SYSLOG_ACTION_READ_ALL question

SYSLOG_ACTION_READ_ALL question

[William Douglas]

Hello all,

I'm trying to figure out why the do_syslog code in kernel/printk.c is using put_user instead of copy_to_user.

There is even a comment about how because of sleep, the buffer is read from back to front to minimize the impact of printk overwriting the syslog section we are using.  That wouldn't be necessary if a temporary copy was made and then written out with copy_to_user.

Does that make sense or am I missing something?


Thanks,
William

SYSLOG_ACTION_READ_ALL question

[william douglas]

Hello all,

I'm trying to figure out why the do_syslog code in kernel/printk.c is
using put_user instead of copy_to_user.

There is even a comment about how because of sleep, the buffer is read
from back to front to minimize the impact of printk overwriting the
syslog section we are using.  That wouldn't be necessary if a
temporary copy was made and then written out with copy_to_user.

Does that make sense or am I missing something?


Thanks,
William

SYSLOG_ACTION_READ_ALL question

[Jeff Haran]

> -----Original Message-----
> From: kernelnewbies-bounces at kernelnewbies.org [mailto:kernelnewbies-
> bounces at kernelnewbies.org] On Behalf Of william douglas
> Sent: Thursday, August 11, 2011 1:23 PM
> To: kernelnewbies at kernelnewbies.org
> Subject: SYSLOG_ACTION_READ_ALL question
> 
> Hello all,
> 
> I'm trying to figure out why the do_syslog code in kernel/printk.c is
> using put_user instead of copy_to_user.
> 

I *THINK* this has something to do with copy_to_user not being
well-behaved in other than process context and since printk() gets
called in all contexts, it wouldn't be a safe operation.
But don't take my word for it.

SYSLOG_ACTION_READ_ALL question

[william douglas]

On Thu, Aug 11, 2011 at 1:50 PM, Jeff Haran <jharan@bytemobile.com> wrote:
>> -----Original Message-----
>> From: kernelnewbies-bounces at kernelnewbies.org [mailto:kernelnewbies-
>> bounces at kernelnewbies.org] On Behalf Of william douglas
>> Sent: Thursday, August 11, 2011 1:23 PM
>> To: kernelnewbies at kernelnewbies.org
>> Subject: SYSLOG_ACTION_READ_ALL question
>>
>> Hello all,
>>
>> I'm trying to figure out why the do_syslog code in kernel/printk.c is
>> using put_user instead of copy_to_user.
>>
>
> I *THINK* this has something to do with copy_to_user not being
> well-behaved in other than process context and since printk() gets
> called in all contexts, it wouldn't be a safe operation.
> But don't take my word for it.
>
>

Well from what I understand, copy_to_user doesn't work in interrupt
context for example as there is no user space necessarily to copy to.
In this case we have a __user buffer to copy back to and
__copy_to_user looks sane for the work load so I'm still unsure.

SYSLOG_ACTION_READ_ALL question

[Mulyadi Santosa]

On Fri, Aug 12, 2011 at 03:22, william douglas
<william.r.douglas@gmail.com> wrote:
> Hello all,
>
> I'm trying to figure out why the do_syslog code in kernel/printk.c is
> using put_user instead of copy_to_user.
>
> There is even a comment about how because of sleep, the buffer is read
> from back to front to minimize the impact of printk overwriting the
> syslog section we are using. ?That wouldn't be necessary if a
> temporary copy was made and then written out with copy_to_user.
>
> Does that make sense or am I missing something?

looking at lxr.linux.no, right now I got impression that put_user is
somewhat a wrapper for copy_to_user. What I quickly check is that
put_user employ checking like might_sleep() and might_fault()...might
be handy in certain situation IMHO

-- 
regards,

Mulyadi Santosa
Freelance Linux trainer and consultant

blog: the-hydra.blogspot.com
training: mulyaditraining.blogspot.com