On 6/28/21 7:14 AM, Pavel Dovgalyuk wrote:
> On 25.06.2021 20:01, Philippe Mathieu-Daudé wrote:
>> Hi Pavel,
>>
>> FYI as of commit 050cee12315 ("Merge remote-tracking branch
>> 'remotes/stsquad/tags/pull-testing-updates-250621-1' into staging")
>
> Doesn't it mean, that the real culprit is hidden and we should bisect?
Could be, I have no idea, I don't run this test often, only when I do
changes related to aa64. I could bisect if you can't, but it won't be
soon.
>> the ReverseDebugging_AArch64 test is failing:
>
> Shouldn't the merge be postponed in such cases?
This test isn't run on CI (you marked it @skip in commit be52eca3097).
>> (28/37)
>> tests/acceptance/reverse_debugging.py:ReverseDebugging_AArch64.test_aarch64_virt:
>>
>> ERROR (0.96 s)
>>
>> INFO | recorded log with 1690570+ steps
>> INFO | replaying the execution...
>> INFO | connecting to gdbstub
>> INFO | stepping forward
>> INFO | saving position 40000000
>> ERROR|
>> ERROR| Reproduced traceback from:
>> lib/python3.8/site-packages/avocado/core/test.py:770
>> ERROR| Traceback (most recent call last):
>> ERROR| File "acceptance/reverse_debugging.py", line 206, in
>> test_aarch64_virt
>> ERROR| self.reverse_debugging(
>> ERROR| File "acceptance/reverse_debugging.py", line 140, in
>> reverse_debugging
>> ERROR| pc = self.get_pc(g)
>> ERROR| File "acceptance/reverse_debugging.py", line 77, in get_pc
>> ERROR| return self.get_reg(g, self.REG_PC)
>> ERROR| File "acceptance/reverse_debugging.py", line 72, in get_reg
>> ERROR| return self.get_reg_le(g, reg)
>> ERROR| File "acceptance/reverse_debugging.py", line 58, in get_reg_le
>> ERROR| res = g.cmd(b'p%x' % reg)
>> ERROR| File "lib/python3.8/site-packages/avocado/utils/gdb.py", line
>> 783, in cmd
>> ERROR| response_payload = self.decode(result)
>> ERROR| File "lib/python3.8/site-packages/avocado/utils/gdb.py", line
>> 738, in decode
>> ERROR| raise InvalidPacketError
>> ERROR| avocado.utils.gdb.InvalidPacketError
