| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <CAK7LNATCtoCEs4-S6kgGCH8XRkX+sKT3zuP-awdFJr+E1bkZww@mail.gmail.com>
Date: Sun, 11 Feb 2024 08:24:50 +0900
From: Masahiro Yamada <masahiroy@...nel.org>
To: sunying@...c.iscas.ac.cn
Cc: linux-kbuild@...r.kernel.org, linux-kernel@...r.kernel.org,
pengpeng@...as.ac.cn, Siyuan Guo <zy21df106@...a.edu.cn>
Subject: Re: [PATCHv3 next] kconfig: add dependency warning print about
invalid values in verbose mode
"in verbose mode" is stale.
Please rephrase the subject.
On Tue, Jan 2, 2024 at 11:11 AM <sunying@...c.iscas.ac.cn> wrote:
>
> From: Ying Sun <sunying@...c.iscas.ac.cn>
>
> Warning in verbose mode about incorrect causes and
> mismatch dependency of invalid values to help users correct them.
Same here, "verbose mode" does not exist in this patch.
>
> Detailed warning messages are printed only when the environment variable
> is set like "KCONFIG_WARN_CHANGED_INPUT=1".
> By default, the current behavior is not changed.
>
> Signed-off-by: Siyuan Guo <zy21df106@...a.edu.cn>
> Signed-off-by: Ying Sun <sunying@...c.iscas.ac.cn>
> ---
> v2 -> v3:
> * Fixed warning message that mess up the ncurses display.
> * Fixed memory leak where str_new() was called but str_free() was not used.
> * Integrated the code to avoid excessive dispersion.
> * Use the environment variable KCONFIG_WARN_CHANGED_INPUT as the switch
This checker prints wrong reports.
I just attached one test case.
[test Kconfig]
config MODULES
def_bool y
modules
config FOO
tristate "foo"
depends on BAR
config BAR
tristate "bar"
config BAZ
tristate "baz"
select FOO
[test .config]
CONFIG_FOO=m
# CONFIG_BAR is not set
CONFIG_BAZ=y
[test command]
$ KCONFIG_WARN_CHANGED_INPUT=1 make olddefconfig
[result]
Kconfig:8:warning: 'MODULES' set to y due to option constraints
WARNING: unmet direct dependencies detected for FOO
Depends on [n]: BAR [=n]
Selected by [y]:
- BAZ [=y]
Kconfig:12:warning: 'FOO' set to n for it unmet direct dependencies
Depends on [n]: BAR [=n]
$ cat .config
#
# Automatically generated file; DO NOT EDIT.
# Linux/x86 6.8.0-rc3 Kernel Configuration
#
CONFIG_MODULES=y
CONFIG_FOO=y
# CONFIG_BAR is not set
CONFIG_BAZ=y
The first report
Kconfig:8:warning: 'MODULES' set to y due to option constraints
should not be printed.
CONFIG options without prompt can be omitted.
The second report
Kconfig:12:warning: 'FOO' set to n for it unmet direct dependencies
is completely wrong.
CONFIG_FOO was changed to 'y' due to the select.
One more thing,
Add
export KCONFIG_WARN_CHANGED_INPUT=1
to scripts/kconfig/Makefile
> v1 -> v2:
> * Reduced the number of code lines by refactoring and simplifying the logic.
> * Changed the print "ERROR" to "WARNING".
> * Focused on handling dependency errors: dir_dep and rev_dep, and range error.
> - A downgrade from 'y' to 'm' has be warned.
> - A new CONFIG option should not be warned.
> - Overwriting caused by default value is not an error and is no longer printed.
> * Fixed style issues.
> ---
> ---
> scripts/kconfig/confdata.c | 76 ++++++++++++++++++++++++++++++++++++++
> 1 file changed, 76 insertions(+)
>
> diff --git a/scripts/kconfig/confdata.c b/scripts/kconfig/confdata.c
> index f1197e672431..352774928558 100644
> --- a/scripts/kconfig/confdata.c
> +++ b/scripts/kconfig/confdata.c
> @@ -195,6 +195,52 @@ static void conf_message(const char *fmt, ...)
> va_end(ap);
> }
>
> +static void conf_warn_unmet_rev_dep(struct symbol *sym)
> +{
> + struct gstr gs = str_new();
> + char value = 'n';
> +
> + switch (sym->curr.tri) {
> + case no:
> + value = 'n';
> + break;
> + case mod:
> + sym_calc_value(modules_sym);
> + value = (modules_sym->curr.tri == no) ? 'n' : 'm';
> + break;
> + case yes:
> + value = 'y';
> + }
> +
> + str_printf(&gs,
> + "'%s' set to %c for it is selected\n",
> + sym->name, value);
> + expr_gstr_print_revdep(sym->rev_dep.expr, &gs, yes,
> + " Selected by [y]:\n");
> + expr_gstr_print_revdep(sym->rev_dep.expr, &gs, mod,
> + " Selected by [m]:\n");
> +
> + conf_warning("%s", str_get(&gs));
> + str_free(&gs);
> +}
> +
> +static void conf_warn_dep_error(struct symbol *sym)
> +{
> + struct gstr gs = str_new();
> +
> + str_printf(&gs,
> + "'%s' set to n for it unmet direct dependencies\n",
> + sym->name);
> +
> + str_printf(&gs,
> + " Depends on [%c]: ",
> + sym->dir_dep.tri == mod ? 'm' : 'n');
> + expr_gstr_print(sym->dir_dep.expr, &gs);
> +
> + conf_warning("%s\n", str_get(&gs));
> + str_free(&gs);
> +}
> +
> const char *conf_get_configname(void)
> {
> char *name = getenv("KCONFIG_CONFIG");
> @@ -568,6 +614,36 @@ int conf_read(const char *name)
> continue;
> conf_unsaved++;
> /* maybe print value in verbose mode... */
> + if (getenv("KCONFIG_WARN_CHANGED_INPUT") && (sym->prop)) {
> + conf_filename = sym->prop->file->name;
> + conf_lineno = sym->prop->menu->lineno;
> +
> + switch (sym->type) {
> + case S_BOOLEAN:
> + case S_TRISTATE:
> + if (sym->def[S_DEF_USER].tri != sym_get_tristate_value(sym)) {
> + if (sym->dir_dep.tri < sym->def[S_DEF_USER].tri)
> + conf_warn_dep_error(sym);
> + else if (sym->rev_dep.tri > sym->def[S_DEF_USER].tri)
> + conf_warn_unmet_rev_dep(sym);
> + else
> + conf_warning("'%s' set to %s due to option constraints\n",
> + sym->name, sym_get_string_value(sym));
> + }
> + break;
> + case S_INT:
> + case S_HEX:
> + case S_STRING:
> + if (sym->dir_dep.tri == no && sym_has_value(sym))
> + conf_warn_dep_error(sym);
> + else
> + conf_warning("'%s' set to %s due to option constraints\n",
> + sym->name, sym_get_string_value(sym));
> + break;
> + default:
> + break;
> + }
> + }
> }
>
> for_all_symbols(i, sym) {
> --
> 2.43.0
>
>
--
Best Regards
Masahiro Yamada
Powered by blists - more mailing lists