On 11/03/2020 15:22, Jan Beulich wrote: > On 11.03.2020 14:09, Andrew Cooper wrote: >> On 10/03/2020 15:48, Jan Beulich wrote: >>> Drop #include-s not needed by the header itself. Put the ones needed >>> into whichever other files actually need them. >>> >>> Signed-off-by: Jan Beulich <jbeul...@suse.com> >>> --- >>> v2: Also make things build with XSM=y. >> Looking better, but still got problems. >> >> xen_pv_console.c: In function ‘pv_console_init’: >> xen_pv_console.c:51:37: error: ‘HVM_PARAM_CONSOLE_PFN’ undeclared (first >> use in this function) >> r = xen_hypercall_hvm_get_param(HVM_PARAM_CONSOLE_PFN, &raw_pfn); >> ^~~~~~~~~~~~~~~~~~~~~ >> >> and >> >> shim.c: In function ‘pv_shim_fixup_e820’: >> shim.c:148:20: error: ‘HVM_PARAM_STORE_PFN’ undeclared (first use in >> this function) >> MARK_PARAM_RAM(HVM_PARAM_STORE_PFN); > Oh, so that's an XSM+shim config aiui; Among the sets of what > I regularly test I have only an XSM one and a shim one. The funs > of allowing too wide a variety of different configs ... In cases > like this series here I really don't see how one is supposed to > cover _all_ possible configs; randconfig builds won't help with > this unless one would let them run until they've covered all > possible ones.
This happens to be my default config. I wouldn't expect to be perfect even under weird combinations, but these aren't weird, and this is the kind of change that `make allyesconfig` is intended for. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel