From: Petr Mladek <pmladek@suse.com>
To: John Ogness <john.ogness@linutronix.de>
Cc: Sergey Senozhatsky <senozhatsky@chromium.org>,
Steven Rostedt <rostedt@goodmis.org>,
Marcos Paulo de Souza <mpdesouza@suse.com>,
Chris Down <chris@chrisdown.name>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 3/8] printk: Separate code for adding/updating preferred console metadata
Date: Thu, 19 Feb 2026 13:46:34 +0100 [thread overview]
Message-ID: <aZcF9Rya9OUShV8c@pathway.suse.cz> (raw)
In-Reply-To: <87v7fwdbqw.fsf@jogness.linutronix.de>
On Mon 2026-02-16 15:11:59, John Ogness wrote:
> Hi Petr,
>
> On 2026-02-06, Petr Mladek <pmladek@suse.com> wrote:
> > diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
> > index 3f856a438e74..ee57c7ac9d02 100644
> > --- a/kernel/printk/printk.c
> > +++ b/kernel/printk/printk.c
> > @@ -2491,18 +2491,82 @@ asmlinkage __visible void early_printk(const char *fmt, ...)
> > }
> > #endif
> >
> > -static void set_user_specified(struct preferred_console *pc, bool user_specified)
> > +static int update_preferred_console(int i, const char *name, const short idx,
>
> Perhaps @i should be unsigned in order to guarantee no possibility of
> negative array indexing.
>
> It would need to be defined that way in __add_preferred_console() as well.
Makes sense. I'll do it in v2.
> > + const char *devname, char *options,
> > + char *brl_options, bool user_specified)
> > {
> > - if (!user_specified)
> > - return;
> > + struct preferred_console *pc;
> > +
> > + if (i >= MAX_PREFERRED_CONSOLES)
> > + return -E2BIG;
> > +
> > + pc = &preferred_consoles[i];
> > +
> > + if (!name && !devname)
> > + return -EINVAL;
> > +
> > + if (devname) {
> > + /*
> > + * A valid console name and index will get assigned when
> > + * a matching device gets registered.
> > + */
> > + if (name) {
> > + pr_err("Adding a preferred console devname with a hard-coded console name: %s, %s\n",
> > + devname, name);
> > + return -EINVAL;
> > + }
> > + if (idx != -1) {
> > + pr_err("Adding a preferred console devname with a hard-coded index: %s, %d\n",
> > + devname, idx);
> > + return -EINVAL;
> > + }
> > +
> > + if (!pc->devname[0]) {
> > + strscpy(pc->devname, devname);
> > + pc->index = idx;
> > + } else if (strcmp(pc->devname, devname) != 0) {
> > + pr_err("Updating a preferred console with an invalid devname: %s vs. %s\n",
> > + pc->devname, devname);
> > + return -EINVAL;
> > + }
> > + }
> > +
> > + if (name) {
> > + /* A console name must be defined with a valid index. */
> > + if (idx < 0) {
> > + pr_err("Adding a preferred console with an invalid index: %s, %d\n",
> > + name, idx);
> > + return -EINVAL;
> > + }
> > +
> > + if (!pc->name[0]) {
> > + strscpy(pc->name, name);
> > + pc->index = idx;
> > + } else if (strcmp(pc->name, name) != 0 || pc->index != idx) {
> > + pr_err("Updating a preferred console with an invalid name or index: %s%d vs. %s%d\n",
> > + pc->name, pc->index, name, idx);
> > + return -EINVAL;
> > + }
> > + }
> > +
> > + if (!pc->options || (user_specified && options))
> > + pc->options = options;
> > +
> > + braille_update_options(pc, brl_options);
> > +
> > + if (!brl_options)
> > + preferred_dev_console = i;
> >
> > /*
> > * @c console was defined by the user on the command line.
> > * Do not clear when added twice also by SPCR or the device tree.
> > */
> > - pc->user_specified = true;
> > - /* At least one console defined by the user on the command line. */
> > - console_set_on_cmdline = 1;
> > + if (user_specified) {
> > + pc->user_specified = true;
> > + console_set_on_cmdline = 1;
> > + }
> > +
> > + return 0;
> > }
> >
> > static int __add_preferred_console(const char *name, const short idx,
>
> There are a lot of rules to arguments of __add_preferred_console(). Can
> you add some comment above the __add_preferred_console() function
> definition about these rules? I think this patch is an appropriate place
> to do that since the rules are quite visible with your changes to
> update_preferred_console(). For example, mentioning:
>
> - required: either @name and a valid @idx OR @devname and idx=-1
> - specify @brl_options if it is a Braille console
> - Braille consoles will never be associated with /dev/console
>
> And a simple description like "Add a new preferred console or update the
> options of an already registered preferred console."
I am preparing v2 and added this:
/** update_preferred_console - Update a given entry in the preferred_consoles[]
* table.
* @i: index of the entry in @preferred_consoles table which should get updated.
* @name: The name of the preferred console driver.
* @idx: Preferred console index, e.g. port number.
* @devname: The name of the preferred physical device.
* @options: Options used when setting up the console driver.
* @brl_options: Options used when setting up the console driver
* as a braille console.
* @user_specified: True if preferred via the kernel command line.
*
* The function ensures that the given values are consistent. Also
* it updates some global variables which are used to make the right
* decisions in register_console().
*
* Rules:
*
* 1. Either @name and valid @idx OR @devname and @idx=-1 are allowed.
* Note that a valid @name and @idx will get assigned later when
* @devname matches during the device initialization.
* 2. Specify @brl_options if the console should be enabled as
* a Braille console [*]
* 3. Only matching entries can be updated.
* 4. @options passed via the command line are used when the same
* console is is preferred also by some platform-specific code.
*
* [*] Braille console is using the mechanism for registering consoles
* but it is very special. It is primary used for an user interaction
* with the system. It neither gets printk() messages nor is
* associated with /dev/cosnole.
*/
Best Regards,
Petr
next prev parent reply other threads:[~2026-02-19 12:46 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-02-06 16:49 [PATCH 0/8] printk: Clean up preferred console handling Petr Mladek
2026-02-06 16:49 ` [PATCH 1/8] printk: Rename struct console_cmdline to preferred_console Petr Mladek
2026-02-19 14:40 ` Chris Down
2026-02-19 18:34 ` Marcos Paulo de Souza
2026-02-06 16:49 ` [PATCH 2/8] printk: Rename preferred_console to preferred_dev_console Petr Mladek
2026-02-19 14:41 ` Chris Down
2026-02-19 18:37 ` Marcos Paulo de Souza
2026-02-06 16:49 ` [PATCH 3/8] printk: Separate code for adding/updating preferred console metadata Petr Mladek
2026-02-16 14:05 ` John Ogness
2026-02-19 12:46 ` Petr Mladek [this message]
2026-02-19 14:06 ` John Ogness
2026-02-19 14:48 ` Chris Down
2026-02-19 16:51 ` Petr Mladek
2026-02-06 16:49 ` [PATCH 4/8] printk: Cleanup _braille_(un)register_console() wrappers Petr Mladek
2026-02-19 14:49 ` Chris Down
2026-02-19 18:50 ` Marcos Paulo de Souza
2026-02-06 16:49 ` [PATCH 5/8] printk: Try to register each console as Braille first Petr Mladek
2026-02-19 14:59 ` Chris Down
2026-02-19 16:59 ` Petr Mladek
2026-02-20 4:52 ` Chris Down
2026-02-20 11:43 ` Petr Mladek
2026-02-06 16:50 ` [PATCH 6/8] printk: Do not set Braille console as preferred_console Petr Mladek
2026-02-16 16:07 ` John Ogness
2026-02-19 14:55 ` Petr Mladek
2026-02-19 15:35 ` John Ogness
2026-02-19 15:03 ` Chris Down
2026-02-06 16:50 ` [PATCH 7/8] printk: Handle pre-enabled consoles directly in register_console() Petr Mladek
2026-02-19 15:03 ` Chris Down
2026-02-06 16:50 ` [PATCH 8/8] printk: Try enable preferred consoles only when there are any Petr Mladek
2026-02-19 15:16 ` Chris Down
2026-02-17 8:56 ` [PATCH 0/8] printk: Clean up preferred console handling John Ogness
2026-02-19 15:20 ` Chris Down
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=aZcF9Rya9OUShV8c@pathway.suse.cz \
--to=pmladek@suse.com \
--cc=chris@chrisdown.name \
--cc=john.ogness@linutronix.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mpdesouza@suse.com \
--cc=rostedt@goodmis.org \
--cc=senozhatsky@chromium.org \
/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