LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   *BSD (https://www.linuxquestions.org/questions/%2Absd-17/)
-   -   OpenBSD 6.5 Released (https://www.linuxquestions.org/questions/%2Absd-17/openbsd-6-5-released-4175652704/)

hitest 04-24-2019 10:21 AM

OpenBSD 6.5 Released
 
Upgrade time! :)

https://undeadly.org/cgi?action=arti...20190424132429

Turbocapitalist 04-24-2019 11:36 AM

I've read only a little about OpenRsync, which is included in 6.5. I also see that it entered the tree back in February.

Is it enough of a drop-in replacement for old rsync, such that I can make a hard link from the name 'rsync' to 'openrsync' and it will work?

jggimi 04-24-2019 11:54 AM

Thanks for posting it, hitest!
Quote:

Originally Posted by Turbocapitalist (Post 5988183)
Is it enough of a drop-in replacement for old rsync, such that I can make a hard link from the name 'rsync' to 'openrsync' and it will work?

Check the man page and see if your use-cases would work: http://man.openbsd.org/openrsync

If it does, you might prefer a shell alias, a symbolic link in $HOME/bin, or a symbolic link in /usr/local/bin to avoid local changes in /usr/bin.

fatmac 04-24-2019 01:24 PM

Thanks for the heads up. :)

hitest 04-24-2019 01:54 PM

So far so good. Have some reading to do on doas.conf. Might have changed.

Code:

bsd$ uname -a
OpenBSD bsd.home.local 6.5 GENERIC.MP#3 amd64


cynwulf 05-10-2019 06:58 AM

Not upgraded as yet - 6.5 install kernel is spitting athn(4) "device timeout" errors and no time to diagnose as yet...

Not had these errors under the 6.4 kernel.

https://www.openbsd.org/65.html
Quote:

Improved radio hardware calibration in the athn(4) driver.
(for some perhaps)

So far as I can tell athn-firmware has not changed between 6.4-release and 6.5-release.

hitest 05-10-2019 07:43 AM

Quote:

Originally Posted by cynwulf (Post 5993631)
Not upgraded as yet - 6.5 install kernel is spitting athn(4) "device timeout" errors and no time to diagnose as yet...

I found this migration to be a bit rougher than earlier versions. There are changes in doas.conf configuration so I have moved back to using sudo which is okay.
Also, I was able to get help from my friends and Landry, the XFCE port developer, to solve an XFCE shut down issue.
Now that I am up and running with 6.5 I am quite happy with it. That is very odd with the athn firmware.

cynwulf 05-10-2019 10:14 AM

Found a bug report with the same hardware involved from a few days ago: https://marc.info/?l=openbsd-bugs&m=155704811002722&w=2

I will try reversing the commits to athn.c and recompiling the kernel, when I've got the time.

cynwulf 05-12-2019 06:06 PM

OK here we go:
https://cvsweb.openbsd.org/src/sys/dev/ic/athn.c

So creating my own patch from a diff of revisions 1.101 of sys/dev/ic/athn.c against the BASE source tree obtained via CVS (1.102):

Code:

$ diff -u /usr/src/sys/dev/ic/athn.c athn.txt > athn.patch
# cp athn.patch /usr/src/
# cd /usr/src
# patch -p0 < athn.patch
Hmm...  Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|--- /usr/src/sys/dev/ic/athn.c Fri Mar  1 07:39:56 2019
|+++ athn.txt  Sun May 12 23:48:30 2019
--------------------------
Patching file /usr/src/sys/dev/ic/athn.c using Plan A...
Hunk #1 succeeded at 1.
Hunk #2 succeeded at 136.
Hunk #3 succeeded at 2403.
Hunk #4 succeeded at 2713.
Hunk #5 succeeded at 2729.
Hunk #6 succeeded at 2739.

So off I go to rebuild the kernel...

//edit: Nope, still the device timeouts after reverting...
Code:

$ dmesg|grep athn
athn0 at pci2 dev 0 function 0 "Atheros AR9281" rev 0x01: apic 2 int 19
athn0: AR9280 rev 2 (1T2R), ROM rev 16, address 00:24:2b:09:2f:1b
athn0: radio is disabled by hardware switch
athn0: unable to reset hardware; reset status 1
athn0: radio is disabled by hardware switch
athn0: unable to reset hardware; reset status 1
athn0: device timeout
athn0: device timeout
athn0: device timeout
athn0 at pci2 dev 0 function 0 "Atheros AR9281" rev 0x01: apic 2 int 19
athn0: AR9280 rev 2 (1T2R), ROM rev 16, address 00:24:2b:09:2f:1b
athn0: device timeout
athn0: device timeout
athn0: device timeout
athn0: device timeout
athn0: device timeout
athn0: radio switch turned off
athn0 at pci2 dev 0 function 0 "Atheros AR9281" rev 0x01: apic 2 int 19
athn0: AR9280 rev 2 (1T2R), ROM rev 16, address 00:24:2b:09:2f:1b

(ignore "the hardware switch" messages, as I had turned it off at that stage)

Code:

$ ls -al /usr/src/sys/dev/ic/athn.*
-rw-r--r--  1 root  wsrc  81829 May 12 23:58 /usr/src/sys/dev/ic/athn.c
-rw-r--r--  1 root  wsrc  84043 Mar  1 07:39 /usr/src/sys/dev/ic/athn.c.orig
$ ls -al /bsd                     
-rwx------  1 root  wheel  15626006 May 13 00:41 /bsd

Reverting to 1.100 next, then if that fails back to 1.99 - which is equivalent to the athn(1) driver in 6.4-release kernel.

Resuming tomorrow...

//edit: Reversed the commits all the way back to 1.00, resulting in the athn.c from 6.4-release which should work (1.99) - but does not. From reverting commit 1.101, the problem is even worse and no association with the AP at all and still the device timeouts.

A few more files to look at however, especially header file and others relating to this specific chip...

Or could try following -current and this patch, etc...

https://marc.info/?l=openbsd-bugs&m=155775488320235&w=2

cynwulf 05-13-2019 03:44 PM

Backed out the commits on the following from athn(4) :
Code:

$ grep \$OpenBSD *
ar5008.c:/*  $OpenBSD: ar5008.c,v 1.46 2017/11/28 04:35:39 stsp Exp $        */
ar5416.c:/*  $OpenBSD: ar5416.c,v 1.20 2017/01/12 16:32:28 stsp Exp $        */
ar5416reg.h:/*        $OpenBSD: ar5416reg.h,v 1.6 2016/01/05 18:41:15 stsp Exp $      */
ar9003.c:/*  $OpenBSD: ar9003.c,v 1.46 2017/05/19 11:42:48 stsp Exp $        */
ar9280.c:/*  $OpenBSD: ar9280.c,v 1.26 2017/01/12 16:32:28 stsp Exp $        */
ar9280reg.h:/*        $OpenBSD: ar9280reg.h,v 1.7 2016/01/05 18:41:15 stsp Exp $      */
ar9285.c:/*  $OpenBSD: ar9285.c,v 1.27 2017/01/12 16:32:28 stsp Exp $        */
ar9285reg.h:/*        $OpenBSD: ar9285reg.h,v 1.8 2016/01/05 18:41:15 stsp Exp $      */
ar9287.c:/*  $OpenBSD: ar9287.c,v 1.25 2017/01/12 16:32:28 stsp Exp $        */
ar9287reg.h:/*        $OpenBSD: ar9287reg.h,v 1.4 2016/01/05 18:41:15 stsp Exp $      */
athn.c:/*    $OpenBSD: athn.c,v 1.99 2018/04/26 12:50:07 pirofti Exp $      */
athnreg.h:/*  $OpenBSD: athnreg.h,v 1.21 2017/11/28 04:35:39 stsp Exp $      */
athnvar.h:/*  $OpenBSD: athnvar.h,v 1.37 2017/01/12 16:32:28 stsp Exp $      */

The common theme here is the "noisefloor calibration". Reverting back to the commits before that was introduced, gives me working wifi again.

//Update: bsd.rd from the 16/05/19 snapshot of -current includes this patch and when booted maintained a connection when tested, without any "device timeout" message.

https://cvsweb.openbsd.org/src/sys/dev/ic/athn.c

Will update via cvs, to reverse out the above patches and try that patch next.

//Update: No good, patch is only good against -current.

////// 05-16-2019 08:14 AM

just upgraded my bridge :)
feels solid.

hitest 05-16-2019 09:12 AM

Quote:

Originally Posted by ////// (Post 5995642)
just upgraded my bridge :)
feels solid.

Agreed! Solid indeed. Fixed my doas.conf issue this morning with the help of my friends. :)


All times are GMT -5. The time now is 07:40 AM.