Re: [PATCH sched_ext/for-7.2] sched_ext: Document the ops compat strategy in compat.h/compat.bpf.h
From: Andrea Righi
Date: Sun Apr 19 2026 - 12:38:06 EST
Hi Tejun,
On Sun, Apr 19, 2026 at 06:18:52AM -1000, Tejun Heo wrote:
> The comments around SCX_OPS_DEFINE() and SCX_OPS_OPEN() were vague about
> how backward compatibility actually works. Expand them to describe the two
> mechanisms: load-time BTF fix-up for additive changes, and multi-variant
> struct_ops for incompatible ones.
Thanks for clarifying these comments.
Acked-by: Andrea Righi <arighi@xxxxxxxxxx>
-Andrea
>
> Signed-off-by: Tejun Heo <tj@xxxxxxxxxx>
> ---
> tools/sched_ext/include/scx/compat.bpf.h | 6 ++++--
> tools/sched_ext/include/scx/compat.h | 23 +++++++++++++++++++----
> 2 files changed, 23 insertions(+), 6 deletions(-)
>
> --- a/tools/sched_ext/include/scx/compat.bpf.h
> +++ b/tools/sched_ext/include/scx/compat.bpf.h
> @@ -423,8 +423,10 @@ static inline void scx_bpf_dsq_reenq(u64
> }
>
> /*
> - * Define sched_ext_ops. This may be expanded to define multiple variants for
> - * backward compatibility. See compat.h::SCX_OPS_LOAD/ATTACH().
> + * Define sched_ext_ops. See compat.h::SCX_OPS_OPEN() for how backward
> + * compatibility is handled (this macro can be expanded to emit multiple
> + * variants for incompatible op changes; SCX_OPS_OPEN() handles purely
> + * additive changes at load time).
> */
> #define SCX_OPS_DEFINE(__name, ...) \
> SEC(".struct_ops.link") \
> --- a/tools/sched_ext/include/scx/compat.h
> +++ b/tools/sched_ext/include/scx/compat.h
> @@ -149,10 +149,24 @@ static inline long scx_hotplug_seq(void)
> }
>
> /*
> - * struct sched_ext_ops can change over time. If compat.bpf.h::SCX_OPS_DEFINE()
> - * is used to define ops and compat.h::SCX_OPS_LOAD/ATTACH() are used to load
> - * and attach it, backward compatibility is automatically maintained where
> - * reasonable.
> + * Open the sched_ext_ops skeleton.
> + *
> + * struct sched_ext_ops can change over time. Two complementary mechanisms
> + * keep BPF schedulers built against newer headers running on older kernels:
> + *
> + * 1. Load-time fix-up (this macro). For each optional ops callback or field
> + * added to struct sched_ext_ops, an explicit stanza below probes the
> + * running kernel's BTF via __COMPAT_struct_has_field() and, if the field
> + * is missing, clears it in the in-memory struct_ops (with a warning to
> + * stderr) before load. Handles additive changes - a new stanza must be
> + * added here for each new optional field.
> + *
> + * 2. Multi-variant struct_ops via compat.bpf.h::SCX_OPS_DEFINE(). That
> + * macro can be expanded to emit several variants of struct sched_ext_ops,
> + * and SCX_OPS_LOAD()/ATTACH() can pick the right one based on what the
> + * kernel supports. Needed when an existing operation has to change
> + * incompatibly (e.g. a callback signature changes); the load-time
> + * fix-up above only handles purely additive changes.
> *
> * ec7e3b0463e1 ("implement-ops") in https://github.com/sched-ext/sched_ext is
> * the current minimum required kernel version.
> @@ -225,6 +239,7 @@ static inline void __scx_ops_assoc_prog(
> }
> #endif
>
> +/* See SCX_OPS_OPEN() above for backward-compatibility handling. */
> #define SCX_OPS_LOAD(__skel, __ops_name, __scx_name, __uei_name) ({ \
> struct bpf_program *__prog; \
> UEI_SET_SIZE(__skel, __ops_name, __uei_name); \