On 21/04/2022 17:09, Jan Beulich wrote:
> On 21.04.2022 18:00, Andrew Cooper wrote:
>> Now that `make MAP` might rebuild $(TARGET), it needs removing from
>> no-dot-config-targets.
> Which raises the question whether the MAP target originally was
> meant to be used only on an already built tree, which would
> explain the missing dependency that you added in the earlier
> commit.
I can't comment on the previous expectation, but the fact you couldn't
previously do `make -j4 build MAP` was simply broken.
There are also several other breakages in Xen's build system which it
turns out is forcing all downstreams to bodge around in ways that range
from gross to dangerous, and which need fixing for everyone's benefit.
(Emails will be appearing for those shortly)
The current debug target (just runs objdump -d) is similar to MAP,
except I don't find it a credible target to use and was planning to
delete it.
>> Otherwise the build eventually fails with:
>>
>> CPP arch/x86/asm-macros.i
>> arch/x86/asm-macros.c:1:10: fatal error: asm/asm-defns.h: No such file or
>> directory
>> 1 | #include <asm/asm-defns.h>
>> | ^~~~~~~~~~~~~~~~~
> Why would this be? asm/ isn't a symlink anymore, so should always
> exist. I would have expected an error about .config or
> autoconf.h not existing, ...
>
>> Fixes: e1e72198213b ("xen/build: Fix dependency for the MAP rule")
>> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com>
>> ---
>> CC: Jan Beulich <jbeul...@suse.com>
>> CC: Anthony PERARD <anthony.per...@citrix.com>
>>
>> Backporting notes. The original, e1e72198213b, wants backporting as far as
>> we
>> can go. This patch is only relevant from 6c122d3984a5e onwards, so 4.14,
>> which is why my main developing/testing on 4.13 didn't spot it.
> ... and the commit you reference seems to support that expectation.
I haven't spent very long investigating exactly how it breaks. I don't
think it's important, because the purpose of no-dot-config-targets is to
skip large swathes of the logic.
~Andrew
