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

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

[Stephen Rothwell]

Hi Jens,

After merging the block tree, today's linux-next build (x86_64 allmodconfig)
failed like this:

drivers/staging/lustre/lustre/llite/../include/obd_support.h:42:0,
                 from drivers/staging/lustre/lustre/llite/../include/lustre_cfg.h:225,
                 from drivers/staging/lustre/lustre/llite/../include/lustre_lib.h:55,
                 from drivers/staging/lustre/lustre/llite/lloop.c:105:
drivers/staging/lustre/lustre/llite/lloop.c: In function 'loop_handle_bio':
drivers/staging/lustre/lustre/llite/../include/linux/lustre_compat25.h:89:35: error: too many arguments to function 'bio_endio'
 #define cfs_bio_endio(a, b, c)    bio_endio((a), (c))
                                   ^
drivers/staging/lustre/lustre/llite/lloop.c:379:3: note: in expansion of macro 'cfs_bio_endio'
   cfs_bio_endio(bio, bio->bi_iter.bi_size, ret);
   ^
In file included from include/linux/blkdev.h:18:0,
                 from drivers/staging/lustre/lustre/llite/lloop.c:91:
include/linux/bio.h:442:13: note: declared here
 extern void bio_endio(struct bio *);
             ^

Caused by commit

  4246a0b63bd8 ("block: add a bi_error field to struct bio")

I added the following fix patch, but more is probably required:

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Thu, 30 Jul 2015 14:10:57 +1000
Subject: [PATCH] block: lustre: temporary fix for bio_endio API change

Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 drivers/staging/lustre/lustre/include/linux/lustre_compat25.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
index 513c81f43d6e..1aac5a77c500 100644
--- a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
+++ b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
@@ -86,7 +86,7 @@
 		vfs_rename(old, old_dir, new, new_dir, NULL, 0)
 
 #define cfs_bio_io_error(a, b)   bio_io_error((a))
-#define cfs_bio_endio(a, b, c)    bio_endio((a), (c))
+#define cfs_bio_endio(a, b, c)    bio_endio((a))
 
 #define cfs_path_put(nd)     path_put(&(nd)->path)
 
-- 
2.4.6

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

Re: linux-next: build failure after merge of the block tree

[Christoph Hellwig]

On Thu, Jul 30, 2015 at 02:17:13PM +1000, Stephen Rothwell wrote:
> Hi Jens,
> 
> After merging the block tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:

Can you please drop staging and especially lustre from these runs?
Conditions of the staging tree are they don't need to update.  Nevermind that
it never should have grown something as big and stale as lustre without
any prospects of being merged.

Re: linux-next: build failure after merge of the block tree

[Stephen Rothwell]

Hi Christoph,

On Thu, 30 Jul 2015 08:19:51 +0200 Christoph Hellwig <hch@lst.de> wrote:
>
> Can you please drop staging and especially lustre from these runs?

One problem with that is that staging is expected to at least build.
and when Linus eventually merges this code in the block tree, he will
do an allmodconfig build and it will fail.  I don't really care about
how things get fixed in the staging tree (as you can see from the patch
I applied to get it to build) and sometimes mostly I just disable an
affected driver.  This was just as easy.

At least this way people know ahead of time that it needs an update.

> Conditions of the staging tree are they don't need to update.

See above - it is expected to build.  Linus himself asked me to make
sure that remains true.

> Nevermind that
> it never should have grown something as big and stale as lustre without
> any prospects of being merged.

You can have that discussion with Greg and the Lustre maintainers. :-)

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au

Re: linux-next: build failure after merge of the block tree

[Drokin, Oleg]

On Jul 30, 2015, at 12:17 AM, Stephen Rothwell wrote:

> Hi Jens,
> 
> After merging the block tree, today's linux-next build (x86_64 allmodconfig)
> failed like this:
> 
> drivers/staging/lustre/lustre/llite/../include/obd_support.h:42:0,
>                 from drivers/staging/lustre/lustre/llite/../include/lustre_cfg.h:225,
>                 from drivers/staging/lustre/lustre/llite/../include/lustre_lib.h:55,
>                 from drivers/staging/lustre/lustre/llite/lloop.c:105:
> drivers/staging/lustre/lustre/llite/lloop.c: In function 'loop_handle_bio':
> drivers/staging/lustre/lustre/llite/../include/linux/lustre_compat25.h:89:35: error: too many arguments to function 'bio_endio'
> #define cfs_bio_endio(a, b, c)    bio_endio((a), (c))
>                                   ^
> drivers/staging/lustre/lustre/llite/lloop.c:379:3: note: in expansion of macro 'cfs_bio_endio'
>   cfs_bio_endio(bio, bio->bi_iter.bi_size, ret);
>   ^
> In file included from include/linux/blkdev.h:18:0,
>                 from drivers/staging/lustre/lustre/llite/lloop.c:91:
> include/linux/bio.h:442:13: note: declared here
> extern void bio_endio(struct bio *);
>             ^
> 
> Caused by commit
> 
>  4246a0b63bd8 ("block: add a bi_error field to struct bio")
> 
> I added the following fix patch, but more is probably required:

It looks like this in itself is enough, but we probably should be good to
kill those cfs_bio_* compat macros while we are at it.
I'll submit a patch, but not sure where to base it on.
If basing on current block tree's linux-next, then I get a bunch of errors about how
kstrto* are undefined, so I cannot really test this yet.

Thanks.

> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Thu, 30 Jul 2015 14:10:57 +1000
> Subject: [PATCH] block: lustre: temporary fix for bio_endio API change
> 
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
> ---
> drivers/staging/lustre/lustre/include/linux/lustre_compat25.h | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
> index 513c81f43d6e..1aac5a77c500 100644
> --- a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
> +++ b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
> @@ -86,7 +86,7 @@
> 		vfs_rename(old, old_dir, new, new_dir, NULL, 0)
> 
> #define cfs_bio_io_error(a, b)   bio_io_error((a))
> -#define cfs_bio_endio(a, b, c)    bio_endio((a), (c))
> +#define cfs_bio_endio(a, b, c)    bio_endio((a))
> 
> #define cfs_path_put(nd)     path_put(&(nd)->path)
> 
> -- 
> 2.4.6
> 
> -- 
> Cheers,
> Stephen Rothwell                    sfr@canb.auug.org.au
> http://www.canb.auug.org.au/~sfr/

[PATCH] staging/lustre/lloop: Update call to bio_endio() to match new prototype

[green]

From: Oleg Drokin <green@linuxhacker.ru>

Also while we are at it, drop all cfs_bio_* wrappers.

Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
---
 drivers/staging/lustre/lustre/include/linux/lustre_compat25.h | 3 ---
 drivers/staging/lustre/lustre/llite/lloop.c                   | 4 ++--
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
index 513c81f..c2a4ffd 100644
--- a/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
+++ b/drivers/staging/lustre/lustre/include/linux/lustre_compat25.h
@@ -85,9 +85,6 @@
 #define ll_vfs_rename(old, old_dir, mnt, new, new_dir, mnt1) \
 		vfs_rename(old, old_dir, new, new_dir, NULL, 0)
 
-#define cfs_bio_io_error(a, b)   bio_io_error((a))
-#define cfs_bio_endio(a, b, c)    bio_endio((a), (c))
-
 #define cfs_path_put(nd)     path_put(&(nd)->path)
 
 
diff --git a/drivers/staging/lustre/lustre/llite/lloop.c b/drivers/staging/lustre/lustre/llite/lloop.c
index cc00fd1..d614234 100644
--- a/drivers/staging/lustre/lustre/llite/lloop.c
+++ b/drivers/staging/lustre/lustre/llite/lloop.c
@@ -365,7 +365,7 @@ static void loop_make_request(struct request_queue *q, struct bio *old_bio)
 	loop_add_bio(lo, old_bio);
 	return;
 err:
-	cfs_bio_io_error(old_bio, old_bio->bi_iter.bi_size);
+	bio_io_error(old_bio);
 }
 
 
@@ -376,7 +376,7 @@ static inline void loop_handle_bio(struct lloop_device *lo, struct bio *bio)
 	while (bio) {
 		struct bio *tmp = bio->bi_next;
 		bio->bi_next = NULL;
-		cfs_bio_endio(bio, bio->bi_iter.bi_size, ret);
+		bio_endio(bio);
 		bio = tmp;
 	}
 }
-- 
2.1.0

Re: [PATCH] staging/lustre/lloop: Update call to bio_endio() to match new prototype

[Sudip Mukherjee]

On Thu, Jul 30, 2015 at 11:03:35AM -0400, green@linuxhacker.ru wrote:
> From: Oleg Drokin <green@linuxhacker.ru>
> 
> Also while we are at it, drop all cfs_bio_* wrappers.
> 
> Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
> ---
Is it dependent on some other patch? This applies cleanly but is
breaking the build. "error: too few arguments to function bio_endio"

regards
sudip

Re: [PATCH] staging/lustre/lloop: Update call to bio_endio() to match new prototype

[Oleg Drokin]

On Jul 31, 2015, at 7:27 AM, Sudip Mukherjee wrote:

> On Thu, Jul 30, 2015 at 11:03:35AM -0400, green@linuxhacker.ru wrote:
>> From: Oleg Drokin <green@linuxhacker.ru>
>> 
>> Also while we are at it, drop all cfs_bio_* wrappers.
>> 
>> Signed-off-by: Oleg Drokin <oleg.drokin@intel.com>
>> ---
> Is it dependent on some other patch? This applies cleanly but is
> breaking the build. "error: too few arguments to function bio_endio"

Yes, this is for block/linux-next specifically.
For current staging tree a corresponding one is 
http://www.gossamer-threads.com/lists/linux/kernel/2227687