On Mon, Feb 26, 2018 at 04:08:44PM -0700, Jim Fehlig wrote:
On 02/08/2018 03:58 PM, Marek Marczykowski-Górecki wrote:
+const char * +xenTranslateCPUFeature(const char *feature_name, bool from_libxl) +{ + static const char *translation_table[][2] = { + /* libvirt name, libxl name */ + { "cx16", "cmpxchg16" }, + { "cid", "cntxid" }, + { "ds_cpl", "dscpl" }, + { "pclmuldq", "pclmulqdq" }, + { "pni", "sse3" }, + { "ht", "htt" }, + { "pn", "psn" }, + { "clflush", "clfsh" }, + { "sep", "sysenter" }, + { "cx8", "cmpxchg8" }, + { "nodeid_msr", "nodeid" }, + { "cr8legacy", "altmovcr8" }, + { "lahf_lm", "lahfsahf" }, + { "cmp_legacy", "cmplegacy" }, + { "fxsr_opt", "ffxsr" }, + { "pdpe1gb", "page1gb" }, + };
In the case of ibrsb and stibp added in xen commit 0d703a701cc and spec-ctrl added in libvirt commit 8b605530e80, have we already encountered the drawback of this approach discussed in V3?
Great... I'll add those names to the list... Anyway I think this is still better approach than the alternative. Unless someone want to extend the CPU API with single bit lookup... -- Best Regards, Marek Marczykowski-Górecki Invisible Things Lab A: Because it messes up the order in which people normally read text. Q: Why is top-posting such a bad thing?