linux-kernel - Re: [PATCH bpf] bpf: update bpf_get_smp_processor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 14 Sep 2021 16:01:36 -0700
From:   Yonghong Song <yhs@...com>
To:     Matteo Croce <mcroce@...ux.microsoft.com>, <bpf@...r.kernel.org>
CC:     <linux-kernel@...r.kernel.org>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>
Subject: Re: [PATCH bpf] bpf: update bpf_get_smp_processor_id() documentation



On 9/14/21 3:23 PM, Matteo Croce wrote:
> From: Matteo Croce <mcroce@...rosoft.com>
> 
> Since commit 1e6c62a88215 ("bpf: Introduce sleepable BPF programs"), BPF
> programs can sleep if the BPF_F_SLEEPABLE flag is set.
> Update the documentation accordingly.
> 
> Fixes: 1e6c62a88215 ("bpf: Introduce sleepable BPF programs")

I think we don't need this Fixes tag as this patch should target to
bpf-next tree. There is no need to backport to bpf tree since it is
just a documentation change.

Based on *current* implementation, referring to sleepable BPF
program commit is not needed any more. See below for the suggested
change.

> Signed-off-by: Matteo Croce <mcroce@...rosoft.com>
> ---
>   include/uapi/linux/bpf.h | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/include/uapi/linux/bpf.h b/include/uapi/linux/bpf.h
> index d21326558d42..5e3b2fb62d84 100644
> --- a/include/uapi/linux/bpf.h
> +++ b/include/uapi/linux/bpf.h
> @@ -1629,9 +1629,9 @@ union bpf_attr {
>    * u32 bpf_get_smp_processor_id(void)
>    * 	Description
>    * 		Get the SMP (symmetric multiprocessing) processor id. Note that
> - * 		all programs run with preemption disabled, which means that the
> - * 		SMP processor id is stable during all the execution of the
> - * 		program.
> + * 		programs run with preemption disabled unless BPF_F_SLEEPABLE is
> + * 		set, which means that the SMP processor id is stable during all
> + * 		the execution of the program.

Currently, migrate_disable() means true migration disable and preemption 
is possible, and BPF_F_SLEEPABLE programs are also protected by 
migrate_disable().

So the patch should just change "with preemption disabled" to
"with migration disabled" to reflect the new implementation.

>    * 	Return
>    * 		The SMP id of the processor running the program.
>    *
>