Message-ID: <20210204223653.23dijma7d526btqb@google.com>
Date:   Thu, 4 Feb 2021 14:36:53 -0800
From:   Fangrui Song <maskray@...gle.com>
To:     Mark Wielaard <mark@...mp.org>,
        Masahiro Yamada <masahiroy@...nel.org>
Cc:     Nick Desaulniers <ndesaulniers@...gle.com>,
        Nathan Chancellor <natechancellor@...il.com>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Sedat Dilek <sedat.dilek@...il.com>,
        LKML <linux-kernel@...r.kernel.org>,
        clang-built-linux <clang-built-linux@...glegroups.com>,
        Linux Kbuild mailing list <linux-kbuild@...r.kernel.org>,
        linux-arch <linux-arch@...r.kernel.org>,
        Jakub Jelinek <jakub@...hat.com>,
        Caroline Tice <cmtice@...gle.com>,
        Nick Clifton <nickc@...hat.com>, Yonghong Song <yhs@...com>,
        Jiri Olsa <jolsa@...nel.org>,
        Andrii Nakryiko <andrii@...nel.org>,
        Arnaldo Carvalho de Melo <acme@...nel.org>,
        Arvind Sankar <nivedita@...m.mit.edu>,
        Nathan Chancellor <nathan@...nel.org>,
        Guenter Roeck <linux@...ck-us.net>
Subject: Re: [PATCH v7 1/2] Kbuild: make DWARF version a choice

On 2021-02-04, Nick Desaulniers wrote:
>On Thu, Feb 4, 2021 at 12:28 PM Mark Wielaard <mark@...mp.org> wrote:
>>
>> On Thu, 2021-02-04 at 12:04 -0800, Nick Desaulniers wrote:
>> > On Thu, Feb 4, 2021 at 11:56 AM Mark Wielaard <mark@...mp.org> wrote:
>> > > I agree with Jakub. Now that GCC has defaulted to DWARF5 all the
>> > > tools
>> > > have adopted to the new default version. And DWARF5 has been out
>> > > for
>> >
>> > "all of the tools" ?
>>
>> I believe so yes, we did a mass-rebuild of all of Fedora a few weeks
>> back with a GCC11 pre-release and did find some tools which weren't
>> ready, but as far as I know all have been fixed now. I did try to
>
>Congrats, I'm sure that was _a lot_ of work.  Our toolchain folks have
>been pouring a lot of effort over getting our internal code all moved
>over, and it doesn't look like it's been easy from my perspective.
>
>> coordinate with the Suse and Debian packagers too, so all the major
>> distros should have all the necessary updates once switching to GCC11.
>
>That's great for users of the next Fedora release who can and will
>upgrade, but I wouldn't assume kernel developers can, or will (or are
>even using those distros).
>
>Most recently, there was discussion on the list about upgrading the
>minimally required version of GCC for building the kernel to GCC 5.1;
>we still had developers complain about abandoning GCC 4.9.  And
>Guenter shared with me a list of architectures he tests with where he
>cannot upgrade the version of GCC in order to keep building them.
>https://github.com/groeck/linux-build-test/blob/master/bin/stable-build-arch.sh
>(I hope someone sent bug reports for those)
>
>My intent is very much to allow for users of toolchains that have not
>switched the implicit default (such as all of the supported versions
>of GCC that have been released ie. up to GCC 10.2, and Clang; so all
>toolchains the kernel still supports) to enjoy the size saving of
>DWARF v5, and find what other tooling needs to be updated.
>
>> > > more than 4 years already. It isn't unreasonable to assume that people
>> > > using GCC11 will also be using the rest of the toolchain that has moved
>> > > on. Which DWARF consumers are you concerned about not being ready for
>> > > GCC defaulting to DWARF5 once GCC11 is released?
>> >
>> > Folks who don't have top of tree pahole or binutils are the two that
>> > come to mind.
>>
>> I believe pahole just saw a 1.20 release. I am sure it will be widely
>> available once GCC11 is released (which will still be 1 or 2 months)
>> and people are actually using it. Or do you expect distros/people are
>> going to upgrade to GCC11 without updating their other toolchain tools?
>
>Does no one test linux kernel builds with top of tree GCC built from
>source?  Or does that require "updating their other toolchain tools?"
>If I build ToT GCC from source, do I need to do the same for
>binutils-gdb in order to build the kernel?  Pretty sure I don't.
>
>https://bugzilla.redhat.com/show_bug.cgi?id=1922707 and
>https://bugzilla.redhat.com/show_bug.cgi?id=1922698 look like user
>reports to me, but hopefully some CI system reported earlier that
>builds of the Linux kernel with GCC 11 pre release would produce the
>warnings from those bug report.  Otherwise it looks like evidence that
>users "are going to upgrade to GCC11 without updating their other
>toolchain tools."  In the case of pahole, they could not, because
>fixes were not yet written.  "Just upgrade" doesn't work if there's no
>fix yet upstream.  (pahole is reported fixed for that specific issue,
>FWIW).
>
>> BTW. GCC11 doesn't need top of tree binutils, it will detect the
>> binutils capabilities (bugs) and adjust its DWARF output based on it.
>
>Yes, I saw https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=6923255e35a3d54f2083ad0f67edebb3f1b86506
>and https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=1aeb7d7d67d167297ca2f4a97ef20f68e7546b4c.
>It's nice that GCC can tightly couple to a version of binutils. Clang
>without its integrated assembler can make no such assumptions about
>which assembler the user will prefer to use instead at runtime, and
>without binutils 2.35.1 being widely available as we all would like,
>leads to issues shipping DWARF v5 by default.
>
>> >   I don't have specifics on out of tree consumers, but
>> > some Aarch64 extensions which had some changes to DWARF for ARMv8.3
>> > PAC support broke some debuggers.
>>
>> It would be really helpful if you could provide some specifics. I did
>> fix some consumers to handle the PAC operands in CFI last year, but I
>> don't believe that had anything to do with the default DWARF version,
>> just with dealing with DW_CFA_AARCH64_negate_ra_state.
>
>Yep, that's the one.  I suspect that the same out of tree consumers of
>DWARF that did not see that coming will similarly be stumped when
>presented with DWARF v5, but it's hypothetical, so not much of an
>argument I'll admit.  I just wouldn't bet that an upgrade to DWARF v5
>will be painless at this point in time, as evidenced by how much blood
>has been poured into finding what tools out there were broken and
>needed to be fixed.  I also recognize we can't drag our heels forever
>over this.  It's not the intent of the patch series to do so.
>
>> > I don't doubt a lot of work has gone into fixing many downstream
>> > projects and then when building everything from ToT that there are no
>> > issues with DWARF v5.  The issue is getting upgrades into developers
>> > hands, and what to default to until then.
>>
>> I would suggest you simply default to what you already do when the
>> compiler is given -g. Just like you do already for the implicit default
>> -std=gnuc*. Once GCC11 is actually released and people upgrade their
>> toolchain to use it the tools will be ready and in developers hands.
>
>Hmmm...thinking about this more over lunch.
>
>I think the linker script additions for the new DWARF v5 sections are
>most important.  There's no need to hold those hostage over what we do
>with the configs. They are orthogonal, and can go in first.  There's
>already one section in the linker script already related to v5, and we
>know that DWARF v5 is coming, so there's no reason not to just add
>them.  That would resolve
>https://bugzilla.redhat.com/show_bug.cgi?id=1922707.
>
>For configs, today the kernel has an explicit opt in for DWARF v4
>called CONFIG_DEBUG_INFO_DWARF4.  When it was introduced, I suspect
>DWARF v2 (or v3) was the default, so it let kernel developers opt in
>to newer versions (say, for testing) than what was the implicit
>default version.  With the advent of DWARF v5 and toolchains moving to
>producing that by default, maybe there is still a place for
>CONFIG_DEBUG_INFO_DWARF4, but as a way to opt in to older versions.
>Your installed version of pahole or binutils is too old for DWARF v5?
>Great, CONFIG_DEBUG_INFO_DWARF4 is your friend until you can update
>your tools.  CONFIG_DEBUG_INFO_DWARF4 becomes "opt backwards" rather
>than it's original "opt forwards" (if that makes sense).
>
>One small issue I have with that is that it doesn't help users of GCC
>5 through 10.2 (or clang) opt into DWARF v5 (to test, or for the size
>savings) similar to the original intent of CONFIG_DEBUG_INFO_DWARF4.
>That was my intent with this series; to opt into new versions so we
>can begin testing them and finding kinks in other tools.  It was not
>intentionally to hold back compilers once they upgrade their default
>versions, even if I still think explicit is better than implicit.  And
>it would be nice to have the framework to continue to do so for v6
>someday.
>
>So opting in to v4 explicitly provides a "pressure relief valve" for
>developers that don't have the latest tools.  Opting into v5
>explicitly allows for testing of those tools with the latest DWARF
>version (and size savings that are worthwhile) for folks with older
>GCC and Clang.
>
>What are your thoughts on this?
>
>What if I modified the config to have 3 options:
>1. (default) CONFIG_DEBUG_INFO_DWARF_VERSION_DONT_CARE (I'm open to a
>better color for the bikeshed). Does not set a -gdwarf-*.  This is the
>status quo today when you build the kernel.  You allow toolchain
>developers to decide your fate. If that makes you unhappy, congrats,
>you have the below options.
>2. CONFIG_DEBUG_INFO_DWARF4 (the ship has already sailed on the name
>here, sorry) Explicitly sets -gdwarf-4. Use this if you have a good
>reason not to upgrade to the recommended -gdwarf-5.  We might
>disappear this option soon. (already exists in tree today, and doesn't
>need to be removed; instead repurposed)
>3. CONFIG_DEBUG_INFO_DWARF5 Explicitly sets -gdwarf-5.  Use this if
>you would like to opt into a newer version than what's provided by
>default.  Provides significant size savings over DWARF v4. Avoid if
>you have broken shit you can't upgrade for whatever reason?
>
>Then I think everyone can be happy? Toolchain vendors can to continue
>pursuing updates of implicit default dwarf version (aggressive, but
>necessary force adoption, perhaps), BTF/pahole/binutils users can
>downgrade until they have the necessary updates, and developers with
>older GCC and clang can opt in new features without having to upgrade
>the toolchain's implicit default (since that requires significant work
>for other codebases, but is making progress) and still realize the
>binary size savings.

The points raised by Nick are very reasonable.

A few colleagues and I are involved in upgrading our internal systems to be
DWARF v5 ready. It required to fix a dozen of things.

Externally I watch binutils mailing list closely and in recent months DWARF
fixes are quite regular.  (Mark and H.J. contributed lot of fixes. Big thanks!)
For example, objdump -S did not display source for gcc-11 compiled object
files.  It was just fixed a few days ago
https://sourceware.org/bugzilla/show_bug.cgi?id=27231 , which is included in
binutils 2.35.2
(https://sourceware.org/pipermail/binutils/2021-January/115150.html).

As we all know DWARF v5 has huge size savings and we want to make it the
default, but today is probably a bit early.


(Note about Clang and binutils: I added -fbinutils-version= to Clang 12.
If there are workaround needs we can use that option.)

Powered by blists - more mailing lists