On Wed, Jun 19, 2019 at 12:23:51PM +0300, Sam Eiderman wrote:
> Adding the following utility functions:
>
> * boot_lchs_find_pci_device
> * boot_lchs_find_scsi_device
> * boot_lchs_find_ata_device
FWIW, this leads to a bit of code duplication. I think it would be
preferable to
On Thu, Jun 20, 2019 at 03:12:33PM +0100, David Woodhouse wrote:
> On Thu, 2019-06-20 at 09:43 -0400, Kevin O'Connor wrote:
> > On Thu, Jun 20, 2019 at 01:07:45PM +0100, David Woodhouse wrote:
> > > For CSM, the highest priority for a boot entry is zero. SeaBIOS doesn't
> > > use zero, and the
On Thu, 2019-06-20 at 09:43 -0400, Kevin O'Connor wrote:
> On Thu, Jun 20, 2019 at 01:07:45PM +0100, David Woodhouse wrote:
> > For CSM, the highest priority for a boot entry is zero. SeaBIOS doesn't
> > use zero, and the highest priority is 1.
>
> FYI, SeaBIOS does treat zero as the highest
On Thu, Jun 20, 2019 at 01:07:45PM +0100, David Woodhouse wrote:
> For CSM, the highest priority for a boot entry is zero. SeaBIOS doesn't
> use zero, and the highest priority is 1.
FYI, SeaBIOS does treat zero as the highest priority. And a negative
priority means "use default priority".
I'm
For CSM, the highest priority for a boot entry is zero. SeaBIOS doesn't
use zero, and the highest priority is 1.
Make the results of csm_bootprio_*() conform to that convention. Also
explicitly handle the BBS_DO_NOT_BOOT_FROM and BBS_IGNORE_ENTRY values.
Signed-off-by: David Woodhouse
---
v2:
On Thu, Jun 20, 2019 at 11:52:01AM +0300, Sam Eiderman wrote:
>
>
> > On 20 Jun 2019, at 8:42, Gerd Hoffmann wrote:
> >
> >> +static int
> >> +overriden_lchs_supplied(struct drive_s *drive)
> >> +{
> >> +return drive->lchs.cylinder || drive->lchs.head || drive->lchs.sector;
> >> +}
> >
>
On Wed, 2019-06-19 at 12:27 +0100, David Woodhouse wrote:
> For CSM, the highest priority is zero. In SeaBIOS that means "don't", and
> the highest priority is 1.
>
> So we end up with the fun outcome that booting from NVMe worked only
> when it *wasn't* selected as the primary boot target,
> On 20 Jun 2019, at 8:42, Gerd Hoffmann wrote:
>
>> +static int
>> +overriden_lchs_supplied(struct drive_s *drive)
>> +{
>> +return drive->lchs.cylinder || drive->lchs.head || drive->lchs.sector;
>> +}
>
>> +case TRANSLATION_MACHINE:
>
> Hmm, why this name? Doesn't look intuitive
> On 20 Jun 2019, at 8:50, Gerd Hoffmann wrote:
>
>> === OUTPUT BEGIN ===
>> 1/8 Checking commit cf2168a66dbd (block: Refactor macros - fix tabbing)
>> ERROR: Macros with complex values should be enclosed in parenthesis
>> #55: FILE: include/hw/block/block.h:65:
>> +#define