RE: epics on FreeBSD: broadcast problem

["Jeff Hill" <johill@lanl.gov>]

Hello again Gerrit,

This macro, in src/libCom/osi/os/freebsd/osdSock.h, should be well
scrutinized, especially in its contrast with its implementation for the
other flavors of unix.

#define ifreq_size(pifreq) (sizeof(pifreq->ifr_name))

In particular, you may, on this version of free bsd, need this

#define ifreq_size(pifreq) ((pifreq)->ifr_addr.sa_len +
sizeof((pifreq)->ifr_name))

to support a range of free bsd versions you will most likely need a BSD
version switch - this is what we have currently for vxWorks

#if ( defined (BSD) && ( BSD >= 44 ) )
#   define ifreq_size(pifreq) (pifreq->ifr_addr.sa_len +
sizeof(pifreq->ifr_name))
#else
#   define ifreq_size(pifreq) sizeof(*pifreq)
#endif

let us know how this goes so we can fix the distribution

Jeff
______________________________________________________
Jeffrey O. Hill           Email        johill@lanl.gov
LANL MS H820              Voice        505 665 1831
Los Alamos NM 87545 USA   FAX          505 665 5107

Message content: TSPA

With sufficient thrust, pigs fly just fine. However, this is
not necessarily a good idea. It is hard to be sure where they
are going to land, and it could be dangerous sitting under them
as they fly overhead. -- RFC 1925


> -----Original Message-----
> From: Jeff Hill [mailto:johill@lanl.gov]
> Sent: Thursday, February 24, 2011 9:19 AM
> To: 'gerrit.kuehn@aei.mpg.de'
> Cc: 'tech-talk@aps.anl.gov'
> Subject: RE: epics on FreeBSD: broadcast problem
> 
> > EPICS_CA_AUTO_ADDR_LIST=YES
> > When EPICS_CAS_AUTO_BEACON_ADDR_LIST is YES
> 
> ...
> 
> > Channel Access Address List
> > 10.15.254.2:5065
> 
> > This last address belongs to some vpn connections the machine has, not
> to
> > the physical interfaces (em0 and em1). The IOC I actually want to talk
> to
> > sits on 192.168.64.9 and is reachable via em1.
> 
> suspect that something is amiss with the code in EPICS base which we use
> to interrogate the broadcast address used by each network interface,
> see src/libCom/osi/os/default/osdNetIntf.c.
> 
> > I have a fairly recent FreeBSD 8.1-something
> 
> There is maybe something new about ioctl SIOCGIFCONF in recent Free BSD.
> 
> It might help out with fault isolation to enable the debug messages
> at the top of the source file.
> 
> #ifdef DEBUG
> #   define ifDepenDebugPrintf(argsInParen) printf argsInParen
> #else
> #   define ifDepenDebugPrintf(argsInParen)
> #endif
> 
> Jeff
> ______________________________________________________
> Jeffrey O. Hill           Email        johill@lanl.gov
> LANL MS H820              Voice        505 665 1831
> Los Alamos NM 87545 USA   FAX          505 665 5107
> 
> Message content: TSPA
> 
> With sufficient thrust, pigs fly just fine. However, this is
> not necessarily a good idea. It is hard to be sure where they
> are going to land, and it could be dangerous sitting under them
> as they fly overhead. -- RFC 1925
> 
> 
> > -----Original Message-----
> > From: Gerrit Kühn [mailto:gerrit@pmp.uni-hannover.de]
> > Sent: Thursday, February 24, 2011 8:34 AM
> > To: Jeff Hill
> > Cc: gerrit.kuehn@aei.mpg.de; tech-talk@aps.anl.gov
> > Subject: Re: epics on FreeBSD: broadcast problem
> >
> > On Wed, 23 Feb 2011 12:08:12 -0700 "Jeff Hill" <johill@lanl.gov> wrote
> > about RE: epics on FreeBSD: broadcast problem:
> >
> > JH> Make certain that the netmasks for the network interfaces in your
> > JH> FreeBSD system match the networks to which they are attached. See
> also
> > JH> the "Troubleshooting" portion of the CA reference manual.
> >
> > They do. Actually this particular machine is serving as router and
> > dhcp-server for the subnet:
> >
> > ---
> > firefly1# ifconfig em1
> > em1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> metric 0 mtu
> 1500
> >
> >
> options=219b<RXCSUM,TXCSUM,VLAN_MTU,VLAN_HWTAGGING,VLAN_HWCSUM,TSO4,WOL_MA
> > GIC>
> >         ether 00:25:90:02:16:31
> >         inet 192.168.64.254 netmask 0xffffe000 broadcast 192.168.95.255
> >         media: Ethernet autoselect (1000baseT <full-duplex>)
> >         status: active
> > ---
> >
> > A Linux-system in the network looks quite the same:
> >
> > ---
> > psl@psl-fe4 ~ $ /sbin/ifconfig eth0
> > eth0      Link encap:Ethernet  HWaddr 00:25:90:15:3e:16
> >           inet addr:192.168.64.9  Bcast:192.168.95.255
> Mask:255.255.224.0
> >           inet6 addr: fe80::225:90ff:fe15:3e16/64 Scope:Link
> >           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
> >           RX packets:2994432254 errors:0 dropped:0 overruns:0 frame:0
> >           TX packets:4677428405 errors:0 dropped:0 overruns:0 carrier:0
> >           collisions:0 txqueuelen:1000
> >           RX bytes:277941608130 (258.8 GiB)  TX bytes:3767794266485 (3.4
> > TiB) Memory:d9020000-d9040000
> > ---
> >
> > JH> If there are not routing issues, then I suspect that something is
> > JH> amiss with the code in EPICS base which we use to interrogate the
> > JH> broadcast address used by each network interface.
> >
> > There is no routing involved, all is happening inside this network.
> >
> > JH> This code runs when
> > JH> EPICS_CA_AUTO_ADDR_LIST is YES. For POSIX, reference
> > JH> src/libCom/osi/os/default/osdNetIntf.c. We have occasionally needed
> to
> > JH> make some minor changes here before because socket library interface
> > JH> changes were not backwards compatible.
> >
> > Even setting this explicitely to "YES" does not help:
> >
> > ---
> > firefly1# setenv | grep EPICS
> > EPICS_HOST_ARCH=freebsd-x86_64
> > EPICS_CA_AUTO_ADDR_LIST=YES
> > firefly1# ./caget G3:PSL-OSC_PZTVOLT
> > Channel connect timed out: 'G3:PSL-OSC_PZTVOLT' not found.
> > ---
> >
> >
> > Only setting the broadcast directly does make it work:
> >
> > ---
> > firefly1# setenv | grep EPICS
> > EPICS_HOST_ARCH=freebsd-x86_64
> > EPICS_CA_AUTO_ADDR_LIST=YES
> > EPICS_CA_ADDR_LIST=192.168.95.255
> > firefly1# ./caget G3:PSL-OSC_PZTVOLT
> > G3:PSL-OSC_PZTVOLT             4.78195
> > ---
> >
> > JH> When running a soft IOC on this FreeBSD system, what does casr, with
> a
> > JH> high interest level (i.e. casr 100), show? There should be a list of
> > JH> network addresses listed there - which are the server's beacon
> > JH> destinations. When EPICS_CAS_AUTO_BEACON_ADDR_LIST is YES (the
> > JH> default) the code in src/libCom/osi/os/default/osdNetIntf.c is used
> to
> > JH> populate this list of beacon destination addresses.
> >
> > ---
> > epics> casr 100
> > Channel Access Server V4.13
> > No clients connected.
> > UDP Server:
> > UDP 0.0.0.0:0(): User="", V4.0, 0 Channels, Priority=0
> >         Task Id=0x80204f760, Socket FD=6
> >         Secs since last send   3.02, Secs since last receive   3.02
> >         Unprocessed request bytes=0, Undelivered response bytes=16
> >         State=up
> >         272 bytes allocated
> >
> >         Send Lock
> > epicsMutexId 0x802236b80 source ../caservertask.c line 730
> > ownerTid 0x0 count 0 owned 0
> >         Put Notify Lock
> > epicsMutexId 0x802236b20 source ../caservertask.c line 731
> > ownerTid 0x0 count 0 owned 0
> >         Address Queue Lock
> > epicsMutexId 0x802236ac0 source ../caservertask.c line 732
> > ownerTid 0x0 count 0 owned 0
> >         Event Queue Lock
> > epicsMutexId 0x802236a60 source ../caservertask.c line 733
> > ownerTid 0x0 count 0 owned 0
> >         Block Semaphore
> > There are currently 1904 bytes on the server's free list
> > 7 client(s), 0 channel(s), 0 event(s) (monitors) 0 putNotify(s)
> > 0 small buffers (16384 bytes ea), and 0 jumbo buffers (16408 bytes ea)
> > The server's array size limit is 16408 bytes max
> > Channel Access Address List
> > 10.15.254.2:5065
> > ---
> >
> > This last address belongs to some vpn connections the machine has, not
> to
> > the physical interfaces (em0 and em1). The IOC I actually want to talk
> to
> > sits on 192.168.64.9 and is reachable via em1.
> >
> >
> > cu
> >   Gerrit