Xen devel - Nov 2010 - [PATCH] fix "xm block-detach 0 ..." for extended-ID devices

Simply taking stat()''s st_rdev doesn''t work here, as the minor
is
split into two parts, the major is present, and the "extended" bit
isn''t set.

Not sure whether non-Linux needs different decoding here.

Signed-off-by: Jan Beulich <jbeulich@novell.com>

--- a/tools/python/xen/util/blkif.py
+++ b/tools/python/xen/util/blkif.py
@@ -20,7 +20,11 @@ def blkdev_name_to_number(name):
     devnum = None
 
     try:
-        return (devname, os.stat(n).st_rdev)
+        rdev = os.stat(n).st_rdev
+        if rdev > 0xffff and (rdev & 0xff00) == (202 << 8):
+            devnum = (1 << 28) + ((rdev >> 12) & ~0xff) + (rdev
& 0xff)
+            return (devname, devnum)
+        return (devname, rdev)
     except Exception, ex:
         pass
 





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

Jan Beulich writes ("[Xen-devel] [PATCH] fix "xm block-detach 0
..." for extended-ID devices"):
> Simply taking stat()''s st_rdev doesn''t work here, as the
minor is
> split into two parts, the major is present, and the "extended"
bit
> isn''t set.
...
> @@ -20,7 +20,11 @@ def blkdev_name_to_number(name):
> -        return (devname, os.stat(n).st_rdev)
This seems to me to be entirely wrong.  When you block detach you
should be providing the device name according to the Xen guest device
naming scheme, not a local device path (which may have different major
and minor numbers).

If you just remove the try/except block and always use the following
code, does it work correctly ?

Ian.

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

>>> On 19.11.10 at 18:56, Ian Jackson <Ian.Jackson@eu.citrix.com>
wrote:
> Jan Beulich writes ("[Xen-devel] [PATCH] fix "xm block-detach 0
..." for
> extended-ID devices"):
>> Simply taking stat()''s st_rdev doesn''t work here, as
the minor is
>> split into two parts, the major is present, and the
"extended" bit
>> isn''t set.
> ...
>> @@ -20,7 +20,11 @@ def blkdev_name_to_number(name):
>> -        return (devname, os.stat(n).st_rdev)
> 
> This seems to me to be entirely wrong.  When you block detach you
> should be providing the device name according to the Xen guest device
> naming scheme, not a local device path (which may have different major
> and minor numbers).
>
> If you just remove the try/except block and always use the following
> code, does it work correctly ?
Yes.

Since the code is there presumably for a reason, I didn''t dare
to suggest removing it altogether.

Jan


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

Jan Beulich writes ("Re: [Xen-devel] [PATCH] fix "xm block-detach 0
..." for extended-ID devices"):
> On 19.11.10 at 18:56, Ian Jackson <Ian.Jackson@eu.citrix.com> wrote:
> > If you just remove the try/except block and always use the following
> > code, does it work correctly ?
> 
> Yes.
Good.  Let''s do that then :-).  Would you care to submit it as a
patch and I''ll put my acked-by on it ?
> Since the code is there presumably for a reason, I didn''t dare
> to suggest removing it altogether.
Historically, there has been a different (not to say ENTIRELY
WRONGHEADED) idea of what the vbd device numbers are supposed to
mean.

[lib]xl doesn''t ever stat devices in the dom0 filesystem for this
purpose.

Ian.

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