netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: John Fastabend <john.fastabend@gmail.com>
To: "Daniel T. Lee" <danieltimlee@gmail.com>,
	Daniel Borkmann <daniel@iogearbox.net>,
	Alexei Starovoitov <ast@kernel.org>,
	Andrii Nakryiko <andrii.nakryiko@gmail.com>
Cc: bpf@vger.kernel.org, netdev@vger.kernel.org
Subject: RE: [PATCH] samples: bpf: fix tracex2 due to empty sys_write count argument
Date: Fri, 10 Dec 2021 10:00:26 -0800	[thread overview]
Message-ID: <61b395ba5cc03_2032520824@john.notmuch> (raw)
In-Reply-To: <20211210111918.4904-1-danieltimlee@gmail.com>

Daniel T. Lee wrote:
> Currently from syscall entry, argument can't be fetched correctly as a
> result of register cleanup.
> 
>     commit 6b8cf5cc9965 ("x86/entry/64/compat: Clear registers for compat syscalls, to reduce speculation attack surface")
> 
> For example in upper commit, registers are cleaned prior to syscall.
> To be more specific, sys_write syscall has count size as a third argument.
> But this can't be fetched from __x64_sys_enter/__s390x_sys_enter due to
> register cleanup. (e.g. [x86] xorl %r8d, %r8d / [s390x] xgr %r7, %r7)
> 
> This commit fix this problem by modifying the trace event to ksys_write
> instead of sys_write syscall entry.
> 
>     # Wrong example of 'write()' syscall argument fetching
>     # ./tracex2
>     ...
>     pid 50909 cmd dd uid 0
>            syscall write() stats
>      byte_size       : count     distribution
>        1 -> 1        : 4968837  |************************************* |
> 
>     # Successful example of 'write()' syscall argument fetching
>     # (dd's write bytes at a time defaults to 512)
>     # ./tracex2
>     ...
>     pid 3095 cmd dd uid 0
>            syscall write() stats
>      byte_size       : count     distribution
>     ...
>      256 -> 511      : 0        |                                      |
>      512 -> 1023     : 4968844  |************************************* |
> 
> Signed-off-by: Daniel T. Lee <danieltimlee@gmail.com>
> ---
>  samples/bpf/tracex2_kern.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/samples/bpf/tracex2_kern.c b/samples/bpf/tracex2_kern.c
> index 5bc696bac27d..96dff3bea227 100644
> --- a/samples/bpf/tracex2_kern.c
> +++ b/samples/bpf/tracex2_kern.c
> @@ -78,7 +78,7 @@ struct {
>  	__uint(max_entries, 1024);
>  } my_hist_map SEC(".maps");
>  
> -SEC("kprobe/" SYSCALL(sys_write))
> +SEC("kprobe/ksys_write")
>  int bpf_prog3(struct pt_regs *ctx)
>  {
>  	long write_size = PT_REGS_PARM3(ctx);
> -- 
> 2.32.0
> 

LGTM

Acked-by: John Fastabend <john.fastabend@gmail.com>

  reply	other threads:[~2021-12-10 18:00 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-10 11:19 [PATCH] samples: bpf: fix tracex2 due to empty sys_write count argument Daniel T. Lee
2021-12-10 18:00 ` John Fastabend [this message]
2021-12-10 18:28 ` Yonghong Song

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=61b395ba5cc03_2032520824@john.notmuch \
    --to=john.fastabend@gmail.com \
    --cc=andrii.nakryiko@gmail.com \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=danieltimlee@gmail.com \
    --cc=netdev@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).