- This topic has 59 replies, 24 voices, and was last updated Oct 30-6:28 am by blur13.
-
Posts
-
Brian Masinick
Regarding the network manager, I’m like roki; I have been using only Connman in the releases since we introduced it. Once properly configured I’ve not had any issues with either the antiX 19 series or the Bullseye test versions.
--
Brian Masinicknorm
Hi,
I am new to Antix, I had to re-install Antix 19.3 into my old laptop due to failed wi-fi issue.
My experience connman works from installation USB, and at initial HD laptop setup, but wi-fi freezes after Antix laptop updates.
I am not a programmer, I tried the following work around and it worked for me:1. Start with installation USB, setup wi-fi using connmam;
2. Fresh Install Antix into my computer;
3. Restart computer 1st time after Antix installation, check wi-fi, it should work;
4. After installation, no Antix updates yet! Otherwise wi-fi will freeze.
5. $ sudo apt-mark hold connman and $ sudo apt-mark hold connman-vpn
(I saw these two connman related programs from $ sudo apt list –upgradable available packages list)
6. Confirm that these two files are hold (not to be updated) by issuing command $ apt-mark showhold
7. Now you can do all other updates $ sudo apt update and $ sudo apt upgrade or from antix control centreI think connman update is causing wi-fi freezing problem, the original connman files from installation USB are OK.
Hope it is helpful,
Normdavidrnewman
Thanks for the work around. But the problem isn’t fixed until the Conman updates apply one of the workarounds.
At the moment, you download an AntiX image, install it, do a software update and DNS stops working.
That is a usability anti-pattern that makes people give up AntiX as soon as they install it.
Even better, update the downloadable ISOs to include the fix.
andfree
You have 2 options (…)
Option 2 worked for both of the two old laptops I use this period. Many thanks.
anticapitalista
There will be a connman update (buster only ie antiX-19 series) coming soon that basically reverts to using the ‘old’ way ie solution 1.
We’re testing before uploading to the repos to make sure that whatever fix a user may have applied (or not), this fix will not break anything.Those using testing/sid repos should use fix 2 as this is the ‘new’ way and will be default connman on antiX-21.
Apologies for any inconvenience caused.
BTW – A heads up to you all, there will be security upgrades to our kernels available as soon as possible. (building them as I type).
Philosophers have interpreted the world in many ways; the point is to change it.
antiX with runit - leaner and meaner.
oops
There will be a connman update (buster only ie antiX-19 series) coming soon that basically reverts to using the ‘old’ way ie solution 1.
We’re testing before uploading to the repos to make sure that whatever fix a user may have applied (or not), this fix will not break anything.Those using testing/sid repos should use fix 2 as this is the ‘new’ way and will be default connman on antiX-21.
Apologies for any inconvenience caused.
BTW – A heads up to you all, there will be security upgrades to our kernels available as soon as possible. (building them as I type).
Last time, almost same network issue with network-manager test repository (into sysVinit) for MX19 😉 … the network, is like the kernel, the keybord&mouse&LCD, the terminal and the web browser, very very important.
FI: I have some errors with the kernel 5.10.52-antix.1-amd64-smp into my PC (no problem for my eepc 32 bit 5.10.52-antix 32b pae )
$ cat /proc/cmdline BOOT_IMAGE=/boot/vmlinuz-5.10.52-antix.1-amd64-smp root=UUID= XXXXXXX dmesg ... [ 4841.160994] ------------[ cut here ]------------ [ 4841.161003] WARNING: CPU: 0 PID: 3391 at kernel/rcu/tree_plugin.h:297 rcu_note_context_switch+0x4b/0x520 [ 4841.161003] Modules linked in: bnep(E) bluetooth(E) ecdh_generic(E) ecc(E) rfkill(E) cpufreq_userspace(E) cpufreq_powersave(E) cpufreq_conservative(E) vboxnetadp(OE) vboxnetflt(OE) vboxdrv(OE) uinput(E) nfsd(E) auth_rpcgss(E) nfs_acl(E) nfs(E) lockd(E) grace(E) nfs_ssc(E) fscache(E) sunrpc(E) v4l2loopback(OE) videodev(E) mc(E) lp(E) intel_rapl_msr(E) snd_hda_codec_realtek(E) snd_hda_codec_hdmi(E) snd_hda_codec_generic(E) ledtrig_audio(E) snd_hda_intel(E) intel_rapl_common(E) snd_intel_dspcfg(E) snd_hda_codec(E) x86_pkg_temp_thermal(E) intel_powerclamp(E) snd_hda_core(E) snd_hwdep(E) coretemp(E) snd_pcm(E) snd_timer(E) kvm_intel(E) snd(E) kvm(E) soundcore(E) irqbypass(E) crct10dif_pclmul(E) crc32_pclmul(E) ghash_clmulni_intel(E) aesni_intel(E) ppdev(E) mei_wdt(E) mei_hdcp(E) crypto_simd(E) iTCO_wdt(E) intel_pmc_bxt(E) mei_me(E) iTCO_vendor_support(E) mei(E) cryptd(E) parport_pc(E) parport(E) intel_smartconnect(E) glue_helper(E) at24(E) evdev(E) regmap_i2c(E) tiny_power_button(E) vmd(E) [ 4841.161042] hid_generic(E) radeon(E) i2c_algo_bit(E) ttm(E) drm_kms_helper(E) syscopyarea(E) sysfillrect(E) sysimgblt(E) fb_sys_fops(E) cec(E) rc_core(E) i2c_i801(E) r8169(E) xhci_pci(E) i2c_smbus(E) crc32c_intel(E) lpc_ich(E) realtek(E) drm(E) xhci_pci_renesas(E) fan(E) thermal(E) wmi(E) video(E) button(E) [ 4841.161065] CPU: 0 PID: 3391 Comm: alsa-sink-ALC88 Tainted: G OE 5.10.52-antix.1-amd64-smp #1 [ 4841.161066] Hardware name: MSI MS-7808/B75MA-E33 (MS-7808), BIOS V1.7 09/30/2013 [ 4841.161068] RIP: 0010:rcu_note_context_switch+0x4b/0x520 [ 4841.161070] Code: 00 48 83 ec 10 65 48 03 1d fa 45 f1 7e 66 66 66 66 90 45 84 ed 75 15 65 48 8b 04 25 00 6d 01 00 8b 90 f0 02 00 00 85 d2 7e 02 <0f> 0b 65 48 8b 04 25 00 6d 01 00 8b 80 f0 02 00 00 85 c0 7e 0f 41 [ 4841.161071] RSP: 0000:ffffc9000158fc98 EFLAGS: 00010002 [ 4841.161073] RAX: ffff88810e628000 RBX: ffff888197c29880 RCX: ffffffff82607c78 [ 4841.161074] RDX: 0000000000000001 RSI: ffffffff81c1a561 RDI: 0000000000000000 [ 4841.161075] RBP: ffffc9000158fd28 R08: ffffffff82607c78 R09: 0000000000010a10 [ 4841.161075] R10: 0000000000000004 R11: 0000000000000001 R12: ffff88810e628000 [ 4841.161076] R13: 0000000000000000 R14: 00000000000000f0 R15: ffffc9000158fd98 [ 4841.161077] FS: 00007f58e1a25700(0000) GS:ffff888197c00000(0000) knlGS:0000000000000000 [ 4841.161078] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 4841.161079] CR2: 00007f58e73b39d0 CR3: 000000010d498005 CR4: 00000000000626f0 [ 4841.161080] Call Trace: [ 4841.161086] __schedule+0x78/0x660 [ 4841.161088] schedule+0x51/0xc0 [ 4841.161091] io_schedule+0x12/0x40 [ 4841.161093] __lock_page_or_retry+0x1c9/0x4b0 [ 4841.161095] ? __page_cache_alloc+0x80/0x80 [ 4841.161098] do_swap_page+0x566/0x960 [ 4841.161099] handle_mm_fault+0xd02/0x16f0 [ 4841.161102] exc_page_fault+0x29c/0x530 [ 4841.161104] ? asm_exc_page_fault+0x8/0x30 [ 4841.161105] asm_exc_page_fault+0x1e/0x30 [ 4841.161107] RIP: 0033:0x7f58e73443a0 [ 4841.161109] Code: 68 34 01 00 00 e9 a0 ec ff ff ff 25 3a f6 06 00 68 35 01 00 00 e9 90 ec ff ff ff 25 32 f6 06 00 68 36 01 00 00 e9 80 ec ff ff <ff> 25 2a f6 06 00 68 37 01 00 00 e9 70 ec ff ff ff 25 22 f6 06 00 [ 4841.161110] RSP: 002b:00007f58e1a21ed8 EFLAGS: 00010246 [ 4841.161111] RAX: 0000000000000000 RBX: 000100000000fcee RCX: 0000000000000921 [ 4841.161112] RDX: 0000000000000921 RSI: 000100000000fcee RDI: 00007f58e73b75c0 [ 4841.161113] RBP: 00007f58d8060040 R08: 0000564599451698 R09: 0000000000000000 [ 4841.161113] R10: 0000000000000010 R11: 0000000000000043 R12: 0000000000000921 [ 4841.161114] R13: 00007f58e8062890 R14: 00007f58e1a22240 R15: 00007f58e1a22420 [ 4841.161116] ---[ end trace 23b1f2342c48ed11 ]---- This reply was modified 1 year, 9 months ago by oops.
Malcolm Reynolds
Hi I upgraded A19 yesterday (Aug 9) and saw that connman was on the list of upgrades.
As upgrade proceeded I was asked if I wanted the package maintainers config or keep my own. I selected keep my own.
After the upgrade connman would not start and start/stop complained about a non-existent argument “—-nodnsproxy” ( 4 hyphens was a clue )
I found that by editing /etc/init.d/connman I was able to fix the problem. I added a ‘$’ character at line 22 and a space at line 28.
All working again now. Did I do the right thing ?
alex5612
I think so.
Made the same changes and now connman works just fine.
Xecure
All working again now. Did I do the right thing ?
Yep. you did very well. Updating the init script for connman was one of the reasons for this update.
Thanks for reporting.
antiX Live system enthusiast.
General Live Boot Parameters for antiX.stevesr0
Hi all,
The reason I didn’t have a problem with the connman upgrade on my sid system is because connman is NOT installed on that system. It is a “roll my own” attempt at a “minimal” antiX install. The only internet connection “tool” installed is ceni.
stevesr0
Hi all,
The newest update to connman (a reversion as promised by anticapitalista) did, as expected, shut off DNS access. However, once again, deleting the /etc/resolv.conf file and setting up a symbolic link from the /run/resolvconf/resolv.conf file to /etc/resolv.conf reestablished the connection.
(I first killed connman, since I have wanted to make sure that wifi connections were working with wicd.)
So, at this point, I am wondering – will future upgrades to connman will routinely mess up internet connections?
Parenthetically, on reboot I note that all remains well – and that BOTH connman and wicd appear active without an apparent conflict. Somehow that surprises me <g>.
stevesr0
- This reply was modified 1 year, 9 months ago by stevesr0.
Xecure
The newest update to connman (a reversion as promised by anticapitalista) did, as expected, shut off DNS access. However, once again, deleting the /etc/resolv.conf file and setting up a symbolic link from the /run/resolvconf/resolv.conf file to /etc/resolv.conf reestablished the connection.
I cannot understand this. I checked the connman source anticapitalista uploaded on gitlab, built it myself, and installed it and didn’t have any problems after extensive testing.
I will have a look to see why the package in the repo has brought back the problem. It should have been a perfect upgrade without issues.
antiX Live system enthusiast.
General Live Boot Parameters for antiX.Xecure
Somehow, the connman_1.36-2.3.0antix2.debian.tar.xz content is the same as the Debian’s official connman version except for the changelog. It is missing the init patch and small other changes (including the revert back to original).
The bad news: /etc/init.d/connman is back to the buggy version in Debian. All updates require solution 1 (remove /etc/default/connman)
sudo rm /etc/default/connmandeleting the /etc/resolv.conf file and setting up a symbolic link from the /run/resolvconf/resolv.conf file to /etc/resolv.conf reestablished the connection.
This solution (solution 2) is not an option. Each reboot will overwrite /etc/resolv.conf (old aggressive connman behavior), so doing this is useless.
The good news: Anyone that had not updated before will not experience any change with this update.
antiX Live system enthusiast.
General Live Boot Parameters for antiX.Hi Xecure,
I just rebooted and didn’t have a problem with connecting to the internet.
Do you mean that FUTURE updates will recreate the problem, rather than REBOOTING with the latest version of connman?
stevesr0
- This reply was modified 1 year, 9 months ago by stevesr0.
Xecure
I just rebooted and didn’t have a problem with connecting to the internet.
if you are still using connman, see
readlink /etc/resolv.conf
if it still points to /run/resolvconf/resolv.conf or /run/connman/resolv.confWhile connman is running it will always take control of /etc/resolv.conf, which was the same as before all the update debacle. But I understand people prefer to let the software decide what is best for them instead of deciding themselves. That is probably why this init bug hasn’t been resolved in YEARS (still in connman’s latest 1.40 version),as most users never make any other choice than let the software “work”.
Anyway, people can be happy with the antiX 19 release returning to the previous behavior. If they experience problems, they just need to remove /etc/default/connman and then it will be exactly as before.
antiX Live system enthusiast.
General Live Boot Parameters for antiX.
- You must be logged in to reply to this topic.
