Re: [PATCH v3 bpf-next 5/7] x86/alternative: introduce text_poke_jit

From: Peter Zijlstra
Date: Tue Jan 11 2022 - 07:14:14 EST


On Wed, Jan 05, 2022 at 06:25:31PM -0800, Song Liu wrote:

> diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
> index 23fb4d51a5da..02c35725cc62 100644
> --- a/arch/x86/kernel/alternative.c
> +++ b/arch/x86/kernel/alternative.c
> @@ -1102,6 +1102,34 @@ void *text_poke_kgdb(void *addr, const void *opcode, size_t len)
> return __text_poke(addr, opcode, len);
> }
>
> +/**
> + * text_poke_jit - Update instructions on a live kernel by jit engine
> + * @addr: address to modify
> + * @opcode: source of the copy
> + * @len: length to copy, could be more than 2x PAGE_SIZE
> + *
> + * Only module memory taking jit text (e.g. for bpf) should be patched.
> + */

Maybe:

text_poke_copy() - Copy instructions into (an unused part of) RX memory
@args...

Not safe against concurrent execution; useful for JITs to dump
new code blocks into unused regions of RX memory. Can be used in
conjunction with synchronize_rcu_tasks() to wait for existing
execution to quiesce after having made sure no existing
functions pointers are life.

or something along those lines?

> +void *text_poke_jit(void *addr, const void *opcode, size_t len)
> +{
> + unsigned long start = (unsigned long)addr;
> + size_t patched = 0;
> +
> + if (WARN_ON_ONCE(core_kernel_text(start)))
> + return NULL;
> +
> + while (patched < len) {
> + unsigned long ptr = start + patched;
> + size_t s;
> +
> + s = min_t(size_t, PAGE_SIZE * 2 - offset_in_page(ptr), len - patched);
> +
> + __text_poke((void *)ptr, opcode + patched, s);
> + patched += s;
> + }
> + return addr;
> +}
> +
> static void do_sync_core(void *info)
> {
> sync_core();
> --
> 2.30.2
>