Re: [PATCH v10 1/8] memory: prevent dma-reentracy issues

[Thomas Huth <thuth@redhat.com>]

On 08/05/2023 15.03, Song Gao wrote:
> Hi, Thomas
> 
> 在 2023/5/8 下午5:33, Thomas Huth 写道:
>> On 06/05/2023 11.25, Song Gao wrote:
>>>   Hi Alexander
>>>
>>> 在 2023/4/28 下午5:14, Thomas Huth 写道:
>>>> On 28/04/2023 11.11, Alexander Bulekov wrote:
>>>>> On 230428 1015, Thomas Huth wrote:
>>>>>> On 28/04/2023 10.12, Daniel P. Berrangé wrote:
>>>>>>> On Thu, Apr 27, 2023 at 05:10:06PM -0400, Alexander Bulekov wrote:
>>>>>>>> Add a flag to the DeviceState, when a device is engaged in 
>>>>>>>> PIO/MMIO/DMA.
...
>>> This patch causes the loongarch virtual machine to fail to start the 
>>> slave cpu.
>>>
>>>      ./build/qemu-system-loongarch64 -machine virt -m 8G -cpu la464 \
>>>               -smp 4 -bios QEMU_EFI.fd -kernel vmlinuz.efi -initrd 
>>> ramdisk   \
>>>                 -serial stdio   -monitor 
>>> telnet:localhost:4495,server,nowait  \
>>>                 -append "root=/dev/ram rdinit=/sbin/init 
>>> console=ttyS0,115200"   --nographic
>>>
>>>
>>> ....
>>> qemu-system-loongarch64: warning: Blocked re-entrant IO on MemoryRegion: 
>>> loongarch_ipi_iocsr at addr: 0x24
>>
>> Oh, another spot that needs special handling ... I see Alexander already 
>> sent a patch (thanks!), but anyway, this is a good indication that we're 
>> missing some test coverage in the CI.
>>
>> Are there any loongarch kernel images available for public download 
>> somewhere? If so, we really should add an avocado regression test for this 
>> - since as far as I can see, we don't have any  tests for loongarch in 
>> tests/avocado yet?
>>
> we can get  some binarys  at:
> https://github.com/yangxiaojuan-loongson/qemu-binary
 >
> I'm not sure that avacodo testing can be done using just the kernel.
> 
> Is a full loongarch system required?

No, you don't need a full distro installation, just a kernel with ramdisk 
(which is also available there) is good enough for a basic test, e.g. just 
check whether the kernel boots to a certain point is good enough to provide 
a basic sanity test. If you then can also get even into a shell (of the 
ramdisk), you can check some additional stuff in the sysfs or "dmesg" 
output, see for example tests/avocado/machine_s390_ccw_virtio.py which does 
such checks with a kernel and initrd on s390x.

  Thomas