public inbox for bpf@vger.kernel.org
 help / color / mirror / Atom feed
From: Jiri Olsa <olsajiri@gmail.com>
To: Yonghong Song <yhs@fb.com>
Cc: bpf@vger.kernel.org, Alexei Starovoitov <ast@kernel.org>,
	Andrii Nakryiko <andrii@kernel.org>,
	Daniel Borkmann <daniel@iogearbox.net>,
	kernel-team@fb.com
Subject: Re: [PATCH bpf-next v2 3/6] bpf: x86: Support in-register struct arguments
Date: Sun, 14 Aug 2022 22:24:24 +0200	[thread overview]
Message-ID: <YvlZ+ETaaTD3hwrM@krava> (raw)
In-Reply-To: <20220812052435.523068-1-yhs@fb.com>

On Thu, Aug 11, 2022 at 10:24:35PM -0700, Yonghong Song wrote:

SNIP

>  }
>  
>  static int invoke_bpf_prog(const struct btf_func_model *m, u8 **pprog,
> @@ -2020,6 +2081,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  	struct bpf_tramp_links *fentry = &tlinks[BPF_TRAMP_FENTRY];
>  	struct bpf_tramp_links *fexit = &tlinks[BPF_TRAMP_FEXIT];
>  	struct bpf_tramp_links *fmod_ret = &tlinks[BPF_TRAMP_MODIFY_RETURN];
> +	int struct_val_off, extra_nregs = 0;
>  	u8 **branches = NULL;
>  	u8 *prog;
>  	bool save_ret;
> @@ -2028,6 +2090,20 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  	if (nr_args > 6)
>  		return -ENOTSUPP;
>  
> +	for (i = 0; i < MAX_BPF_FUNC_ARGS; i++) {
> +		if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG) {
> +			/* Only support up to 16 bytes struct which should keep
> +			 * values in registers.
> +			 */

it seems that if the struct contains 'double' field, it's passed in
SSE register, which we don't support is save/restore

we should probably check struct's BTF in btf_distill_func_proto and
fail if we found anything else than regular regs types?

> +			if (m->arg_size[i] > 16)
> +				return -ENOTSUPP;
> +
> +			extra_nregs += (m->arg_size[i] + 7) / 8 - 1;
> +		}
> +	}
> +	if (nr_args + extra_nregs > 6)

should this value be minus the number of actually found struct arguments?

> +		return -ENOTSUPP;
> +
>  	/* Generated trampoline stack layout:
>  	 *
>  	 * RBP + 8         [ return address  ]
> @@ -2066,6 +2142,13 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  	stack_size += (sizeof(struct bpf_tramp_run_ctx) + 7) & ~0x7;
>  	run_ctx_off = stack_size;
>  
> +	/* For structure argument */
> +	for (i = 0; i < MAX_BPF_FUNC_ARGS; i++) {
> +		if (m->arg_flags[i] & BTF_FMODEL_STRUCT_ARG)
> +			stack_size += (m->arg_size[i] + 7) & ~0x7;
> +	}
> +	struct_val_off = stack_size;

could you please update the 'Generated trampoline stack layout' table
above with this offset

thanks,
jirka

> +
>  	if (flags & BPF_TRAMP_F_SKIP_FRAME) {
>  		/* skip patched call instruction and point orig_call to actual
>  		 * body of the kernel function.
> @@ -2101,7 +2184,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  		emit_stx(&prog, BPF_DW, BPF_REG_FP, BPF_REG_0, -ip_off);
>  	}
>  
> -	save_regs(m, &prog, nr_args, regs_off);
> +	save_regs(m, &prog, nr_args, regs_off, struct_val_off);
>  
>  	if (flags & BPF_TRAMP_F_CALL_ORIG) {
>  		/* arg1: mov rdi, im */
> @@ -2131,7 +2214,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  	}
>  
>  	if (flags & BPF_TRAMP_F_CALL_ORIG) {
> -		restore_regs(m, &prog, nr_args, regs_off);
> +		restore_regs(m, &prog, nr_args, regs_off, struct_val_off);
>  
>  		if (flags & BPF_TRAMP_F_ORIG_STACK) {
>  			emit_ldx(&prog, BPF_DW, BPF_REG_0, BPF_REG_FP, 8);
> @@ -2172,7 +2255,7 @@ int arch_prepare_bpf_trampoline(struct bpf_tramp_image *im, void *image, void *i
>  		}
>  
>  	if (flags & BPF_TRAMP_F_RESTORE_REGS)
> -		restore_regs(m, &prog, nr_args, regs_off);
> +		restore_regs(m, &prog, nr_args, regs_off, struct_val_off);
>  
>  	/* This needs to be done regardless. If there were fmod_ret programs,
>  	 * the return value is only updated on the stack and still needs to be
> -- 
> 2.30.2
> 

  reply	other threads:[~2022-08-14 20:24 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-12  5:24 [PATCH bpf-next v2 0/6] bpf: Support struct argument for trampoline base progs Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 1/6] bpf: Add struct argument info in btf_func_model Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 2/6] bpf: x86: Rename stack_size to regs_off in {save,restore}_regs() Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 3/6] bpf: x86: Support in-register struct arguments Yonghong Song
2022-08-14 20:24   ` Jiri Olsa [this message]
2022-08-15  5:29     ` Yonghong Song
2022-08-15  7:29       ` Jiri Olsa
2022-08-15 15:25         ` Yonghong Song
2022-08-15 22:44   ` Alexei Starovoitov
2022-08-18  4:56     ` Yonghong Song
2022-08-18 20:44       ` Alexei Starovoitov
2022-08-24 19:04         ` Yonghong Song
2022-08-24 22:35           ` Alexei Starovoitov
2022-08-25  4:10             ` Yonghong Song
2022-08-24 19:05         ` Andrii Nakryiko
2022-08-25  4:04           ` Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 4/6] bpf: arm64: No support of struct argument Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 5/6] bpf: Populate struct argument info in btf_func_model Yonghong Song
2022-08-12  5:24 ` [PATCH bpf-next v2 6/6] selftests/bpf: Add struct argument tests with fentry/fexit programs 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=YvlZ+ETaaTD3hwrM@krava \
    --to=olsajiri@gmail.com \
    --cc=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=kernel-team@fb.com \
    --cc=yhs@fb.com \
    /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