Skip to contentSkip to conversation

linux-next: build failure after merge of the origin tree

PATCH Stephen Rothwell 3 participants 4 messages Sep 30, 2024 Linux-next tree

lore
Share & tools
lore
#0 lore raw

Hi all,

After merging the origin tree, today's linux-next build (s390 defconfig) failed like this:

In file included from fs/bcachefs/bset.h:9,
from fs/bcachefs/btree_iter.h:5,
from fs/bcachefs/str_hash.h:5,
from fs/bcachefs/xattr.h:5,
from fs/bcachefs/acl.c:6:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

  557 |         x(BKEY_FIELD_VERSION_HI,        bversion.hi)                    \
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~
fs/bcachefs/bkey.h:558:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
  558 |         x(BKEY_FIELD_VERSION_LO,        bversion.lo)
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~
In file included from fs/bcachefs/bset.h:10:
fs/bcachefs/bkey_methods.h: In function 'bch2_bkey_maybe_mergable':
fs/bcachefs/bkey_methods.h:73:34: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                  ^~~~~~~~
      |                                  version
fs/bcachefs/bkey_methods.h:73:47: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                               ^~~~~~~~
      |                                               version

In file included from fs/bcachefs/extents.h:6,
from fs/bcachefs/buckets.h:12,
from fs/bcachefs/alloc_background.h:7,
from fs/bcachefs/alloc_foreground.c:15:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

  557 |         x(BKEY_FIELD_VERSION_HI,        bversion.hi)                    \
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~
fs/bcachefs/bkey.h:558:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
  558 |         x(BKEY_FIELD_VERSION_LO,        bversion.lo)
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~

In file included from fs/bcachefs/btree_cache.h:7,
from fs/bcachefs/backpointers.h:5,
from fs/bcachefs/alloc_foreground.c:17:
fs/bcachefs/bkey_methods.h: In function 'bch2_bkey_maybe_mergable':
fs/bcachefs/bkey_methods.h:73:34: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                  ^~~~~~~~
      |                                  version
fs/bcachefs/bkey_methods.h:73:47: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                               ^~~~~~~~
      |                                               version

In file included from fs/bcachefs/extents.h:6,
from fs/bcachefs/buckets.h:12,
from fs/bcachefs/alloc_background.h:7,
from fs/bcachefs/alloc_background.c:3:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

  557 |         x(BKEY_FIELD_VERSION_HI,        bversion.hi)                    \
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~
fs/bcachefs/bkey.h:558:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
  558 |         x(BKEY_FIELD_VERSION_LO,        bversion.lo)
      |                                         ^~~~~~~~

fs/bcachefs/bkey.h:578:50: note: in definition of macro 'x'
578 | #define x(id, field) __bkey_format_add(s, id, k->field);
| ^~~~~
fs/bcachefs/bkey.h:579:9: note: in expansion of macro 'bkey_fields'

  579 |         bkey_fields()
      |         ^~~~~~~~~~~

In file included from fs/bcachefs/btree_cache.h:7,
from fs/bcachefs/backpointers.h:5,
from fs/bcachefs/alloc_background.c:5:
fs/bcachefs/bkey_methods.h: In function 'bch2_bkey_maybe_mergable':
fs/bcachefs/bkey_methods.h:73:34: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                  ^~~~~~~~
      |                                  version
fs/bcachefs/bkey_methods.h:73:47: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   73 |                 !bversion_cmp(l->bversion, r->bversion) &&
      |                                               ^~~~~~~~
      |                                               version

In file included from fs/bcachefs/btree_write_buffer.h:6,
from fs/bcachefs/alloc_background.c:13:
fs/bcachefs/disk_accounting.h: In function 'bch2_accounting_accumulate':
fs/bcachefs/disk_accounting.h:39:33: error: 'struct bkey' has no member named 'bversion'; did you mean 'version'?

   39 |         if (bversion_cmp(dst->k.bversion, src.k->bversion) < 0)
      |                                 ^~~~~~~~
      |                                 version
fs/bcachefs/disk_accounting.h:39:50: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   39 |         if (bversion_cmp(dst->k.bversion, src.k->bversion) < 0)
      |                                                  ^~~~~~~~
      |                                                  version
fs/bcachefs/disk_accounting.h:40:24: error: 'struct bkey' has no member named 'bversion'; did you mean 'version'?
   40 |                 dst->k.bversion = src.k->bversion;
      |                        ^~~~~~~~
      |                        version
fs/bcachefs/disk_accounting.h:40:42: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?
   40 |                 dst->k.bversion = src.k->bversion;
      |                                          ^~~~~~~~
      |                                          version

Caused by commit

cf49f8a8c277 ("bcachefs: rename version -> bversion")

I have applied the following fix patch for today.

From: Stephen Rothwell <[email protected]>
Date: Mon, 30 Sep 2024 13:31:16 +1000
Subject: [PATCH] fix up for "bcachefs: rename version -> bversion"

---
 fs/bcachefs/bcachefs_format.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/bcachefs/bcachefs_format.h +1 -1
index 203ee627cab5..84832c2d4df9 100644
@@ -223,7 +223,7 @@ struct bkey {
 #elif __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__
 	struct bpos	p;
 	__u32		size;		/* extent size, in sectors */
-	struct bversion	version;
+	struct bversion	bversion;
 
 	__u8		pad[1];
 #endif
Signature · 5 lines
2.45.2

-- 
Cheers,
Stephen Rothwell
#1 lore raw
Quoted 12 lines · Stephen Rothwell

Hi all,

After merging the origin tree, today's linux-next build (s390 defconfig) failed like this:

In file included from fs/bcachefs/bset.h:9,
from fs/bcachefs/btree_iter.h:5,
from fs/bcachefs/str_hash.h:5,
from fs/bcachefs/xattr.h:5,
from fs/bcachefs/acl.c:6:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

Quoted 2 lines
  557 |         x(BKEY_FIELD_VERSION_HI,        bversion.hi)                    \
      |                                         ^~~~~~~~

Also reported earlier here:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/

But the true problem is here:

commit cf49f8a8c277f9f2b78e2a56189a741a508a9820

Author:     Kent Overstreet <[email protected]>
AuthorDate: Thu Sep 26 15:49:17 2024 -0400
                    ^^^^^^^^^^^
Commit:     Kent Overstreet <[email protected]>
CommitDate: Fri Sep 27 21:46:35 2024 -0400
                    ^^^^^^^^^^^ one day difference!

Last minute commits usually won't receive wide build coverage, at least not instantaneously.

And if you go through the history, I see around 40 commits with authored date ~20-26 September and committed on Sep 27. Plus another ~40 authored earlier but committed on September 21, which is middle of merge window.

Why such commits for RC1 are sent at the end of merge window or committed during merge window?

Best regards,
Krzysztof

Best regards,
Krzysztof

#2 lore raw
Quoted 34 lines · Krzysztof Kozlowski
Quoted 12 lines · Stephen Rothwell

Hi all,

After merging the origin tree, today's linux-next build (s390 defconfig) failed like this:

In file included from fs/bcachefs/bset.h:9,
from fs/bcachefs/btree_iter.h:5,
from fs/bcachefs/str_hash.h:5,
from fs/bcachefs/xattr.h:5,
from fs/bcachefs/acl.c:6:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'?

Quoted 2 lines
  557 |         x(BKEY_FIELD_VERSION_HI,        bversion.hi)                    \
      |                                         ^~~~~~~~

Also reported earlier here:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/

But the true problem is here:

commit cf49f8a8c277f9f2b78e2a56189a741a508a9820

Author:     Kent Overstreet <[email protected]>
AuthorDate: Thu Sep 26 15:49:17 2024 -0400
                    ^^^^^^^^^^^
Commit:     Kent Overstreet <[email protected]>
CommitDate: Fri Sep 27 21:46:35 2024 -0400
                    ^^^^^^^^^^^ one day difference!

Last minute commits usually won't receive wide build coverage, at least not instantaneously.

Quoted 6 lines

And if you go through the history, I see around 40 commits with authored date ~20-26 September and committed on Sep 27. Plus another ~40 authored earlier but committed on September 21, which is middle of merge window.

Why such commits for RC1 are sent at the end of merge window or committed during merge window?

the rename was something I did to track down a bug in the disk accounting rewrite:
https://lore.kernel.org/linux-bcachefs/pvga5sgp4vejnnr5ojgiuwte6qeve4x7ld4dhdmzb625l367fq@q4td2cutlfvu/T/

I do need to get multi-arch build testing going on my CI, but right now I'm busy working on corner cases in the repair code...

#3 lore raw
Quoted 37 lines · Kent Overstreet
Quoted 36 lines · Krzysztof Kozlowski
Quoted 5 lines · Stephen Rothwell

Hi all,

After merging the origin tree, today's linux-next build (s390 defconfig) failed like this:

Quoted 8 lines

In file included from fs/bcachefs/bset.h:9, from fs/bcachefs/btree_iter.h:5, from fs/bcachefs/str_hash.h:5, from fs/bcachefs/xattr.h:5, from fs/bcachefs/acl.c:6:
fs/bcachefs/bkey.h: In function 'bch2_bkey_format_add_key':
fs/bcachefs/bkey.h:557:41: error: 'const struct bkey' has no member named 'bversion'; did you mean 'version'? 557 |
x(BKEY_FIELD_VERSION_HI, bversion.hi)
\ | ^~~~~~~~

Also reported earlier here:
https://lore.kernel.org/all/[email protected]/
https://lore.kernel.org/all/[email protected]/

But the true problem is here:

commit cf49f8a8c277f9f2b78e2a56189a741a508a9820 Author: Kent
Overstreet <[email protected]> AuthorDate: Thu Sep 26
15:49:17 2024 -0400 ^^^^^^^^^^^ Commit: Kent Overstreet
<[email protected]> CommitDate: Fri Sep 27 21:46:35 2024 -0400 ^^^^^^^^^^^ one day difference!

Last minute commits usually won't receive wide build coverage, at least not instantaneously.

And if you go through the history, I see around 40 commits with authored date ~20-26 September and committed on Sep 27. Plus another ~40 authored earlier but committed on September 21, which is middle of merge window.

Quoted 6 lines
Quoted 2 lines

Why such commits for RC1 are sent at the end of merge window or committed during merge window?

the rename was something I did to track down a bug in the disk accounting rewrite:
https://lore.kernel.org/linux-bcachefs/pvga5sgp4vejnnr5ojgiuwte6qeve4x7ld4dhdmzb625l367fq@q4td2cutlfvu/T/

I am sorry, but I do not see how cleanup-like rename is related to tracking down a tricky bug and fixing it. Looks like fixes got inter-mixed with such minor refactoring.

Quoted 3 lines

I do need to get multi-arch build testing going on my CI, but right
now I'm busy working on corner cases in the repair code...

Well, I claim it is easy, if you wait one day for the notification from LKP - I got notification after every push, also on success, so I am sure that my kernel (and development, but that's unrelated) builds fine on wide configuration. And this is free as in free beer...

Best regards,
Krzysztof