Kevin, Rethinking this change (where we construct the device path from outside and call boot_prio_find()), this is pretty tricky to implement since we need to take care of csm_bootprio_ata() and csm_bootprio_pci() which do not work with device path. In addition, bootprio_find_fdc_device bootprio_find_pci_rom bootprio_find_named_rom bootprio_find_usb Will not require this change and this will just result in too much refactoring.
Maybe simply introduce build_scsi_path() and build_ata_path() functions and then, for instance, make booprio_find_scsi_device() and boot_lchs_find_scsi_device() call the same build_scsi_path() function, resulting in less code duplication. Sam > On 22 Jun 2019, at 20:33, Sam Eiderman <shmuel.eider...@oracle.com> wrote: > > > >> On 22 Jun 2019, at 18:27, Kevin O'Connor <ke...@koconnor.net> wrote: >> >> On Sat, Jun 22, 2019 at 11:51:48AM +0300, Sam Eiderman wrote: >>> But maybe someone wants bootorder but doesn’t want to override legacy disk >>> translations… >>> >>> I’m thinking of maybe adding >>> >>> if (!CONFIG_BOOTORDER || !CONFIG_BIOS_GEOMETRY) >>> return NULL; >> >> That's fine - though it's (!CONFIG_BOOTORDER && !CONFIG_BIOS_GEOMETRY). > > Yes of course, my bad > >> >> FYI, I think BIOS_GEOMETRY is a little confusing - maybe >> CUSTOM_DISK_GEOMETRY. > > The thing is that disk geometry is actually (physical geometry, reported by > the disk controller) and here > bios geometry stands for the geometry reported from bios int13. > Also “bios geometry” === “logical geometry” === “lchs”. > So following previous discussion with Gerd, maybe CONFIG_HOST_BIOS_GEOMETRY > is better? > > Sam > >> >> -Kevin > _______________________________________________ SeaBIOS mailing list -- seabios@seabios.org To unsubscribe send an email to seabios-le...@seabios.org