| Message ID | 20241121165806.476008-1-alex.bennee@linaro.org |
|---|---|
| Headers | show |
| Series | maintainer updates for -rc2 pre-PR | expand |
On Thu, 21 Nov 2024 at 16:58, Alex Bennée <alex.bennee@linaro.org> wrote: > > This is a mostly testing focused set of patches but a few bug fixes as > well. I plan to send the PR in on Monday. I can drop any patches that > are objected to but I think its pretty safe. > > Contains: > > - Daniel's clean-up of functional tests > - Another avocado->function conversion from Thomas > - Update the tuxrun baseline images > - Minor fix to the rust pl011 device > - Documentation clarification on identity Should we really be updating the tuxrun baseline images in the middle of a release freeze period? Unless the old images are going to go away and break the tests, I think it would be better to stick with what we're currently testing. thanks -- PMM
Peter Maydell <peter.maydell@linaro.org> writes: > On Thu, 21 Nov 2024 at 16:58, Alex Bennée <alex.bennee@linaro.org> wrote: >> >> This is a mostly testing focused set of patches but a few bug fixes as >> well. I plan to send the PR in on Monday. I can drop any patches that >> are objected to but I think its pretty safe. >> >> Contains: >> >> - Daniel's clean-up of functional tests >> - Another avocado->function conversion from Thomas >> - Update the tuxrun baseline images >> - Minor fix to the rust pl011 device >> - Documentation clarification on identity > > Should we really be updating the tuxrun baseline images > in the middle of a release freeze period? Unless the old images > are going to go away and break the tests, I think it would > be better to stick with what we're currently testing. Well the arm64be fixed a real problem and while I was at it I figured might as well keep the rest in sync. I have tested them so they all pass (although I'm waiting on the CI run now). > > thanks > -- PMM
Hello Alex, On 11/21/24 17:57, Alex Bennée wrote: > This is a mostly testing focused set of patches but a few bug fixes as > well. I plan to send the PR in on Monday. I can drop any patches that > are objected to but I think its pretty safe. > > Contains: > > - Daniel's clean-up of functional tests > - Another avocado->function conversion from Thomas > - Update the tuxrun baseline images > - Minor fix to the rust pl011 device > - Documentation clarification on identity > > The following could do with some review: > > tests/functional: update the x86_64 tuxrun tests > tests/functional: update the sparc64 tuxrun tests > tests/functional: update the s390x tuxrun tests > tests/functional: update the riscv64 tuxrun tests > tests/functional: update the riscv32 tuxrun tests > tests/functional: update the ppc64 tuxrun tests > tests/functional: update the ppc32 tuxrun tests > tests/functional: update the mips64el tuxrun tests > tests/functional: update the mips64 tuxrun tests > tests/functional: update the mips32el tuxrun tests > tests/functional: update the mips32 tuxrun tests > tests/functional: add a m68k tuxrun tests > tests/functional: update the i386 tuxrun tests > tests/functional: update the aarch64 tuxrun tests > tests/functional: update the arm tuxrun tests > tests/functional: Convert the Avocado aarch64 tuxrun tests Do you think we could include patches 2-4 from this series ? https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ Only patch 3 lacks a R-b. Thanks, C.
On 21/11/2024 20.03, Cédric Le Goater wrote: > Hello Alex, > > On 11/21/24 17:57, Alex Bennée wrote: >> This is a mostly testing focused set of patches but a few bug fixes as >> well. I plan to send the PR in on Monday. I can drop any patches that >> are objected to but I think its pretty safe. >> >> Contains: >> >> - Daniel's clean-up of functional tests >> - Another avocado->function conversion from Thomas >> - Update the tuxrun baseline images >> - Minor fix to the rust pl011 device >> - Documentation clarification on identity >> >> The following could do with some review: >> >> tests/functional: update the x86_64 tuxrun tests >> tests/functional: update the sparc64 tuxrun tests >> tests/functional: update the s390x tuxrun tests >> tests/functional: update the riscv64 tuxrun tests >> tests/functional: update the riscv32 tuxrun tests >> tests/functional: update the ppc64 tuxrun tests >> tests/functional: update the ppc32 tuxrun tests >> tests/functional: update the mips64el tuxrun tests >> tests/functional: update the mips64 tuxrun tests >> tests/functional: update the mips32el tuxrun tests >> tests/functional: update the mips32 tuxrun tests >> tests/functional: add a m68k tuxrun tests >> tests/functional: update the i386 tuxrun tests >> tests/functional: update the aarch64 tuxrun tests >> tests/functional: update the arm tuxrun tests >> tests/functional: Convert the Avocado aarch64 tuxrun tests > > Do you think we could include patches 2-4 from this series ? > > https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ > > Only patch 3 lacks a R-b. I had a question on patch 3 and a suggestion on patch 4 ... could you maybe address them first? Thanks! Thomas
On 11/21/24 20:10, Thomas Huth wrote: > On 21/11/2024 20.03, Cédric Le Goater wrote: >> Hello Alex, >> >> On 11/21/24 17:57, Alex Bennée wrote: >>> This is a mostly testing focused set of patches but a few bug fixes as >>> well. I plan to send the PR in on Monday. I can drop any patches that >>> are objected to but I think its pretty safe. >>> >>> Contains: >>> >>> - Daniel's clean-up of functional tests >>> - Another avocado->function conversion from Thomas >>> - Update the tuxrun baseline images >>> - Minor fix to the rust pl011 device >>> - Documentation clarification on identity >>> >>> The following could do with some review: >>> >>> tests/functional: update the x86_64 tuxrun tests >>> tests/functional: update the sparc64 tuxrun tests >>> tests/functional: update the s390x tuxrun tests >>> tests/functional: update the riscv64 tuxrun tests >>> tests/functional: update the riscv32 tuxrun tests >>> tests/functional: update the ppc64 tuxrun tests >>> tests/functional: update the ppc32 tuxrun tests >>> tests/functional: update the mips64el tuxrun tests >>> tests/functional: update the mips64 tuxrun tests >>> tests/functional: update the mips32el tuxrun tests >>> tests/functional: update the mips32 tuxrun tests >>> tests/functional: add a m68k tuxrun tests >>> tests/functional: update the i386 tuxrun tests >>> tests/functional: update the aarch64 tuxrun tests >>> tests/functional: update the arm tuxrun tests >>> tests/functional: Convert the Avocado aarch64 tuxrun tests >> >> Do you think we could include patches 2-4 from this series ? >> >> https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ >> >> Only patch 3 lacks a R-b. > > I had a question on patch 3 and a suggestion on patch 4 ... could you maybe address them first? Thanks! Oh Sorry I forgot. I did address patch 4 in my tree ... In Patch 3, EXTRA_BOOTARGS was added by Alex IIRC to work around the console issue. We ended up adding the sleep call. I can resend tomorrow. Thanks, C.
On 21/11/2024 17.57, Alex Bennée wrote: > This is a mostly testing focused set of patches but a few bug fixes as > well. I plan to send the PR in on Monday. I can drop any patches that > are objected to but I think its pretty safe. > > Contains: > > - Daniel's clean-up of functional tests > - Another avocado->function conversion from Thomas > - Update the tuxrun baseline images > - Minor fix to the rust pl011 device > - Documentation clarification on identity > > The following could do with some review: > > tests/functional: update the x86_64 tuxrun tests > tests/functional: update the sparc64 tuxrun tests > tests/functional: update the s390x tuxrun tests > tests/functional: update the riscv64 tuxrun tests > tests/functional: update the riscv32 tuxrun tests > tests/functional: update the ppc64 tuxrun tests > tests/functional: update the ppc32 tuxrun tests > tests/functional: update the mips64el tuxrun tests > tests/functional: update the mips64 tuxrun tests > tests/functional: update the mips32el tuxrun tests > tests/functional: update the mips32 tuxrun tests > tests/functional: add a m68k tuxrun tests > tests/functional: update the i386 tuxrun tests > tests/functional: update the aarch64 tuxrun tests > tests/functional: update the arm tuxrun tests If you've got some spare time: Maybe also update the sh4 tuxrun test to see whether we could drop the FLAKY marker there now? Thomas
On 21/11/2024 18.31, Alex Bennée wrote: > Peter Maydell <peter.maydell@linaro.org> writes: > >> On Thu, 21 Nov 2024 at 16:58, Alex Bennée <alex.bennee@linaro.org> wrote: >>> >>> This is a mostly testing focused set of patches but a few bug fixes as >>> well. I plan to send the PR in on Monday. I can drop any patches that >>> are objected to but I think its pretty safe. >>> >>> Contains: >>> >>> - Daniel's clean-up of functional tests >>> - Another avocado->function conversion from Thomas >>> - Update the tuxrun baseline images >>> - Minor fix to the rust pl011 device >>> - Documentation clarification on identity >> >> Should we really be updating the tuxrun baseline images >> in the middle of a release freeze period? Unless the old images >> are going to go away and break the tests, I think it would >> be better to stick with what we're currently testing. > > Well the arm64be fixed a real problem and while I was at it I figured > might as well keep the rest in sync. I have tested them so they all pass > (although I'm waiting on the CI run now). But there could be new intermittent problems in the new images ... so if we'd face such a problem, we would not know whether it is the image or whether it is QEMU. Thus maybe let's better keep the old versions for 9.2 (except for the arm64 patch that fixes a real problem), and use the new versions for 10.0 ? Thomas
On 21/11/2024 22.46, Cédric Le Goater wrote: > On 11/21/24 20:10, Thomas Huth wrote: >> On 21/11/2024 20.03, Cédric Le Goater wrote: >>> Hello Alex, >>> >>> On 11/21/24 17:57, Alex Bennée wrote: >>>> This is a mostly testing focused set of patches but a few bug fixes as >>>> well. I plan to send the PR in on Monday. I can drop any patches that >>>> are objected to but I think its pretty safe. >>>> >>>> Contains: >>>> >>>> - Daniel's clean-up of functional tests >>>> - Another avocado->function conversion from Thomas >>>> - Update the tuxrun baseline images >>>> - Minor fix to the rust pl011 device >>>> - Documentation clarification on identity >>>> >>>> The following could do with some review: >>>> >>>> tests/functional: update the x86_64 tuxrun tests >>>> tests/functional: update the sparc64 tuxrun tests >>>> tests/functional: update the s390x tuxrun tests >>>> tests/functional: update the riscv64 tuxrun tests >>>> tests/functional: update the riscv32 tuxrun tests >>>> tests/functional: update the ppc64 tuxrun tests >>>> tests/functional: update the ppc32 tuxrun tests >>>> tests/functional: update the mips64el tuxrun tests >>>> tests/functional: update the mips64 tuxrun tests >>>> tests/functional: update the mips32el tuxrun tests >>>> tests/functional: update the mips32 tuxrun tests >>>> tests/functional: add a m68k tuxrun tests >>>> tests/functional: update the i386 tuxrun tests >>>> tests/functional: update the aarch64 tuxrun tests >>>> tests/functional: update the arm tuxrun tests >>>> tests/functional: Convert the Avocado aarch64 tuxrun tests >>> >>> Do you think we could include patches 2-4 from this series ? >>> >>> https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ >>> >>> Only patch 3 lacks a R-b. >> >> I had a question on patch 3 and a suggestion on patch 4 ... could you >> maybe address them first? Thanks! > > Oh Sorry I forgot. > > I did address patch 4 in my tree ... In Patch 3, EXTRA_BOOTARGS was > added by Alex IIRC to work around the console issue. We ended up > adding the sleep call. I can resend tomorrow. IMHO v4 looks fine now ( https://lore.kernel.org/qemu-devel/20241122090322.1934697-1-clg@redhat.com/ ) in case you want to pick it up, Alex. Thomas
On 11/22/24 11:47, Thomas Huth wrote: > On 21/11/2024 22.46, Cédric Le Goater wrote: >> On 11/21/24 20:10, Thomas Huth wrote: >>> On 21/11/2024 20.03, Cédric Le Goater wrote: >>>> Hello Alex, >>>> >>>> On 11/21/24 17:57, Alex Bennée wrote: >>>>> This is a mostly testing focused set of patches but a few bug fixes as >>>>> well. I plan to send the PR in on Monday. I can drop any patches that >>>>> are objected to but I think its pretty safe. >>>>> >>>>> Contains: >>>>> >>>>> - Daniel's clean-up of functional tests >>>>> - Another avocado->function conversion from Thomas >>>>> - Update the tuxrun baseline images >>>>> - Minor fix to the rust pl011 device >>>>> - Documentation clarification on identity >>>>> >>>>> The following could do with some review: >>>>> >>>>> tests/functional: update the x86_64 tuxrun tests >>>>> tests/functional: update the sparc64 tuxrun tests >>>>> tests/functional: update the s390x tuxrun tests >>>>> tests/functional: update the riscv64 tuxrun tests >>>>> tests/functional: update the riscv32 tuxrun tests >>>>> tests/functional: update the ppc64 tuxrun tests >>>>> tests/functional: update the ppc32 tuxrun tests >>>>> tests/functional: update the mips64el tuxrun tests >>>>> tests/functional: update the mips64 tuxrun tests >>>>> tests/functional: update the mips32el tuxrun tests >>>>> tests/functional: update the mips32 tuxrun tests >>>>> tests/functional: add a m68k tuxrun tests >>>>> tests/functional: update the i386 tuxrun tests >>>>> tests/functional: update the aarch64 tuxrun tests >>>>> tests/functional: update the arm tuxrun tests >>>>> tests/functional: Convert the Avocado aarch64 tuxrun tests >>>> >>>> Do you think we could include patches 2-4 from this series ? >>>> >>>> https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ >>>> >>>> Only patch 3 lacks a R-b. >>> >>> I had a question on patch 3 and a suggestion on patch 4 ... could you maybe address them first? Thanks! >> >> Oh Sorry I forgot. >> >> I did address patch 4 in my tree ... In Patch 3, EXTRA_BOOTARGS was >> added by Alex IIRC to work around the console issue. We ended up >> adding the sleep call. I can resend tomorrow. > > IMHO v4 looks fine now ( https://lore.kernel.org/qemu-devel/20241122090322.1934697-1-clg@redhat.com/ ) in case you want to pick it up, Alex. > > Thomas > Here is a check-functional run with this series and the v4 I sent : https://paste.debian.net/1336417/ This system is fast so the timeouts are more certainly tests failures : func-sh4-sh4_tuxrun ? func-rx-rx_gdbsim https://gitlab.com/qemu-project/qemu/-/issues/2691 Thanks, C.
On 22/11/2024 12.59, Cédric Le Goater wrote: > On 11/22/24 11:47, Thomas Huth wrote: >> On 21/11/2024 22.46, Cédric Le Goater wrote: >>> On 11/21/24 20:10, Thomas Huth wrote: >>>> On 21/11/2024 20.03, Cédric Le Goater wrote: >>>>> Hello Alex, >>>>> >>>>> On 11/21/24 17:57, Alex Bennée wrote: >>>>>> This is a mostly testing focused set of patches but a few bug fixes as >>>>>> well. I plan to send the PR in on Monday. I can drop any patches that >>>>>> are objected to but I think its pretty safe. >>>>>> >>>>>> Contains: >>>>>> >>>>>> - Daniel's clean-up of functional tests >>>>>> - Another avocado->function conversion from Thomas >>>>>> - Update the tuxrun baseline images >>>>>> - Minor fix to the rust pl011 device >>>>>> - Documentation clarification on identity >>>>>> >>>>>> The following could do with some review: >>>>>> >>>>>> tests/functional: update the x86_64 tuxrun tests >>>>>> tests/functional: update the sparc64 tuxrun tests >>>>>> tests/functional: update the s390x tuxrun tests >>>>>> tests/functional: update the riscv64 tuxrun tests >>>>>> tests/functional: update the riscv32 tuxrun tests >>>>>> tests/functional: update the ppc64 tuxrun tests >>>>>> tests/functional: update the ppc32 tuxrun tests >>>>>> tests/functional: update the mips64el tuxrun tests >>>>>> tests/functional: update the mips64 tuxrun tests >>>>>> tests/functional: update the mips32el tuxrun tests >>>>>> tests/functional: update the mips32 tuxrun tests >>>>>> tests/functional: add a m68k tuxrun tests >>>>>> tests/functional: update the i386 tuxrun tests >>>>>> tests/functional: update the aarch64 tuxrun tests >>>>>> tests/functional: update the arm tuxrun tests >>>>>> tests/functional: Convert the Avocado aarch64 tuxrun tests >>>>> >>>>> Do you think we could include patches 2-4 from this series ? >>>>> >>>>> https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ >>>>> >>>>> Only patch 3 lacks a R-b. >>>> >>>> I had a question on patch 3 and a suggestion on patch 4 ... could you >>>> maybe address them first? Thanks! >>> >>> Oh Sorry I forgot. >>> >>> I did address patch 4 in my tree ... In Patch 3, EXTRA_BOOTARGS was >>> added by Alex IIRC to work around the console issue. We ended up >>> adding the sleep call. I can resend tomorrow. >> >> IMHO v4 looks fine now ( https://lore.kernel.org/qemu- >> devel/20241122090322.1934697-1-clg@redhat.com/ ) in case you want to pick >> it up, Alex. >> >> Thomas >> > > Here is a check-functional run with this series and the v4 I sent : > > https://paste.debian.net/1336417/ > > This system is fast so the timeouts are more certainly tests failures : > > func-sh4-sh4_tuxrun ? > func-rx-rx_gdbsim https://gitlab.com/qemu-project/qemu/-/issues/2691 Both tests are marked with QEMU_TEST_FLAKY_TESTS, so yes, it's likely the flakiness of the tests. Thomas
On 11/22/24 13:32, Thomas Huth wrote: > On 22/11/2024 12.59, Cédric Le Goater wrote: >> On 11/22/24 11:47, Thomas Huth wrote: >>> On 21/11/2024 22.46, Cédric Le Goater wrote: >>>> On 11/21/24 20:10, Thomas Huth wrote: >>>>> On 21/11/2024 20.03, Cédric Le Goater wrote: >>>>>> Hello Alex, >>>>>> >>>>>> On 11/21/24 17:57, Alex Bennée wrote: >>>>>>> This is a mostly testing focused set of patches but a few bug fixes as >>>>>>> well. I plan to send the PR in on Monday. I can drop any patches that >>>>>>> are objected to but I think its pretty safe. >>>>>>> >>>>>>> Contains: >>>>>>> >>>>>>> - Daniel's clean-up of functional tests >>>>>>> - Another avocado->function conversion from Thomas >>>>>>> - Update the tuxrun baseline images >>>>>>> - Minor fix to the rust pl011 device >>>>>>> - Documentation clarification on identity >>>>>>> >>>>>>> The following could do with some review: >>>>>>> >>>>>>> tests/functional: update the x86_64 tuxrun tests >>>>>>> tests/functional: update the sparc64 tuxrun tests >>>>>>> tests/functional: update the s390x tuxrun tests >>>>>>> tests/functional: update the riscv64 tuxrun tests >>>>>>> tests/functional: update the riscv32 tuxrun tests >>>>>>> tests/functional: update the ppc64 tuxrun tests >>>>>>> tests/functional: update the ppc32 tuxrun tests >>>>>>> tests/functional: update the mips64el tuxrun tests >>>>>>> tests/functional: update the mips64 tuxrun tests >>>>>>> tests/functional: update the mips32el tuxrun tests >>>>>>> tests/functional: update the mips32 tuxrun tests >>>>>>> tests/functional: add a m68k tuxrun tests >>>>>>> tests/functional: update the i386 tuxrun tests >>>>>>> tests/functional: update the aarch64 tuxrun tests >>>>>>> tests/functional: update the arm tuxrun tests >>>>>>> tests/functional: Convert the Avocado aarch64 tuxrun tests >>>>>> >>>>>> Do you think we could include patches 2-4 from this series ? >>>>>> >>>>>> https://lore.kernel.org/all/20241112130246.970281-1-clg@redhat.com/ >>>>>> >>>>>> Only patch 3 lacks a R-b. >>>>> >>>>> I had a question on patch 3 and a suggestion on patch 4 ... could you maybe address them first? Thanks! >>>> >>>> Oh Sorry I forgot. >>>> >>>> I did address patch 4 in my tree ... In Patch 3, EXTRA_BOOTARGS was >>>> added by Alex IIRC to work around the console issue. We ended up >>>> adding the sleep call. I can resend tomorrow. >>> >>> IMHO v4 looks fine now ( https://lore.kernel.org/qemu- devel/20241122090322.1934697-1-clg@redhat.com/ ) in case you want to pick it up, Alex. >>> >>> Thomas >>> >> >> Here is a check-functional run with this series and the v4 I sent : >> >> https://paste.debian.net/1336417/ >> >> This system is fast so the timeouts are more certainly tests failures : >> >> func-sh4-sh4_tuxrun ? >> func-rx-rx_gdbsim https://gitlab.com/qemu-project/qemu/-/issues/2691 > > Both tests are marked with QEMU_TEST_FLAKY_TESTS, so yes, it's likely the flakiness of the tests. > > Thomas func-sh4-sh4_tuxrun runs fine with : @@ -46,10 +46,8 @@ class TuxRunSh4Test(TuxRunBaselineTest): console_index=1) self.vm.launch() - self.wait_for_console_pattern("Welcome to TuxTest") - time.sleep(0.1) - exec_command(self, 'root') - time.sleep(0.1) + self.wait_for_console_pattern("tuxtest login:") + exec_command_and_wait_for_pattern(self, 'root', 'root@tuxtest:~#') exec_command_and_wait_for_pattern(self, 'halt', "reboot: System halted")
Thomas Huth <thuth@redhat.com> writes: > On 21/11/2024 18.31, Alex Bennée wrote: >> Peter Maydell <peter.maydell@linaro.org> writes: >> >>> On Thu, 21 Nov 2024 at 16:58, Alex Bennée <alex.bennee@linaro.org> wrote: >>>> >>>> This is a mostly testing focused set of patches but a few bug fixes as >>>> well. I plan to send the PR in on Monday. I can drop any patches that >>>> are objected to but I think its pretty safe. >>>> >>>> Contains: >>>> >>>> - Daniel's clean-up of functional tests >>>> - Another avocado->function conversion from Thomas >>>> - Update the tuxrun baseline images >>>> - Minor fix to the rust pl011 device >>>> - Documentation clarification on identity >>> >>> Should we really be updating the tuxrun baseline images >>> in the middle of a release freeze period? Unless the old images >>> are going to go away and break the tests, I think it would >>> be better to stick with what we're currently testing. >> Well the arm64be fixed a real problem and while I was at it I >> figured >> might as well keep the rest in sync. I have tested them so they all pass >> (although I'm waiting on the CI run now). > > But there could be new intermittent problems in the new images ... so > if we'd face such a problem, we would not know whether it is the image > or whether it is QEMU. Thus maybe let's better keep the old versions > for 9.2 (except for the arm64 patch that fixes a real problem), and > use the new versions for 10.0 ? Fair enough. Will drop the others for now. > > Thomas