[U-Boot] include/autoconf.mk:212: *** missing separator. Stop.
Hi all...
When running "MAKEALL", I see occasional
/home/wd/git/u-boot/work/include/autoconf.mk:212: *** missing separator. Stop.
errors for some boards. The line numbers are changing (57, 85, 113, 129, 183, 212, 215 etc.), and so are the affected boards (barco, kilauea, MPC8544DS, neo, PM856, sequoia). But the build continues to run and completes successfully without additional warnigns or errors. Repeating the build will show the problem with other line numbers and other boards.
The problem is obviously related to running parallel build (this happens on a 4 core system); if I use a single make job it goes away.
It looks as if some build step was using autoconf.mk before the task building it has completed.
I've searched, but did not find any place where this could happen.
Is anybody else experiencing this problem?
Any ideas where to look at?
Best regards,
Wolfgang Denk
On Dec 16, 2008, at 6:35 AM, Wolfgang Denk wrote:
Hi all...
When running "MAKEALL", I see occasional
/home/wd/git/u-boot/work/include/autoconf.mk:212: *** missing separator. Stop.
errors for some boards. The line numbers are changing (57, 85, 113, 129, 183, 212, 215 etc.), and so are the affected boards (barco, kilauea, MPC8544DS, neo, PM856, sequoia). But the build continues to run and completes successfully without additional warnigns or errors. Repeating the build will show the problem with other line numbers and other boards.
The problem is obviously related to running parallel build (this happens on a 4 core system); if I use a single make job it goes away.
It looks as if some build step was using autoconf.mk before the task building it has completed.
I've searched, but did not find any place where this could happen.
Is anybody else experiencing this problem?
Now that you mention this, I've occasionally seen this as well on various ppc boards, and I always run a parallel build (4-way). In my case, repeating the build usually results in no errors.
But I haven't looked into it at all, so I'm afraid I'm not too helpful here except to say that it's not just you.
Cheers, B
Dear Becky,
In message 168B6481-BF70-4E86-B766-85FACA790A4B@kernel.crashing.org you wrote:
Is anybody else experiencing this problem?
Now that you mention this, I've occasionally seen this as well on various ppc boards, and I always run a parallel build (4-way). In my case, repeating the build usually results in no errors.
Same here. It happened about 1...2 times out of 20 on a 4 core system. Much, much less frequently when using only 3 parallel jobs. I didn't see it yet on a 2 core system.
But I haven't looked into it at all, so I'm afraid I'm not too helpful here except to say that it's not just you.
I think I fixed it. Maybe you can test the patch?
Best regards,
Wolfgang Denk
participants (2)
-
Becky Bruce -
Wolfgang Denk