All of lore.kernel.org
 help / color / mirror / Atom feed
From: Luis Chamberlain <mcgrof@kernel.org>
To: Allen Webb <allenwebb@google.com>
Cc: "linux-modules@vger.kernel.org" <linux-modules@vger.kernel.org>,
	"linux-usb@vger.kernel.org" <linux-usb@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	gregkh@linuxfoundation.org, christophe.leroy@csgroup.eu,
	nick.alcock@oracle.com
Subject: Re: [PATCH v10 05/11] modpost: Track module name for built-in modules
Date: Tue, 23 May 2023 23:50:40 -0700	[thread overview]
Message-ID: <ZG2zwMTdJryhGrbU@bombadil.infradead.org> (raw)
In-Reply-To: <20230406190030.968972-6-allenwebb@google.com>

On Thu, Apr 06, 2023 at 02:00:24PM -0500, Allen Webb wrote:
> Keep track of the module name when processing match table symbols.

This should mention why this would be good. Otherwise, think about it,
ok, it's done but why? If the reason is that it will be needed later
you need to say that in this commit log entry. If its not used now, it
also needs to say that in this commit log so it is easier to review
and set expecataions correctly for the reviewer.

  Luis

> Signed-off-by: Allen Webb <allenwebb@google.com>
> ---
>  scripts/mod/file2alias.c | 39 +++++++++++++++++++++++++++++++++++----
>  scripts/mod/modpost.h    |  1 +
>  2 files changed, 36 insertions(+), 4 deletions(-)
> 
> diff --git a/scripts/mod/file2alias.c b/scripts/mod/file2alias.c
> index 91c2e7ba5e52..b392d51c3b06 100644
> --- a/scripts/mod/file2alias.c
> +++ b/scripts/mod/file2alias.c
> @@ -28,6 +28,7 @@ typedef Elf64_Addr	kernel_ulong_t;
>  #include <stdint.h>
>  #endif
>  
> +#include <assert.h>
>  #include <ctype.h>
>  #include <stdbool.h>
>  
> @@ -1540,9 +1541,9 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
>  			Elf_Sym *sym, const char *symname)
>  {
>  	void *symval;
> -	char *zeros = NULL;
> -	const char *name, *identifier;
> -	unsigned int namelen;
> +	char *zeros = NULL, *modname_str = NULL;
> +	const char *name, *identifier, *modname;
> +	unsigned int namelen, modnamelen;
>  
>  	/* We're looking for a section relative symbol */
>  	if (!sym->st_shndx || get_secindex(info, sym) >= info->num_sections)
> @@ -1552,7 +1553,12 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
>  	if (ELF_ST_TYPE(sym->st_info) != STT_OBJECT)
>  		return;
>  
> -	/* All our symbols are of form __mod_<name>__<identifier>_device_table. */
> +	/*
> +	 * All our symbols are either of form
> +	 *   __mod_<name>__<identifier>_device_table
> +	 * or
> +	 *   __mod_<name>__<identifier>__kmod_<builtin-name>_device_table
> +	 */
>  	if (strncmp(symname, "__mod_", strlen("__mod_")))
>  		return;
>  	name = symname + strlen("__mod_");
> @@ -1564,8 +1570,30 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
>  	identifier = strstr(name, "__");
>  	if (!identifier)
>  		return;
> +	modnamelen = namelen;
>  	namelen = identifier - name;
>  
> +	/*
> +	 * In the vmlinuz.o case we want to handle __kmod_ so aliases from
> +	 * builtin modules are attributed correctly.
> +	 */
> +	modname = strstr(identifier + 2, "__kmod_");
> +	if (modname) {
> +		modname += strlen("__kmod_");
> +		modnamelen -= (modname - name) + strlen("_device_table");
> +		modname_str = malloc(modnamelen + 1);
> +		/* We don't want to continue if the allocation fails. */
> +		assert(modname_str);
> +		memcpy(modname_str, modname, modnamelen);
> +		modname_str[modnamelen] = '\0';
> +	}
> +
> +	if (modname_str)
> +		modname = modname_str;
> +	else
> +		modname = mod->name;
> +	mod->builtin_name = modname;
> +
>  	/* Handle all-NULL symbols allocated into .bss */
>  	if (info->sechdrs[get_secindex(info, sym)].sh_type & SHT_NOBITS) {
>  		zeros = calloc(1, sym->st_size);
> @@ -1597,6 +1625,9 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
>  		}
>  	}
>  	free(zeros);
> +	mod->builtin_name = NULL;
> +	if (modname_str)
> +		free(modname_str);
>  }
>  
>  /* Now add out buffered information to the generated C source */
> diff --git a/scripts/mod/modpost.h b/scripts/mod/modpost.h
> index 1178f40a73f3..34fe5fc0b02c 100644
> --- a/scripts/mod/modpost.h
> +++ b/scripts/mod/modpost.h
> @@ -128,6 +128,7 @@ struct module {
>  	struct list_head missing_namespaces;
>  	// Actual imported namespaces
>  	struct list_head imported_namespaces;
> +	const char *builtin_name;
>  	char name[];
>  };
>  
> -- 
> 2.39.2
> 

  parent reply	other threads:[~2023-05-24  6:51 UTC|newest]

Thread overview: 102+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CAJzde06+FXNpyBzT+NfS2GCfqEERMkGDpdsmHQj=v1foLJW4Cw@mail.gmail.com>
2022-11-29 22:43 ` [PATCH v3] modules: add modalias file to sysfs for modules Allen Webb
2022-11-30  7:06   ` Greg Kroah-Hartman
2022-11-30 22:14     ` [PATCH v4] " Allen Webb
2022-12-01  4:33       ` kernel test robot
2022-12-01  6:06       ` Greg Kroah-Hartman
2022-12-01  9:46       ` kernel test robot
2022-12-08  2:34   ` [PATCH v3] " Luis Chamberlain
2022-12-08 14:22     ` Allen Webb
2022-12-08 15:20       ` Greg Kroah-Hartman
2022-12-16 22:16         ` [PATCH v7 0/5] Generate modules.builtin.alias from match ids Allen Webb
2022-12-16 22:16           ` [PATCH v7 1/5] module.h: MODULE_DEVICE_TABLE for built-in modules Allen Webb
2022-12-17  3:49             ` kernel test robot
2022-12-17  3:59             ` kernel test robot
2022-12-17  4:50             ` kernel test robot
2022-12-17 10:05             ` Christophe Leroy
2022-12-19 15:56               ` Allen Webb
2022-12-16 22:17           ` [PATCH v7 2/5] modpost: Track module name " Allen Webb
2022-12-17 10:08             ` Christophe Leroy
2022-12-16 22:17           ` [PATCH v7 3/5] modpost: Add -b option for emitting built-in aliases Allen Webb
2022-12-17 10:10             ` Christophe Leroy
2022-12-16 22:17           ` [PATCH v7 4/5] file2alias.c: Implement builtin.alias generation Allen Webb
2022-12-17  0:47             ` kernel test robot
2022-12-17  3:09             ` kernel test robot
2022-12-17 10:13             ` Christophe Leroy
2022-12-16 22:17           ` [PATCH v7 5/5] build: Add modules.builtin.alias Allen Webb
2022-12-19 19:18           ` [PATCH v8 0/9] Generate modules.builtin.alias from match ids Allen Webb
2022-12-19 19:18             ` [PATCH v8 1/9] imx: Fix typo Allen Webb
2022-12-19 19:21               ` Greg Kroah-Hartman
2022-12-19 19:55                 ` Allen Webb
2022-12-19 19:18             ` [PATCH v8 2/9] rockchip-mailbox: " Allen Webb
2022-12-19 19:18             ` [PATCH v8 3/9] scsi/BusLogic: Always include device id table Allen Webb
2022-12-19 19:18             ` [PATCH v8 4/9] stmpe-spi: Fix typo Allen Webb
2022-12-19 19:18             ` [PATCH v8 5/9] module.h: MODULE_DEVICE_TABLE for built-in modules Allen Webb
2022-12-19 19:18             ` [PATCH v8 6/9] modpost: Track module name " Allen Webb
2022-12-19 19:18             ` [PATCH v8 7/9] modpost: Add -b option for emitting built-in aliases Allen Webb
2022-12-19 19:18             ` [PATCH v8 8/9] file2alias.c: Implement builtin.alias generation Allen Webb
2022-12-19 19:18             ` [PATCH v8 9/9] build: Add modules.builtin.alias Allen Webb
2022-12-19 20:06             ` [PATCH v8 0/9] Generate modules.builtin.alias from match ids Luis Chamberlain
2022-12-19 20:42               ` Allen Webb
2022-12-19 20:46             ` [PATCH v9 00/10] " Allen Webb
2022-12-19 20:46               ` [PATCH v9 01/10] imx: Fix typo Allen Webb
2022-12-19 21:23                 ` Luis Chamberlain
2022-12-20  6:42                 ` Greg Kroah-Hartman
2022-12-20 14:26                   ` Allen Webb
2022-12-20 14:32                     ` Greg Kroah-Hartman
2022-12-20 14:45                       ` Allen Webb
2022-12-19 20:46               ` [PATCH v9 02/10] rockchip-mailbox: " Allen Webb
2022-12-20  6:46                 ` Greg Kroah-Hartman
2022-12-20 14:58                   ` Allen Webb
2022-12-20 18:12                     ` Luis Chamberlain
2022-12-20 18:19                       ` Allen Webb
2022-12-20 18:47                         ` Luis Chamberlain
2022-12-20 19:49                           ` Allen Webb
2022-12-20 20:03                             ` Luis Chamberlain
2022-12-20 21:57                               ` Allen Webb
2022-12-20 23:09                                 ` Luis Chamberlain
2022-12-27 17:42                                   ` Allen Webb
2023-01-10  0:25                                     ` Luis Chamberlain
2023-01-09 11:54                           ` Nick Alcock
2023-01-10 18:20                             ` Allen Webb
2022-12-19 20:46               ` [PATCH v9 03/10] scsi/BusLogic: Always include device id table Allen Webb
2022-12-19 20:46               ` [PATCH v9 04/10] stmpe-spi: Fix typo Allen Webb
2022-12-19 20:46               ` [PATCH v9 05/10] module.h: MODULE_DEVICE_TABLE for built-in modules Allen Webb
2022-12-20  6:45                 ` Greg Kroah-Hartman
2022-12-20 16:36                   ` Allen Webb
2022-12-19 20:46               ` [PATCH v9 06/10] modpost: Track module name " Allen Webb
2022-12-19 20:46               ` [PATCH v9 07/10] modpost: Add -b option for emitting built-in aliases Allen Webb
2022-12-20  6:43                 ` Greg Kroah-Hartman
2022-12-20 17:32                   ` Allen Webb
2022-12-19 20:46               ` [PATCH v9 08/10] file2alias.c: Implement builtin.alias generation Allen Webb
2022-12-19 20:46               ` [PATCH v9 09/10] build: Add modules.builtin.alias Allen Webb
2022-12-19 20:46               ` [PATCH v9 10/10] docs: Include modules.builtin.alias Allen Webb
2022-12-19 20:49                 ` Allen Webb
2022-12-19 21:23                 ` Luis Chamberlain
2022-12-19 21:40                   ` Allen Webb
2022-12-19 22:07                     ` Luis Chamberlain
2022-12-19 22:20                       ` Allen Webb
2022-12-19 22:51                         ` Luis Chamberlain
2022-12-19 20:46               ` [PATCH v9 10/10] Documentation: " Allen Webb
2023-04-06 19:00               ` [PATCH v10 00/11] Generate modules.builtin.alias from match ids Allen Webb
2023-04-06 19:00                 ` [PATCH v10 01/11] rockchip-mailbox: Remove unneeded MODULE_DEVICE_TABLE Allen Webb
2023-04-06 19:00                 ` [PATCH v10 02/11] scsi/BusLogic: Always include device id table Allen Webb
2023-04-06 19:00                 ` [PATCH v10 03/11] stmpe-spi: Fix MODULE_DEVICE_TABLE entries Allen Webb
2023-05-24  6:52                   ` Luis Chamberlain
2023-05-24  6:52                   ` Luis Chamberlain
2023-04-06 19:00                 ` [PATCH v10 04/11] module.h: MODULE_DEVICE_TABLE for built-in modules Allen Webb
2023-05-24  6:44                   ` Luis Chamberlain
2023-04-06 19:00                 ` [PATCH v10 05/11] modpost: Track module name " Allen Webb
2023-04-20  9:47                   ` Greg KH
2023-05-24  6:50                   ` Luis Chamberlain [this message]
2023-04-06 19:00                 ` [PATCH v10 06/11] modpost: Add -b option for emitting built-in aliases Allen Webb
2023-05-24  6:54                   ` Luis Chamberlain
2023-04-06 19:00                 ` [PATCH v10 07/11] file2alias.c: Implement builtin.alias generation Allen Webb
2023-05-24  7:00                   ` Luis Chamberlain
2023-04-06 19:00                 ` [PATCH v10 08/11] build: Add modules.builtin.alias Allen Webb
2023-05-24  7:02                   ` Luis Chamberlain
2023-07-19 19:51                     ` Allen Webb
2023-07-26 18:30                       ` Luis Chamberlain
2023-04-06 19:00                 ` [PATCH v10 09/11] Documentation: Include modules.builtin.alias Allen Webb
2023-04-06 19:00                 ` [PATCH v10 10/11] Documentation: Update writing_usb_driver for built-in modules Allen Webb
2023-04-06 19:00                 ` [PATCH v10 11/11] Documentation: add USB authorization document to driver-api Allen Webb
2023-04-20  9:51                   ` Greg KH

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=ZG2zwMTdJryhGrbU@bombadil.infradead.org \
    --to=mcgrof@kernel.org \
    --cc=allenwebb@google.com \
    --cc=christophe.leroy@csgroup.eu \
    --cc=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-modules@vger.kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=nick.alcock@oracle.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.