On Wed, Jun 21, 2017 at 09:10:51AM +0100, Andrew Cooper wrote:
> On 20/06/2017 21:14, Daniel Kiper wrote:
> > Current approach, wholesale efi struct initialization from efi_xen, is not
> > good. Usually if new member is defined then it is properly initialized in
> > drivers/firmware/efi/efi.c but not in arch/x86/xen/efi.c. As I saw it happened
> > a few times until now. So, let's initialize only efi struct members used by
> > Xen to avoid such issues in the future.
> >
> > Signed-off-by: Daniel Kiper <daniel.kiper@xxxxxxxxxx>
> > ---
> > arch/x86/xen/efi.c | 45 ++++++++++++---------------------------------
> > 1 file changed, 12 insertions(+), 33 deletions(-)
> >
> > diff --git a/arch/x86/xen/efi.c b/arch/x86/xen/efi.c
> > index 30bb2e8..01b9faf 100644
> > --- a/arch/x86/xen/efi.c
> > +++ b/arch/x86/xen/efi.c
> > @@ -54,38 +54,6 @@
> > .tables = EFI_INVALID_TABLE_ADDR /* Initialized later. */
> > };
> >
> > -static const struct efi efi_xen __initconst = {
> > - .systab = NULL, /* Initialized later. */
> > - .runtime_version = 0, /* Initialized later. */
> > - .mps = EFI_INVALID_TABLE_ADDR,
> > - .acpi = EFI_INVALID_TABLE_ADDR,
> > - .acpi20 = EFI_INVALID_TABLE_ADDR,
> > - .smbios = EFI_INVALID_TABLE_ADDR,
> > - .smbios3 = EFI_INVALID_TABLE_ADDR,
> > - .sal_systab = EFI_INVALID_TABLE_ADDR,
> > - .boot_info = EFI_INVALID_TABLE_ADDR,
> > - .hcdp = EFI_INVALID_TABLE_ADDR,
> > - .uga = EFI_INVALID_TABLE_ADDR,
> > - .uv_systab = EFI_INVALID_TABLE_ADDR,
> > - .fw_vendor = EFI_INVALID_TABLE_ADDR,
> > - .runtime = EFI_INVALID_TABLE_ADDR,
> > - .config_table = EFI_INVALID_TABLE_ADDR,
> > - .get_time = xen_efi_get_time,
> > - .set_time = xen_efi_set_time,
> > - .get_wakeup_time = xen_efi_get_wakeup_time,
> > - .set_wakeup_time = xen_efi_set_wakeup_time,
> > - .get_variable = xen_efi_get_variable,
> > - .get_next_variable = xen_efi_get_next_variable,
> > - .set_variable = xen_efi_set_variable,
> > - .query_variable_info = xen_efi_query_variable_info,
> > - .update_capsule = xen_efi_update_capsule,
> > - .query_capsule_caps = xen_efi_query_capsule_caps,
> > - .get_next_high_mono_count = xen_efi_get_next_high_mono_count,
> > - .reset_system = xen_efi_reset_system,
> > - .set_virtual_address_map = NULL, /* Not used under Xen. */
> > - .flags = 0 /* Initialized later. */
> > -};
> > -
> > static efi_system_table_t __init *xen_efi_probe(void)
> > {
> > struct xen_platform_op op = {
> > @@ -102,7 +70,18 @@ static efi_system_table_t __init *xen_efi_probe(void)
> >
> > /* Here we know that Xen runs on EFI platform. */
> >
> > - efi = efi_xen;
> > + efi.get_time = xen_efi_get_time;
> > + efi.set_time = xen_efi_set_time;
> > + efi.get_wakeup_time = xen_efi_get_wakeup_time;
> > + efi.set_wakeup_time = xen_efi_set_wakeup_time;
> > + efi.get_variable = xen_efi_get_variable;
> > + efi.get_next_variable = xen_efi_get_next_variable;
> > + efi.set_variable = xen_efi_set_variable;
> > + efi.query_variable_info = xen_efi_query_variable_info;
> > + efi.update_capsule = xen_efi_update_capsule;
> > + efi.query_capsule_caps = xen_efi_query_capsule_caps;
> > + efi.get_next_high_mono_count = xen_efi_get_next_high_mono_count;
> > + efi.reset_system = xen_efi_reset_system;
>
> This presumably means that the system default values are already present
> in efi at the point that we overwrite some Xen specifics?
More or less.
> If so, surely you need to retain the clobbering of set_virtual_address_map ?
Nope, by default efi.set_virtual_address_map is NULL (please take a look
at efi struct initialization in drivers/firmware/efi/efi.c). And it is
not touched if efi_enabled(EFI_PARAVIRT).
Daniel