linux-kernel - Re: [PATCH v2] x86: Call setup_early_printk before first call to early

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>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <z2j86802c441004072244ma8035defi49a8e729ea9fba9c@mail.gmail.com>
Date:	Wed, 7 Apr 2010 22:44:14 -0700
From:	Yinghai Lu <yinghai@...nel.org>
To:	Guenter Roeck <guenter.roeck@...csson.com>
Cc:	linux-kernel@...r.kernel.org, hpa@...or.com, mingo@...hat.com,
	x86@...nel.org
Subject: Re: [PATCH v2] x86: Call setup_early_printk before first call to 
	early_printk

On Tue, Apr 6, 2010 at 6:42 PM, Guenter Roeck
<guenter.roeck@...csson.com> wrote:
> Current code calls setup_early_printk() too late to catch the initial calls
> to early_printk(). This results in those early messages not to be printed on a
> serial console. Also, if there is no VGA device in the system, it will cause
> bad writes into VGA memory space. This can cause problems if the VGA memory
> space is used by another device in such a system.
>
> Fix the problem by calling setup_early_printk() immediately after the boot
> command line is available, and by preventing early_printk() to print anything
> before setup_early_printk() was executed.
>
> Also removed conditional output "Kernel alive".
>
> Signed-off-by: Guenter Roeck <guenter.roeck@...csson.com>
> ---
> The first version of this patch used the headline
>    x86: Do not write to VGA memory space if CONFIG_VGA_CONSOLE is undefined
> I changed the headline to reflect its expanded scope.
>
> I removed the "Kernel alive" message because I don't see how console_loglevel
> can ever be equal to 10 when the code is executed. An alternative might be
> (if that possibility exists after all) to move it after the call to
> setup_early_printk().
>
>  arch/x86/include/asm/setup.h   |    2 ++
>  arch/x86/kernel/early_printk.c |    5 ++++-
>  arch/x86/kernel/head32.c       |    4 ++++
>  arch/x86/kernel/head64.c       |    7 ++++---
>  4 files changed, 14 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/include/asm/setup.h b/arch/x86/include/asm/setup.h
> index 86b1506..83847af 100644
> --- a/arch/x86/include/asm/setup.h
> +++ b/arch/x86/include/asm/setup.h
> @@ -93,6 +93,8 @@ void *extend_brk(size_t size, size_t align);
>                        : : "i" (sz));                                  \
>        }
>
> +int __init setup_early_printk(char *buf);
> +
>  #ifdef __i386__
>
>  void __init i386_start_kernel(void);
> diff --git a/arch/x86/kernel/early_printk.c b/arch/x86/kernel/early_printk.c
> index b9c830c..873f1cd 100644
> --- a/arch/x86/kernel/early_printk.c
> +++ b/arch/x86/kernel/early_printk.c
> @@ -170,6 +170,9 @@ asmlinkage void early_printk(const char *fmt, ...)
>        int n;
>        va_list ap;
>
> +       if (!early_console_initialized)
> +               return;
> +

if someone specify "earlyprintk", early_console will still point to
early_vga_console...

prefer to set early_console to NULL,

and here change to

if (!early_console) {
    printk(fmt, vargs)
    return;
}


>        va_start(ap, fmt);
>        n = vscnprintf(buf, sizeof(buf), fmt, ap);
>        early_console->write(early_console, buf, n);
> @@ -191,7 +194,7 @@ static inline void early_console_register(struct console *con, int keep_early)
>        register_console(early_console);
>  }
>
> -static int __init setup_early_printk(char *buf)
> +int __init setup_early_printk(char *buf)
>  {
>        int keep;
>
> diff --git a/arch/x86/kernel/head32.c b/arch/x86/kernel/head32.c
> index b2e2460..e2e9c2b 100644
> --- a/arch/x86/kernel/head32.c
> +++ b/arch/x86/kernel/head32.c
> @@ -30,6 +30,10 @@ static void __init i386_default_early_setup(void)
>
>  void __init i386_start_kernel(void)
>  {
> +#ifdef CONFIG_EARLY_PRINTK
> +       setup_early_printk(boot_command_line);
> +#endif
> +
>  #ifdef CONFIG_X86_TRAMPOLINE
>        /*
>         * But first pinch a few for the stack/trampoline stuff
> diff --git a/arch/x86/kernel/head64.c b/arch/x86/kernel/head64.c
> index 7147143..1b4d6b2 100644
> --- a/arch/x86/kernel/head64.c
> +++ b/arch/x86/kernel/head64.c
> @@ -88,9 +88,6 @@ void __init x86_64_start_kernel(char * real_mode_data)
>        }
>        load_idt((const struct desc_ptr *)&idt_descr);
>
> -       if (console_loglevel == 10)
> -               early_printk("Kernel alive\n");
> -
>        x86_64_start_reservations(real_mode_data);
>  }
>
> @@ -98,6 +95,10 @@ void __init x86_64_start_reservations(char *real_mode_data)
>  {
>        copy_bootdata(__va(real_mode_data));
>
> +#ifdef CONFIG_EARLY_PRINTK
> +       setup_early_printk(boot_command_line);
> +#endif

you may need to search "earlyprintk" and cut the "earlyprintk=....."
string out at first, and then use that
string to call setup_early_printk()

YH
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/