Re: [PATCH] fs: efs: Fix compilation bug and use pr_debug()

[Maxwell Doose]

On Thu, Jun 4, 2026 at 4:30 PM Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
>
> (cc linux-fsdevel)
>

Whoops, I guess I'm too reliant on --cc-cmd :(

> On Thu,  4 Jun 2026 15:24:40 -0500 Maxwell Doose <m32285159@xxxxxxxxx> wrote:
>
> > Firstly, the current code uses formatters that are incompatible with the
> > most recent GCC for x86_64. Replace them and explicitly cast the
> > formatted variables to their respective forms of long long.
>
> Please quote the compiler output in the changelog.
>

It's rather lengthy:
fs/efs/file.c: In function ‘efs_get_block’:
fs/efs/file.c:26:35: error: ‘block’ undeclared (first use in this
function); did you mean ‘iblock’?
  26 |                         __func__, block, inode->i_blocks, inode->i_size);
     |                                   ^~~~~
./include/linux/printk.h:483:33: note: in definition of macro
‘printk_index_wrap’
 483 |                 _p_func(_fmt, ##__VA_ARGS__);                           \
     |                                 ^~~~~~~~~~~
./include/linux/printk.h:564:9: note: in expansion of macro ‘printk’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |         ^~~~~~
fs/efs/file.c:25:17: note: in expansion of macro ‘pr_warn’
  25 |                 pr_warn("%s(): block %d >= %ld (filesize %ld)\n",
     |                 ^~~~~~~
fs/efs/file.c:26:35: note: each undeclared identifier is reported only
once for each function it appears i
n
./include/linux/printk.h:483:33: note: in definition of macro
‘printk_index_wrap’
 483 |                 _p_func(_fmt, ##__VA_ARGS__);                           \
     |                                 ^~~~~~~~~~~
./include/linux/printk.h:564:9: note: in expansion of macro ‘printk’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |         ^~~~~~
fs/efs/file.c:25:17: note: in expansion of macro ‘pr_warn’
  25 |                 pr_warn("%s(): block %d >= %ld (filesize %ld)\n",
     |                 ^~~~~~~
fs/efs/file.c: In function ‘efs_bmap’:
./include/linux/kern_levels.h:5:25: error: format ‘%ld’ expects
argument of type ‘long int’, but argument
4 has type ‘blkcnt_t’ {aka ‘long long unsigned int’} [-Werror=format=]
   5 | #define KERN_SOH        "\001"          /* ASCII Start Of Header */
     |                         ^~~~~~
./include/linux/printk.h:483:25: note: in definition of macro
‘printk_index_wrap’
 483 |                 _p_func(_fmt, ##__VA_ARGS__);                           \
     |                         ^~~~
./include/linux/printk.h:564:9: note: in expansion of macro ‘printk’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |         ^~~~~~
./include/linux/kern_levels.h:12:25: note: in expansion of macro ‘KERN_SOH’
  12 | #define KERN_WARNING    KERN_SOH "4"    /* warning conditions */
     |                         ^~~~~~~~
./include/linux/printk.h:564:16: note: in expansion of macro ‘KERN_WARNING’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |                ^~~~~~~~~~~~
fs/efs/file.c:47:17: note: in expansion of macro ‘pr_warn’
  47 |                 pr_warn("%s(): block %d >= %ld (filesize %ld)\n",
     |                 ^~~~~~~
./include/linux/kern_levels.h:5:25: error: format ‘%ld’ expects
argument of type ‘long int’, but argument
5 has type ‘loff_t’ {aka ‘long long int’} [-Werror=format=]
   5 | #define KERN_SOH        "\001"          /* ASCII Start Of Header */
     |                         ^~~~~~
./include/linux/printk.h:483:25: note: in definition of macro
‘printk_index_wrap’
 483 |                 _p_func(_fmt, ##__VA_ARGS__);                           \
     |                         ^~~~
./include/linux/printk.h:564:9: note: in expansion of macro ‘printk’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |         ^~~~~~
./include/linux/kern_levels.h:12:25: note: in expansion of macro ‘KERN_SOH’
  12 | #define KERN_WARNING    KERN_SOH "4"    /* warning conditions */
     |                         ^~~~~~~~
./include/linux/printk.h:564:16: note: in expansion of macro ‘KERN_WARNING’
 564 |         printk(KERN_WARNING pr_fmt(fmt), ##__VA_ARGS__)
     |                ^~~~~~~~~~~~
fs/efs/file.c:47:17: note: in expansion of macro ‘pr_warn’
  47 |                 pr_warn("%s(): block %d >= %ld (filesize %ld)\n",
     |                 ^~~~~~~

(excuse the misalignment of some of this stuff, I'm using gmail to send this.)

> > Secondly, replace the legacy conditional compilation for DEBUG with the
> > more preferred pr_debug() to condense the code, while keeping the final
> > output functionally identical.
> >
> > Fixes: f403d1dbac6d ("fs/efs: add pr_fmt / use __func__")
> >
> > ...
> >
>
> >  Test compiled only with gcc 16.1.1 for x86_64.
>
> Compiled how?  DEBUG was defined in some fashion?
>

I cheated a little and removed the #ifdef DEBUG to compile-test it.

> > --- a/fs/efs/file.c
> > +++ b/fs/efs/file.c
> > @@ -19,13 +19,12 @@ int efs_get_block(struct inode *inode, sector_t iblock,
> >       if (create)
> >               return error;
> >       if (iblock >= inode->i_blocks) {
> > -#ifdef DEBUG
> >               /*
> >                * i have no idea why this happens as often as it does
> >                */
> > -             pr_warn("%s(): block %d >= %ld (filesize %ld)\n",
> > -                     __func__, block, inode->i_blocks, inode->i_size);
>
> That wouldn't have compiled anyway - `block' is undefined?
>

Correct, so we're fixing two bugs in one patch (this one is so trivial
to fix that I just included it).

> > -#endif
> > +             pr_debug("EFS: block %llu >= %llu (filesize %lld)\n",
> > +                      (unsigned long long)iblock, (unsigned long long)inode->i_blocks,
> > +                      (long long)inode->i_size);
>
> It seems that nowadays sector_t and blkcnt_t are unconditionally u64
> and loff_t is unconditionally `long long'.  I think, please check.
>
> If so, no casts are needed.
>

Looks like I got a bit sloppy:

fs/efs/file.c: In function ‘efs_bmap’:
./include/linux/kern_levels.h:5:25: error: format ‘%llu’ expects
argument of type ‘long long unsigned int’
, but argument 3 has type ‘efs_block_t’ {aka ‘int’} [-Werror=format=]
   5 | #define KERN_SOH        "\001"          /* ASCII Start Of Header */
     |                         ^~~~~~

This is the one I'm worried about, it seems I casted to ULL in error
and the formatter should be %d. And I think we can remove the casting,
just double checked (all compilations were HEAD~2 (so e8c2f9fdadee
("Merge tag 'for-7.1/hpfs-fixes' of
git://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm")))

Will send a v2 tomorrow with that fix (unless you want to tweak while
applying, of course, but obviously bad practice to rely on that).

-- 
best regards,
max