[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH 6 of 9 v2] libxl: execute hotplug scripts directly from libxl
2011/11/24 Ian Campbell <Ian.Campbell@xxxxxxxxxx>: > I was just looking at this again because someone was talking about > hotplug issues on Linux and something new occurred to me. > > On Fri, 2011-11-18 at 11:59 +0000, Roger Pau Monne wrote: >> [...] >> +++ b/tools/libxl/hotplug_linux.c    Fri Sep 30 14:38:55 2011 +0200 >> [...] >> +int libxl_disk_hotplug(libxl__gc *gc, libxl__device *dev) >> [...] >> +int libxl_nic_hotplug_connect(libxl__gc *gc, libxl__device *dev) >> [...] >> +++ b/tools/libxl/hotplug_netbsd.c   ÂFri Sep 30 14:38:55 2011 +0200 >> [...] >> +int libxl__disk_hotplug(libxl__gc *gc, libxl__device *dev) >> [...] >> +int libxl__nic_hotplug(libxl__gc *gc, libxl__device *dev) > > Although Linux doesn't do hotplug this way now I think it will/should in > the future and the code will look very much like the netbsd one. So I > think that the code which actually handles calling the script should > probably be a generic function, only the selector on whether to do so > needs to be platform specific. Sorry for the delay, I've been ill most of this week. From what I saw, the problem is that NetBSD and Linux hotplug scripts have different parameters, that's why I've split the calling functions into two different files, and made them OS-dependent. I think I said that before, but I would be good to agree on the calling parameters passed to hotplug scripts. Currently NetBSD has the following parameters: * block: <backend path> <status> <type> where <type> can be "file" or "phy" * vif-bridge: <backend path> <status> * vif-ip: <backend path> <status> Also Linux hotplug scripts need a good clean up, hard tabs and spaces are all mixed, and makes the code really hard to understand. Regards, Roger. _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxxxxxxxx http://lists.xensource.com/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |