Hi, I built 1.4.4 from source and loaded it on a GL-AR300M (NAND), which happens to be providing IPv6 via 6in4 tunnel to HE tunnel broker. After upgrade, log was getting hammered by errors like so:
Wed Jul 25 08:55:20 2018 kern.info kernel: [ 5511.429707] sit: non-ECT from 160.240.0.0 with TOS=0x7
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.416060] sit: non-ECT from 64.9.8.10 with TOS=0x2
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.716129] sit: non-ECT from 64.9.8.16 with TOS=0x2
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.818672] sit: non-ECT from 64.9.8.6 with TOS=0xd
Thankfully that post provides some info on how to quiet down the log messages. But I am wondering if some patch from that bug report needs to be backported into librecmc's kernel version, or something like that, to deal with the actual bug.
Hi, I built 1.4.4 from source and loaded it on a GL-AR300M (NAND), which happens to be providing IPv6 via 6in4 tunnel to HE tunnel broker. After upgrade, log was getting hammered by errors like so:
Wed Jul 25 08:55:20 2018 kern.info kernel: [ 5511.429707] sit: non-ECT from 160.240.0.0 with TOS=0x7
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.416060] sit: non-ECT from 64.9.8.10 with TOS=0x2
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.716129] sit: non-ECT from 64.9.8.16 with TOS=0x2
Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.818672] sit: non-ECT from 64.9.8.6 with TOS=0xd
This looks like the same bug mentioned here:
https://bugs.openwrt.org/index.php?do=details&task_id=1541
Thankfully that post provides some info on how to quiet down the log messages. But I am wondering if some patch from that bug report needs to be backported into librecmc's kernel version, or something like that, to deal with the actual bug.
Can you try to pull and build again. See commit a7ce8a7e360f6f1dfd593ee4e6220a10a6f5f932
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/net/ipv6?h=v4.9.112&id=5b8fcc075714b86fb8fe194bb463fed2998a8e85
Hello, I rebuilt the firmware using that commit, and loaded it, but am still seeing the same errors flooding the logs.
While I was careful, nevertheless I was going rebuild and reload again with clean source, to remove any trace of doubt that I might have loaded the wrong image, or that the new image did not load. The system is in a production environment, so I can only do this once per workday. (I actually load onto a swap unit.) So I probably won't be able to do that until Monday or Tuesday.
Hello, I rebuilt the firmware using that commit, and loaded it, but am still seeing the same errors flooding the logs.
While I was careful, nevertheless I was going rebuild and reload again with clean source, to remove any trace of doubt that I might have loaded the wrong image, or that the new image did not load. The system is in a production environment, so I can only do this once per workday. (I actually load onto a swap unit.) So I probably won't be able to do that until Monday or Tuesday.
I loaded firmware built from commit a7ce8a7e36, using clean source, on a GL-AR300M, and swapped it into production. The system is running a 6in4-pd type interface, i.e., IPv6 tunnel. I browsed to multiple IPv6 and IPv4 sites for 30 minutes or so and am not seeing in the logs any of the errors I saw previously.
I loaded firmware built from commit a7ce8a7e36, using clean source, on a GL-AR300M, and swapped it into production. The system is running a 6in4-pd type interface, i.e., IPv6 tunnel. I browsed to multiple IPv6 and IPv4 sites for 30 minutes or so and am not seeing in the logs any of the errors I saw previously.
RISCI_ATOM
Hi, I built 1.4.4 from source and loaded it on a GL-AR300M (NAND), which happens to be providing IPv6 via 6in4 tunnel to HE tunnel broker. After upgrade, log was getting hammered by errors like so:
Wed Jul 25 08:55:20 2018 kern.info kernel: [ 5511.429707] sit: non-ECT from 160.240.0.0 with TOS=0x7 Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.416060] sit: non-ECT from 64.9.8.10 with TOS=0x2 Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.716129] sit: non-ECT from 64.9.8.16 with TOS=0x2 Wed Jul 25 08:55:23 2018 kern.info kernel: [ 5514.818672] sit: non-ECT from 64.9.8.6 with TOS=0xd
This looks like the same bug mentioned here:
https://bugs.openwrt.org/index.php?do=details&task_id=1541
Thankfully that post provides some info on how to quiet down the log messages. But I am wondering if some patch from that bug report needs to be backported into librecmc's kernel version, or something like that, to deal with the actual bug.
I will look into patching this.
Thank you. I'm a fairly good test case if that is helpful. Can rebuild from any libreCMC commit you point me to, development branch or otherwise.
Can you try to pull and build again. See commit
a7ce8a7e36https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/net/ipv6?h=v4.9.112&id=5b8fcc075714b86fb8fe194bb463fed2998a8e85
Will do. Please give me a day or so.
Hello, I rebuilt the firmware using that commit, and loaded it, but am still seeing the same errors flooding the logs.
While I was careful, nevertheless I was going rebuild and reload again with clean source, to remove any trace of doubt that I might have loaded the wrong image, or that the new image did not load. The system is in a production environment, so I can only do this once per workday. (I actually load onto a swap unit.) So I probably won't be able to do that until Monday or Tuesday.
Work has been brutal, but hoping to test this again tomorrow morning.
I loaded firmware built from commit
a7ce8a7e36, using clean source, on a GL-AR300M, and swapped it into production. The system is running a 6in4-pd type interface, i.e., IPv6 tunnel. I browsed to multiple IPv6 and IPv4 sites for 30 minutes or so and am not seeing in the logs any of the errors I saw previously.Still no errors in the log. Looks good. Not sure if I'm supposed to close this ticket.