Message-ID: <CABVgOSnBq0QE+Cq+SDeV-LxOQYbGZ6Bqbjix6h-UpNj0GMicPA@mail.gmail.com>
Date:   Sat, 19 Feb 2022 16:00:04 +0800
From:   David Gow <davidgow@...gle.com>
To:     Johannes Berg <johannes@...solutions.net>
Cc:     Jeff Dike <jdike@...toit.com>, Richard Weinberger <richard@....at>,
        Anton Ivanov <anton.ivanov@...bridgegreys.com>,
        Shuah Khan <skhan@...uxfoundation.org>,
        Brendan Higgins <brendanhiggins@...gle.com>,
        Randy Dunlap <rdunlap@...radead.org>,
        linux-um <linux-um@...ts.infradead.org>,
        amd-gfx@...ts.freedesktop.org, dri-devel@...ts.freedesktop.org,
        KUnit Development <kunit-dev@...glegroups.com>,
        "open list:KERNEL SELFTEST FRAMEWORK" 
        <linux-kselftest@...r.kernel.org>, linux-rdma@...r.kernel.org,
        x86@...nel.org, felix.kuehling@....com,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 4/4] kunit: tool: Disable broken options for --alltests

On Fri, Feb 18, 2022 at 8:26 PM Johannes Berg <johannes@...solutions.net> wrote:
>
> On Fri, 2022-02-18 at 15:57 +0800, David Gow wrote:
> >
> > Note that, while this does build again, it still segfaults on startup,
> > so more work remains to be done.
>
> That's probably just a lot more stuff getting included somehow?
>

Yeah: it used to work (a couple of years ago), but something has
broken it in the meantime. It's just a shame that bisecting things
with allyesconfig takes so long...

> > They are:
> > - CONFIG_VFIO_PCI: Needs ioport_map/ioport_unmap.
> > - CONFIG_INFINIBAND_RDMAVT: Needs cpuinfo_x86 and __copy_user_nocache
> > - CONFIG_BNXT: Failing under UML with -Werror
> > ERROR:root:../drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c: In function ‘bnxt_ptp_enable’:
> > ../drivers/net/ethernet/broadcom/bnxt/bnxt_ptp.c:400:43: error: array subscript 255 is above array bounds of ‘struct pps_pin[4]’ [-Werror=array-bounds]
> >   400 |                         ptp->pps_info.pins[pin_id].event = BNXT_PPS_EVENT_EXTERNAL;
> >       |                         ~~~~~~~~~~~~~~~~~~^~~~~~~~
> > - CONFIG_PATA_CS5535: Needs MSR access (__tracepoint_{read,write}_msr)
> > - CONFIG_VDPA: Enables CONFIG_DMA_OPS, which is unimplemented. ('dma_ops' is not defined)
> >
> > These are all issues which should be investigated properly and the
> > corresponding options either fixed or disabled under UML. Having this
> > list of broken options should act as a good to-do list here, and will
> > allow these issues to be worked on independently, and other tests to
> > work in the meantime.
> >
>
> I'm not really sure it makes sense to even do anything other than
> disabling these.
>
> It looks like all of them are just exposed by now being able to build
> PCI drivers on UML. Surely the people writing the driver didn't expect
> their drivers to run over simulated PCI (which is what the UML PCI
> support is all about).
>
> Now from a PCI driver point of view you can't really tell the difference
> (and anyway the driver won't be probed), but the issues (at least the
> build time ones) come from having
>
>     UML && PCI && X86_64
>
> or
>
>     UML && PCI && X86_32
>
> because drivers typically depend on X86_64 or X86_32, rather than on
> "X86 && X86_64" or "X86 && X86_32". In a sense thus, the issue is those
> drivers don't know that "!X86 && (X86_32 || X86_64)" can happen (with
> UML).
>
>
> Now you could say that's the driver bug, or you could say that they
> should just add "depends on !UML" (though that's basically equivalent to
> adding "depends on X86" and the latter may be preferable in some cases).
>
> Or actually in the three patches you have (1-3) it's in the code, but
> same thing, you can either add && !UML (like you did) or add && X86.
>

I didn't realise X86 wasn't defined in UML: that's definitely a bit
cleaner than !UML in a number of these cases.

Not all of those issues are fundamentally solved by "depends on X86",
though: there are a few which might be other missing things in UML
(maybe the 'dma_ops' issues?), and/or might be the result of -Werror
being enabled.

>
> Arguably, however, building PCI drivers by default is somewhat
> questionable in the first place?

We do want the ability to build PCI drivers under UML, as it makes
running KUnit tests for PCI drivers much simpler and more pleasant.
And indeed, it does work for KUnit in general, it's just that some
drivers do have the issues mentioned above, so allyesconfig picks up
every broken driver.

We don't actually build the PCI drivers by default, only if the
"--alltests" option is passed, which does include them, as we do have
tests which depend on PCI we'd like to run (like the thunderbolt
test).

>
> So maybe you should just add
>
>     # CONFIG_UML_PCI_OVER_VIRTIO is not set
>
> to the broken_on_uml.config since it exposes all these issues, and
> really is not very useful since you're not going to actually run with
> any simulated PCI devices anyway, so drivers will not be probed.

I did try this as well, and it just got us a different set of issues
(there are a bunch of drivers which depend on IOMEM but don't state it
-- I'll try to send fixes for those out next week). And we'd lose
things like the thunderbolt test if PCI

Ultimately, the 'broken_on_uml.config' file is just there to pare back
allyesconfig a bit for KUnit's purposes, but we still definitely want
as many options (and hence tests) enabled as possible long-term. So I
think actual fixes to either the code or Kconfig do make sense.

Is 'make ARCH=um allyesconfig' something we actually want to be able
to build? If so, no amount of adding things to KUnit's
broken_on_uml.config will solve the underlying issues, and we'll need
to at least update the Kconfig entries.

Cheers,
-- David

Download attachment "smime.p7s" of type "application/pkcs7-signature" (4003 bytes)

Powered by blists - more mailing lists