[M] Change in coreboot[master]: soc/amd/spi: Mmap 32M/64M flash

[Nico Huber (Code Review)]

Attention is currently required from: Bao Zheng, Jason Glenesk, Raul Rangel, Marshall Dawson, Matt DeVillier, Zheng Bao, Martin Roth, Fred Reitberger, Felix Held.

Nico Huber has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/74260 )

Change subject: soc/amd/spi: Mmap 32M/64M flash
......................................................................

Patch Set 7:

(1 comment)

Patchset:

PS5:
> For the newest SOC and 32M or 64M SPI flash.

I meant more, where will the mmap API be called? AIUI, the current implementation
will only work in very simple cases, i.e. a single mapping at a time. While the API
currently allows things like this:

const struct region_device *rd = boot_device_ro();
    void *a = rdev_mmap(rd, 0, 1234);
    void *b = rdev_mmap(rd, 34567890, 1234);
    memcmp(a, b, 1234);

Such uses are not possible with the one-mapping-at-a-time SPI_ROM_PAGE register.
We also leak mappings of boot_device_ro() on purpose, because the mapping is
expected to work indefinitely. For instance elog_flash_offset_to_address(), or
spi_flash_get_mmap_windows(). These are expected to be used by the payload and OS
without knowing any SPI_ROM_PAGE index.

To be fully compatible with existing uses of boot_device_ro(), I would investigate
if working on a copy is feasible for `index != 0`, i.e. allocate memory when
`.mmap` gets called and pre-fill it with the data at the given offset.