[Buildroot] [PATCH 2024.02.x] package/systemd: fix bcachefs magic patch

[Buildroot] [PATCH 2024.02.x] package/systemd: fix bcachefs magic patch

[Thomas Petazzoni via buildroot]

Commit 29409c9d345a2536f6b72c605cda0d13f306200f ("package/systemd: fix
build with recent kernels") added a patch from upstream that adds the
BCACHEFS_SUPER_MAGIC definition if not provided by the kernel headers.

However, this commit from upstream was bogus: in the case
BCACHEFS_SUPER_MAGIC was defined, it does an assert_cc() to verify
that the value matches what we expect, but this assert_cc() statement
lacks the final semi-colon, causing build breakage on host systems
that do have BCACHEFS_SUPER_MAGIC defined in their kernel headers,
with a weird error occuring in another header file... because only the
next statement after assert_cc() was considered as having a syntax
error.

Fixes:

  http://autobuild.buildroot.net/results/9f99a97df2efd5e4ae4ad5cc5882607f0c8766bc/

This issue only exists in 2024.02.x at this point, because in master
we have moved to systemd 256.x, which has dropped entirely this piece
of code related to defining BCACHEFS_SUPER_MAGIC (and we wonder how
systemd can then build with slightly older kernel headers, as
BCACHEFS_SUPER_MAGIC is only in the public kernel headers since Linux
6.10).

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
---
 package/systemd/0002-basic-add-bcachefs-magic.patch | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/package/systemd/0002-basic-add-bcachefs-magic.patch b/package/systemd/0002-basic-add-bcachefs-magic.patch
index e7f4b1cbb7..372066ebef 100644
--- a/package/systemd/0002-basic-add-bcachefs-magic.patch
+++ b/package/systemd/0002-basic-add-bcachefs-magic.patch
@@ -1,4 +1,4 @@
-From e5ff9b585d7dc54f50085d518d7d588c250b0690 Mon Sep 17 00:00:00 2001
+From 48c1cd62de0339722fd3b72ef86aaaf1c38ff3dd Mon Sep 17 00:00:00 2001
 From: Chris Packham <chris.packham@alliedtelesis.co.nz>
 Date: Fri, 31 May 2024 09:51:38 +1200
 Subject: [PATCH] basic: Add BCACHEFS magic
@@ -34,7 +34,7 @@ index 1cd66b5a5f..c82fe98227 100644
  binder,          {BINDERFS_SUPER_MAGIC}
  binfmt_misc,     {BINFMTFS_MAGIC}
 diff --git a/src/basic/missing_magic.h b/src/basic/missing_magic.h
-index 82d71c8ad1..b3d2ee9f57 100644
+index 82d71c8ad1..c8ecf44357 100644
 --- a/src/basic/missing_magic.h
 +++ b/src/basic/missing_magic.h
 @@ -197,3 +197,10 @@
@@ -46,8 +46,8 @@ index 82d71c8ad1..b3d2ee9f57 100644
 +#ifndef BCACHEFS_SUPER_MAGIC
 +#  define BCACHEFS_SUPER_MAGIC 0xca451a4e
 +#else
-+assert_cc(BCACHEFS_SUPER_MAGIC == 0xca451a4e)
++assert_cc(BCACHEFS_SUPER_MAGIC == 0xca451a4e);
 +#endif
---
-2.45.1
+-- 
+2.47.0
 
-- 
2.47.0

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 2024.02.x] package/systemd: fix bcachefs magic patch

[Peter Korsgaard]

>>>>> "Thomas" == Thomas Petazzoni via buildroot <buildroot@buildroot.org> writes:

 > Commit 29409c9d345a2536f6b72c605cda0d13f306200f ("package/systemd: fix
 > build with recent kernels") added a patch from upstream that adds the
 > BCACHEFS_SUPER_MAGIC definition if not provided by the kernel headers.

 > However, this commit from upstream was bogus: in the case
 > BCACHEFS_SUPER_MAGIC was defined, it does an assert_cc() to verify
 > that the value matches what we expect, but this assert_cc() statement
 > lacks the final semi-colon, causing build breakage on host systems
 > that do have BCACHEFS_SUPER_MAGIC defined in their kernel headers,
 > with a weird error occuring in another header file... because only the
 > next statement after assert_cc() was considered as having a syntax
 > error.

 > Fixes:

 >   http://autobuild.buildroot.net/results/9f99a97df2efd5e4ae4ad5cc5882607f0c8766bc/

 > This issue only exists in 2024.02.x at this point, because in master
 > we have moved to systemd 256.x, which has dropped entirely this piece
 > of code related to defining BCACHEFS_SUPER_MAGIC (and we wonder how
 > systemd can then build with slightly older kernel headers, as
 > BCACHEFS_SUPER_MAGIC is only in the public kernel headers since Linux
 > 6.10).

Hmm, odd indeed..

 > Signed-off-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>

Committed to 2024.02.x, thanks.

I see there are newer 254.x bugfix releases since v254.13 (latest is
v254.21). Maybe it makes sense to update the version we use? I don't see
anything directly relevant for this though.

-- 
Bye, Peter Korsgaard
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 2024.02.x] package/systemd: fix bcachefs magic patch

[Thomas Petazzoni via buildroot]

On Sun, 01 Dec 2024 20:41:10 +0100
Peter Korsgaard <peter@korsgaard.com> wrote:

>  > This issue only exists in 2024.02.x at this point, because in master
>  > we have moved to systemd 256.x, which has dropped entirely this piece
>  > of code related to defining BCACHEFS_SUPER_MAGIC (and we wonder how
>  > systemd can then build with slightly older kernel headers, as
>  > BCACHEFS_SUPER_MAGIC is only in the public kernel headers since Linux
>  > 6.10).  
> 
> Hmm, odd indeed..

Yes, I didn't have the time to do an actual test... but I assume this
would have caused some breakage in the autobuilders, so I'm really not
sure to understand why it works.

> I see there are newer 254.x bugfix releases since v254.13 (latest is
> v254.21). Maybe it makes sense to update the version we use? I don't see
> anything directly relevant for this though.

No, the issue is not fixed upstream. I didn't test 254.21 because it
wasn't there, but I tested 254.20 and it definitely was still broken.

Best regards,

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot