On Tue, Jan 17, 2017 at 06:16:36PM +0000, Andrew Cooper wrote: > On 17/01/17 18:05, Konrad Rzeszutek Wilk wrote: > > On Mon, Jan 16, 2017 at 01:04:09PM +0000, Andrew Cooper wrote: > >> The chageset/version/compile information is currently exported as a set of > >> function calls into a separate translation unit, which is inefficient for > >> all > >> callers. > >> > >> Replace the function calls with externs pointing appropriately into > >> .rodata, > >> which allows all users to generate code referencing the data directly. > >> > >> No functional change, but causes smaller and more efficient compiled code. > >> > >> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com> > > Ah crud. That breaks the livepatch test-cases (they patch the > > xen_extra_version > > function). > > Lucky I haven't pushed it then, (although the livepatch build seems to > still work fine for me, despite this change.)
make tests should fail. > > > Are there some other code that can be modified that is reported > > by 'xl info' on which the test-cases can run (and reported easily?). > > Patch do_version() itself to return the same difference of information? Ugh. That is going to make the building of test-cases quite complex. I guess it can just do it and .. return only one value :-) > > ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel
