[Buildroot] [PATCH 0/3] Gummiboot fixup

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Esben Haabendal]

From: Esben Haabendal <esben@haabendal.dk>

These patches fixes gummiboot for building with toolchains using
newer gcc and glibc versions.

Esben Haabendal (3):
  gummiboot: use new official upstream git repository
  gummiboot: upgrade to last commit before being removed
  gummiboot: fix compatibility with newer glibc versions

 boot/gummiboot/0001-glibc-2.25-fix.patch | 16 ++++++++++++++++
 boot/gummiboot/gummiboot.mk              |  4 ++--
 2 files changed, 18 insertions(+), 2 deletions(-)
 create mode 100644 boot/gummiboot/0001-glibc-2.25-fix.patch

-- 
2.21.0

[Buildroot] [PATCH 1/3] gummiboot: use new official upstream git repository

[Esben Haabendal]

From: Esben Haabendal <esben@haabendal.dk>

Signed-off-by: Esben Haabendal <esben@haabendal.dk>
---
 boot/gummiboot/gummiboot.mk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/boot/gummiboot/gummiboot.mk b/boot/gummiboot/gummiboot.mk
index 85a5daaeab69..0e67e5e8ce0e 100644
--- a/boot/gummiboot/gummiboot.mk
+++ b/boot/gummiboot/gummiboot.mk
@@ -4,7 +4,7 @@
 #
 ################################################################################
 
-GUMMIBOOT_SITE = http://cgit.freedesktop.org/gummiboot
+GUMMIBOOT_SITE = https://gitlab.freedesktop.org/archived-projects/gummiboot.git
 GUMMIBOOT_SITE_METHOD = git
 GUMMIBOOT_VERSION = 43
 GUMMIBOOT_LICENSE = LGPL-2.1+
-- 
2.21.0

[Buildroot] [PATCH 2/3] gummiboot: upgrade to last commit before being removed

[Esben Haabendal]

From: Esben Haabendal <esben@haabendal.dk>

This upgrades to version 48 plus the last commits done after that.
Note, that after that, gummiboot is officially dead, so there will
be no more upstream commits after this.

For various reasons, there are still users of gummiboot, so let's base our
work on the latest gummiboot "version".

Signed-off-by: Esben Haabendal <esben@haabendal.dk>
---
 boot/gummiboot/gummiboot.mk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/boot/gummiboot/gummiboot.mk b/boot/gummiboot/gummiboot.mk
index 0e67e5e8ce0e..748e87030e7b 100644
--- a/boot/gummiboot/gummiboot.mk
+++ b/boot/gummiboot/gummiboot.mk
@@ -6,7 +6,7 @@
 
 GUMMIBOOT_SITE = https://gitlab.freedesktop.org/archived-projects/gummiboot.git
 GUMMIBOOT_SITE_METHOD = git
-GUMMIBOOT_VERSION = 43
+GUMMIBOOT_VERSION = 2bcd919c681c952eb867ef1bdb458f1bc49c2d55
 GUMMIBOOT_LICENSE = LGPL-2.1+
 GUMMIBOOT_LICENSE_FILES = LICENSE
 
-- 
2.21.0

[Buildroot] [PATCH 3/3] gummiboot: fix compatibility with newer glibc versions

[Esben Haabendal]

From: Esben Haabendal <esben@haabendal.dk>

Signed-off-by: Esben Haabendal <esben@haabendal.dk>
---
 boot/gummiboot/0001-glibc-2.25-fix.patch | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
 create mode 100644 boot/gummiboot/0001-glibc-2.25-fix.patch

diff --git a/boot/gummiboot/0001-glibc-2.25-fix.patch b/boot/gummiboot/0001-glibc-2.25-fix.patch
new file mode 100644
index 000000000000..377098f9ea60
--- /dev/null
+++ b/boot/gummiboot/0001-glibc-2.25-fix.patch
@@ -0,0 +1,16 @@
+Allow building with glibc with sys/sysmacros.h removed from sys/types.h.
+
+See https://sourceware.org/ml/libc-alpha/2015-11/msg00253.html for why
+this was done.
+
+diff -urN gummiboot-43~orig/src/setup/setup.c gummiboot-43/src/setup/setup.c
+--- gummiboot-43~orig/src/setup/setup.c	2019-03-14 22:28:34.533206018 +0100
++++ gummiboot-43/src/setup/setup.c	2019-03-15 06:40:07.895651992 +0100
+@@ -37,6 +37,7 @@
+ #include <ftw.h>
+ #include <stdbool.h>
+ #include <blkid.h>
++#include <sys/sysmacros.h>
+ 
+ #include "efivars.h"
+ 
-- 
2.21.0

[Buildroot] [PATCH 3/3] gummiboot: fix compatibility with newer glibc versions

[Thomas Petazzoni]

Hello Esben,

On Tue, 19 Mar 2019 12:28:52 +0100
Esben Haabendal <esben.haabendal@gmail.com> wrote:

> From: Esben Haabendal <esben@haabendal.dk>
> 
> Signed-off-by: Esben Haabendal <esben@haabendal.dk>

Since gummiboot is maintained in a Git repository upstream, we really
like to have patches generated using "git format-patch", as it makes
sense easier to handle. Also, we require all patches to have a
Signed-off-by.

Thanks!

Thomas
-- 
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Peter Korsgaard]

>>>>> "Esben" == Esben Haabendal <esben.haabendal@gmail.com> writes:

 > From: Esben Haabendal <esben@haabendal.dk>
 > These patches fixes gummiboot for building with toolchains using
 > newer gcc and glibc versions.

 > Esben Haabendal (3):
 >   gummiboot: use new official upstream git repository
 >   gummiboot: upgrade to last commit before being removed
 >   gummiboot: fix compatibility with newer glibc versions

Out of interest, what is the use case for using gummiboot? We recently
added systemd-boot support, so I was considering removing gummiboot as
it is dead upstream.

Notice that you currently cannot enable systemd-boot without the rest of
systemd, so that may be a disadvantage.

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Peter Korsgaard]

>>>>> "Esben" == Esben Haabendal <esbenhaabendal@gmail.com> writes:

 > On Tue, Mar 19, 2019 at 7:56 PM Peter Korsgaard <peter@korsgaard.com> wrote:
 >> Out of interest, what is the use case for using gummiboot? We recently
 >> added systemd-boot support, so I was considering removing gummiboot as
 >> it is dead upstream.
 >> 
 >> Notice that you currently cannot enable systemd-boot without the rest of
 >> systemd, so that may be a disadvantage.
 >> 

 > In this case, I am upgrading a system which is currently using gummiboot.
 > Switching to systemd-boot does not seem to bring anything meaningful
 > improvements,
 > as gummiboot does the job of booting the system.

Ok. Out of interest, what is gummiboot adding over just directly booting
a Linux kernel with CONFIG_EFI_STUB?

 > And yes, this is not a systemd system, and transforming the entire
 > system to systemd is clearly a big thing compared to just fixing
 > building of gummiboot.

 > I would prefer if we could keep gummiboot support in buildroot for now.

Fine by me. Will you add yourself to DEVELOPERS for gummiboot?

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Peter Korsgaard]

>>>>> "Esben" == Esben Haabendal <esbenhaabendal@gmail.com> writes:

Hi,

 >> Ok. Out of interest, what is gummiboot adding over just directly booting
 >> a Linux kernel with CONFIG_EFI_STUB?
 >> 

 > With gummiboot, I can point to a separate initramfs file. I haven't figured
 > out
 > how to do that with CONFIG_EFI_STUB, as the CONFIG_CMDLINE args
 > are not processed by the EFI stub.

I have never tried, but it looks as if you can just set
initrd=<path/to/initramfs/in/esp>

https://www.kernel.org/doc/Documentation/efi-stub.txt

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Esben Haabendal]

On Wed, Mar 20, 2019 at 3:36 PM Peter Korsgaard <peter@korsgaard.com> wrote:

> >>>>> "Esben" == Esben Haabendal <esbenhaabendal@gmail.com> writes:
>
>   >> Ok. Out of interest, what is gummiboot adding over just directly
> booting
>  >> a Linux kernel with CONFIG_EFI_STUB?
>
>  > With gummiboot, I can point to a separate initramfs file. I haven't
> figured
>  > out
>  > how to do that with CONFIG_EFI_STUB, as the CONFIG_CMDLINE args
>  > are not processed by the EFI stub.
>
> I have never tried, but it looks as if you can just set
> initrd=<path/to/initramfs/in/esp>
>
> https://www.kernel.org/doc/Documentation/efi-stub.txt
>

Yes, you can use initrd=... if you provide the cmdline to the EFI stub.
But if you compile in the cmdline using CONFIG_CMDLINE, it will
not work, as it will only be available for Linux kernel, and not for
EFI stub.

/Esben
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20190320/64a132a5/attachment.html>

[Buildroot] [PATCH 0/3] Gummiboot fixup

[Peter Korsgaard]

>>>>> "Esben" == Esben Haabendal <esbenhaabendal@gmail.com> writes:

 > Yes, you can use initrd=... if you provide the cmdline to the EFI stub.
 > But if you compile in the cmdline using CONFIG_CMDLINE, it will
 > not work, as it will only be available for Linux kernel, and not for
 > EFI stub.

I guess Even with CONFIG_CMDLINE, you could configure the boot entry to
pass initrd= to EFI, but OK.

-- 
Bye, Peter Korsgaard