public inbox for dwarves@vger.kernel.org
 help / color / mirror / Atom feed
From: Eduard Zingerman <eddyz87@gmail.com>
To: Ihor Solodrai <ihor.solodrai@pm.me>,
	dwarves@vger.kernel.org,  acme@kernel.org
Cc: bpf@vger.kernel.org, alan.maguire@oracle.com, andrii@kernel.org,
	 mykolal@fb.com
Subject: Re: [RFC PATCH 3/9] btf_encoder: separate elf function, saved function representations
Date: Fri, 29 Nov 2024 12:37:57 -0800	[thread overview]
Message-ID: <39b3f6f04838b96e858effc09e01d7b29c529f2e.camel@gmail.com> (raw)
In-Reply-To: <20241128012341.4081072-4-ihor.solodrai@pm.me>

On Thu, 2024-11-28 at 01:23 +0000, Ihor Solodrai wrote:
> From: Alan Maguire <alan.maguire@oracle.com>
> 
> Have saved function representation point back at immutable ELF function
> table.  This will make sharing the ELF function table across encoders
> easier.  Simply accumulate saved functions for each encoder, and on
> completion combine them into a name-sorted list.  Then carry out
> comparisons to check for inconsistent representations, skipping functions
> that are inconsistent in their representation.
> 
> Thre is a small growth in maximum resident set size due to saving
> more functions; it grows from
> 
> 	Maximum resident set size (kbytes): 701888
> 
> to:
> 
> 	Maximum resident set size (kbytes): 704168
> 
> ...with this patch for -j1.
> 
> Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
> Signed-off-by: Ihor Solodrai <ihor.solodrai@pm.me>
> ---

Acked-by: Eduard Zingerman <eddyz87@gmail.com>

I like what this patch does, a few nits below.

Note:
this patch leads to 58 less functions being generated,
compared to a previous patch, for my test configuration.
For example, functions like:
- hid_map_usage_clear
- jhash
- nlmsg_parse_deprecated_strict
Are not in the BTF anymore. It would be good if patch message could
explain why this happens.

[...]

> +static int btf_encoder__add_saved_funcs(struct btf_encoder *encoder)
> +{
> +	struct btf_encoder_func_state **saved_fns, *s;
> +	struct btf_encoder *e = NULL;
> +	int i = 0, j, nr_saved_fns = 0;
> +
> +	/* Retrieve function states from each encoder, combine them
> +	 * and sort by name, addr.
> +	 */
> +	btf_encoders__for_each_encoder(e) {
> +		list_for_each_entry(s, &e->func_states, node)
> +			nr_saved_fns++;
> +	}
> +	/* Another thread already did this work */
> +	if (nr_saved_fns == 0) {
> +		printf("nothing to do for encoder...\n");
> +		return 0;
> +	}

Nit: this function is called from pahole_threads_collect():

	static int pahole_threads_collect(...)
		for (i = 0; i < nr_threads; i++)
			...
			err = btf_encoder__add_encoder(btf_encoder, threads[i]->encoder);
			...

	int32_t btf_encoder__add_encoder(struct btf_encoder *encoder, struct btf_encoder *other)
		...
		btf_encoder__add_saved_funcs(other);
		...
	
      maybe move call to btf_encoder__add_saved_funcs() to pahole_threads_collect()
      outside of the loop? So that comment about another thread won't be necessary.

> +
> +	printf("got %d saved functions...\n", nr_saved_fns);
> +	saved_fns = calloc(nr_saved_fns, sizeof(*saved_fns));
> +	btf_encoders__for_each_encoder(e) {
> +		list_for_each_entry(s, &e->func_states, node)
> +			saved_fns[i++] = s;
> +	}
> +	printf("added %d saved fns\n", i);
> +	qsort(saved_fns, nr_saved_fns, sizeof(*saved_fns), saved_functions_cmp);
> +
> +	for (i = 0; i < nr_saved_fns; i = j) {
> +		struct btf_encoder_func_state *state = saved_fns[i];
> +
> +		/* Compare across sorted functions that match by name/prefix;
> +		 * share inconsistent/unexpected reg state between them.
> +		 */
> +		j = i + 1;
> +
> +		while (j < nr_saved_fns &&
> +		       saved_functions_combine(saved_fns[i], saved_fns[j]) == 0)
> +				j++;
> +
> +		/* do not exclude functions with optimized-out parameters; they
> +		 * may still be _called_ with the right parameter values, they
> +		 * just do not _use_ them.  Only exclude functions with
> +		 * unexpected register use or multiple inconsistent prototypes.
> +		 */
> +		if (!encoder->skip_encoding_inconsistent_proto ||
> +		    (!state->unexpected_reg && !state->inconsistent_proto)) {
> +			if (btf_encoder__add_func(state->encoder, state)) {
> +				free(saved_fns);
> +				return -1;
> +			}
> +		}
> +	}
> +	/* Now that we are done with function states, free them. */
> +	free(saved_fns);
> +	btf_encoders__for_each_encoder(e)
> +		btf_encoder__delete_saved_funcs(e);
> +	return 0;
> +}

[...]

> @@ -2437,16 +2470,8 @@ out_delete:
>  	return NULL;
>  }
>  
> -void btf_encoder__delete_func(struct elf_function *func)
> -{
> -	free(func->alias);

Nit: it looks like func->alias is never freed after this change.

> -	zfree(&func->state.annots);
> -	zfree(&func->state.parms);
> -}

[...]


  reply	other threads:[~2024-11-29 20:38 UTC|newest]

Thread overview: 26+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-28  1:23 [RFC PATCH 0/9] pahole: shared ELF and faster reproducible BTF encoding Ihor Solodrai
2024-11-28  1:23 ` [RFC PATCH 1/9] btf_encoder: simplify function encoding Ihor Solodrai
2024-11-29  8:16   ` Eduard Zingerman
2024-12-02 13:55   ` Jiri Olsa
2024-11-28  1:23 ` [RFC PATCH 2/9] btf_encoder: store,use section-relative addresses in ELF function representation Ihor Solodrai
2024-11-29  9:07   ` Eduard Zingerman
2024-12-02 14:34     ` Alan Maguire
2024-11-28  1:23 ` [RFC PATCH 3/9] btf_encoder: separate elf function, saved function representations Ihor Solodrai
2024-11-29 20:37   ` Eduard Zingerman [this message]
2024-12-09 21:19     ` Ihor Solodrai
2024-11-28  1:24 ` [RFC PATCH 4/9] dwarf_loader: introduce pre_load_module hook to conf_load Ihor Solodrai
2024-11-29 21:03   ` Eduard Zingerman
2024-11-28  1:24 ` [RFC PATCH 5/9] btf_encoder: introduce elf_functions struct type Ihor Solodrai
2024-11-29 21:50   ` Eduard Zingerman
2024-11-28  1:24 ` [RFC PATCH 6/9] btf_encoder: collect elf_functions in btf_encoder__pre_load_module Ihor Solodrai
2024-11-29 22:27   ` Eduard Zingerman
2024-11-28  1:24 ` [RFC PATCH 7/9] btf_encoder: switch to shared elf_functions table Ihor Solodrai
2024-11-29 22:35   ` Eduard Zingerman
2024-12-09 23:55     ` Ihor Solodrai
2024-11-28  1:24 ` [RFC PATCH 8/9] btf_encoder: introduce btf_encoding_context Ihor Solodrai
2024-11-29 23:12   ` Eduard Zingerman
2024-11-28  1:24 ` [RFC PATCH 9/9] pahole: faster reproducible BTF encoding Ihor Solodrai
2024-11-30  0:17   ` Eduard Zingerman
2024-12-02 13:55 ` [RFC PATCH 0/9] pahole: shared ELF and " Jiri Olsa
2024-12-06 18:19   ` Ihor Solodrai
2024-12-06 18:30     ` Arnaldo Carvalho de Melo

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=39b3f6f04838b96e858effc09e01d7b29c529f2e.camel@gmail.com \
    --to=eddyz87@gmail.com \
    --cc=acme@kernel.org \
    --cc=alan.maguire@oracle.com \
    --cc=andrii@kernel.org \
    --cc=bpf@vger.kernel.org \
    --cc=dwarves@vger.kernel.org \
    --cc=ihor.solodrai@pm.me \
    --cc=mykolal@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