BPF List
 help / color / mirror / Atom feed
From: Jiri Olsa <olsajiri@gmail.com>
To: Alan Maguire <alan.maguire@oracle.com>
Cc: acme@kernel.org, yhs@fb.com, ast@kernel.org, olsajiri@gmail.com,
	timo@incline.eu, daniel@iogearbox.net, andrii@kernel.org,
	songliubraving@fb.com, john.fastabend@gmail.com,
	kpsingh@chromium.org, sdf@google.com, haoluo@google.com,
	martin.lau@kernel.org, bpf@vger.kernel.org
Subject: Re: [PATCH dwarves 5/5] btf_encoder: skip BTF encoding of static functions with inconsistent prototypes
Date: Wed, 25 Jan 2023 14:39:02 +0100	[thread overview]
Message-ID: <Y9Ew9iwd3Jg5vk9c@krava> (raw)
In-Reply-To: <1674567931-26458-6-git-send-email-alan.maguire@oracle.com>

On Tue, Jan 24, 2023 at 01:45:31PM +0000, Alan Maguire wrote:

SNIP

>  static int32_t btf_encoder__save_func(struct btf_encoder *encoder, struct function *fn)
>  {
> @@ -819,13 +837,51 @@ static int32_t btf_encoder__save_func(struct btf_encoder *encoder, struct functi
>  	}
>  	/* If we find an existing entry, we want to merge observations
>  	 * across both functions, checking that the "seen optimized-out
> -	 * parameters" status is reflected in our tree entry.
> +	 * parameters"/inconsistent proto status is reflected in tree entry.
>  	 * If the entry is new, record encoder state required
>  	 * to add the local function later (encoder + type_id_off)
> -	 * such that we can add the function later.
> +	 * such that we can add the function later.  Parameter names are
> +	 * also stored in state to speed up multiple static function
> +	 * comparisons.
>  	 */
>  	if (*nodep != fn) {
> -		(*nodep)->proto.optimized_parms |= fn->proto.optimized_parms;
> +		struct function *ofn = *nodep;
> +
> +		ofn->proto.optimized_parms |= fn->proto.optimized_parms;
> +		/* compare parameters to see if signatures match */
> +
> +		if (ofn->proto.inconsistent_proto)
> +			goto out;
> +
> +		if (ofn->proto.nr_parms != fn->proto.nr_parms) {
> +			ofn->proto.inconsistent_proto = 1;
> +			goto out;
> +		}
> +		if (ofn->proto.nr_parms > 0) {
> +			struct btf_encoder_state *state = ofn->priv;
> +			const char *parameter_names[BTF_ENCODER_MAX_PARAMETERS];
> +			int i;
> +
> +			if (!state->got_parameter_names) {
> +				parameter_names__get(&ofn->proto, BTF_ENCODER_MAX_PARAMETERS,
> +						     state->parameter_names);
> +				state->got_parameter_names = true;
> +			}
> +			parameter_names__get(&fn->proto, BTF_ENCODER_MAX_PARAMETERS,
> +					     parameter_names);
> +			for (i = 0; i < ofn->proto.nr_parms; i++) {
> +				if (!state->parameter_names[i]) {
> +					if (!parameter_names[i])
> +						continue;
> +				} else if (parameter_names[i]) {
> +					if (strcmp(state->parameter_names[i],
> +						   parameter_names[i]) == 0)
> +						continue;

I guess we can't check type easily? tag has type field,
but I'm not sure if we can get reasonable type info from that

jirka

> +				}
> +				ofn->proto.inconsistent_proto = 1;
> +				goto out;
> +			}
> +		}
>  	} else {
>  		struct btf_encoder_state *state = zalloc(sizeof(*state));
>  
> @@ -898,10 +954,12 @@ static void btf_encoder__add_saved_func(const void *nodep, const VISIT which,
>  	/* we can safely free encoder state since we visit each node once */
>  	free(fn->priv);
>  	fn->priv = NULL;
> -	if (fn->proto.optimized_parms) {
> +	if (fn->proto.optimized_parms || fn->proto.inconsistent_proto) {
>  		if (encoder->verbose)
> -			printf("skipping addition of '%s' due to optimized-out parameters\n",
> -			       function__name(fn));
> +			printf("skipping addition of '%s' due to %s\n",
> +			       function__name(fn),
> +			       fn->proto.optimized_parms ? "optimized-out parameters" :
> +							   "multiple inconsistent function prototypes");
>  	} else {
>  		btf_encoder__add_func(encoder, fn);
>  	}
> @@ -1775,6 +1833,8 @@ int btf_encoder__encode_cu(struct btf_encoder *encoder, struct cu *cu, struct co
>  		 */
>  		if (fn->declaration)
>  			continue;
> +		if (!fn->external)
> +			save = true;
>  		if (!ftype__has_arg_names(&fn->proto))
>  			continue;
>  		if (encoder->functions.cnt) {
> @@ -1790,7 +1850,8 @@ int btf_encoder__encode_cu(struct btf_encoder *encoder, struct cu *cu, struct co
>  			if (func) {
>  				if (func->generated)
>  					continue;
> -				func->generated = true;
> +				if (!save)
> +					func->generated = true;
>  			} else if (encoder->functions.suffix_cnt) {
>  				/* falling back to name.isra.0 match if no exact
>  				 * match is found; only bother if we found any
> diff --git a/dwarves.h b/dwarves.h
> index 1ad1b3b..9b80262 100644
> --- a/dwarves.h
> +++ b/dwarves.h
> @@ -830,6 +830,7 @@ struct ftype {
>  	uint16_t	 nr_parms;
>  	uint8_t		 unspec_parms:1; /* just one bit is needed */
>  	uint8_t		 optimized_parms:1;
> +	uint8_t		 inconsistent_proto:1;
>  };
>  
>  static inline struct ftype *tag__ftype(const struct tag *tag)
> -- 
> 1.8.3.1
> 

  reply	other threads:[~2023-01-25 13:39 UTC|newest]

Thread overview: 27+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-24 13:45 [PATCH dwarves 0/5] dwarves: support encoding of optimized-out parameters, removal of inconsistent static functions Alan Maguire
2023-01-24 13:45 ` [PATCH dwarves 1/5] dwarves: help dwarf loader spot functions with optimized-out parameters Alan Maguire
2023-01-25 16:53   ` Jiri Olsa
2023-01-25 17:47   ` Eduard Zingerman
2023-01-25 18:28     ` Alan Maguire
2023-01-25 21:34       ` Eduard Zingerman
2023-01-25 22:52         ` Alan Maguire
2023-01-25 23:42           ` Eduard Zingerman
2023-01-26  0:20             ` Eduard Zingerman
2023-01-26 14:02               ` Alan Maguire
2023-01-26 15:02                 ` Eduard Zingerman
2023-01-24 13:45 ` [PATCH dwarves 2/5] btf_encoder: refactor function addition into dedicated btf_encoder__add_func Alan Maguire
2023-01-24 13:45 ` [PATCH dwarves 3/5] btf_encoder: child encoders should have a reference to parent encoder Alan Maguire
2023-01-24 13:45 ` [PATCH dwarves 4/5] btf_encoder: represent "."-suffixed optimized functions (".isra.0") in BTF Alan Maguire
2023-01-25 17:54   ` Kui-Feng Lee
2023-01-25 18:56     ` Arnaldo Carvalho de Melo
2023-01-26 18:37       ` Kui-Feng Lee
2023-01-25 18:59     ` Alan Maguire
2023-01-26 17:43       ` Kui-Feng Lee
2023-01-24 13:45 ` [PATCH dwarves 5/5] btf_encoder: skip BTF encoding of static functions with inconsistent prototypes Alan Maguire
2023-01-25 13:39   ` Jiri Olsa [this message]
2023-01-25 14:18     ` Alan Maguire
2023-01-25 16:53   ` Jiri Olsa
2023-01-26 14:12     ` Alan Maguire
2023-01-24 15:14 ` [PATCH dwarves 0/5] dwarves: support encoding of optimized-out parameters, removal of inconsistent static functions Jiri Olsa
2023-01-24 16:11   ` Alan Maguire
2023-01-25 13:59     ` Jiri Olsa

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=Y9Ew9iwd3Jg5vk9c@krava \
    --to=olsajiri@gmail.com \
    --cc=acme@kernel.org \
    --cc=alan.maguire@oracle.com \
    --cc=andrii@kernel.org \
    --cc=ast@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=daniel@iogearbox.net \
    --cc=haoluo@google.com \
    --cc=john.fastabend@gmail.com \
    --cc=kpsingh@chromium.org \
    --cc=martin.lau@kernel.org \
    --cc=sdf@google.com \
    --cc=songliubraving@fb.com \
    --cc=timo@incline.eu \
    --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