| 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: <d429b1af-b7c2-42d6-85c6-ce318d058cb6@paulmck-laptop>
Date: Mon, 8 Apr 2024 17:53:02 -0700
From: "Paul E. McKenney" <paulmck@...nel.org>
To: Masami Hiramatsu <mhiramat@...nel.org>
Cc: linux-kernel@...r.kernel.org, linux-trace-kernel@...r.kernel.org,
linux-fsdevel@...r.kernel.org,
Zhenhua Huang <quic_zhenhuah@...cinc.com>
Subject: Re: [PATCH fs/proc/bootconfig] remove redundant comments from
/proc/bootconfig
On Tue, Apr 09, 2024 at 09:25:40AM +0900, Masami Hiramatsu wrote:
> On Mon, 8 Apr 2024 12:18:19 -0700
> "Paul E. McKenney" <paulmck@...nel.org> wrote:
>
> > On Sat, Apr 06, 2024 at 11:11:08AM +0900, Masami Hiramatsu wrote:
> > > On Thu, 4 Apr 2024 21:25:41 -0700
> > > "Paul E. McKenney" <paulmck@...nel.org> wrote:
> > >
> > > > On Fri, Apr 05, 2024 at 11:57:45AM +0900, Masami Hiramatsu wrote:
> > > > > On Fri, 5 Apr 2024 10:23:24 +0900
> > > > > Masami Hiramatsu (Google) <mhiramat@...nel.org> wrote:
> > > > >
> > > > > > On Thu, 4 Apr 2024 10:43:14 -0700
> > > > > > "Paul E. McKenney" <paulmck@...nel.org> wrote:
> > > > > >
> > > > > > > On Thu, Apr 04, 2024 at 08:55:22AM +0900, Masami Hiramatsu wrote:
> > > > > > > > On Wed, 3 Apr 2024 12:16:28 -0700
> > > > > > > > "Paul E. McKenney" <paulmck@...nel.org> wrote:
> > > > > > > >
> > > > > > > > > commit 717c7c894d4b ("fs/proc: Add boot loader arguments as comment to
> > > > > > > > > /proc/bootconfig") adds bootloader argument comments into /proc/bootconfig.
> > > > > > > > >
> > > > > > > > > /proc/bootconfig shows boot_command_line[] multiple times following
> > > > > > > > > every xbc key value pair, that's duplicated and not necessary.
> > > > > > > > > Remove redundant ones.
> > > > > > > > >
> > > > > > > > > Output before and after the fix is like:
> > > > > > > > > key1 = value1
> > > > > > > > > *bootloader argument comments*
> > > > > > > > > key2 = value2
> > > > > > > > > *bootloader argument comments*
> > > > > > > > > key3 = value3
> > > > > > > > > *bootloader argument comments*
> > > > > > > > > ...
> > > > > > > > >
> > > > > > > > > key1 = value1
> > > > > > > > > key2 = value2
> > > > > > > > > key3 = value3
> > > > > > > > > *bootloader argument comments*
> > > > > > > > > ...
> > > > > > > > >
> > > > > > > > > Fixes: 717c7c894d4b ("fs/proc: Add boot loader arguments as comment to /proc/bootconfig")
> > > > > > > > > Signed-off-by: Zhenhua Huang <quic_zhenhuah@...cinc.com>
> > > > > > > > > Signed-off-by: Paul E. McKenney <paulmck@...nel.org>
> > > > > > > > > Cc: Masami Hiramatsu <mhiramat@...nel.org>
> > > > > > > > > Cc: <linux-trace-kernel@...r.kernel.org>
> > > > > > > > > Cc: <linux-fsdevel@...r.kernel.org>
> > > > > > > >
> > > > > > > > OOps, good catch! Let me pick it.
> > > > > > > >
> > > > > > > > Acked-by: Masami Hiramatsu (Google) <mhiramat@...nel.org>
> > > > > > >
> > > > > > > Thank you, and I have applied your ack and pulled this into its own
> > > > > > > bootconfig.2024.04.04a.
> > > > > > >
> > > > > > > My guess is that you will push this via your own tree, and so I will
> > > > > > > drop my copy as soon as yours hits -next.
> > > > > >
> > > > > > Thanks! I would like to make PR this soon as bootconfig fixes for v6.9-rc2.
> > > > >
> > > > > Hmm I found that this always shows the command line comment in
> > > > > /proc/bootconfig even without "bootconfig" option.
> > > > > I think that is easier for user-tools but changes the behavior and
> > > > > a bit redundant.
> > > > >
> > > > > We should skip showing this original argument comment if bootconfig is
> > > > > not initialized (no "bootconfig" in cmdline) as it is now.
> > > >
> > > > So something like this folded into that patch?
> > >
> > > Hm, I expected just checking it in the loop as below.
> > >
> > > ------------------------------------------------------------------------
> > > diff --git a/fs/proc/bootconfig.c b/fs/proc/bootconfig.c
> > > index e5635a6b127b..98e0780f7e07 100644
> > > --- a/fs/proc/bootconfig.c
> > > +++ b/fs/proc/bootconfig.c
> > > @@ -27,6 +27,7 @@ static int __init copy_xbc_key_value_list(char *dst, size_t size)
> > > {
> > > struct xbc_node *leaf, *vnode;
> > > char *key, *end = dst + size;
> > > + bool empty = true;
> > > const char *val;
> > > char q;
> > > int ret = 0;
> > > @@ -62,8 +63,9 @@ static int __init copy_xbc_key_value_list(char *dst, size_t size)
> > > break;
> > > dst += ret;
> > > }
> > > + empty = false;
> > > }
> > > - if (ret >= 0 && boot_command_line[0]) {
> > > + if (!empty && ret >= 0 && boot_command_line[0]) {
> > > ret = snprintf(dst, rest(dst, end), "# Parameters from bootloader:\n# %s\n",
> > > boot_command_line);
> > > if (ret > 0)
> > >
> > > ------------------------------------------------------------------------
> > >
> > > The difference is checking "bootconfig" cmdline option or checking
> > > the "bootconfig" is actually empty. So the behaviors are different
> > > when the "bootconfig" is specified but there is no bootconfig data.
> >
> > Ah, understood, the point is to avoid the comment in cases where its
> > content would be identical to /proc/cmdline.
> >
> > > Another idea is to check whether the cmdline is actually updated by
> > > bootconfig and show original one only if it is updated.
> > > (I think this fits the purpose of the original patch better.)
> > >
> > > ------------------------------------------------------------------------
> > > diff --git a/fs/proc/bootconfig.c b/fs/proc/bootconfig.c
> > > index e5635a6b127b..95d6a231210c 100644
> > > --- a/fs/proc/bootconfig.c
> > > +++ b/fs/proc/bootconfig.c
> > > @@ -10,6 +10,9 @@
> > > #include <linux/bootconfig.h>
> > > #include <linux/slab.h>
> > >
> > > +/* defined in main/init.c */
> > > +bool __init cmdline_has_extra_options(void);
> > > +
> > > static char *saved_boot_config;
> > >
> > > static int boot_config_proc_show(struct seq_file *m, void *v)
> > > @@ -63,7 +66,7 @@ static int __init copy_xbc_key_value_list(char *dst, size_t size)
> > > dst += ret;
> > > }
> > > }
> > > - if (ret >= 0 && boot_command_line[0]) {
> > > + if (cmdline_has_extra_options() && ret >= 0 && boot_command_line[0]) {
> > > ret = snprintf(dst, rest(dst, end), "# Parameters from bootloader:\n# %s\n",
> > > boot_command_line);
> > > if (ret > 0)
> > > diff --git a/init/main.c b/init/main.c
> > > index 2ca52474d0c3..881f6230ee59 100644
> > > --- a/init/main.c
> > > +++ b/init/main.c
> > > @@ -487,6 +487,11 @@ static int __init warn_bootconfig(char *str)
> > >
> > > early_param("bootconfig", warn_bootconfig);
> > >
> > > +bool __init cmdline_has_extra_options(void)
> > > +{
> > > + return extra_command_line || extra_init_args;
> > > +}
> > > +
> > > /* Change NUL term back to "=", to make "param" the whole string. */
> > > static void __init repair_env_string(char *param, char *val)
> > > {
> > > ------------------------------------------------------------------------
> >
> > This one looks good to me!
> >
> > I had to move the declaration from /fs/proc/bootconfig.c to
> > include/linux/bootconfig.h in order to avoid a build error. (The
> > build system wants the declaration and definition to be visible as
> > a cross-check.)
> >
> > Does the resulting patch below work for you?
>
> Yes, this looks good to me. BTW, shouldn't we push this with the
> previous fix (from the viewpoint of memory usage, this is a kind
> of fix for redundant memory usage)?
Makes sense to me! Let me run a few tests on it.
For whatever it is worth, I have them both queued on their own branch
on -rcu:
4facfdfaec7a ("fs/proc: remove redundant comments from /proc/bootconfig")
8d95b50c523f ("fs/proc: Skip bootloader comment if no embedded kernel parameters")
Thanx, Paul
> Thank you!
>
> >
> > Thanx, Paul
> >
> > ------------------------------------------------------------------------
> >
> > commit 8d95b50c523fba7133368650b3b5f71b169c76b5
> > Author: Masami Hiramatsu <mhiramat@...nel.org>
> > Date: Mon Apr 8 12:10:38 2024 -0700
> >
> > fs/proc: Skip bootloader comment if no embedded kernel parameters
> >
> > If the "bootconfig" kernel command-line argument was specified or if
> > the kernel was built with CONFIG_BOOT_CONFIG_FORCE, but if there are
> > no embedded kernel parameter, omit the "# Parameters from bootloader:"
> > comment from the /proc/bootconfig file. This will cause automation
> > to fall back to the /proc/cmdline file, which will be identical to the
> > comment in this no-embedded-kernel-parameters case.
> >
> > Signed-off-by: Masami Hiramatsu <mhiramat@...nel.org>
> > Signed-off-by: Paul E. McKenney <paulmck@...nel.org>
> >
> > diff --git a/fs/proc/bootconfig.c b/fs/proc/bootconfig.c
> > index e5635a6b127b0..87dcaae32ff87 100644
> > --- a/fs/proc/bootconfig.c
> > +++ b/fs/proc/bootconfig.c
> > @@ -63,7 +63,7 @@ static int __init copy_xbc_key_value_list(char *dst, size_t size)
> > dst += ret;
> > }
> > }
> > - if (ret >= 0 && boot_command_line[0]) {
> > + if (cmdline_has_extra_options() && ret >= 0 && boot_command_line[0]) {
> > ret = snprintf(dst, rest(dst, end), "# Parameters from bootloader:\n# %s\n",
> > boot_command_line);
> > if (ret > 0)
> > diff --git a/include/linux/bootconfig.h b/include/linux/bootconfig.h
> > index ca73940e26df8..e5ee2c694401e 100644
> > --- a/include/linux/bootconfig.h
> > +++ b/include/linux/bootconfig.h
> > @@ -10,6 +10,7 @@
> > #ifdef __KERNEL__
> > #include <linux/kernel.h>
> > #include <linux/types.h>
> > +bool __init cmdline_has_extra_options(void);
> > #else /* !__KERNEL__ */
> > /*
> > * NOTE: This is only for tools/bootconfig, because tools/bootconfig will
> > diff --git a/init/main.c b/init/main.c
> > index 2ca52474d0c30..881f6230ee59e 100644
> > --- a/init/main.c
> > +++ b/init/main.c
> > @@ -487,6 +487,11 @@ static int __init warn_bootconfig(char *str)
> >
> > early_param("bootconfig", warn_bootconfig);
> >
> > +bool __init cmdline_has_extra_options(void)
> > +{
> > + return extra_command_line || extra_init_args;
> > +}
> > +
> > /* Change NUL term back to "=", to make "param" the whole string. */
> > static void __init repair_env_string(char *param, char *val)
> > {
>
>
> --
> Masami Hiramatsu (Google) <mhiramat@...nel.org>
Powered by blists - more mailing lists