[PATCH] printk: finalize records with trailing newlines

[PATCH] printk: finalize records with trailing newlines

[John Ogness]

Any record with a trailing newline (LOG_NEWLINE flag) cannot
be continued because the newline has been stripped and will
not be visible if the message is appended. This was already
handled correctly when committing in log_output() but was
not handled correctly when committing in log_store().

Fixes: f5f022e53b87 ("printk: reimplement log_cont using record extension")
Signed-off-by: John Ogness <john.ogness@linutronix.de>
---
 This is an important bugfix for continuous messages and should
 be part of the 5.10 release.

 If not applied, newlines will vanish when concatenating
 continuous with trailing newlines.

 kernel/printk/printk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index fe64a49344bf..bc1e3b5a97bd 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -528,8 +528,8 @@ static int log_store(u32 caller_id, int facility, int level,
 	if (dev_info)
 		memcpy(&r.info->dev_info, dev_info, sizeof(r.info->dev_info));
 
-	/* insert message */
-	if ((flags & LOG_CONT) || !(flags & LOG_NEWLINE))
+	/* A message without a trailing newline can be continued. */
+	if (!(flags & LOG_NEWLINE))
 		prb_commit(&e);
 	else
 		prb_final_commit(&e);
-- 
2.20.1

Re: [PATCH] printk: finalize records with trailing newlines

[Kefeng Wang]

On 2020/11/26 19:48, John Ogness wrote:
> Any record with a trailing newline (LOG_NEWLINE flag) cannot
> be continued because the newline has been stripped and will
> not be visible if the message is appended. This was already
> handled correctly when committing in log_output() but was
> not handled correctly when committing in log_store().
>
> Fixes: f5f022e53b87 ("printk: reimplement log_cont using record extension")
> Signed-off-by: John Ogness <john.ogness@linutronix.de>
> ---
>   This is an important bugfix for continuous messages and should
>   be part of the 5.10 release.
>
>   If not applied, newlines will vanish when concatenating
>   continuous with trailing newlines.
Thanks for you quick fix.

Reported-and-tested-by:  Kefeng Wang <wangkefeng.wang@huawei.com>

>   kernel/printk/printk.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
> index fe64a49344bf..bc1e3b5a97bd 100644
> --- a/kernel/printk/printk.c
> +++ b/kernel/printk/printk.c
> @@ -528,8 +528,8 @@ static int log_store(u32 caller_id, int facility, int level,
>   	if (dev_info)
>   		memcpy(&r.info->dev_info, dev_info, sizeof(r.info->dev_info));
>   
> -	/* insert message */
> -	if ((flags & LOG_CONT) || !(flags & LOG_NEWLINE))
> +	/* A message without a trailing newline can be continued. */
> +	if (!(flags & LOG_NEWLINE))
>   		prb_commit(&e);
>   	else
>   		prb_final_commit(&e);

Re: [PATCH] printk: finalize records with trailing newlines

[Petr Mladek]

On Thu 2020-11-26 20:32:18, Kefeng Wang wrote:
> 
> On 2020/11/26 19:48, John Ogness wrote:
> > Any record with a trailing newline (LOG_NEWLINE flag) cannot
> > be continued because the newline has been stripped and will
> > not be visible if the message is appended. This was already
> > handled correctly when committing in log_output() but was
> > not handled correctly when committing in log_store().
> > 
> > Fixes: f5f022e53b87 ("printk: reimplement log_cont using record extension")
> > Signed-off-by: John Ogness <john.ogness@linutronix.de>
> > ---
> >   This is an important bugfix for continuous messages and should
> >   be part of the 5.10 release.
> > 
> >   If not applied, newlines will vanish when concatenating
> >   continuous with trailing newlines.
> Thanks for you quick fix.
> 
> Reported-and-tested-by:  Kefeng Wang <wangkefeng.wang@huawei.com>

The patch is committed in printk/linux.git, branch
for-5.10-pr_cont-fixup.

I do not have much experience with pushing fixes this late in the
release cycle.

Normally, I would prefer to give it few days in linux-next. But
the fix is quite trivial. And it would be better to have it
in 10-rc6 rather than in rc7 that might be the last rc before
release.

Thanks a lot for report, fix, and testing.

Best Regards,
Petr