[systemd-devel] [PATCH] journalctl: support /usr/bin/nginx, etc

[shawn]

On Wed, 2012-05-30 at 14:06 -0700, shawn wrote: 
> Regarding the commited version:
> 
> <snip>
> +                        if (S_ISREG(st.st_mode) && (0111 & st.st_mode))
> {
> </snip>
> 
> isn't it conceivable that some programs which would log would only be
> executable by the owner (generally root)? because they might be suid
> root? or then so we just revert to requiring the full _EXE= syntax?

errr, executable by owner+group, but not other, and suid whoever. 
> 
> On Wed, 2012-05-30 at 20:18 +0200, Kay Sievers wrote: 
> > On Wed, May 30, 2012 at 7:17 PM, shawn <shawnlandden at gmail.com> wrote:
> > > With relative path handling.
> > 
> > Should probably do a stat() on the file in the fs and check for the
> > executable bit?
> > 
> > > I'm not sure what is meant by the other selectors, like /dev/sda
> > 
> > Will be based on this:
> >   http://lwn.net/Articles/492125/
> >   http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=e11fea92e13fb91c50bacca799a6131c81929986
> >   http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=c4e00daaa96d3a0786f1f4fe6456281c60ef9a16
> > 
> > We will kill the current /proc/kmsg code and the ImportKernel option
> > and use the /dev/kmsg interface which will allow us to properly
> > recognize messages from the kernel.
> > 
> > We can stat() /dev/sda find the dev_t and look up the device in the journal.
> > 
> > The magic --device= strings in the TODO are explained in the
> > dev_printk() patch above.
> > 
> > Kay
> 
> 


-- 
-Shawn Landden