[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH v2 07/17] libxc/restore: STATIC_DATA_END inference for v2 compatibility

  • To: Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>
  • From: Ian Jackson <ian.jackson@xxxxxxxxxx>
  • Date: Fri, 29 May 2020 16:53:44 +0100
  • Authentication-results: esa2.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none
  • Cc: Xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxxx>, Wei Liu <wl@xxxxxxx>
  • Delivery-date: Fri, 29 May 2020 15:53:58 +0000
  • Ironport-sdr: VgNztI4E0pMS45/t1Eue0i0VhZ8BPddJ995Bb7c5IW+YwIGDfxm1S9e9Apfq0qh1H5OSLa1fIs OTiPImdgUeaUxIAjjlXVo3u5nXtg0ygzgy7UPns3sbo/TvnL8rZHLNqxv4rhuH24QPwkMmzuL2 dQZxwyepc8mKlvJBJVhWAh8+5PnL41U36VEwc86lTtjUXSZxgXYply75N6mFWtuZCzHd0r1fea Ze+vBS6SM0WpPvsflqUdCWSGx21AIyRFclj7Nk8V6Ry/HGXlXwuujSEL2pwVdObfqEvF1Er+z3 hTE=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

Andrew Cooper writes ("Re: [PATCH v2 07/17] libxc/restore: STATIC_DATA_END 
inference for v2 compatibility"):
> On 05/03/2020 16:24, Ian Jackson wrote:
> > Andrew Cooper writes ("Re: [PATCH v2 07/17] libxc/restore: STATIC_DATA_END 
> > inference for v2 compatibility"):
> >> More importantly however, by design, common code can't call
> >> arch-specific code without a restore_ops hook.  Deduping these would
> >> require breaking the restriction which is currently doing a decent job
> >> of keeping x86-isms out of common code.
> > I'm afraid you're going to have to explain that to me at a bit greater
> > length.  The biggest thing that is confusing me about your statement
> > here is that your patch is indeed adding x86-specific code to a common
> > file.  But also I don't understand the comment about restore_ops
> > hooks - do you mean something in restore_callbacks ?
> No.  restore_callbacks are plumbing between libxl-save-helper and libxc.
> restore_ops are internal to the restore code, and come in x86_pv and
> x86_hvm flavours, with ARM existing in some theoretical future.  The
> design of the common save/restore code had deliberate measures put in
> place to make it hard to get arch-specific details slip into common
> code, so porting to different architectures didn't have to start by
> doing a bunch of cleanup.
> tl;dr I could put an #ifdef x86'd static inline in the root common
> header (xc_sr_common.h), but the overall complexity is greater than what
> is presented here.

Well, I still don't quite follow but as you point out on irc I haven't
replied for too long.  I don't think I should withhold my ack at this

Acked-by: Ian Jackson <ian.jackson@xxxxxxxxxxxxx>




Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.