[U-Boot] Apparent conflict between CONFIG_BLK and CONFIG_API

[fuz at fuz.su <fuz@fuz.su>]

Hello Simon,

CONFIG_BLK is set by default for all Rockchip boards. Turning it off was
the first thing I tried, but I actually wasn't able to figure out how to
turn it off at all and was then advised elsewhere that this is probably
not a good idea either.

Yours,
Robert Clausecker

On Mon, Mar 13, 2017 at 06:33:16AM -0600, Simon Glass wrote:
> Hi,
> 
> On 7 March 2017 at 14:53,  <fuz@fuz.su> wrote:
> > Good evening,
> >
> > I am trying to port FreeBSD to the ASUS Tinker Board, a computer based
> > on the Rockchip 3288 SoC. FreeBSD's boot loader (named loader(8)) needs
> > CONFIG_API to be enabled, but trying to build an U-Boot from trunk with
> > both CONFIG_API and CONFIG_BLK (as required for Rockchip SoC's?) leads
> > to the following build failure:
> >
> > $ CROSS_COMPILE=arm-none-eabi- gmake tinker-rk3288_defconfig all
> > ...
> >   CC      api/api_storage.o
> > api/api_storage.c: In function 'dev_read_stor':
> > api/api_storage.c:334:9: error: 'struct blk_desc' has no member named 'block_read'
> >   if ((dd->block_read) == NULL) {
> >          ^~
> > api/api_storage.c:339:11: error: 'struct blk_desc' has no member named 'block_read'
> >   return dd->block_read(dd, start, len, buf);
> >            ^~
> > api/api_storage.c:340:1: warning: control reaches end of non-void function [-Wreturn-type]
> >  }
> >  ^
> > gmake[2]: *** [scripts/Makefile.build:281: api/api_storage.o] Fehler 1
> > gmake[1]: *** [Makefile:1229: api] Fehler 2
> > gmake: *** [Makefile:460: __build_one_by_one] Error 2
> >
> > I applied the following fix, but the product doesn't boot. Perhaps
> > that's not a property of the fix though:
> >
> > diff --git a/api/api_storage.c b/api/api_storage.c
> > index e80818df1c..815ed1128d 100644
> > --- a/api/api_storage.c
> > +++ b/api/api_storage.c
> > @@ -331,10 +331,14 @@ lbasize_t dev_read_stor(void *cookie, void *buf, lbasize_t len, lbastart_t start
> >         if (!dev_stor_is_valid(type, dd))
> >                 return 0;
> >
> > +#ifdef CONFIG_BLK
> > +       return blk_dread(dd, start, len, buf);
> > +#else
> >         if ((dd->block_read) == NULL) {
> >                 debugf("no block_read() for device 0x%08x\n", cookie);
> >                 return 0;
> >         }
> >
> >         return dd->block_read(dd, start, len, buf);
> > +#endif /* defined(CONFIG_BLK) */
> 
> This fix looks right to me. There may be something else wrong. I have
> not used API very much but it might be worth checking if the board
> boots OK without CONFIG_BLK?
> 
> >  }
> >
> > Yours,
> > Robert Clausecker
> >
> > --
> > ()  ascii ribbon campaign - for an 8-bit clean world
> > /\  - against html email  - against proprietary attachments
> > _______________________________________________
> > U-Boot mailing list
> > U-Boot at lists.denx.de
> > https://lists.denx.de/listinfo/u-boot
> 
> Regards,
> Simon

-- 
()  ascii ribbon campaign - for an 8-bit clean world 
/\  - against html email  - against proprietary attachments