On Thu, 31 Mar 2011 11:51:30 -0500 Kim Phillips kim.phillips@freescale.com wrote:
On Thu, 31 Mar 2011 11:21:09 -0500 Scott Wood scottwood@freescale.com wrote:
On Thu, 31 Mar 2011 17:40:17 +0200 Wolfgang Denk wd@denx.de wrote:
In message 20110331103141.80162901.kim.phillips@freescale.com you wrote:
Scott, please push this commit upstream to fix out of source tree NAND builds.
See my previous mail with a fix.
Why put such fiuxes into a NEXT branch when we're trying to get a release ot?
That's ...
The patch Kim quoted, which is in my next branch, is unrelated to the above error. It is a cleanup patch.
it prepends the build directory $(obj) to the nand_spl target.
The build directory was always prepended to the nand_spl target. That patch just moves where the target was defined.
The problem is with $(nandobj) in the nand_spl makefiles.
But now WD has posted another patch (that seems more invasive than the one I quoted).
It's solving a different problem, and it's the right fix -- it's better to avoid depending on deferred evaluation if we can do so reasonably.
I don't see the above error with top-of-tree:
$ CROSS_COMPILE=~/ppc-tc-4.5/bin/powerpc-linux-gnu- ./MAKEALL MPC8313ERDB_NAND_33
try BUILD_DIR=/tmp/hah ./MAKEALL MPC8313ERDB_NAND_33
Right, I saw afterward that it was an out-of-tree build. I see the same thing in my next branch, once I rebase onto Wolfgang's current master which has the LDFLAGS_FINAL patch.
"Fix NAND_SPL and ONENAND_IPL in Makefile" has nothing to do with this.
-Scott