Re: [Qemu-devel] [PATCH 3/3] TPM2 ACPI table support

[Michael S. Tsirkin]

On Mon, Apr 13, 2015 at 10:29:32PM -0400, Stefan Berger wrote:
> On 04/13/2015 02:27 AM, Michael S. Tsirkin wrote:
> >>@@ -1428,12 +1444,22 @@ void acpi_build(PcGuestInfo *guest_info, 
> >>AcpiBuildTables *tables)
> >>          acpi_add_table(table_offsets, tables_blob);
> >>          build_hpet(tables_blob, tables->linker);
> >>      }
> >>-    if (misc.has_tpm) {
> >>+    if (misc.tpm_version != TPMVersion_Unspec) {
> >>          acpi_add_table(table_offsets, tables_blob);
> >>          build_tpm_tcpa(tables_blob, tables->linker, tables->tcpalog);
> >>          acpi_add_table(table_offsets, tables_blob);
> >>-        build_tpm_ssdt(tables_blob, tables->linker);
> >>+        switch (misc.tpm_version) {
> >>+        case TPMVersion1_2:
> >>+            build_tpm_ssdt(tables_blob, tables->linker);
> >>+            break;
> >>+        case TPMVersion2_0:
> >>+            build_tpm2(tables_blob, tables->linker);
> >>+            break;
> >>+        case TPMVersion_Unspec:
> >>+            /* not possible */
> >>+            break;
> >>+        }
> >>      }
> >>      if (guest_info->numa_nodes) {
> >>          acpi_add_table(table_offsets, tables_blob);
> >I think we should fix QEMU coding style violations in TPM code.
> >In particular, mixing of _ and camel case, enum and define values
> >not all upper case, local variables not all lower case.
> >This was less of an issue when the violations were contained
> >within hw/tpm/, but it's more of an issue now that this affects
> >other code.
> >
> 
> The above enums look like the only violation to me. Fixed.
> 
>     Stefan


there's also needIrq in existing code, let's fix that.