On 28/08/18 15:48, Wei Liu wrote: > On Tue, Aug 28, 2018 at 08:36:37AM -0600, Jan Beulich wrote: >>>>> On 28.08.18 at 15:19, <wei.l...@citrix.com> wrote: >>> On Tue, Aug 28, 2018 at 05:44:51AM -0600, Jan Beulich wrote: >>>>>>> On 26.08.18 at 14:19, <wei.l...@citrix.com> wrote: >>>>> Move the file to x86 common code and change its name to emul-i8254.c. >>>>> >>>>> Put HVM only code under CONFIG_HVM or is_hvm_domain. >>>>> >>>>> Signed-off-by: Wei Liu <wei.l...@citrix.com> >>>>> --- >>>>> v2: move the whole file. >>>>> --- >>>>> xen/arch/x86/Makefile | 1 +- >>>>> xen/arch/x86/emul-i8254.c | 609 +++++++++++++++++++++++++++++++++++++++- >>>>> xen/arch/x86/hvm/Makefile | 1 +- >>>>> xen/arch/x86/hvm/i8254.c | 597 +-------------------------------------- >>>>> 4 files changed, 610 insertions(+), 598 deletions(-) >>>>> create mode 100644 xen/arch/x86/emul-i8254.c >>>>> delete mode 100644 xen/arch/x86/hvm/i8254.c >>>> On the assumption that the changes to the file are only CONFIG_HVM >>>> additions >>> No only that. There are a few is_hvm_domain calls as well. >>> >>>> Acked-by: Jan Beulich <jbeul...@suse.com> >>> Does this still stand? >> Yes, albeit looking at the full diff of the files I'm then a little >> puzzled about is_hvm_domain() uses: Why does pit_load_count() >> need an #ifdef added at all, with the is_hvm_vcpu() check >> right before it? > Because if I don't do that we get "defined but not used" warnings.
(hopefully) not with the newer is_hvm_vcpu() which evaluates its arguments properly. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
