On 11 May 2018 at 18:21, Rowland Penny via samba-technical < samba-technical at lists.samba.org> wrote:> On Fri, 11 May 2018 17:56:44 +0200 > "Timur I. Bakeyev" <timur at com.bat.ru> wrote: > > > On 11 May 2018 at 17:00, Rowland Penny via samba-technical < > > samba-technical at lists.samba.org> wrote: > > > > > On Fri, 11 May 2018 16:46:55 +0200 > > > "Timur I. Bakeyev" <timur at freebsd.org> wrote: > > > > static const struct enum_list enum_log_facilities[] = { > > > > { LOG_USER, "USER" }, > > > > { LOG_LOCAL0, "LOCAL0" }, > > > > { LOG_LOCAL1, "LOCAL1" }, > > > > { LOG_LOCAL2, "LOCAL2" }, > > > > { LOG_LOCAL3, "LOCAL3" }, > > > > { LOG_LOCAL4, "LOCAL4" }, > > > > { LOG_LOCAL5, "LOCAL5" }, > > > > { LOG_LOCAL6, "LOCAL6" }, > > > > { LOG_LOCAL7, "LOCAL7" }, > > > > > > > > > > > Clearly, the only possible facilities are LOCAL* and USER. That > > > > raises the question why it's not mentioned in the man page and > > > > why is this limitation in place? > > > > > > Good questions and I don't know the answers to them, the other > > > question is, why does the man page say you can use any syslog > > > facility when obviously you cannot ? > > > > > > > > Everyone lies! (c) Seems it's a day of quotes for me :) > > > > something has to be fixed here - either man page or the code. I don't > > see much of the reason to limit facilities to those low-end, general > > purpose ones. AUTH, actually, makes sense for audit logs. > > > > Question to developers. > > Well, I can prepare a patch to fix the man page, but would it be better > to fix the code to do what the man page says it can do now ? > >Fixing it in both ways is easy, but it would be nice to hear from the developers, why did they choose such a limited set of syslog facilities. Timur.
On Fri, May 11, 2018 at 09:56:47PM +0200, Timur I. Bakeyev via samba wrote:> > > Fixing it in both ways is easy, but it would be nice to hear from the > developers, why did they choose such a limited set of syslog facilities.I don't know. The limited set was done a long time ago (2005) by Deryck (no longer a Team member) in commit 435295f1840aa8bd4c04f20a2348c6d701c6b7dc so I think we've lost the institutional memory here. How about we fix the code to do what the man page says ? Jeremy.
On 11 May 2018 at 22:17, Jeremy Allison <jra at samba.org> wrote:> On Fri, May 11, 2018 at 09:56:47PM +0200, Timur I. Bakeyev via samba wrote: > > > > > Fixing it in both ways is easy, but it would be nice to hear from the > > developers, why did they choose such a limited set of syslog facilities. > > I don't know. The limited set was done a long time ago > (2005) by Deryck (no longer a Team member) in commit > 435295f1840aa8bd4c04f20a2348c6d701c6b7dc so I think > we've lost the institutional memory here. > > How about we fix the code to do what the man page > says ?I'm all for it. I don't see any security implications of allowing full set of supported syslog facilities. Just to keep in mind that the same code is shared by vfs_audit, vfs_ext_audit and vfs_full_audit, so all of them have to be fixed. With regards, Timur.
On Fri, 11 May 2018 13:17:52 -0700 Jeremy Allison <jra at samba.org> wrote:> On Fri, May 11, 2018 at 09:56:47PM +0200, Timur I. Bakeyev via samba > wrote: > > > > > Fixing it in both ways is easy, but it would be nice to hear from > > the developers, why did they choose such a limited set of syslog > > facilities. > > I don't know. The limited set was done a long time ago > (2005) by Deryck (no longer a Team member) in commit > 435295f1840aa8bd4c04f20a2348c6d701c6b7dc so I think > we've lost the institutional memory here. > > How about we fix the code to do what the man page > says ? > > Jeremy.What a good idea and thank you for volunteering ;-) Rowland