. It currently all goes via sysctl. There are examples of readonly sysctls: $ sudo find /proc/sys -perm 0444 | wc -l 90 There are no examples of sysctls with mode 0400 however: $ sudo find /proc/sys -perm 0400 | wc -l 0 I find it kind of weird that the bpf sysctls are so tightly locked down (CAP_SYS_ADMIN && root) even for reading. Maybe something I can change? -- Lorenz Bauer | Systems Engineer 6th Floor, County Hall/The Riverside Building, SE1 7PB, UK www.cloudflare.com
