linux-kernel - Re: [RFC] Add Alternative Log Buffer Support for printk Messages

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] [thread-next>] [day] [month] [year] [list]

Message-Id: <20090130142825.e63ee797.akpm@linux-foundation.org>
Date:	Fri, 30 Jan 2009 14:28:25 -0800
From:	Andrew Morton <akpm@...ux-foundation.org>
To:	Grant Erickson <gerickson@...vations.com>
Cc:	linux-kernel@...r.kernel.org, wd@...x.de
Subject: Re: [RFC] Add Alternative Log Buffer Support for printk Messages

On Fri, 30 Jan 2009 14:01:49 -0800
Grant Erickson <gerickson@...vations.com> wrote:

> Your summary is precisely the intent. As I
> mentioned in check-in notice, this isn't my work per se, but a fix of prior
> unsubmitted work. Consequently, I can only speculate on the original
> implementation choices of the primary authors (Yuri Tikhonov
> <yur@...raft.com> on behalf of Denx from what I can ascertain from GIT
> logs). I've added Wolfgang in the event he can add background where
> appropriate.
> 
> I believe this to be a useful feature for the embedded system space;
> however, I am not wedded in any way to the current implementation and am
> happy to evolve it as needed to meet the needs of key stakeholders.
> 
> > Why not just copy the external buffer's contents into the normal buffer during
> > bootup?  ie: printk("%s", external_buffer).
> 
> Again, conjecturing, I suspect the goal was to ensure that the boot messages
> appear on the console or in the syslog before kernel messages of that
> particular boot cycle. Depending on the board and architecture in question,
> the kernel log buffer may have several messages in it by the time the log
> coalescing code runs. For example, on the board I tested this on:
> 
>     # dmesg
>     U-Boot 1.3.3 (Jan  9 2009 - 10:24:45)
>     POST memory PASSED
>     POST ecc PASSED
>     POST cache PASSED
>     POST ethernet PASSED
>     Using Kilauea machine description
>     log_buf=fdffc000
>     Linux version 2.6.27.7 (gerickson@...ntu-fusion) (gcc version 4.0.0
> (DENX ELDK 4.1 4.0.0)) #1 Fri Jan 9 10:30:30 PST 2009

It shouldn't be too hard to cook up some way of copying some data into
th log buffer before the kernel itself has put anything in there, if
that's all which is needed.


something similar to this?

 include/linux/kernel.h |    2 ++
 init/main.c            |    6 ++++++
 kernel/printk.c        |    8 ++++++++
 3 files changed, 16 insertions(+)

diff -puN init/main.c~a init/main.c
--- a/init/main.c~a
+++ a/init/main.c
@@ -535,11 +535,17 @@ void __init __weak thread_info_cache_ini
 {
 }
 
+void __init __weak arch_copy_boot_messages(void)
+{
+}
+
 asmlinkage void __init start_kernel(void)
 {
 	char * command_line;
 	extern struct kernel_param __start___param[], __stop___param[];
 
+	arch_copy_boot_messages();
+
 	smp_setup_processor_id();
 
 	/*
diff -puN kernel/printk.c~a kernel/printk.c
--- a/kernel/printk.c~a
+++ a/kernel/printk.c
@@ -493,6 +493,14 @@ static void emit_log_char(char c)
 		logged_chars++;
 }
 
+void emit_log_string(const char *s)
+{
+	spin_lock_irqsave(&logbuf_lock);
+	while (*s)
+		emit_log_char(*s++);
+	spin_unlock_irqsave(&logbuf_lock);
+}
+
 /*
  * Zap console related locks when oopsing. Only zap at most once
  * every 10 seconds, to leave time for slow consoles to print a
diff -puN include/linux/kernel.h~a include/linux/kernel.h
--- a/include/linux/kernel.h~a
+++ a/include/linux/kernel.h
@@ -257,6 +257,8 @@ static inline bool printk_timed_ratelimi
 
 extern int printk_needs_cpu(int cpu);
 extern void printk_tick(void);
+extern void __init arch_copy_boot_messages(void);
+extern void emit_log_string(const char *s);
 
 extern void asmlinkage __attribute__((format(printf, 1, 2)))
 	early_printk(const char *fmt, ...);
_

--
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/