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
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [day] [month] [year] [list]
Message-ID: <CAJfuBxyeKz3bsc=WfjJZDKgAHScC80_irQvmsecxPukjM-J8gw@mail.gmail.com>
Date:   Sun, 19 Mar 2023 23:05:52 -0600
From:   jim.cromie@...il.com
To:     LKML <linux-kernel@...r.kernel.org>, linux-kbuild@...r.kernel.org
Cc:     Greg KH <gregkh@...uxfoundation.org>,
        Jason Baron <jbaron@...mai.com>,
        Luis Chamberlain <mcgrof@...nel.org>
Subject: RFC - KBUILD_MODNAME is misleading in builtins, as seen in /proc/dynamic_debug/control

dynamic-debug METADATA uses KBUILD_MODNAME as:

#define DEFINE_DYNAMIC_DEBUG_METADATA_CLS(name, cls, fmt)       \
        static struct _ddebug  __aligned(8)                     \
        __section("__dyndbg") name = {                          \
                .modname = KBUILD_MODNAME,                      \

This is going amiss for some builtins, ie those enabled here, by:

    echo module main +pmf > /proc/dynamic_debug_control
    grep =pmf /proc/dynamic_debug/control

init/main.c:1187 [main]initcall_blacklist =pmf "blacklisting initcall %s\n"
init/main.c:1226 [main]initcall_blacklisted =pmf "initcall %s blacklisted\n"
init/main.c:1432 [main]run_init_process =pmf "  with arguments:\n"
init/main.c:1434 [main]run_init_process =pmf "    %s\n"
init/main.c:1435 [main]run_init_process =pmf "  with environment:\n"
init/main.c:1437 [main]run_init_process =pmf "    %s\n"
kernel/module/main.c:336 [main]find_symbol =pmf "Failed to find symbol %s\n"
kernel/module/main.c:567 [main]already_uses =pmf "%s uses %s!\n"
kernel/module/main.c:571 [main]already_uses =pmf "%s does not use %s!\n"
kernel/module/main.c:586 [main]add_module_usage =pmf "Allocating new
usage for %s.\n"
kernel/module/main.c:627 [main]module_unload_free =pmf "%s unusing %s\n"
kernel/module/main.c:733 [main]__do_sys_delete_module =pmf "%s already dying\n"
kernel/module/main.c:1345 [main]simplify_symbols =pmf "Common symbol: %s\n"
kernel/module/main.c:1353 [main]simplify_symbols =pmf "Absolute
symbol: 0x%08lx\n"
kernel/module/main.c:1508 [main]__layout_sections =pmf "\t%s\n"
kernel/module/main.c:1526 [main]layout_sections =pmf "Core section
allocation order:\n"
kernel/module/main.c:1529 [main]layout_sections =pmf "Init section
allocation order:\n"
kernel/module/main.c:2168 [main]move_module =pmf "final section addresses:\n"
kernel/module/main.c:2183 [main]move_module =pmf "\t0x%lx %s\n"
kernel/module/main.c:2921 [main]__do_sys_init_module =pmf
"init_module: umod=%p, len=%lu, uargs=%p\n"
kernel/module/main.c:2942 [main]__do_sys_finit_module =pmf
"finit_module: fd=%d, uargs=%p, flags=%i\n"
drivers/base/power/main.c:135 [main]device_pm_add =pmf "Adding info for %s:%s\n"
drivers/base/power/main.c:156 [main]device_pm_remove =pmf "Removing
info for %s:%s\n"
drivers/base/power/main.c:175 [main]device_pm_move_before =pmf "Moving
%s:%s before %s:%s\n"
drivers/base/power/main.c:189 [main]device_pm_move_after =pmf "Moving
%s:%s after %s:%s\n"
drivers/base/power/main.c:202 [main]device_pm_move_last =pmf "Moving
%s:%s to end of list\n"
drivers/base/power/main.c:441 [main]pm_dev_dbg =pmf "PM: %s%s%s driver
flags: %x\n"
drivers/base/power/main.c:467 [main]dpm_show_time =pmf "%s%s%s of
devices %s after %ld.%03ld msecs\n"
bash-5.2#

Basically, KBUILD_MODNAME appears to get basename,
not something specific, set in some config (such as pm, or module)
or cleverly picked out of the path (power is possible)

this compromises the clarity of  dyndbg's module keyword

is some heuristic possible to improve this,
with a manual setting to fix when heuristic fails ?

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ