Re: [PATCH] OpenVPN: Prevent internal server error cause of bad header wrapper

[ummeegge <ummeegge@ipfire.org>]

[-- Attachment #1: Type: text/plain, Size: 1851 bytes --]

Hi Michael,

Am Mittwoch, den 04.07.2018, 14:59 +0100 schrieb Michael Tremer:

> > 
> > You mean to mark the first patch as v1 ? In that case i need to
> > setup
> > the old patch again as a new one and send it as answer to the v2
> > patch.
> 
> No, not as v1, but in Patchwork, when you log in manually, you can
> set a patch
> as superseeded. It is a bit annoying to do this manually, but I do
> not know
> about any better way.
OK, new ways :-). But i do have currently no access with my
credentials.


> 
> > > > 
> > > > P.S. I do have some more OpenVPN patches (extensions no bugs),
> > > > should i commit
> > > > some more or should we wait until the next release ?
> > > 
> > > What are those?
> > 
> > Wanted to finish the 2.4 OpenVPN project in the course which we did
> > discussed some time ago. So i thought about this order:
> > 
> > 1) Automatic cipher negotiation for RWs only (checkbox in advanced
> > section)
> 
> Isn't that something you would always want?
Might be a good opportunity for people with lot´s of clients and old
configuration files but an updated OpenVPN client. No new config
transfer is needed in that case but AES-GCM can nevertheless be used,
if too old (< 2.3.x), the before configured algorithms will be used.


> 
> > 2) tls-crypt for N2N only (checkbox in N2N main menu).
> > 3) LZ4 compression possibility for N2N and RW (menu with possiblity
> > for
> > none, lzo, lz4v2)
> 
> Yes, that should be a dropdown then instead of a checkbox.
Done already.

> 
> > 4) Clean up ovpnmain.cgi from mtu-discovery since there are some
> > old
> > code blocks left.
> 
> Okay, cool.
> 
> > There is more but to get the old list shorter for the first.
> 
> I guess it is best to start with the cleanup and then send in the
> other things
> one patch, or one patchset at a time.
This is how we do it.

Best,


Erik