[U-Boot-Users] Can we drop I386 specific hack from do_bootm()?

[U-Boot-Users] Can we drop I386 specific hack from do_bootm()?

[Bartlomiej Sieka]

Hello,

While working on adapting do_bootm() to new uImage format, I've came 
across this I386-specific hack:

#ifdef __I386__ /* correct image format not implemented yet - fake it */
                 if (fake_header(hdr, (void*)addr, -1) != NULL) {
                         /* to compensate for the addition below */
                         addr -= sizeof(image_header_t);
                         /* turnof verify,
                          * fake_header() does not fake the data crc
                          */
                         verify = 0;
                 } else
#endif  /* __I386__ */

It looks like I386 targets provide the kernel as an zImage, instead of
an uImage, and this special case is needed to add a faked uImage header,
so that the rest of the booting code works without modification.

How about dropping this hack altogether? This would make the code
cleaner, but would require I386 targets (sc520_cdp, sc520_spunk) to use
kernels in uImage format.

What do people think? What is the opinion of the maintainer of the 
affected platforms (Daniel Engstr?m, copied)?

Regards,
Bartlomiej

[U-Boot-Users] Can we drop I386 specific hack from do_bootm()?

[Bartlomiej Sieka]

Bartlomiej Sieka wrote:
[...]
> While working on adapting do_bootm() to new uImage format, I've came 
> across this I386-specific hack:
[...]
> How about dropping this hack altogether? This would make the code
> cleaner, but would require I386 targets (sc520_cdp, sc520_spunk) to use
> kernels in uImage format.
> 
> What do people think? What is the opinion of the maintainer of the 
> affected platforms (Daniel Engstr?m, copied)?

Just a quick update: Daniel's email address given in MAINTAINERS is no
longer actual, so he's not likely to voice his opinion... Anybody else
would like to comment?

Regards,
Bartlomiej