public inbox for u-boot@lists.denx.de
 help / color / mirror / Atom feed
From: AKASHI Takahiro <takahiro.akashi@linaro.org>
To: u-boot@lists.denx.de
Subject: [PATCH 1/1] sandbox: allow cross-compiling sandbox
Date: Tue, 9 Feb 2021 20:31:16 +0900	[thread overview]
Message-ID: <20210209113116.GA48142@laputa> (raw)
In-Reply-To: <2EAA47CB-2649-42A6-B341-8A2DE689E80C@gmx.de>

On Tue, Feb 09, 2021 at 07:40:23AM +0100, Heinrich Schuchardt wrote:
> Am 9. Februar 2021 06:30:31 MEZ schrieb AKASHI Takahiro <takahiro.akashi@linaro.org>:
> >On Mon, Feb 08, 2021 at 09:29:43PM -0700, Simon Glass wrote:
> >> Hi Heinrich,
> >> 
> >> On Mon, 8 Feb 2021 at 15:06, Heinrich Schuchardt <xypron.glpk@gmx.de>
> >wrote:
> >> >
> >> > UEFI test files like helloworld.efi require an architecture
> >specific
> >> > PE-COFF header.
> >> 
> >> architecture-specific
> >> 
> >> >
> >> > For non-sandbox the PE-COFF header is chosen by the target
> >architecture.
> >> > For the sandbox we use the host architecture. This is not helpful
> >for cross
> >> > compiling. Allow specifying the target architecture of the sandbox
> >via
> >> > environment variable MK_ARCH, e.g.
> >> 
> >> What exactly is the target arch? Is the the one that sandbox is
> >> running on, or something else?
> >> 
> >> >
> >> >     make sandbox_defconfig NO_SDL=1
> >> >     CROSS_COMPILE=aarch64-linux-gnu- NO_SDL=1 MK_ARCH=aarch64 make
> >> >
> >> > Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
> >> > ---
> >> >  Makefile             |  2 +-
> >> >  doc/arch/sandbox.rst | 14 ++++++++++++++
> >> >  2 files changed, 15 insertions(+), 1 deletion(-)
> >> >
> >> > diff --git a/Makefile b/Makefile
> >> > index 23dd11f723..286e5148ae 100644
> >> > --- a/Makefile
> >> > +++ b/Makefile
> >> > @@ -19,7 +19,7 @@ MAKEFLAGS += -rR --include-dir=$(CURDIR)
> >> >
> >> >  # Determine host architecture
> >> >  include include/host_arch.h
> >> > -MK_ARCH="${shell uname -m}"
> >> > +MK_ARCH?="${shell uname -m}"
> >> >  unexport HOST_ARCH
> >> >  ifeq ("x86_64", $(MK_ARCH))
> >> >    export HOST_ARCH=$(HOST_ARCH_X86_64)
> >> > diff --git a/doc/arch/sandbox.rst b/doc/arch/sandbox.rst
> >> > index 60ee1e0741..0dd9edc8cb 100644
> >> > --- a/doc/arch/sandbox.rst
> >> > +++ b/doc/arch/sandbox.rst
> >> > @@ -33,6 +33,20 @@ integers can only be built on 64-bit hosts.
> >> >
> >> >  Note that standalone/API support is not available at present.
> >> >
> >> > +Cross compiling
> >> > +---------------
> >> > +
> >> > +When cross compiling the U-Boot sandbox with CONFIG_EFI_LOADER=y
> >the target
> >> > +architecture must be specified using the MK_ARCH environment
> >variable using one
> >> > +of the values aarch64, armv7l, i386, riscv32, riscv64, x86_64,
> >e.g.
> >> 
> >> OK, but what is the target architecture? I'm just not sure from your
> >> comments what this actually means.
> >
> >I'm not quite sure if it is very useful to be able to cross-compile
> >the sandbox application. But any way, it is not very convenient
> 
> Building the sandbox on ARMv7 is broken due to patch introduced in 2020. We get loads of warnings when building for 32bit.

Are you using "cross-compiling" sandbox on armv7 machine?
For what target architecture?

> We should integrate building the sandbox for different architectures into Gitlab CI.
> 
> >to always enforce users to explicitly specify "target" architecture.
> >How about trying the following first?
> >$(CC) -v | grep "Target: " | sed -e 's/-*$//'
> >
> >It will give us a hint about the architecture.
> >("--print-multiarch" shows a similar triple, but this option won't
> >work for some pre-built gcc, like aarch64 gcc from Arm.)
> 
> Did you check clang?

I don't know, but we must specify "--target" option explicitly
in clang case, and according to U-Boot's Makefile, it will be uniquely
determined from "CROSS_COMPILE".

-Takahiro Akashi

> Is the output format specified? Will this be stable for compiler versions to come? Given the --print-multiarch failure probably not.
> 
> When compiling Linux you also have to supply CROSS_COMPILER and ARCH.
> 
> For me it is anyway a Gitlab CI thing. Nothing a user would typically use.
> 
> Best regards
> 
> Heinrich
> 
> 
> >
> >> Does MK mean make?
> >
> >I prefer another name for clarity.
> >
> >-Takahiro Akashi
> >
> >> 
> >> > +
> >> > +.. code-block:: bash
> >> > +
> >> > +    NO_SDL=1 make sandbox_defconfig
> >> > +    CROSS_COMPILE=aarch64-linux-gnu- MK_ARCH=aarch64 NO_SDL=1 make
> >> > +
> >> > +NO_SDL=1 was chosen in the example to remove the dependency on the
> >SDL library.
> >> > +The graphical console will not be available.
> >> >
> >> >  Prerequisites
> >> >  -------------
> >> > --
> >> > 2.30.0
> >> >
> >> 
> >> Regards,
> >> Simon
> 

  reply	other threads:[~2021-02-09 11:31 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-02-08 22:06 [PATCH 1/1] sandbox: allow cross-compiling sandbox Heinrich Schuchardt
2021-02-09  4:29 ` Simon Glass
2021-02-09  5:30   ` AKASHI Takahiro
2021-02-09  6:40     ` Heinrich Schuchardt
2021-02-09 11:31       ` AKASHI Takahiro [this message]
2021-02-09 12:25         ` Heinrich Schuchardt
2021-02-09  6:26   ` Heinrich Schuchardt
2021-02-09 15:25     ` Simon Glass
2021-02-09 17:52       ` Heinrich Schuchardt

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=20210209113116.GA48142@laputa \
    --to=takahiro.akashi@linaro.org \
    --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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox