by Willy Tarreau

[permalink] [raw]

Subject: Re: [PATCH v3 08/12] selftests/nolibc: allow quit qemu-system when poweroff fails

On Sat, Jul 29, 2023 at 08:05:00PM +0800, Zhangjin Wu wrote:
(...)
> The new added lines are just want to keep the console active to tell
> users the test is running at background, even there is potential hang,
> users will learn what the progress it is, but if we allow print the
> running log to screen (may still have the issues mentioned by Willy in
> another email thread), these lines can be simply removed to get a
> cleaner version.
>
> The old version didn't add such logic, beside timeout logic, it simply
> try to match the expected 'reboot: ' string, but we do need to match the
> 'Kernel panic' string too to catch a crash nolibc-test case, so, a
> simplified version looks like this:
>
> TIMEOUT_CMD = t=$(QEMU_TIMEOUT); finish=0; \
> echo "Running $(IMAGE_NAME) on qemu-system-$(QEMU_ARCH)"; \
> while [ $$t -gt 0 ]; do \
> sleep 1; t=$$(expr $$t - 1); \
> grep -E "reboot: System halted|reboot: Power down|Kernel panic" "$(RUN_OUT)"; then finish=1; sleep 1; break; fi; \
> done; \
> if [ $$finish -eq 1 ]; then echo "test finish"; else echo "test timeout"; fi; \
> pkill -15 qemu-system-$(QEMU_ARCH) || true
>
> Now, the lines are not too many, and the left expected strings should be
> stable enough, do you like this one?

Do you realize how unreadable and undebuggable this is for the casual
reader who just tries to figure what command line to pass to qemu to
run the test manually ? You seem to focus a lot on "if the test hangs"
but nobody ever reported such a problem in the last few years. However
we've been spending weeks discussing how to add support for extra
features or architectures and such hacky stuff definitely steps in the
way and complicates everything.

> I do expect the timeout command has
> a "-m" option like this:
>
> timeout --foreground 10 -m "reboot: |Kernel panic" qemu-system-$(QEMU_ARCH) ...

I suggest that we give up on this timeout thing completely. You've shown
that it significantly complicates everything and we yet have to find a
single valid use case. Other subsystems also use QEMU without this.
rcutorture does perform some process management but it's way more advanced
and complete, and there's a reason: the tests are usually not supposed to
end by themselves.

> > > Before we complicated nolibc-test to handle the no-procfs case to save a
> > > few seconds building the kernel and now we have fairly big timeouts.
>
> In reality, the answer is NO to "now we have fairly big timeouts".
>
> Firstly, If there is no hang or no poweroff failure, every run will quit
> normally.

Which is one good reason for not complexifying everything.

> Secondly, even there is a poweroff failure or kernel panic, a 'reboot: '
> line will be always printed as expected and quit normally.

This is not supposed to happen in automated tests, and archs that fail
at this will simply be either excluded from automated tests, or will
be run through whatever external timeout mechanism.

> Thirdly, just only when there is a hang (not specific to tinyconfig),
> such as wrong bios version or missing firmware or even kernel hang with
> new changes (for example, the irqstack related riscv kernel hang I
> reported), blabla. such hangs will be detected by the fixed timeout
> value (like a watchdog).

These are local tools issues, we can't fix them all and it's not our
job. We're just providing an easy and hopefully convenient framework
to test syscalls. We're not supposed to require users to have to go
through complex debugging in this. And if they face a failure due to
their local tools (like I had with my old qemu version), they'll just
work around it by rebuilding it and that will be done. In the worst
case, some archs might require a Ctrl-C once in a while during a
manual test. No big deal. Definitely not as big as spending 10 minutes
trying to figure how to find one's way through a complicated makefile,
or wondering what's that runaway qemu process in the background that
refuses to die, etc.

> When poweroff fails on a target qemu, it still prints the 'reboot: '
> line as below, but will hang there after this line, our timeout logic
> will detect this line and tell qemu quit as a normal poweroff.
>
> Total number of errors: 0
> Leaving init with final status: 0
> reboot: System halted /* This line which be printed when reboot() issues, not depends on qemu or kernel driver */
> ---> detecting reboot: System halted|reboot: Power down|Kernel panic - not syncing: Attempted to kill init|Rebooting ...
> test finish /* Even qemu not poweroff successfully, we will kill it while the above line detected */
> LOG: Boot run successfully.
> Running boot-finish
>
> The oldest method I used locally is the 'timeout' command itself only,
> it is really a dead wait for a fixed timeout, the new method we used
> here to match the expected string does help a lot to simulate a always
> sucessful qemu poweroff support, no kernel and qemu dependent,
> lightweight enough.

I'm still definitely not fond of it because it turns something super simple
into something complex and likely unreliable, for very little benefit.

> Is the above explaination clearer? ;-)

At least given the numerous approaches we've seen, now I'm convinced I
don't want to see that anymore. Too much complication, a feeling of
hackish and unreliable solution that will very likely cause lots of
fixes in the future.

> Based on the kernel versions from v2.6.10 - v6.x which I have used on
> more than 7 architectures in my own Linux Lab project, the poweroff
> support of qemu + kernel is never deterministic, some versions ok, some
> versions fail, even with the defconfig, that's why an external timeout
> is always required.

Again, never ever experienced such a problem with the default configs
for the supported archs, with one of my machines having a qemu version
as old as 2.7. Paul always runs all the tests as well and never reported
this problem either. Thus I would like that we stick to precise facts
about problems that occur rather than just papering over them in a
generic way.

> > > "Since the low-level poweroff support is heavily kernel & qemu dependent"
> > >
> > > The kernel we can control.
> > >
> > > How common are qemus with that are missing poweroff support?
> > > As this worked before I guess the only architecture where this could
> > > pose a problem would be ppc.
> > >
>
> Yes, as explained above, based on the experience I have on tons of
> kernel versions of different architecture, it is really hard to make
> poweroff work as expected all the time, as the kernel and qemu changes,
> it may fail at any version randomly.

Please could you provide us with a reproducer for this problem, with
the mainline commit ID, arch, toolchain used, qemu version, because I
think you're generalizing over a few cases that happened during your
tinyconfig tests, for various possible reasons, but which are likely
not good reasons for complicating everything.

> Beside ppc, all of my local tinyconfig config files of every
> architecture are ready for boot and print and also of course for the
> 'reboot: ' line print. but it is 'hard' to find and enable the left
> options to just further enable 'poweroff' support.

If tinyconfig is not fixable, it's not usable, period. Right now all
archs stop fine for many of us with defconfig. If only a few tinyconfig
fail we'd rather invest time on these specific ones trying to figure
what options you need to add to the extra_config.

> Firstly, I have tried to enable some of them, but it deviates the
> tinyconfig goal, for example, x86 requires to enable both ACPI and PCI
> to just let poweroff work, so, I'm not planning to really enable them.
>
> Secondly, the time cost to just find and enable the poweroff options for
> every architecture (and even for the new nolibc portings) is huge, I
> give up after several tries, and they may fail in some future versions
> randomly, I do think we may be not really interested in fixing up such
> bugs in kernel drivers side ;-)

OK so better just give up on tinyconfig if it's not suitable for the
tests ? The more you present the shortcomings that come with them, the
less appealing it looks now.

> Thirdly, as Thomas mentioned before, the wireguard test use tinyconfig
> too, just found it also gives up the poweroff support in its config
> options and it use a simple raw timeout command, but the timeout is
> really 'huge', 20m v.s. 10 seconds ;-)
>
> grep timeout tools/testing/selftests/wireguard/qemu/Makefile
> timeout --foreground 20m qemu-system-$(QEMU_ARCH) \
> timeout --foreground 20m $< \

As I previously mentioned, I'm not fundamentally against a simple
timeout command *iff* we can validate the exact problem and the
conditions where it happens and we decide that it's the best workaround.
And I suspect that even once we find it we'll prefer to just not run
that specific setup by default and that's all.

> > I think I have a much simpler idea: we don't care about PPC32. I mean
> > OK it can be supported if it happens to work, we will just not include
> > it in default runs, because it will require Ctrl-C to finish, and so
> > what ? nolibc has been in the kernel for 5 years or so, nobody ever
> > cared about PPC, why should we suddenly break or complicate everything
> > just to support a sub-arch that nobody found interesting to add till
> > now?
> >
>
> Yes, this timeout logic is ok to be removed from this patchset, till we
> get a better solution.

Thanks.

It would really help if your series could focus on *one thing* at a
time. Currently I feel like in almost all patch series you've sent
there are good stuff that could have been merged already, but which
are mixed with hacks or unacceptable massive reworks that just result
in the rest being kept on hold.

I would really appreciate it if you thought about clearly presenting
the problems you're trying to solve before sending patch series, so
that we can collectively decide whether these problems deserve being
fixed or can be ignored, and if the cost of addressing them outweigh
their cost. It would save many hours of review of patches whose goal
is not always very clear. A good rule of thumb is that something that
is only added and that provides a specific feature generally suffers
not much discussion (beyond the usual style/naming etc). But patches
that modify existing process, code organization or affect reliability
should be discussed and argumented before even being created. It's
easier to discuss a purpose than to try to review a patch for a context
that is not very clear.

Thanks,
Willy

2023-07-30 23:49:53

by Zhangjin Wu

[permalink] [raw]

Subject: Re: [PATCH v3 08/12] selftests/nolibc: allow quit qemu-system when poweroff fails

Hi, Willy

[...]
> You seem to focus a lot on "if the test hangs"
> but nobody ever reported such a problem in the last few years. However

Willy, as Thomas shared in another reply too, there is really a poweroff failure
(even with defconfig) in the default ppc/g3beige 32-bit PowerPC Qemu we are
adding support for.

Btw, It is possible to switch another 32-bit PowerPC board who support poweroff
in qemu+kernel side, but need more investigate and survey, I have tried another
two, no lucky, the ppce500 webpage [1] even claims there is 'Power-off
functionality via one GPIO pin', but tried to enable the related gpios and
reset options, still not work, perhaps I need help from more PowerPc users,
will ask somebody to help me ;-)

[1]: https://qemu.readthedocs.io/en/latest/system/ppc/ppce500.html

> we've been spending weeks discussing how to add support for extra
> features or architectures and such hacky stuff definitely steps in the
> way and complicates everything.

The background is, based on your feedback in another reply, printing the log to
screen (another patch I have removed in this revision) may have some issues
before. But if without any log and just hang after a poweroff failure and even
further no timeout ... then, dead hang, that is really hard to debug, edit of
Makefile and rerun and then we will know it fails at poweroff or just during
qemu start ..., as you replied below, CTRL+C may help to cat the log file too,
but it waste a CTRL+C + cat during the early development stage, especially when
we require frequenty modify and test.

[...]
>
> This is not supposed to happen in automated tests, and archs that fail
> at this will simply be either excluded from automated tests, or will
> be run through whatever external timeout mechanism.
>

Agree with external timeout mechanism or even excluded from automated tests,
but we at least to allow to do basic test without editing the Makefile or
without forever Ctrl-C + cat run.out manually ;-)

So, what about at least restore part of my old 'print running log to screen'
patch like this:

# run the tests after building the kernel
run: kernel
$(Q)$(QEMU_SYSTEM_RUN) | tee "$(RUN_OUT)"
$(Q)$(REPORT) "$(RUN_OUT)"

# re-run the tests from an existing kernel
rerun:
$(Q)$(QEMU_SYSTEM_RUN) | tee "$(RUN_OUT)"
$(Q)$(REPORT) "$(RUN_OUT)"

Or even better, allow use a swtich to control it (the one we also removed from
an old patch).

ifneq ($(QUIET_RUN),1)
LOG_OUT= | tee "$(RUN_OUT)"
else
LOG_OUT= > "$(RUN_OUT)"
endif

If missing this, the early stages of a new architecture porting is definitely
very hard, not think about poweroff failure, just think about some other
potential exceptions which simply stop the booting or testing (not continue and
not exit).

> > Thirdly, just only when there is a hang (not specific to tinyconfig),
> > such as wrong bios version or missing firmware or even kernel hang with
> > new changes (for example, the irqstack related riscv kernel hang I
> > reported), blabla. such hangs will be detected by the fixed timeout
> > value (like a watchdog).
>
> These are local tools issues, we can't fix them all and it's not our
> job. We're just providing an easy and hopefully convenient framework
> to test syscalls. We're not supposed to require users to have to go
> through complex debugging in this. And if they face a failure due to
> their local tools (like I had with my old qemu version), they'll just
> work around it by rebuilding it and that will be done. In the worst
> case, some archs might require a Ctrl-C once in a while during a
> manual test. No big deal. Definitely not as big as spending 10 minutes
> trying to figure how to find one's way through a complicated makefile,
> or wondering what's that runaway qemu process in the background that
> refuses to die, etc.

Again Willy, without log and without timeout (with a tail of last 10 lines in
the old version), just a 'forever' hang, Ctrl-C helps but not much, still
require to check the log to see what happens or editing the Makefile to enable
the logging temply, it is realy time-waste during development ;-)

[...]
>
> Again, never ever experienced such a problem with the default configs
> for the supported archs, with one of my machines having a qemu version
> as old as 2.7. Paul always runs all the tests as well and never reported
> this problem either. Thus I would like that we stick to precise facts
> about problems that occur rather than just papering over them in a
> generic way.
>

Willy, before, the old versions of this patchset did only configure
QEMU_TIMEOUT for the failed 32-bit PowerPC and not enabled for the other
boards, but later after a dicussion (perhaps I misunderstood one of your
suggestions?), this revision configure a default one for all, I also thought
about it is not good for all boards, but still a choice to a board really have
no poweroff support (external one is ok if with explicitly log printing to
screen).

we have used something like this before:

ifneq ($(QEMU_TIMEOUT),)
...
endif

> > > > "Since the low-level poweroff support is heavily kernel & qemu dependent"
> > > >
> > > > The kernel we can control.
> > > >
> > > > How common are qemus with that are missing poweroff support?
> > > > As this worked before I guess the only architecture where this could
> > > > pose a problem would be ppc.
> > > >
> >
> > Yes, as explained above, based on the experience I have on tons of
> > kernel versions of different architecture, it is really hard to make
> > poweroff work as expected all the time, as the kernel and qemu changes,
> > it may fail at any version randomly.
>
> Please could you provide us with a reproducer for this problem, with
> the mainline commit ID, arch, toolchain used, qemu version, because I
> think you're generalizing over a few cases that happened during your
> tinyconfig tests, for various possible reasons, but which are likely
> not good reasons for complicating everything.
>

It is very hard to list all, the following one is a section [1] (not updated
for a long time) of the document of my 'Linux Lab' project:

### 6.2.2 Poweroff hang

Both of the `poweroff` and `reboot` commands not work on these boards currently (LINUX=v5.1):

* mipsel/malta (exclude LINUX=v2.6.36)
* mipsel/ls232
* mipsel/ls1b
* mips64el/ls2k
* mips64el/ls3a7a
* aarch64/raspi3
* arm/versatilepb

System will directly hang there while running `poweroff` or `reboot`, to exit qemu, please pressing `CTRL+a x` or using `pkill qemu`.

To test such boards automatically, please make sure setting `TEST_TIMEOUT`, e.g. `make test TEST_TIMEOUT=50`.

Welcome to fix up them.

IMHO, since my project is mainly for kernel learning and development, I even
don't care about why it not poweroff successfully at last although I have tried
my best to find the realted options but failed at last, the reasons may be
complex, from kernel side or from qemu side, both possible, or even simply a
kernel option changed and the config options are not updated timely.

We have so many boards with kernel version from v2.6.10 to v6.4, many poweroff
failures before:

$ ls -1 -d */*
aarch64/raspi3
aarch64/virt
arm/ebf-imx6ull
arm/mcimx6ul-evk
arm/versatilepb
arm/vexpress-a9
arm/virt
i386/pc
loongarch64/virt
mips64el/loongson3-virt
mips64el/ls2k
mips64el/ls3a7a
mipsel/ls1b
mipsel/ls232
mipsel/malta
ppc64le/powernv
ppc64le/pseries
ppc64/powernv
ppc64/pseries
ppc/g3beige
ppc/ppce500
riscv32/virt
riscv64/virt
s390x/s390-ccw-virtio
x86_64/pc

The kernel we used for i386/pc:

v2.6.10
v2.6.11.12
v2.6.12.6
v2.6.21.5
v2.6.24.7
v2.6.34.9
v2.6.35.14
v2.6.36
v3.10.108
v4.6.7
v5.1
v5.13
v5.2
v6.1.1

[1]: https://github.com/tinyclub/linux-lab/blob/master/README.md#622-poweroff-hang

> > Beside ppc, all of my local tinyconfig config files of every
> > architecture are ready for boot and print and also of course for the
> > 'reboot: ' line print. but it is 'hard' to find and enable the left
> > options to just further enable 'poweroff' support.
>
> If tinyconfig is not fixable, it's not usable, period. Right now all
> archs stop fine for many of us with defconfig. If only a few tinyconfig
> fail we'd rather invest time on these specific ones trying to figure
> what options you need to add to the extra_config.
>

Yes, that's why I plan to send the left patches by architecture, it will give
us more time to configure and confirm the missing poweroff options if really
required, I will ask somebody else to work with me together.

> > Firstly, I have tried to enable some of them, but it deviates the
> > tinyconfig goal, for example, x86 requires to enable both ACPI and PCI
> > to just let poweroff work, so, I'm not planning to really enable them.
> >
> > Secondly, the time cost to just find and enable the poweroff options for
> > every architecture (and even for the new nolibc portings) is huge, I
> > give up after several tries, and they may fail in some future versions
> > randomly, I do think we may be not really interested in fixing up such
> > bugs in kernel drivers side ;-)
>
> OK so better just give up on tinyconfig if it's not suitable for the
> tests ? The more you present the shortcomings that come with them, the
> less appealing it looks now.
>

The only shortcoming is some 'poweroff' failures currently, I do think the
time-saved is huge, with tinyconfig, I even don't want to try the time-consumer
defconfig anymore (although it is a requirement of many kernel features for the
last release), especially for nolibc development ;-)

I'm using tinyconfig+nolibc for most of the boards supported by my Linux Lab
project, I'm really happy with it, I use similar timeout + expected string
logic now, it works better than the old pure timeout logic, then, I don't care
about which poweroff options the target board want, just ignore the
requirement. It is a very good kernel learning and development experience, I
have used a defconfig based config and also a small config customized for
nolibc for them before, but with the new tinyconfig (extra boot/print options)
suggested by Thomas, it is really a good and a whole new experience.

To be honest, from my side, it is ok to add tinyconfig for nolibc or not,
because I can use the test script from Linux Lab project to test new nolibc
features, but I do think this will help a lot for both the kernel and nolibc
developers, so, I'm happy to continue, even we are facing some real
difficulties currently.

[...]
> > >
> >
> > Yes, this timeout logic is ok to be removed from this patchset, till we
> > get a better solution.
>

But again as explained above, we do need something to cope with 'dead' hang
without any output.

> Thanks.
>
> It would really help if your series could focus on *one thing* at a
> time. Currently I feel like in almost all patch series you've sent
> there are good stuff that could have been merged already, but which
> are mixed with hacks or unacceptable massive reworks that just result
> in the rest being kept on hold.
>

Agree, thanks, I will try my best to split a huge stuff to smaller ones, I was
a little hurried to want to upstream the non-rv32 related parts quickly, then,
we can back to rv32 asap ;-)

> I would really appreciate it if you thought about clearly presenting
> the problems you're trying to solve before sending patch series, so
> that we can collectively decide whether these problems deserve being
> fixed or can be ignored, and if the cost of addressing them outweigh
> their cost. It would save many hours of review of patches whose goal
> is not always very clear.

I like this suggestion, since you have mentioned before, based on your
suggestions, like the syscall.h stuff, a disucss or a proposal is fired before
a real patchset, although I think it worths a real RFC patchset, but let's
delay it and I do need more work to solve the questions asked from you, this
may be also related to some ideas mentioned by Arnd before, seems he have
planed to generate some syscall templates from the kernel .tbls, let's discuss
this in another thread, please ignore this one.

> A good rule of thumb is that something that
> is only added and that provides a specific feature generally suffers
> not much discussion (beyond the usual style/naming etc). But patches
> that modify existing process, code organization or affect reliability
> should be discussed and argumented before even being created. It's
> easier to discuss a purpose than to try to review a patch for a context
> that is not very clear.

That's reasonable, a good rule is 'subtraction', not 'addition', sometimes, the
implementation of new idea does add more impulsion to modify anything, it
require calmness to go back to the core issue we want to solve.

Thanks a lot.

Best regards,
Zhangjin

>
> Thanks,
> Willy