Re: [RFC PATCH 2/2] tests/qtest: Enable spapr dma with linear iommu map

[Nicholas Piggin]

On Wed Apr 16, 2025 at 4:07 AM AEST, Fabiano Rosas wrote:
> Nicholas Piggin <npiggin@gmail.com> writes:
>
>> qtests spapr dma was broken because the iommu was not set up.
>>
>> spapr requires hypercalls to set up the iommu (TCE tables), but
>> there is no support for that or a side-channel to the iommu in
>> qtests at the moment, so add a quick workaround in QEMU to have
>> the spapr iommu provide a linear map to memory when running
>> qtests.
>
> That's fine.
>
> But what would it take to add support? Add another callback such as
> qtest_rtas_call() to handle hcalls and call H_PUT_TCE from the test? Or
> is there some other complication?

Yeah, exactly. qtest_ppc_spapr_rtas_call() and/or _spapr_hcall() in
the qtest interface I think would do it.

It would be nice to do that in general to be able to test various spapr
bits with qtests. It's just getting all the pieces together would be a
bit more work than this simple bandaid.

>
>>
>> The buggy msix checks can all be removed since the tests all work
>> now.
>>
>> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
>> ---
>>  tests/qtest/libqos/pci.h             |  4 ----
>>  hw/ppc/spapr_iommu.c                 |  9 ++++++++-
>>  tests/qtest/e1000e-test.c            | 23 +++--------------------
>>  tests/qtest/igb-test.c               | 21 ---------------------
>>  tests/qtest/libqos/generic-pcihost.c |  1 -
>>  tests/qtest/libqos/pci-pc.c          |  3 ---
>>  tests/qtest/libqos/pci-spapr.c       |  7 ++++---
>>  tests/qtest/libqos/pci.c             | 14 --------------
>>  tests/qtest/vhost-user-blk-test.c    |  6 ------
>>  tests/qtest/virtio-blk-test.c        | 12 ------------
>>  10 files changed, 15 insertions(+), 85 deletions(-)
>>
>
> ...
>
>> @@ -173,13 +159,10 @@ static void test_e1000e_multiple_transfers(void *obj, 
>> void *data,
>>  
>>      QE1000E_PCI *e1000e = obj;
>>      QE1000E *d = &e1000e->e1000e;
>> -    QOSGraphObject *e_object = obj;
>> -    QPCIDevice *dev = e_object->get_driver(e_object, "pci-device");
>>  
>> -    /* FIXME: add spapr support */
>> -    if (qpci_check_buggy_msi(dev)) {
>> -        return;
>> -    }
>> +    /* Use EITR for one irq and disable it for the other, for testing */
>> +    e1000e_macreg_write(d, E1000_EITR + E1000E_RX0_MSG_ID * 4, 500);
>> +    e1000e_macreg_write(d, E1000_EITR + E1000E_TX0_MSG_ID * 4, 0);
>
> What's this about?

It's a rebase bug, thanks for catching it.

Thanks,
Nick