qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PULL v2 00/28] Kconfig for Arm machines


From: Thomas Huth
Subject: Re: [Qemu-devel] [PULL v2 00/28] Kconfig for Arm machines
Date: Fri, 10 May 2019 09:52:32 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.6.1

On 09/05/2019 18.54, Peter Maydell wrote:
> On Wed, 8 May 2019 at 16:33, Thomas Huth <address@hidden> wrote:
>>
>> On 08/05/2019 17.09, Peter Maydell wrote:
>>> On Tue, 7 May 2019 at 14:45, Thomas Huth <address@hidden> wrote:
>>>> ----------------------------------------------------------------
>>>> Kconfig settings for the Arm machines
>>>> (v2: Fix the dependency of q35 to AHCI_ICH9 in the second patch)
>>>> ----------------------------------------------------------------
>>>
>>> Hi -- this is still failing in the build test where I 'make clean'
>>
>> Very weird. What is running before the "make clean"? Could you provide
>> me with the content of i386-softmmu/config-devices.mak please?
> 
> Nothing runs before make clean -- my scripts effectively just do
> a git merge, then make clean, then make, then make check.

I tried to reproduce it for a while, but for me, switching between
master and the kconfig-for-arm tree always triggers a "configure" run
and thus the dependencies get regenerated right...

> I notice that "make clean" does not delete config-devices.mak,
> and "make" doesn't cause anything to update it, either.

They are only removed during "distclean", but not for "clean" (which is
what we want, IMHO).

> Further, if I "touch hw/i386/Kconfig" and then run make, nothing
> is rebuilt at all, so something seems to be wrong with our
> makefile dependencies somewhere.

Now that's a good hint. There is definitely something wrong with the
dependencies here. I can see that I've got a file called
"x86_64-softmmu-config.devices.mak.d" in my build directory, but
apparently the Makefile tries to include
x86_64-softmmu/config-devices.mak.d via the SUBDIR_DEVICES_MAK_DEP
variable instead. Since that file does not exist, the dependencies are
completely ignored... Paolo, what was the intention here? Should the
dependencies always be generated in the *softmmu folders instead?

 Thomas



reply via email to

[Prev in Thread] Current Thread [Next in Thread]