Change in xm interface

Change in xm interface

[Anthony Liguori]

[-- Attachment #1: Type: text/plain, Size: 407 bytes --]

This is a great patch, I just want to make sure everyone who has written 
a tool against xm that is perhaps screen scrapping the output of xm 
block-list or xm network-list that the interface has changed.

You will now have to use --long to get the same S-Expression output.

Perhaps this sort of changes should be announced on xen-announce in the 
future?  Is there such a list?

Regards,

Anthony Liguori

[-- Attachment #2: [Xen-changelog] This patch displays the	network-list/block-list/vtpm-list using the 'xm' --]
[-- Type: message/rfc822, Size: 10899 bytes --]

From: Xen patchbot -unstable <patchbot-unstable@lists.xensource.com>
To: xen-changelog@lists.xensource.com
Subject: [Xen-changelog] This patch displays the network-list/block-list/vtpm-list using the 'xm'
Date: Fri, 17 Feb 2006 19:44:07 +0000
Message-ID: <E1FABWR-0001gR-KN@xenbits.xensource.com>

# HG changeset patch
# User emellor@leeni.uk.xensource.com
# Node ID 34f6a1efe52d8c5c722a2d423a0b26b469540a3a
# Parent  aeeeedc6c9b754dc9d26b4bc12ab9d025578da6e
This patch displays the network-list/block-list/vtpm-list using the 'xm'
command. It supports the '--long' option.

Signed-off-by: Stefan Berger <stefanb@us.ibm.com>

diff -r aeeeedc6c9b7 -r 34f6a1efe52d tools/python/xen/xm/main.py
--- a/tools/python/xen/xm/main.py	Fri Feb 17 17:22:23 2006
+++ b/tools/python/xen/xm/main.py	Fri Feb 17 18:35:38 2006
@@ -90,18 +90,18 @@
                                     where <DevId> may either be the device ID
                                     or the device name as mounted in the guest"""
 
-block_list_help = "block-list <DomId>               List virtual block devices for a domain"
+block_list_help = "block-list <DomId> [--long]      List virtual block devices for a domain"
 network_attach_help = """network-attach  <DomID> [script=<script>] [ip=<ip>] [mac=<mac>]
                            [bridge=<bridge>] [backend=<backDomID>]
                                     Create a new virtual network device """
 network_detach_help = """network-detach  <DomId> <DevId>  Destroy a domain's virtual network
                                     device, where <DevId> is the device ID."""
 
-network_list_help = "network-list <DomId>             List virtual network interfaces for a domain"
+network_list_help = "network-list <DomId> [--long]    List virtual network interfaces for a domain"
 vnet_list_help = "vnet-list [-l|--long]            list vnets"
 vnet_create_help = "vnet-create <config>             create a vnet from a config file"
 vnet_delete_help = "vnet-delete <vnetid>             delete a vnet"
-vtpm_list_help = "vtpm-list <DomId>                list virtual TPM devices"
+vtpm_list_help = "vtpm-list <DomId> [--long]       list virtual TPM devices"
 
 short_command_list = [
     "console",
@@ -684,29 +684,133 @@
     from xen.xend.XendClient import server
     print server.xend_node_log()
 
+def parse_dev_info(info):
+    def get_info(n, t, d):
+        i = 0
+        while i < len(info):
+            if (info[i][0] == n):
+                return t(info[i][1])
+            i = i + 1
+        return t(d)
+    return {
+        #common
+        'backend-id' : get_info('backend-id',   int,   -1),
+        'handle'     : get_info('handle',       int,    0),
+        'state'      : get_info('state',        int,   -1),
+        'be-path'    : get_info('backend',      str,   '??'),
+        'event-ch'   : get_info('event-channel',int,   -1),
+        #network specific
+        'virtual-device' : get_info('virtual-device', str, '??'),
+        'tx-ring-ref': get_info('tx-ring-ref',  int,   -1),
+        'rx-ring-ref': get_info('rx-ring-ref',  int,   -1),
+        'mac'        : get_info('mac',          str,   '??'),
+        #block-device specific
+        'ring-ref'   : get_info('ring-ref',     int,   -1),
+        }
+
+def has_long_option(args):
+    use_long = 0
+    try:
+        (options, params) = getopt.gnu_getopt(args, 'l', ['long'])
+    except getopt.GetoptError, opterr:
+        err(opterr)
+        sys.exit(1)
+
+    for (k, v) in options:
+        if k in ['-l', '--long']:
+            use_long = 1
+    return (use_long, params)
+
 def xm_network_list(args):
-    arg_check(args, "network-list", 1)
-    dom = args[0]
-    from xen.xend.XendClient import server
-    for x in server.xend_domain_devices(dom, 'vif'):
-        sxp.show(x)
-        print
+    arg_check(args, "network-list", 1, 2)
+
+    (use_long, params) = has_long_option(args)
+
+    if len(params) == 0:
+        print 'No domain parameter given'
+        sys.exit(1)
+    dom = params[0]
+    from xen.xend.XendClient import server
+    if use_long:
+        devs = server.xend_domain_devices(dom, 'vif')
+        map(PrettyPrint.prettyprint, devs)
+    else:
+        hdr = 0
+        for x in server.xend_domain_devices(dom, 'vif'):
+            if hdr == 0:
+                print 'Idx BE     MAC Addr.     handle state evt-ch tx-/rx-ring-ref BE-path'
+                hdr = 1
+            ni = parse_dev_info(x[1])
+            ni['idx'] = int(x[0])
+            print ("%(idx)-3d "
+                   "%(backend-id)-3d"
+                   "%(mac)-17s    "
+                   "%(handle)-3d   "
+                   "%(state)-3d    "
+                   "%(event-ch)-3d   "
+                   "%(tx-ring-ref)-5d/%(rx-ring-ref)-5d   "
+                   "%(be-path)-30s  "
+                   % ni)
 
 def xm_block_list(args):
-    arg_check(args, "block-list", 1)
-    dom = args[0]
-    from xen.xend.XendClient import server
-    for x in server.xend_domain_devices(dom, 'vbd'):
-        sxp.show(x)
-        print
+    arg_check(args, "block-list", 1, 2)
+
+    (use_long, params) = has_long_option(args)
+
+    if len(params) == 0:
+        print 'No domain parameter given'
+        sys.exit(1)
+    dom = params[0]
+    from xen.xend.XendClient import server
+    if use_long:
+        devs = server.xend_domain_devices(dom, 'vbd')
+        map(PrettyPrint.prettyprint, devs)
+    else:
+        hdr = 0
+        for x in server.xend_domain_devices(dom, 'vbd'):
+            if hdr == 0:
+                print 'Vdev  BE handle state evt-ch ring-ref BE-path'
+                hdr = 1
+            ni = parse_dev_info(x[1])
+            ni['idx'] = int(x[0])
+            print ("%(idx)-3d    "
+                   "%(backend-id)-3d  "
+                   "%(handle)-3d   "
+                   "%(state)-3d    "
+                   "%(event-ch)-3d    "
+                   "%(ring-ref)-5d "
+                   "%(be-path)-30s  "
+                   % ni)
 
 def xm_vtpm_list(args):
-    arg_check(args, "vtpm-list", 1)
-    dom = args[0]
-    from xen.xend.XendClient import server
-    for x in server.xend_domain_devices(dom, 'vtpm'):
-        sxp.show(x)
-        print
+    arg_check(args, "vtpm-list", 1, 2)
+
+    (use_long, params) = has_long_option(args)
+
+    if len(params) == 0:
+        print 'No domain parameter given'
+        sys.exit(1)
+    dom = params[0]
+    from xen.xend.XendClient import server
+    if use_long:
+        devs = server.xend_domain_devices(dom, 'vtpm')
+        map(PrettyPrint.prettyprint, devs)
+    else:
+        hdr = 0
+        for x in server.xend_domain_devices(dom, 'vtpm'):
+            if hdr == 0:
+                print 'Idx  BE handle state evt-ch ring-ref BE-path'
+                hdr = 1
+            ni = parse_dev_info(x[1])
+            ni['idx'] = int(x[0])
+            print ("%(idx)-3d   "
+                   "%(backend-id)-3d  "
+                   "%(handle)-3d   "
+                   "%(state)-3d    "
+                   "%(event-ch)-3d    "
+                   "%(ring-ref)-5d "
+                   "%(be-path)-30s  "
+                   % ni)
 
 def xm_block_attach(args):
     arg_check(args, 'block-attach', 4, 5)

_______________________________________________
Xen-changelog mailing list
Xen-changelog@lists.xensource.com
http://lists.xensource.com/xen-changelog


[-- Attachment #3: Type: text/plain, Size: 138 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

Re: Change in xm interface

[Ewan Mellor]

On Fri, Feb 17, 2006 at 03:06:00PM -0600, Anthony Liguori wrote:

> This is a great patch, I just want to make sure everyone who has written 
> a tool against xm that is perhaps screen scrapping the output of xm 
> block-list or xm network-list that the interface has changed.
> 
> You will now have to use --long to get the same S-Expression output.

Your concern is perfectly reasonable.  On that note though, could I
emphasise

  a) Screen-scraping xm is a bad idea.  There is an interface straight
  into xend that one can use for integration (the same interface that xm
  uses).  This will avoid such hackery, and make it an awful lot easier
  to maintain backwards compatibility.  xm is for humans to use, not
  tools.

  b) There is no guarantee that any interface inside domain 0, be it
  to the tools, or even to Xen, are fixed.  The only interfaces that are fixed
  at the moment are the interfaces used by guests.  Everything else is
  subject to change at any time.  Of course, it's in everyone's
  interests that these interfaces become fixed in the future, but
  there's plenty of work that has to go into them yet before we could
  claim that they are supportable in the long term.

> Perhaps this sort of changes should be announced on xen-announce in the 
> future?  Is there such a list?

That list is xen-devel, for the moment.  I'm not sure there's any value
in having a separate list -- anyone who's developing products that sit
on top of Xen _really_ needs to be reading xen-devel anyway -- the
project simply isn't mature enough for you to get away without doing
that.

Ewan.

Re: Change in xm interface

[Anthony Liguori]

Ewan Mellor wrote:
> Your concern is perfectly reasonable.  On that note though, could I
> emphasise
>
>   a) Screen-scraping xm is a bad idea.  There is an interface straight
>   into xend that one can use for integration (the same interface that xm
>   uses).  This will avoid such hackery, and make it an awful lot easier
>   to maintain backwards compatibility.  xm is for humans to use, not
>   tools.
>   
I agree with you completely and will take the time to point out that 
libvirt uses the S-Expression/HTTP interface so you can just use that :-)
>   b) There is no guarantee that any interface inside domain 0, be it
>   to the tools, or even to Xen, are fixed.  The only interfaces that are fixed
>   at the moment are the interfaces used by guests.  Everything else is
>   subject to change at any time.  Of course, it's in everyone's
>   interests that these interfaces become fixed in the future, but
>   there's plenty of work that has to go into them yet before we could
>   claim that they are supportable in the long term.
>   
Yup.
> That list is xen-devel, for the moment.  I'm not sure there's any value
> in having a separate list -- anyone who's developing products that sit
> on top of Xen _really_ needs to be reading xen-devel anyway -- the
> project simply isn't mature enough for you to get away without doing
> that.
>   
Ok.  Always good to reiterate this occasionally :-)

Regards,

Anthony Liguori
> Ewan.
>
>   

Re: Change in xm interface

[Nivedita Singhvi]

Ewan Mellor wrote:

>>Perhaps this sort of changes should be announced on xen-announce in the 
>>future?  Is there such a list?
> 
> 
> That list is xen-devel, for the moment.  I'm not sure there's any value
> in having a separate list -- anyone who's developing products that sit
> on top of Xen _really_ needs to be reading xen-devel anyway -- the
> project simply isn't mature enough for you to get away without doing
> that.

Heavily seconded. Please don't fork mailing lists - it really
isn't warranted, and tends to exaggerate the problem we're
trying to solve (people not seeing the information they need
because it's hidden somewhere)...

thanks,
Nivedita