mbox series

[v3,00/11] tests/qtest: Allow running boot-serial / migration with TCG disabled

Message ID 20230120082341.59913-1-philmd@linaro.org
Headers show
Series tests/qtest: Allow running boot-serial / migration with TCG disabled | expand

Message

Philippe Mathieu-Daudé Jan. 20, 2023, 8:23 a.m. UTC 
Missing review: #7

Two test were failing on Darwin when testing Fabiano's series
which allows building ARM targets without TCG accelerator:
https://lore.kernel.org/qemu-devel/20230118193518.26433-1-farosas@suse.de/

These patches allow boot-serial / migration tests to run without
TCG / KVM.

Philippe Mathieu-Daudé (11):
  tests/qtest/boot-serial-test: Constify tests[] array
  tests/qtest/boot-serial-test: Simplify test_machine() a bit
  tests/qtest/boot-serial-test: Build command line using GString API
  tests/qtest/boot-serial-test: Only use available accelerators
  tests/qtest/migration-test: Inverse #ifdef'ry ladders
  tests/qtest/migration-test: Reduce 'cmd_source' string scope
  tests/qtest/migration-test: Build command line using GString API (1/4)
  tests/qtest/migration-test: Build command line using GString API (2/4)
  tests/qtest/migration-test: Build command line using GString API (3/4)
  tests/qtest/migration-test: Build command line using GString API (4/4)
  tests/qtest/migration-test: Only use available accelerators

 tests/qtest/boot-serial-test.c | 46 +++++++++-------
 tests/qtest/migration-test.c   | 96 ++++++++++++++++++----------------
 2 files changed, 80 insertions(+), 62 deletions(-)

Comments

Thomas Huth Jan. 20, 2023, 9:18 a.m. UTC | #1 
On 20/01/2023 09.23, Philippe Mathieu-Daudé wrote:
> Missing review: #7
> 
> Two test were failing on Darwin when testing Fabiano's series
> which allows building ARM targets without TCG accelerator:
> https://lore.kernel.org/qemu-devel/20230118193518.26433-1-farosas@suse.de/
> 
> These patches allow boot-serial / migration tests to run without
> TCG / KVM.

I just started reviewing while you were sending v3 ... please see my mails 
in reply to v2 for some comments.

Also, I don't quite understand why this is necessary. If you specify both, 
"-accel kvm -accel tcg" on the command line, QEMU should simply fall back to 
the second accelerator if the first one is not available? What's exactly the 
problem you are trying to solve here?

  Thomas
Philippe Mathieu-Daudé Jan. 20, 2023, 11:09 a.m. UTC | #2 
On 20/1/23 10:18, Thomas Huth wrote:
> On 20/01/2023 09.23, Philippe Mathieu-Daudé wrote:
>> Missing review: #7
>>
>> Two test were failing on Darwin when testing Fabiano's series
>> which allows building ARM targets without TCG accelerator:
>> https://lore.kernel.org/qemu-devel/20230118193518.26433-1-farosas@suse.de/
>>
>> These patches allow boot-serial / migration tests to run without
>> TCG / KVM.
> 
> I just started reviewing while you were sending v3 ... please see my 
> mails in reply to v2 for some comments.

Sure.

> Also, I don't quite understand why this is necessary. If you specify 
> both, "-accel kvm -accel tcg" on the command line, QEMU should simply 
> fall back to the second accelerator if the first one is not available? 
> What's exactly the problem you are trying to solve here?

When testing Fabiano's series two tests hang, i.e. [*]:

  18/20 qemu:qtest+qtest-aarch64 / qtest-aarch64/boot-serial-test
  qemu-system-aarch64: -accel tcg: invalid accelerator tcg

Maybe it is just a matter of adding HVF to the test? I'll see
when I switch back to that. Anyhow the "invalid accelerator"
warning is annoying noise when looking for real problems in the
output.

[*] 
https://lore.kernel.org/qemu-devel/ab32e4db-2a6d-69f8-4b69-62a1f1c0f6ba@linaro.org/