Hollis Blanchard
2006-Aug-14 19:37 UTC
[Xen-devel] Re: [Xen-changelog] [XENBUS] Avoid direct use of xen_start_info.
This is great, because on PowerPC we''re looking to drop struct start_info entirely. However, I notice there are still a fair number of ''xen_start_info'' references in drivers/xen. Are there more of these patches coming Steven, or since HVM guests don''t use those drivers you''re not pursuing them? On Mon, 2006-08-14 at 19:20 +0000, Xen patchbot-unstable wrote:> # HG changeset patch > # User kfraser@localhost.localdomain > # Node ID e061f14161b052b3287a1e41222cfa292a194fc2 > # Parent 9519445d9e9d7b8bd48fcd93d1bba485508ddbff > [XENBUS] Avoid direct use of xen_start_info. It''s unevailable when > building as a separate driver for an HVM guest. > Signed-off-by: Steven Smith <ssmith@xensource.com> > Signed-off-by: Keir Fraser <keir@xensource.com>...> @@ -1014,8 +1017,13 @@ static int __init xenbus_probe_init(void > if (xsd_port_intf) > xsd_port_intf->read_proc = xsd_port_read; > #endif > - } else > + } else { > xenstored_ready = 1; > + xen_store_evtchn = xen_start_info->store_evtchn; > + xen_store_mfn = xen_start_info->store_mfn; > + } > + > + xen_store_interface = mfn_to_virt(xen_store_mfn); > > /* Initialize the interface to xenstore. */ > err = xs_init();Could we abstract this part more please? I think calling into an arch-provided xen_get_store_mfn() function would be good. (And of course the same for store_evtchn and the console_* equivalents.) -- Hollis Blanchard IBM Linux Technology Center _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Keir Fraser
2006-Aug-15 09:01 UTC
Re: [Xen-devel] Re: [Xen-changelog] [XENBUS] Avoid direct use of xen_start_info.
On 14/8/06 8:37 pm, "Hollis Blanchard" <hollisb@us.ibm.com> wrote:> Could we abstract this part more please? I think calling into an > arch-provided xen_get_store_mfn() function would be good. (And of course > the same for store_evtchn and the console_* equivalents.)No, that''s it from Steve''s patch queue. Further cleanup here would be nice though, so please send patches. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Steven Smith
2006-Aug-15 09:04 UTC
[Xen-devel] Re: [Xen-changelog] [XENBUS] Avoid direct use of xen_start_info.
> This is great, because on PowerPC we''re looking to drop struct > start_info entirely. However, I notice there are still a fair number of > ''xen_start_info'' references in drivers/xen. Are there more of these > patches coming Steven, or since HVM guests don''t use those drivers > you''re not pursuing them?I wasn''t planning on doing so since, as you say, none of the others are relevant to HVM guests.> > @@ -1014,8 +1017,13 @@ static int __init xenbus_probe_init(void > > if (xsd_port_intf) > > xsd_port_intf->read_proc = xsd_port_read; > > #endif > > - } else > > + } else { > > xenstored_ready = 1; > > + xen_store_evtchn = xen_start_info->store_evtchn; > > + xen_store_mfn = xen_start_info->store_mfn; > > + } > > + > > + xen_store_interface = mfn_to_virt(xen_store_mfn); > > > > /* Initialize the interface to xenstore. */ > > err = xs_init(); > Could we abstract this part more please? I think calling into an > arch-provided xen_get_store_mfn() function would be good. (And of course > the same for store_evtchn and the console_* equivalents.)This''d probably reduce the number of #ifdef''s in the PV-on-HVM patches as well, so I''d be happy to see it. Steven. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel