On Fri, Dec 30, 2016 at 05:33:44PM +0900, YASUOKA Masahiko wrote:
> Hi,
>
> On Thu, 29 Dec 2016 14:53:20 -0600
> Ryan <pulse.pu...@gmail.com> wrote:
> > I've successfully booted the machine with the following changes:
> >
> >
> > --- sys/arch/amd64/stand/efiboot/efiboot.c 2016-06-10 12:36:06.000000000
> > -0600
> > +++ sys/arch/amd64/stand/efiboot/efiboot.c 2016-12-28 15:57:36.728002168
> > -0600
> > @@ -137,12 +137,16 @@ efi_main(EFI_HANDLE image, EFI_SYSTEM_TA
> > void
> > efi_cleanup(void)
> > {
> > + int retry;
> > EFI_STATUS status;
> >
> > - efi_memprobe_internal(); /* sync the current map */
> > - status = EFI_CALL(BS->ExitBootServices, IH, mmap_key);
> > - if (status != EFI_SUCCESS)
> > - panic("ExitBootServices");
> > + /* retry once in case of failure */
> > + for (retry = 1; retry == 0; retry--) {
> > + efi_memprobe_internal(); /* sync the current map */
> > + status = EFI_CALL(BS->ExitBootServices, IH, mmap_key);
> > + if (status != EFI_SUCCESS && retry == 0)
> > + panic("ExitBootServices failed (%d)\n", status);
> > + }
> > }
>
> See the condition of "for" carefully. It seems that it never become
> true. Can you check whether it will still boot successfull after this
> problem is fixed?
>
> (snip)
> > Oh before I forget, while I was reading through the code I noticed what
> > appears to be a typo in sys/arch/amd64/stand/efiboot/eficall.h:
> >
> > #define EFI_CALL(...) \
> > _efi_call_fn(__VA_ARGS__, _call_9, _call_8, _call_7, _call_6, _call_5, \
> > _call_4, _call_3, _call_2, _call_1, _call_1)(__VA_ARGS__)
> > #endif
> >
> > It looks like the second instance of _call_1 should be _call_0 (which is
> > defined earlier), unless I am totally misreading this.
>
> I commited this already. Thanks.
Shouldn't this change also be made to armv7 and arm64?
