Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <1658681004-132191-1-git-send-email-john.garry@huawei.com>
In-Reply-To: <1658681004-132191-1-git-send-email-john.garry@huawei.com>
From:   Arnd Bergmann <arnd@arndb.de>
Date:   Sun, 24 Jul 2022 22:35:53 +0200
Message-ID: <CAK8P3a0umWWic6LAzxXJ7BACYVE--m-wbynh7Z8F+pGoUBxGTA@mail.gmail.com>
Subject: Re: [PATCH 0/2] arm64 defconfig: Get faddr2line working
To:     John Garry <john.garry@huawei.com>
Cc:     Will Deacon <will@kernel.org>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Linux ARM <linux-arm-kernel@lists.infradead.org>,
        Arnd Bergmann <arnd@arndb.de>, Olof Johansson <olof@lixom.net>,
        SoC Team <soc@kernel.org>, jpoimboe@kernel.org,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sun, Jul 24, 2022 at 6:43 PM John Garry <john.garry@huawei.com> wrote:
>
> scripts/faddr2line has not worked by default for arm64 since 5.17
>
> Firstly, since commit f9b3cd245784 ("Kconfig.debug: make DEBUG_INFO
> selectable from a choice"), CONFIG_DEBUG_INFO was not getting enabled by
> default (and this is required for faddr2line to work).

I just noticed this the other day and applied a patch for it in the
arm/defconfig
branch.

> Secondly, commit dcea997beed6 ("faddr2line: Fix overlapping text section
> failures, the sequel") caused a breakage for arm64, as reported at the
> following:
> https://lore.kernel.org/lkml/3bd9817d-1959-c081-e5d0-8b0e70b3f41e@huawei.com/
>
> Josh has sent fixes/improvements for faddr2line at the following:
> https://lore.kernel.org/lkml/cover.1658426357.git.jpoimboe@kernel.org/
>
> In this series I enable CONFIG_DEBUG_INFO by enabling
> CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT, which seems the sensible
> option.
>
> As for merging this, I am not sure who would pick it up - any takers
> welcome.
>
> Note: this is based on next-20220722 and it may be wiser to sync the
> defconfig manually (instead of using 1/2). Indeed I am not sure what is
> the policy is of sync'ing this anyway.

I only synchronized the 32-bit defconfig files in my tree, not the 64-bit
one. However, I can't really apply your patch 2/2 because you appear
to  mix refreshing the order of the options with changes that remove
options that are gone after a 'savedefconfig', risking that we miss
other bugs as well, as seen from your diffstat:

 1 file changed, 36 insertions(+), 48 deletions(-)

I have refreshed this one as well now, which on my tree gives me

 1 file changed, 31 insertions(+), 31 deletions(-)

for a nonfunction change. I have left the other ones untouched
for the moment:

CONFIG_ARCH_BCMBCA=y
CONFIG_SECCOMP=y
CONFIG_QRTR=m
CONFIG_PINCTRL_MSM=y
CONFIG_SND_SOC_TEGRA210_OPE=m
CONFIG_MAILBOX=y
CONFIG_QCOM_ICC_BWMON=m
CONFIG_SLIMBUS=m
CONFIG_INTERCONNECT=y
CONFIG_CONFIGFS_FS=y

These should be checked manually to find out why savedefconfig
no longer shows them, it could be either a bug (a new dependency,
renamed option, a driver randomly selects another subsystem, etc)
that we need to fix, or a harmless change (driver was removed,
option is now intended to be default-enabled, ...)

If you want to help more, can you check some or all of the above
and send patches to either re-enable the options or remove them
individually with explanations about why they are no longer
part of the savedefconfig output?

          Arnd