Martin -

I haven't given it allot of thought about a plan to squash the fix commits
into the broken commits.  Hasn't been high on my priority :).  I hear you
though.

About the different platforms though.  I think a few more minutes of my
time saves everyones time.  I can intelligently know when to compile on
different platforms by just looking at the diff.  This last problem should
have been a no brainer we changed the api that zebra uses to communicate
with the different platforms.  Basically the question boils down to does it
touch zebra?

Now on to find some work I did on Friday night that I seemed to have
misplaced.

donald

On Sun, Nov 22, 2015 at 2:43 AM, Martin Winter <
[email protected]> wrote:

> Donald,
>
> On 21 Nov 2015, at 4:58, Donald Sharp wrote:
>
> Martin -
>>
>> Pushed a fix for the issue.  Again thanks for finding this.
>>
>
> Yes, fixed (at least the compile issue).
> Will be testing more…
>
> On a side note with these proposed branches:
> Are you planning to fix bad commits with an additional commit or are you
> planning
> to fix the actual bad commit directly before merging to master? (i.e.
> rebase a
> few times - I think Paul did this at least for accepted branch)
> I would prefer the 2nd for the accepted branch (don’t care on proposed
> branch)
>
> I guess I need to get a solaris vm.
>>
>
> Not sure if this makes sense for you. I would rather suggest you spend
> your time
> on the excellent work you already do and let me catch these occasional
> distro issues.
> Running all these different OS’es and keeping them up to date is painful.
>
> Are they freely available?
>>
>
> I use OmniOS which is one OpenSolaris clone and based on Illumos and
> server focused.
> I’ve initially tried OpenIndiana, but could not get it working reliable
> (not quagga
> related the OS itself had issues during package upgrades/install
>
> I’m thinking about adding more distress, but I’m kind of clueless on
> Solaris side, so
> would have no idea which would be a common, but quite different distro
> which would make
> sense to test (Hint: Suggestions on adding distros to CI system welcome)
>
> OmniOS is free (http://omnios.omniti.com/)
> For some hints in installation, see
>
> https://git-us.netdef.org/projects/OSR/repos/ci-files/browse/doc/Quagga_OmniOS.md
>
> And if someone has any idea on how to build packages on Solaris/OmniOS,
> then please
> contact me privately. So far I was unable to find good info.
>
> - Martin
>
> PS: I think I got all of OpenBSD now figured out and will start adding
> this to the CI
> as well. Thanks to great help by Peter Hessler from the OpenBSD team
> during the RIPE
> conference. I plan to add at least OpenBSD 5.8 initially, maybe 5.7 as well
>
>
>
>
> On Sat, Nov 21, 2015 at 6:17 AM, Martin Winter <
>> [email protected]> wrote:
>>
>> On 20 Nov 2015, at 6:14, Donald Sharp wrote:
>>>
>>> Martin -
>>>
>>>>
>>>> I've fixed the issues above, thanks for finding them.
>>>>
>>>>
>>> Thanks, but now failing on the next issue (Solaris only):
>>>
>>> CC       ipforward_solaris.o
>>> CC       if_ioctl_solaris.o
>>> CC       rt_socket.o
>>> CC       rtread_getmsg.o
>>> rtread_getmsg.c: In function ‘handle_route_entry’:
>>> rtread_getmsg.c:95:9: error: too few arguments to function ‘rib_add_ipv4’
>>>       0, 0, SAFI_UNICAST);
>>>       ^
>>> In file included from rtread_getmsg.c:30:0:
>>> ../zebra/rib.h:459:12: note: declared here
>>> extern int rib_add_ipv4 (int type, int flags, struct prefix_ipv4 *p,
>>>          ^
>>> make[2]: Leaving directory
>>> `/export/home/ci/cibuild.3/quagga-source/zebra'
>>> make[2]: *** [rtread_getmsg.o] Error 1
>>> make[1]: Leaving directory `/export/home/ci/cibuild.3/quagga-source'
>>> make[1]: *** [all-recursive] Error 1
>>> make: *** [all] Error 2
>>> make failed
>>>
>>> - Martin
>>>
>>>
>>>
>>> On Mon, Nov 16, 2015 at 5:17 PM, Martin Winter <
>>>
>>>> [email protected]> wrote:
>>>>
>>>> Not sure if you expect this branch to be just collection of patches or
>>>>
>>>>> actually
>>>>> be useable.
>>>>>
>>>>> Trying to build Quagga based on this branch fails building on *BSD
>>>>> systems.
>>>>>
>>>>> See https://ci1.netdef.org/browse/QUAGGA-CUMULUS-13
>>>>>
>>>>> FreeBSD and NetBSD fails similar errors:
>>>>>
>>>>>
>>>>>
>>>>> https://ci1.netdef.org/browse/QUAGGA-CUMULUS-13/artifact/CI003BUILD/ErrorLog/log_make.txt
>>>>>
>>>>> CC       rtadv.o
>>>>>
>>>>> rtadv.c:379:9: error: no member named 'rtadv' in 'struct zebra_vrf'
>>>>>> zvrf->rtadv.ra_timer = NULL;
>>>>>> ~~~~  ^
>>>>>> rtadv.c:380:13: error: no member named 'rtadv' in 'struct zebra_vrf'
>>>>>> if (zvrf->rtadv.adv_msec_if_count == 0)
>>>>>> ~~~~  ^
>>>>>> rtadv.c:406:33: error: no member named 'rtadv' in 'struct zebra_vrf'
>>>>>>       rtadv_send_packet (zvrf->rtadv.sock, ifp);
>>>>>>                          ~~~~  ^
>>>>>> rtadv.c:420:28: error: no member named 'rtadv' in 'struct zebra_vrf'
>>>>>> rtadv_send_packet (zvrf->rtadv.sock, ifp);
>>>>>>              ~~~~  ^
>>>>>> rtadv.c:499:9: error: no member named 'rtadv' in 'struct zebra_vrf'
>>>>>> zvrf->rtadv.ra_read = NULL;
>>>>>> ~~~~  ^
>>>>>>
>>>>>>
>>>>>> and then there is a warning too:
>>>>>
>>>>> CC       log.lo
>>>>> log.c: In function 'zlog_hexdump':
>>>>> log.c:1035:15: warning: array subscript has type 'char'
>>>>> [-Wchar-subscripts]
>>>>>          else if(isprint(((char*)mem)[j])) /* printable char */
>>>>>          ^
>>>>>
>>>>> Regards,
>>>>> Martin Winter
>>>>> [email protected]
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> On 16 Nov 2015, at 11:56, Donald Sharp wrote:
>>>>>
>>>>> Currently everything queued up is at:
>>>>>
>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>> https://github.com/donaldsharp/quagga/tree/volatile/patch-tracking/5/proposed/ff
>>>>>>
>>>>>> This was just a convenient place to put code out there until such time
>>>>>> it
>>>>>> was accepted.
>>>>>>
>>>>>> Patches Queued up:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1399/  - OSPF Fundamentals doc
>>>>>> change
>>>>>>
>>>>>> Daniel's Default changes have been queued up as well.  I see that
>>>>>> there
>>>>>> is
>>>>>> some fresh discussion today as well.  I'll fix the timer down to 1
>>>>>> second:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1404/    - lower the default MRAI
>>>>>> timer
>>>>>> for iBGP peers to 0
>>>>>> http://patchwork.quagga.net/patch/1405/    - Lower the default MRAI
>>>>>> timer
>>>>>> for EBGP peers to 0
>>>>>> http://patchwork.quagga.net/patch/1406/    - Lower the default
>>>>>> 'timers
>>>>>> connect' in BGP to 10 seconds
>>>>>> http://patchwork.quagga.net/patch/1407/    - Remove BGP's asorig
>>>>>> timer,
>>>>>> it
>>>>>> is no longer used
>>>>>> http://patchwork.quagga.net/patch/1408/    - Do not allo a timers
>>>>>> connect
>>>>>> of 0, this can hammer the CPU
>>>>>> http://patchwork.quagga.net/patch/1409/    - Enable "bgp
>>>>>> log-neighbor-changes" by default
>>>>>> http://patchwork.quagga.net/patch/1410/    - Remove
>>>>>> BGP_ERROR_START_TIMER,
>>>>>> it was no longer used
>>>>>> http://patchwork.quagga.net/patch/1411/    - Lower BGP's default
>>>>>> keepalive/holdtime to 3s/9s
>>>>>> http://patchwork.quagga.net/patch/1412/    - Lower the default OSPF
>>>>>> spf
>>>>>> timers to '0 50 5000'
>>>>>> http://patchwork.quagga.net/patch/1413/    - BGP: crash from not
>>>>>> NULLing
>>>>>> freed pointers
>>>>>>
>>>>>> Timo's Fixup to Master in Preparation for NHRP code:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1421/   - zebra: fix rtadv
>>>>>> detection
>>>>>> http://patchwork.quagga.net/patch/1419/   - check rtt later after the
>>>>>> real
>>>>>> peer is known
>>>>>> http://patchwork.quagga.net/patch/1418/   - update rtt on soft clear
>>>>>> http://patchwork.quagga.net/patch/1420/   - send new interface
>>>>>> messages
>>>>>> if
>>>>>> ifindex changes
>>>>>> http://patchwork.quagga.net/patch/1422/   - Add suport for timer
>>>>>> commands
>>>>>> with peer-group syntax
>>>>>>
>>>>>> Some Various Fixes:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1414/   - Add zlog_hexdump() for
>>>>>> debugging
>>>>>> http://patchwork.quagga.net/patch/1415/   - Fix bgp_btoa to compile
>>>>>> http://patchwork.quagga.net/patch/1417/   - Limit pim hello log
>>>>>> messages
>>>>>> http://patchwork.quagga.net/patch/1429/   - lib: Memory reporting
>>>>>> fails
>>>>>> over 2Gb
>>>>>> http://patchwork.quagga.net/patch/1506/   - ripd, isisd: fix warnings
>>>>>> that
>>>>>> make build fail
>>>>>> http://patchwork.quagga.net/patch/1590/   - Update dump to allow
>>>>>> Extended
>>>>>> Time Format
>>>>>> -> Alexis please take a look at this to make sure it is right.
>>>>>>
>>>>>> Timo's zebra fib/rib updates for NHRP:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1432/    - make ZEBRA_FLAG_CHANGED
>>>>>> internal status
>>>>>> http://patchwork.quagga.net/patch/1433     - implement per-route mtu
>>>>>> handling
>>>>>>
>>>>>> Not Added due to issues
>>>>>> http://patchwork.quagga.net/patch/1435/    - atomic FIB updates (
>>>>>> Fails
>>>>>> to compile on some platforms )
>>>>>> http://patchwork.quagga.net/patch/1431/    - use link scope for
>>>>>> interface routes ( CR comments not addressed yet )
>>>>>> http://patchwork.quagga.net/patch/1436/    - Support FIB override
>>>>>> routes ( Patch does not apply cleanly anymore )
>>>>>> http://patchwork.quagga.net/patch/1405/    - remove metric from
>>>>>> kernel
>>>>>> ( Patch does not apply cleanly anymore )
>>>>>>
>>>>>> Donald's static_XXX refactoring:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1441/      - Collapse struct
>>>>>> static_ipv[4|6] into struct static_route
>>>>>> http://patchwork.quagga.net/patch/1438/      - Remove HAVE_IPV6 from
>>>>>> rib.h
>>>>>> and zebra_rib.c
>>>>>> http://patchwork.quagga.net/patch/1440/      - Combine
>>>>>> static_install_ipv4[4|6]
>>>>>> http://patchwork.quagga.net/patch/1439/      - combine
>>>>>> static_ipv[4|6}_nexthop_same into one function
>>>>>> http://patchwork.quagga.net/patch/1442/      - combine
>>>>>> static_uninstall_ipv[4|6] into one function
>>>>>> http://patchwork.quagga.net/patch/1443/      - Fix change of distance
>>>>>> on
>>>>>> ipv6 route creating duplicate routes
>>>>>>
>>>>>> Christian's isis fixes:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1507/      - add function to get
>>>>>> precise
>>>>>> remaining time of timer thread
>>>>>> http://patchwork.quagga.net/patch/1508/      - don't have log
>>>>>> functions
>>>>>> change errno
>>>>>> http://patchwork.quagga.net/patch/1512/      - annotate some
>>>>>> functions
>>>>>> arguments with const
>>>>>> http://patchwork.quagga.net/patch/1514/      - remove superfluous
>>>>>> checks
>>>>>> after XMALLOC etc.
>>>>>> http://patchwork.quagga.net/patch/1509/      - fix a type in a log
>>>>>> message
>>>>>> http://patchwork.quagga.net/patch/1510/      - don't corrupt memory
>>>>>> for
>>>>>> long hostnames
>>>>>> http://patchwork.quagga.net/patch/1511/      - do remove ipv6 routes
>>>>>> from
>>>>>> Zebra
>>>>>> http://patchwork.quagga.net/patch/1513/      - initialize circuit to
>>>>>> match
>>>>>> area is_type
>>>>>> http://patchwork.quagga.net/patch/1515/      - purge on correct level
>>>>>> http://patchwork.quagga.net/patch/1517/      - fix assertion in LSP
>>>>>> refresh
>>>>>> timer calculation
>>>>>> http://patchwork.quagga.net/patch/1518/      - provide more detailed
>>>>>> log
>>>>>> for failed address removal
>>>>>> http://patchwork.quagga.net/patch/1519/      - show interface's ipv6
>>>>>> address
>>>>>> http://patchwork.quagga.net/patch/1520/      - fix IPv6 mask
>>>>>> application
>>>>>> http://patchwork.quagga.net/patch/1521/      - handle lsp confusion
>>>>>> (ISO/IEC 10598: 2002 7.3.16.2)
>>>>>> http://patchwork.quagga.net/patch/1522/      - fix misleading wording
>>>>>> in
>>>>>> log
>>>>>> http://patchwork.quagga.net/patch/1523/      - make send_lsp more
>>>>>> robust
>>>>>> http://patchwork.quagga.net/patch/1524/      - allow to adjust
>>>>>> lsp-mtu
>>>>>> http://patchwork.quagga.net/patch/1525/      - add a debug mode that
>>>>>> traces
>>>>>> LSP construction
>>>>>> http://patchwork.quagga.net/patch/1526/      - add a slight delay to
>>>>>> lsp_regenerate_schedule
>>>>>> http://patchwork.quagga.net/patch/1527/      - add support to import
>>>>>> routes
>>>>>> from other protocols
>>>>>>
>>>>>>
>>>>>> Please note, I have not finished pushing up Christian's patches into
>>>>>> the above branch.  This work will be done tonight.
>>>>>>
>>>>>>
>>>>>> List of patches that need more work or I was unsure of the status:
>>>>>>
>>>>>> http://patchwork.quagga.net/patch/1425/    - Cleanup BGP memory  (Is
>>>>>> this
>>>>>> already in? )
>>>>>> http://patchwork.quagga.net/patch/1426/    - When the config file is
>>>>>> close
>>>>>> to the boundry of size of buffer vtysh hangs. ( CR issue )
>>>>>> http://patchwork.quagga.net/patch/1430/    - Kernel route does not
>>>>>> get
>>>>>> updated when neighbor remove the 'Nex...'  ( I'm just confused here,
>>>>>> there
>>>>>> are multiple versions of this patch )
>>>>>>
>>>>>> Input here would be awesome!
>>>>>>
>>>>>> thanks!
>>>>>>
>>>>>> donald
>>>>>>
>>>>>>
>>>>>>
>>>>>
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev

Reply via email to