On Thu, 10 Oct 2019 21:08:10 +0800
Guoheyi <guoh...@huawei.com> wrote:
> Thanks for your comments.
>
> On 2019/10/10 17:15, Igor Mammedov wrote:
> > On Tue, 24 Sep 2019 23:21:51 +0800
> > Heyi Guo <guoh...@huawei.com> wrote:
> >
> >> Add SDEI table if SDEI is enabled, so that guest OS can get aware and
> >> utilize the interfaces.
> >>
> >> Signed-off-by: Heyi Guo <guoh...@huawei.com>
> >> Cc: Peter Maydell <peter.mayd...@linaro.org>
> >> Cc: Dave Martin <dave.mar...@arm.com>
> >> Cc: Marc Zyngier <marc.zyng...@arm.com>
> >> Cc: Mark Rutland <mark.rutl...@arm.com>
> >> Cc: James Morse <james.mo...@arm.com>
> >> Cc: Shannon Zhao <shannon.zha...@gmail.com>
> >> Cc: "Michael S. Tsirkin" <m...@redhat.com>
> >> Cc: Igor Mammedov <imamm...@redhat.com>
> >> ---
> >> hw/arm/virt-acpi-build.c | 16 ++++++++++++++++
> >> include/hw/acpi/acpi-defs.h | 5 +++++
> >> 2 files changed, 21 insertions(+)
> >>
> >> diff --git a/hw/arm/virt-acpi-build.c b/hw/arm/virt-acpi-build.c
> >> index 6cdf156..1088214 100644
> >> --- a/hw/arm/virt-acpi-build.c
> >> +++ b/hw/arm/virt-acpi-build.c
> >> @@ -32,6 +32,7 @@
> >> #include "trace.h"
> >> #include "hw/core/cpu.h"
> >> #include "target/arm/cpu.h"
> >> +#include "target/arm/sdei.h"
> >> #include "hw/acpi/acpi-defs.h"
> >> #include "hw/acpi/acpi.h"
> >> #include "hw/nvram/fw_cfg.h"
> >> @@ -475,6 +476,16 @@ build_iort(GArray *table_data, BIOSLinker *linker,
> >> VirtMachineState *vms)
> >> }
> >>
> > here should be a comment describing purpose with a reference to spec
> OK.
>
> >
> >> static void
> >> +build_sdei(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
> >> +{
> >> + int sdei_start = table_data->len;
> >> +
> >> + (void)acpi_data_push(table_data, sizeof(AcpiSdei));
> >> + build_header(linker, table_data, (void *)(table_data->data +
> >> sdei_start),
> >> + "SDEI", table_data->len - sdei_start, 1, NULL, NULL);
> >> +}
> > that's dummy table and doesn't match what spec describes,
> > It doesn't look correct SDEI table.
>
> Refer to below document, Appendix C, I see the definition of SDEI is
> really a dummy table, also it is said in the document: The table
> consists only of a basic header with revision 1.
>
> http://infocenter.arm.com/help/topic/com.arm.doc.den0054a/ARM_DEN0054A_Software_Delegated_Exception_Interface.pdf
>
> Do I miss anything?
No you are right and I was mistaken.
It's pretty waste-full way to communicate to OS a Boolean flag
that a new GHEST event is supported. But since they managed to make
it a spec, it should be coded as such.
>
> >
> >> +
> >> +static void
> >> build_spcr(GArray *table_data, BIOSLinker *linker, VirtMachineState *vms)
> >> {
> >> AcpiSerialPortConsoleRedirection *spcr;
> >> @@ -796,6 +807,11 @@ void virt_acpi_build(VirtMachineState *vms,
> >> AcpiBuildTables *tables)
> >> acpi_add_table(table_offsets, tables_blob);
> >> build_spcr(tables_blob, tables->linker, vms);
> >>
> >> + if (sdei_enabled) {
> >> + acpi_add_table(table_offsets, tables_blob);
> >> + build_sdei(tables_blob, tables->linker, vms);
> >> + }
> >> +
> >> if (ms->numa_state->num_nodes > 0) {
> >> acpi_add_table(table_offsets, tables_blob);
> >> build_srat(tables_blob, tables->linker, vms);
> >> diff --git a/include/hw/acpi/acpi-defs.h b/include/hw/acpi/acpi-defs.h
> >> index 57a3f58..0a2265d 100644
> >> --- a/include/hw/acpi/acpi-defs.h
> >> +++ b/include/hw/acpi/acpi-defs.h
> >> @@ -634,4 +634,9 @@ struct AcpiIortRC {
> >> } QEMU_PACKED;
> >> typedef struct AcpiIortRC AcpiIortRC;
> >>
> >> +struct AcpiSdei {
> >> + ACPI_TABLE_HEADER_DEF /* ACPI common table header */
> >> +} QEMU_PACKED;
> >> +typedef struct AcpiSdei AcpiSdei;
> > we don't use packed structures for ACPI anymore, see build_rsdp() for
> > example of how tables are composed and comment style required for each
> > field.
> Sure; will fix it in v2.
>
> Thanks,
>
> Heyi
>
> >
> >> #endif
> >
> > .
> >
>
>
