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

[Thomas Huth]

On 07/05/2019 13.52, Philippe Mathieu-Daudé wrote:
> On 5/7/19 1:01 PM, Peter Maydell wrote:
>> On Sun, 5 May 2019 at 14:57, Thomas Huth <address@hidden> wrote:
>>>
>>>  Hi Peter,
>>>
>>> the following changes since commit a6ae23831b05a11880b40f7d58e332c45a6b04f7:
>>>
>>>   Merge remote-tracking branch 
>>> 'remotes/ehabkost/tags/python-next-pull-request' into staging (2019-05-03 
>>> 15:26:09 +0100)
>>>
>>> are available in the Git repository at:
>>>
>>>   https://gitlab.com/huth/qemu.git tags/pull-request-2019-05-05
>>>
>>> for you to fetch changes up to 55e5578fabf744e62038f7357369a68e460fe205:
>>>
>>>   hw/arm: Remove hard-enablement of the remaining PCI devices (2019-05-03 
>>> 17:06:20 +0200)
>>>
>>> ----------------------------------------------------------------
>>> Kconfig settings for the Arm machines
>>> ----------------------------------------------------------------
>>
>> Hi Thomas. My test setup that does 'make clean' before the build
>> failed in 'make check':
>>
>> MALLOC_PERTURB_=${MALLOC_PERTURB_:-$(( ${RANDOM:-0} % 255 + 1))}
>> QTEST_QEMU_BINARY=i386-softmmu/qemu-system-i38
>> 6 QTEST_QEMU_IMG=qemu-img tests/ahci-test -m=quick -k --tap <
>> /dev/null | ./scripts/tap-driver.pl --test-name="a
>> hci-test"
>> qemu-system-i386: Unknown device 'ich9-ahci' for bus 'PCIE'
>> Broken pipe
>> /home/petmay01/linaro/qemu-for-merges/tests/libqtest.c:143:
>> kill_qemu() detected QEMU death from signal 6 (Abort
>> ed) (core dumped)
>> Aborted (core dumped)
>> ERROR - too few tests run (expected 74, got 0)
>> /home/petmay01/linaro/qemu-for-merges/tests/Makefile.include:903:
>> recipe for target 'check-qtest-i386' failed
>>
>> (all the builds that are just incremental without the 'make clean'
>> worked ok). Any idea what's happened here?
> 
> You found a bug in hw/i386/Kconfig :)
> 
> I noticed this while working on refactoring ICH9 for the Boston board,
> but didn't expect the Q35 machine would suffer the same issue.
> 
> The quick fix is:
> 
> -- >8 --
> diff --git a/hw/i386/Kconfig b/hw/i386/Kconfig
> @@ -80,7 +80,7 @@ config Q35
>      select PC_ACPI
>      select PCI_EXPRESS_Q35
>      select LPC_ICH9
> -    select AHCI
> +    select AHCI_ICH9
>      select DIMM
>      select SMBIOS
>      select VMPORT
> ---
> 
> I'll send a patch.

That config switch is not available in upstream yet, it just got
introduced by my series. But that's the right fix, I think, so I'll
squash that in and try again... (it's just weird that I can not
reproduce the build failure here locally)

 Thomas