On Thu, 2010-12-23 at 10:46 +0100, Karl Relton wrote: > Dear friends >
Hi Karl
Apologies, I should have replied earlier.
The ubuntu package is actually a bit out of date, so I've already added
some of the fixes you've been sending. I'll upload a new one as soon as
I can.
I'll also take a look at the below. I would rather do this at the udev
level, but I couldn't figure out how. If you've cracked it that will be
great.
The main problem that had stopped me releasing the udev stuff was with
pre WM 5 devices. It worked connecting once, but to connect again I had
to kill dccm. I've now managed to lose my WM2003 test device, can you
try connecting multiple times and confirm if it works or not ?
Thanks
Mark
> I've done some digging, and worked out that the device querying code in
> udev-synce-serial is only necessary for trying to handle two-port
> devices. I think this can be better handled in the udev rules, from
> which we can launch udev-synce-serial only when required (i.e. on the
> 2nd port of a two port device), and thus remove all that logic from
> udev-synce-serial itself.
>
> That means udev-synce-serial no longer has to call udevadm or use a
> library to interrogate the udev database, cutting a chunk of code.
>
> The patch below does just that.
>
> Regards
> Karl
>
>
> diff -Naur scripts.orig/85-synce.rules.in scripts/85-synce.rules.in
> --- scripts.orig/85-synce.rules.in 2010-07-19 18:24:17.000000000 +0100
> +++ scripts/85-synce.rules.in 2010-12-23 09:24:35.267741000 +0000
> @@ -8,4 +8,18 @@
> SUBSYSTEM=="net", DRIVERS=="rndis_host", ATTRS{bInterfaceClass}=="ef",
> ATTRS{bInterfaceSubClass}=="01", ATTRS{bInterfaceProtocol}=="01",
> ENV{pocketpc}="env-synce", RUN+="@UDEVDIR@/synce-udev-rndis"
>
> # capabilities = serial, parent driver = ipaq
> -SUBSYSTEM=="tty", DRIVERS=="ipaq", ENV{pocketpc}="env-synce",
> RUN+="@UDEVDIR@/synce-udev-serial"
> +# If the device has more than one port, then we need to only use the last
> +# one.
> +# From ipaq.c we can see the choice is only ever 1 or 2 points
> +# depending on bNumEndpoints being <=3 or >3 respectively.
> +# So we use that knowledge to create rules to only use port_number=1
> +# for 2 port devices
> +SUBSYSTEM!="tty", GOTO="synce_ipaq_end"
> +
> +DRIVERS=="ipaq",ATTRS{bNumEndpoints}=="", ENV{pocketpc}="env-synce"
> +DRIVERS=="ipaq",ATTRS{bNumEndpoints}=="0[0123]", ENV{pocketpc}="env-synce"
> +DRIVERS=="ipaq",ATTRS{port_number}=="1", ENV{pocketpc}="env-synce"
> +
> +ENV{pocketpc}=="env-synce", RUN+="@UDEVDIR@/synce-udev-serial"
> +
> +LABEL="synce_ipaq_end"
> diff -Naur scripts.orig/udev-synce-serial.in scripts/udev-synce-serial.in
> --- scripts.orig/udev-synce-serial.in 2010-08-25 12:44:02.000000000 +0100
> +++ scripts/udev-synce-serial.in 2010-12-23 09:26:51.427741000 +0000
> @@ -35,63 +35,6 @@
> logger.critical("Udev environment not set: DEVPATH missing")
> sys.exit(1)
>
> - # need to check for 2 port devices here
> - # there are no python bindings for udev, so this is a little awkward
> -
> - # get device name
> - cmd_list = ["udevadm", "info", "--path="+devpath, "--query=name"]
> - try:
> - proc = subprocess.Popen(cmd_list, stdout=subprocess.PIPE,
> stderr=subprocess.STDOUT)
> - output_text = proc.communicate()[0]
> - except Exception,e:
> - logger.error("failure running udevadm: %s" % e)
> - sys.exit(1)
> -
> - rc = proc.returncode
> - if rc != 0:
> - logger.error("udevadm failed with return code %d", rc)
> - sys.exit(1)
> -
> - devname = output_text.strip()
> -
> - # get parent device path
> - match_obj = re.match('^(.+)/'+name+"/"+subsystem+'/'+name, devpath)
> - if match_obj == None:
> - logger.error("failed to get parent device path from udev")
> - sys.exit(1)
> -
> - parent_devpath = match_obj.group(1)
> -
> - # find all relevant child devices of the parent
> - cmd_list = ["udevadm", "info", "--export-db"]
> - try:
> - proc = subprocess.Popen(cmd_list, stdout=subprocess.PIPE,
> stderr=subprocess.STDOUT)
> - output_text = proc.communicate()[0]
> - except Exception,e:
> - logger.error("failure getting device db: %s" % e)
> - sys.exit(1)
> -
> - rc = proc.returncode
> - if rc != 0:
> - logger.error("udevadm failed with return code %d when getting
> device db",rc)
> - sys.exit(1)
> -
> - output_list = output_text.split('\n')
> - devpath_re = re.compile('^P:
> ('+parent_devpath+'.+/'+subsystem+'/.+)')
> -
> - devpath_list = []
> - for line in output_list:
> - match_obj = devpath_re.match(line)
> - if match_obj != None:
> - path = match_obj.group(1)
> - devpath_list.append(path)
> -
> - if len(devpath_list) > 1:
> - devpath_list.sort()
> - if device_path != devpath_list[len(devpath_list) - 1]:
> - logger.warning("found inactive port, we need to mark this
> somehow")
> - sys.exit(0)
> -
> # get serial device file
> if platform.system() == "FreeBSD":
> logger.critical("do we support udev/devicekit on FreeBSD ???")
>
>
>
> ------------------------------------------------------------------------------
> Learn how Oracle Real Application Clusters (RAC) One Node allows customers
> to consolidate database storage, standardize their database environment, and,
> should the need arise, upgrade to a full multi-node Oracle RAC database
> without downtime or disruption
> http://p.sf.net/sfu/oracle-sfdevnl
> _______________________________________________
> SynCE-Devel mailing list
> SynCE-Devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/synce-devel
signature.asc
Description: This is a digitally signed message part
------------------------------------------------------------------------------ Learn how Oracle Real Application Clusters (RAC) One Node allows customers to consolidate database storage, standardize their database environment, and, should the need arise, upgrade to a full multi-node Oracle RAC database without downtime or disruption http://p.sf.net/sfu/oracle-sfdevnl
_______________________________________________ SynCE-Devel mailing list SynCE-Devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/synce-devel
