From: Arnout Vandecappelle <arnout@mind.be>
To: buildroot@busybox.net
Subject: [Buildroot] [PATCH 7/9] boot/syslinux: bump version
Date: Mon, 28 Apr 2014 08:14:52 +0200 [thread overview]
Message-ID: <535DF1DC.9090307@mind.be> (raw)
In-Reply-To: <e9a6a68649fc668ebc91f3f307e40af30c2ccb9b.1398378217.git.yann.morin.1998@free.fr>
On 25/04/14 00:30, Yann E. MORIN wrote:
> From: "Yann E. MORIN" <yann.morin.1998@free.fr>
>
> This new version has a very, very weird build system. There are different
> images that syslinux can now build:
> - the plain legacy-bios images we already supported previously
> - two new EFI32 and EFI64 applications
>
> To build one or the other, the Makefile accepts one or more of:
> make (bios|efi32|efi64)
>
> Specify all of them, and it builds all. Specify 'install', and it installs
> all of them, as one may expect.
>
> Still a regular behaviour, is to build only a subset (down to one):
> make bios <-- builds just the legacy-bios images
> make efi32 bios <-- builds just the legacy-bios and efi32 images
>
> Where it gets weird is the install procedure. Can you guess how it's done?
> Hint: the syslinux guys have invented the multiple-argument parsing in
> pure Makefiles. To build then install only the bios images, one would do:
> make bios
> make bios install
>
> Yep, that's it. make bios install. Two arguments, one action.
>
> That makes for some funky workarounds in our install procedure...
>
> 'bios' is the only image we support so far, with efi to come in a future
> patch.
>
> Using MAKE1, as there are issues with highly-parallel builds.
>
> Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
> Cc: Frank Hunleth <fhunleth@troodon-software.com>
> ---
> boot/syslinux/syslinux.mk | 32 ++++++++++++++++++++++++--------
> 1 file changed, 24 insertions(+), 8 deletions(-)
>
> diff --git a/boot/syslinux/syslinux.mk b/boot/syslinux/syslinux.mk
> index a1eb394..bec132f 100644
> --- a/boot/syslinux/syslinux.mk
> +++ b/boot/syslinux/syslinux.mk
> @@ -4,32 +4,48 @@
> #
> ################################################################################
>
> -SYSLINUX_VERSION = 4.07
> -SYSLINUX_SOURCE = syslinux-$(SYSLINUX_VERSION).tar.bz2
> -SYSLINUX_SITE = $(BR2_KERNEL_MIRROR)/linux/utils/boot/syslinux/4.xx/
> +SYSLINUX_VERSION = 6.02
> +SYSLINUX_SOURCE = syslinux-$(SYSLINUX_VERSION).tar.xz
> +SYSLINUX_SITE = $(BR2_KERNEL_MIRROR)/linux/utils/boot/syslinux/
>
> SYSLINUX_LICENSE = GPLv2+
> SYSLINUX_LICENSE_FILES = COPYING
>
> SYSLINUX_INSTALL_IMAGES = YES
>
> -SYSLINUX_DEPENDENCIES = host-nasm host-util-linux
> +SYSLINUX_DEPENDENCIES = host-nasm host-util-linux host-upx
> +
> +# The syslinux tarball comes with pre-compiled binaries.
> +# Since timestamps might not be in the correct order, a rebuild is
> +# not always triggered for all the different images.
> +# Cleanup the mess even before we attempt a build, so we indeed
> +# build everything from source.
Do we really need to build everything from source? We anyway already
check that our host compiler can build for ia32, so the included
executables will work.
> +define SYSLINUX_CLEANUP
> + rm -rf $(@D)/bios $(@D)/efi32 $(@D)/efi64
> +endef
> +SYSLINUX_POST_PATCH_HOOKS += SYSLINUX_CLEANUP
>
> # syslinux build system has no convenient way to pass CFLAGS,
> # and the internal zlib should take precedence so -I shouldn't
> # be used.
Actually, I'm not sure if this comment is still true. I haven't checked
in 6.02, but 4.07 already had a MYCFLAGS variables that can probably be
used. Also, the HOST_LDFLAGS may no longer be needed.
> define SYSLINUX_BUILD_CMDS
> - $(TARGET_MAKE_ENV) $(MAKE) CC="$(HOSTCC) -idirafter $(HOST_DIR)/usr/include $(HOST_LDFLAGS)" AR="$(HOSTAR)" -C $(@D)
> + $(TARGET_MAKE_ENV) $(MAKE1) CC="$(HOSTCC) -idirafter $(HOST_DIR)/usr/include $(HOST_LDFLAGS)" \
> + AR="$(HOSTAR)" -C $(@D) bios
> endef
>
> define SYSLINUX_INSTALL_TARGET_CMDS
> # While the actual bootloader is compiled for the target, several
> # utilities for installing the bootloader are meant for the host.
> - $(TARGET_MAKE_ENV) $(MAKE) -C $(@D) INSTALLROOT=$(HOST_DIR) install
> + # Repeat the target, otherwise syslinux will try to build everything
> + # Repeat CC and AR, since syslinux really wants to check them at
> + # install time
These comments should come before the define, otherwise they'll be
printed while building.
Regards,
Arnout
> + $(TARGET_MAKE_ENV) $(MAKE1) CC="$(HOSTCC) -idirafter $(HOST_DIR)/usr/include $(HOST_LDFLAGS)" \
> + AR="$(HOSTAR)" INSTALLROOT=$(HOST_DIR) \
> + -C $(@D) bios install
> endef
>
> -SYSLINUX_IMAGES-$(BR2_TARGET_SYSLINUX_ISOLINUX) += core/isolinux.bin
> -SYSLINUX_IMAGES-$(BR2_TARGET_SYSLINUX_PXELINUX) += core/pxelinux.bin
> +SYSLINUX_IMAGES-$(BR2_TARGET_SYSLINUX_ISOLINUX) += bios/core/isolinux.bin
> +SYSLINUX_IMAGES-$(BR2_TARGET_SYSLINUX_PXELINUX) += bios/core/pxelinux.bin
>
> define SYSLINUX_INSTALL_IMAGES_CMDS
> for i in $(SYSLINUX_IMAGES-y); do \
>
--
Arnout Vandecappelle arnout at mind be
Senior Embedded Software Architect +32-16-286500
Essensium/Mind http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint: 7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
next prev parent reply other threads:[~2014-04-28 6:14 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-24 22:30 [Buildroot] [PATCH 0/9] Bump and improve syslinux (branch yem/adopted) Yann E. MORIN
2014-04-24 22:29 ` [Buildroot] [PATCH 1/9] boot/syslinux: needs an ia32-capable compiler Yann E. MORIN
2014-04-28 5:55 ` Arnout Vandecappelle
2014-04-28 17:51 ` Yann E. MORIN
2014-04-24 22:29 ` [Buildroot] [PATCH 2/9] boot/syslinux: prepare to install non-core images Yann E. MORIN
2014-04-24 22:29 ` [Buildroot] [PATCH 3/9] boot/syslinux: remove 'default y' in sub-options Yann E. MORIN
2014-04-24 22:55 ` Arnout Vandecappelle
2014-04-25 14:54 ` Yann E. MORIN
2014-04-24 22:29 ` [Buildroot] [PATCH 4/9] boot/syslinux: rewrite options prompts Yann E. MORIN
2014-04-28 5:57 ` Arnout Vandecappelle
2014-04-28 17:36 ` Yann E. MORIN
2014-04-28 19:17 ` Arnout Vandecappelle
2014-04-24 22:30 ` [Buildroot] [PATCH 5/9] package/ucl: new package Yann E. MORIN
2014-04-24 22:30 ` [Buildroot] [PATCH 6/9] package/upx: " Yann E. MORIN
2014-04-28 6:03 ` Arnout Vandecappelle
2014-04-28 17:43 ` Yann E. MORIN
2014-04-24 22:30 ` [Buildroot] [PATCH 7/9] boot/syslinux: bump version Yann E. MORIN
2014-04-28 6:14 ` Arnout Vandecappelle [this message]
2014-04-28 17:26 ` Yann E. MORIN
2014-04-24 22:30 ` [Buildroot] [PATCH 8/9] boot/syslinux: add option to install the EFI image Yann E. MORIN
2014-04-28 6:18 ` Arnout Vandecappelle
2014-04-28 17:15 ` Yann E. MORIN
2014-04-28 19:16 ` Arnout Vandecappelle
2014-04-28 19:45 ` Yann E. MORIN
2014-04-28 19:51 ` Arnout Vandecappelle
2014-04-24 22:30 ` [Buildroot] [PATCH 9/9] boot/syslinux: add an option to install an MBR blob Yann E. MORIN
2014-04-28 16:23 ` Arnout Vandecappelle
2014-04-28 19:52 ` Yann E. MORIN
2014-04-28 19:58 ` Frank Hunleth
2014-04-28 20:39 ` Yann E. MORIN
2014-04-28 20:45 ` Frank Hunleth
2014-04-28 20:57 ` Arnout Vandecappelle
-- strict thread matches above, loose matches on Subject: below --
2014-04-29 16:49 [Buildroot] [PATCH 0/9 v2] Bump and improve syslinux (branch yem/syslinux) Yann E. MORIN
2014-04-29 16:49 ` [Buildroot] [PATCH 7/9] boot/syslinux: bump version Yann E. MORIN
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=535DF1DC.9090307@mind.be \
--to=arnout@mind.be \
--cc=buildroot@busybox.net \
/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