On Wed, Aug 07, 2024 at 06:22:37PM +0300, Vitaliy Makkoveev wrote:
> On Wed, Aug 07, 2024 at 04:51:44PM +0200, Walter Alejandro Iglesias wrote:
> > 
> > In my case it was exactly the opposite, I also use an old machine.
> > Firefox runs slow and torrent applications (transmission-gtk, rtorrent)
> > crash without any message explaining why (network UDP issue?).  No
> > message appears in the system logs either.
> > 
> 
> Hello,
> 
> Can you provide more information? When it was started? The major UDP
> changes were made at Jul 12, Jul 20 and Jul 26.
> 

Now I did the following.  I updated src, I compiled and installed the
kernel and reboot.  The problem was still there.  Then I downloaded this
diff:

   https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/netinet/in_proto.c.diff?r1=1.106&r2=1.107&sortby=date

I reverted the diff in the source, recompiled the kernel, reboot, and
the problem is still there.  So, my guessing about UDP was wrong.


What led me to think it was a network problem was the following.

I've been using transmission-gtk every day without problems since a long
time.  I don't use rtorrent, I downloaded it to find out if it was a bug
in transmission (I don't remember if there was some update in ports
lately affecting transmission).  Both applications quit without message
after about ~10-20 seconds after starting the downloads.  It can't be a
coincidence.  I also tried to download torrents with lftp, it doesn't
crash but it goes so slow that it takes it forever to download a few
megabytes.  The problem is not in my network I can't reproduce it under
Linux.  I also have OpenBSD installed in my laptop where I can also
reproduce the transmission crash with this snapshot:

  OpenBSD 7.5-current (GENERIC.MP) #230: Mon Aug  5 19:35:34 MDT 2024
    deraadt@amd64.openbsd.org:/usr/src/sys/arch/amd64/compile/GENERIC.MP

I can't tell you exactly after which snapshot it started, but since I
use transmission-gtk regularly, I don't think we need to go too far back
in time.  If you think that I'm right in this could be a system problem
I'll try other snapshots until I find out which one introduced the
problem and I'll let you know.


-- 
Walter