From: Jeroen Hofstee <jeroen@myspectrum.nl>
To: u-boot@lists.denx.de
Subject: [U-Boot] [PATCH v2 0/3] Add some missing buildman features and deprecate MAKEALL
Date: Mon, 04 Aug 2014 22:56:20 +0200 [thread overview]
Message-ID: <53DFF374.2060406@myspectrum.nl> (raw)
In-Reply-To: <CAPnjgZ0D_YGHQQf4pFJ6SdmdbkvA6hqPiz7i1FL82fr+DUz_QQ@mail.gmail.com>
Hello Simon,
On 04-08-14 22:41, Simon Glass wrote:
> Hi Tom,
>
> On 4 August 2014 07:23, Tom Rini <trini@ti.com> wrote:
>> On Mon, Jul 28, 2014 at 11:53:27PM -0600, Simon Glass wrote:
>>
>>> Buildman has been around for a little over a year and is used by a fair
>>> number of U-Boot developers. However quite a few people still use MAKEALL.
>>>
>>> Buildman was intended to replace MAKEALL, so perhaps now is a good time to
>>> start that process.
>> OK, lets start. First, I had to make up a special directory with
>> symlinks to the various toolchains, in order to get them to work with
>> buildman, due to the various layouts:
>>
>> $ ls -l ~/work/u-boot/toolchain-links/
>> total 4
>> lrwxrwxrwx 1 trini trini 38 Aug 1 17:09 arc_gnu_4.8-R2_prebuilt_uclibc_be -> /opt/arc_gnu_4.8-R2_prebuilt_uclibc_be
>> lrwxrwxrwx 1 trini trini 38 Aug 1 17:09 arc_gnu_4.8-R2_prebuilt_uclibc_le -> /opt/arc_gnu_4.8-R2_prebuilt_uclibc_le
>> lrwxrwxrwx 1 trini trini 24 Aug 1 17:10 bfin-elf -> /opt/uClinux-45/bfin-elf
>> lrwxrwxrwx 1 trini trini 32 Aug 1 17:09 m68k-linux -> /opt/gcc-4.9.0-nolibc/m68k-linux
>> lrwxrwxrwx 1 trini trini 34 Aug 1 17:09 microblazeel-xilinx-linux-gnu -> /opt/microblazeel-xilinx-linux-gnu
>> lrwxrwxrwx 1 trini trini 32 Aug 1 17:09 microblaze-xilinx-linux-gnu -> /opt/microblaze-xilinx-linux-gnu
>> lrwxrwxrwx 1 trini trini 66 Aug 1 17:15 mips32-linux -> /opt/eldk-5.2.1/mips/sysroots/i686-eldk-linux/usr/bin/mips32-linux
>> lrwxrwxrwx 1 trini trini 36 Aug 1 17:09 nds32le-linux-glibc-v1f -> /opt/nds32le/nds32le-linux-glibc-v1f
>> lrwxrwxrwx 1 trini trini 20 Aug 1 17:09 sparc-elf-3.4.4 -> /opt/sparc-elf-3.4.4
>>
>> I tried a few things but I couldn't get a toolchain section setup that
>> found all of those. Second:
> OK, can you give me a trace of the output of 'buildman
> --list-tool-chains' - I should be able to figure out what is wrong.
Like Tom, I tried this a bit. (and ended up faking clang is gcc). There
are at least some assumptions cc = gcc and strlen(cc) == 3.
>> (testing the MMC pull request)
>> $ ./tools/buildman/buildman -b master
>> 'arc|blackfin|m68k|microblaze|nds32|sparc|mips'
>> No section: 'make-flags'
>> Building 11 commits for 121 boards (6 threads, 1 job per thread)
>> Cloning repo for thread 0
>> Cloning repo for thread 1
>> Cloning repo for thread 2
>> Cloning repo for thread 3
>> Cloning repo for thread 4
>> Cloning repo for thread 5
>> 1155 88 88 /1331 axs101
>>
>> So it's nice that it spit out a thread per commit, but I really only
>> care right now if master builds right. And I've gone blind and can't
>> find the end result logs of what failed to build (88 failures doesn't
>> quite match the warning/errors I get for my wrapper around
>> toolchains+MAKEALL).
> You can use the -s flag to see the broken builds, and -se to see
> errors also. Is that what you mean?
>
> The output is in ../<branch>/<commit>/<board> but if you want to keep
> the binaries, you need -k also.
Like Tom, I saw fancy output, like merged not being a property
or something. I like to see the warnings and errors directly, but -e?
failed to do so. (or at least I didn't spend enough time to get it to work)
>> I also didn't quite see how to plugin an arm64 toolchain.
> I haven't tried - can you please point me to one and I'll give it a go.
>
>> Next up, is there some way to have multiple toolchains for an arch
>> available? On some compute resources I have at work I throw
>> arm/eldk4.2, arm/eldk5.2, arm/linaro-2013.03 (since we've pinned to that
>> for some projects) and some various combos for powerpc as well. I
>> didn't see how to do that with buildman aside from different ~/.buildman
>> files per invocation.
> Yes I remember you talking about this before. One option would be to
> add toolchain as a new dimension in the matrix. It might be painful
> though. So to be clear, you want to define multiple toolchains for an
> arch, and have buildman always build with all of them?
>
I would like this as well. For serious changing I typically build with
MAKEALL with clang and gcc (for ARM) + gcc for powerpc if it is
a overall change.
But since clang always warns it is actually not a viable target for
buildman at all.
Regards,
Jeroen
next prev parent reply other threads:[~2014-08-04 20:56 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-07-29 5:53 [U-Boot] [PATCH v2 0/3] Add some missing buildman features and deprecate MAKEALL Simon Glass
2014-07-29 5:53 ` [U-Boot] [PATCH v2 1/3] buildman: Fix a few typos Simon Glass
2014-07-29 5:53 ` [U-Boot] [PATCH v2 2/3] buildman: Add some notes about moving from MAKEALL Simon Glass
2014-08-01 8:10 ` Masahiro Yamada
2014-08-04 10:15 ` Simon Glass
2014-07-29 5:53 ` [U-Boot] [PATCH v2 3/3] RFC: Deprecate MAKEALL Simon Glass
2014-08-04 13:23 ` [U-Boot] [PATCH v2 0/3] Add some missing buildman features and deprecate MAKEALL Tom Rini
2014-08-04 20:41 ` Simon Glass
2014-08-04 20:56 ` Jeroen Hofstee [this message]
2014-08-05 14:33 ` Simon Glass
2014-08-04 21:23 ` Tom Rini
2014-08-05 14:06 ` Simon Glass
2014-08-05 14:24 ` Simon Glass
2014-08-05 14:31 ` Tom Rini
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=53DFF374.2060406@myspectrum.nl \
--to=jeroen@myspectrum.nl \
--cc=u-boot@lists.denx.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.