XPost: linux.debian.bugs.dist   
   From: tobias@fiebig.nl   
      
   Hello Salvatore,   
      
   I would actually give reverting that one a shot first. Let me see if I   
   can quickly build a kernel without that specific patch.   
      
   With best regards,   
   Tobias   
      
   On Tue, 2026-02-10 at 13:18 +0100, Salvatore Bonaccorso wrote:   
   > Control: tags -1 + moreinfo   
   >    
   > Hi BAstien, hi Tobias,   
   >    
   > On Tue, Feb 10, 2026 at 09:57:15AM +0100, Bastien Durel wrote:   
   > > Source: linux   
   > > Version: 6.12.48+deb13-amd64   
   > > Severity: important   
   > > Tags: ipv6   
   > > X-Debbugs-Cc: debian-amd64@lists.debian.org   
   > > User: debian-amd64@lists.debian.org   
   > > Usertags: amd64   
   > >    
   > > Dear Maintainer,   
   > >    
   > > When I boot on the 6.12.48+deb13-amd64 version of the kernel, my   
   > > GRE6 tunnels stop to work, it seems that decapsulation fail   
   > >    
   > > Rolling back to 6.12.63+deb13-amd64 make them work again, so I   
   > > _think_ this is not a configuration issue.   
   > >    
   > > I have two GRE6 tunnels on which I run BGP sessions.   
   > >    
   > > Here's a capture of the external interface, filtered on ip6[40] =   
   > > 47   
   > >    
   > > 21:56:44.411270 IP6 (flowlabel 0x7759c, hlim 64, next-header   
   > > unknown (60) payload length: 92) 2001:41d0:302:2000::2144 >   
   > > 2a0c:b641:6::: DSTOPT (opt_type 0x04: len=1)(padn) GREv0, Flags   
   > > [none], length 84   
   > > 	IP6 (class 0xc0, flowlabel 0x7759c, hlim 255, next-header   
   > > TCP (6) payload length: 40) 2a0c:b640:fffe::21:2834:2.47098 >   
   > > 2a0c:b640:fffe::21:2834:1.179: Flags [S], cksum 0xc046 (correct),   
   > > seq 2822657693, win 65236, options [mss 1388,sackOK,TS val   
   > > 903776896 ecr 0,nop,wscale 7], length 0   
   > > 21:56:44.421735 IP6 (flowlabel 0xc3931, hlim 242, next-header   
   > > unknown (60) payload length: 92) 2a0c:b641:6:: >   
   > > 2001:41d0:302:2000::2144: DSTOPT (opt_type 0x04: len=1)(padn)   
   > > GREv0, Flags [none], length 84   
   > > 	IP6 (class 0xc0, flowlabel 0xc3931, hlim 255, next-header   
   > > TCP (6) payload length: 40) 2a0c:b640:fffe::21:2834:1.179 >   
   > > 2a0c:b640:fffe::21:2834:2.47098: Flags [S.], cksum 0x03b1   
   > > (correct), seq 1998626730, ack 2822657694, win 64672, options [mss   
   > > 1388,sackOK,TS val 911965588 ecr 903773822,nop,wscale 7], length 0   
   > >    
   > > We can see the SYN sent by bird, and the SYN/ACK returned by the   
   > > peer.   
   > > Meanwile, when capturing on the gre interfact itself, only the SYNs   
   > > sent by my side shows up   
   > >    
   > > 21:56:43.387145 IP6 (class 0xc0, flowlabel 0xd8556, hlim 255, next-   
   > > header TCP (6) payload length: 40) 2a0c:b640:fffe::21:2834:2.47098   
   > > > 2a0c:b640:fffe::21:2834:1.179: Flags [S], cksum 0x1173 (incorrect   
   > > -> 0xc446), seq 2822657693, win 65236, options [mss 1388,sackOK,TS   
   > > val 903775872 ecr 0,nop,wscale 7], length 0   
   > > 21:56:44.411239 IP6 (class 0xc0, flowlabel 0x7759c, hlim 255, next-   
   > > header TCP (6) payload length: 40) 2a0c:b640:fffe::21:2834:2.47098   
   > > > 2a0c:b640:fffe::21:2834:1.179: Flags [S], cksum 0x1173 (incorrect   
   > > -> 0xc046), seq 2822657693, win 65236, options [mss 1388,sackOK,TS   
   > > val 903776896 ecr 0,nop,wscale 7], length 0   
   > > 21:56:45.435116 IP6 (class 0xc0, flowlabel 0x9203f, hlim 255, next-   
   > > header TCP (6) payload length: 40) 2a0c:b640:fffe::21:2834:2.47098   
   > > > 2a0c:b640:fffe::21:2834:1.179: Flags [S], cksum 0x1173 (incorrect   
   > > -> 0xbc46), seq 2822657693, win 65236, options [mss 1388,sackOK,TS   
   > > val 903777920 ecr 0,nop,wscale 7], length 0   
   > >    
   > > ICMP inside the tunnels behave the same (seen from the outer   
   > > interface, not on the GRE one)   
   >    
   > Thanks for your both reports. So this look we introduced a regression   
   > from the preivous 6.12.63-1 to 6.12.69-1 with gre6.    
   >    
   > I will have a look at this shortly.   
   >    
   > If any of you both has a testing environment where you can bisect the   
   > upstream changes between 6.12.63 and 6.12.69 to clearly isolate the   
   > breaking change that would be great, let me know if you need   
   > instructions on how to do it. It might be the ip6_gre change   
   > db5b4e39c4e6 ("ip6_gre: make ip6gre_header() robust") in 6.12.64.   
   >    
   > Regards,   
   > Salvatore   
      
   --    
   Dr.-Ing. Tobias Fiebig   
   T +31 616 80 98 99   
   M tobias@fiebig.nl   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)