ext2fs_bg_flags_clear shouldn't take an unused bg_flags argument
if its purpose is to clear -all- flags. That just makes people
like me call it for the wrong purpose ;)
Signed-off-by: Eric Sandeen <[email protected]>
---
Index: e2fsprogs/lib/ext2fs/blknum.c
===================================================================
--- e2fsprogs.orig/lib/ext2fs/blknum.c
+++ e2fsprogs/lib/ext2fs/blknum.c
@@ -435,9 +435,9 @@ void ext2fs_bg_flags_set(ext2_filsys fs,
}
/*
- * Clear the flags for this block group
+ * Clear all flags for this block group
*/
-void ext2fs_bg_flags_clear(ext2_filsys fs, dgrp_t group, __u16 bg_flags)
+void ext2fs_bg_flags_clear(ext2_filsys fs, dgrp_t group)
{
if (fs->super->s_desc_size >= EXT2_MIN_DESC_SIZE_64BIT) {
struct ext4_group_desc *gdp;
Index: e2fsprogs/lib/ext2fs/csum.c
===================================================================
--- e2fsprogs.orig/lib/ext2fs/csum.c
+++ e2fsprogs/lib/ext2fs/csum.c
@@ -204,7 +204,7 @@ int main(int argc, char **argv)
fs->group_desc[i].bg_free_blocks_count = 31119;
fs->group_desc[i].bg_free_inodes_count = 15701;
fs->group_desc[i].bg_used_dirs_count = 2;
- ext2fs_bg_flags_clear(fs, i, 0);
+ ext2fs_bg_flags_clear(fs, i);
};
csum1 = ext2fs_group_desc_csum(fs, 0);
Index: e2fsprogs/resize/resize2fs.c
===================================================================
--- e2fsprogs.orig/resize/resize2fs.c
+++ e2fsprogs/resize/resize2fs.c
@@ -495,7 +495,7 @@ retry:
sizeof(struct ext2_group_desc));
adjblocks = 0;
- ext2fs_bg_flags_clear(fs, i, 0);
+ ext2fs_bg_flags_clear(fs, i);
if (csum_flag)
ext2fs_bg_flag_set(fs, i, EXT2_BG_INODE_UNINIT | EXT2_BG_INODE_ZEROED)
;
Index: e2fsprogs/lib/ext2fs/ext2fs.h
===================================================================
--- e2fsprogs.orig/lib/ext2fs/ext2fs.h
+++ e2fsprogs/lib/ext2fs/ext2fs.h
@@ -777,8 +777,7 @@ extern void ext2fs_bg_itable_unused_set(
blk64_t blk);
extern __u16 ext2fs_bg_flags(ext2_filsys fs, dgrp_t group);
extern void ext2fs_bg_flags_set(ext2_filsys fs, dgrp_t group, __u16 bg_flags);
-extern void ext2fs_bg_flags_clear(ext2_filsys fs, dgrp_t group,
- __u16 bg_flags);
+extern void ext2fs_bg_flags_clear(ext2_filsys fs, dgrp_t group);
extern int ext2fs_bg_flag_test(ext2_filsys fs, dgrp_t group, __u16 bg_flag);
extern void ext2fs_bg_flag_set(ext2_filsys fs, dgrp_t group, __u16 bg_flag);
extern void ext2fs_bg_flag_clear(ext2_filsys fs, dgrp_t group, __u16 bg_flag);
On Sep 02, 2009 16:43 -0500, Eric Sandeen wrote:
> ext2fs_bg_flags_clear shouldn't take an unused bg_flags argument
> if its purpose is to clear -all- flags. That just makes people
> like me call it for the wrong purpose ;)
I'd pointed this out when the code was originally submitted.
That said, I'd prefer a function which allows clearing
individual flags, rather than all of them. It is possible to
call it with ~0 to clear all of the flags.
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.
On Wed, Sep 02, 2009 at 05:28:11PM -0600, Andreas Dilger wrote:
> On Sep 02, 2009 16:43 -0500, Eric Sandeen wrote:
> > ext2fs_bg_flags_clear shouldn't take an unused bg_flags argument
> > if its purpose is to clear -all- flags. That just makes people
> > like me call it for the wrong purpose ;)
>
> I'd pointed this out when the code was originally submitted.
> That said, I'd prefer a function which allows clearing
> individual flags, rather than all of them. It is possible to
> call it with ~0 to clear all of the flags.
Agreed; if we want to have a function which clears all of the flags,
it should be named something like ext2fs_bg_flags_zap(), or some such.
It would be confusing for ext2fs_bg_flags_clear() and
ext2fs_bg_flags_set() not to be symmetric.
- Ted
Theodore Tso wrote:
> On Wed, Sep 02, 2009 at 05:28:11PM -0600, Andreas Dilger wrote:
>> On Sep 02, 2009 16:43 -0500, Eric Sandeen wrote:
>>> ext2fs_bg_flags_clear shouldn't take an unused bg_flags argument
>>> if its purpose is to clear -all- flags. That just makes people
>>> like me call it for the wrong purpose ;)
>> I'd pointed this out when the code was originally submitted.
>> That said, I'd prefer a function which allows clearing
>> individual flags, rather than all of them. It is possible to
>> call it with ~0 to clear all of the flags.
ext2fs_bg_flag_clear(fs, group, flag) does just that...
TBH "ext2fs_bg_clear_flags(fs, group, ~0)" seems a bit tortured to me.
> Agreed; if we want to have a function which clears all of the flags,
> it should be named something like ext2fs_bg_flags_zap(), or some such.
> It would be confusing for ext2fs_bg_flags_clear() and
> ext2fs_bg_flags_set() not to be symmetric.
>
> - Ted
But they are symmetric; just a bit odd in usage, esp. since the argument
is ignored.
Ok should have looked closer; we have these which overwrite bg_flags:
void ext2fs_bg_flags_set(ext2_filsys fs, dgrp_t group, __u16 bg_flags);
void ext2fs_bg_flags_clear(ext2_filsys fs, dgrp_t group,__u16 bg_flags);
(_set sets exactly bg_flags; _clear clears all and ignores bg_flags)
and these, which can twiddle individual bits in bg_flags:
void ext2fs_bg_flag_set(ext2_filsys fs, dgrp_t group, __u16 bg_flag);
void ext2fs_bg_flag_clear(ext2_filsys fs, dgrp_t group, __u16 bg_flag);
hrmph. Ok, not the most obvious. I still don't think that an unused
variable to "ext2fs_bg_flags_clear" is helpful, regardless of the name.
Oh, and ext2fs_bg_flags_set is unused it seems.
What about perhaps just these 3:
ext2fs_bg_flags_zero(fs, group) /* zeros bg_flags */
ext2fs_bg_flags_set(fs, group, flags) /* adds flags to bg_flags */
ext2fs_bg_flags_clear(fs, group, flags) /* clears flags in bg_flags */
and remove the original ext2fs_bg_flags_set / ext2fs_bg_flags_clear.
-Eric
On Sep 02, 2009 22:00 -0500, Eric Sandeen wrote:
> What about perhaps just these 3:
>
> ext2fs_bg_flags_zero(fs, group) /* zeros bg_flags */
>
> ext2fs_bg_flags_set(fs, group, flags) /* adds flags to bg_flags */
> ext2fs_bg_flags_clear(fs, group, flags) /* clears flags in bg_flags */
>
> and remove the original ext2fs_bg_flags_set / ext2fs_bg_flags_clear.
Yes, this is exactly what I would expect from this kind of interface.
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.
Andreas Dilger wrote:
> On Sep 02, 2009 22:00 -0500, Eric Sandeen wrote:
>> What about perhaps just these 3:
>>
>> ext2fs_bg_flags_zero(fs, group) /* zeros bg_flags */
>>
>> ext2fs_bg_flags_set(fs, group, flags) /* adds flags to bg_flags */
>> ext2fs_bg_flags_clear(fs, group, flags) /* clears flags in bg_flags */
>>
>> and remove the original ext2fs_bg_flags_set / ext2fs_bg_flags_clear.
>
> Yes, this is exactly what I would expect from this kind of interface.
Ok, I'll try to get that sent out tomorrow, and any other similar
interfaces.
Thanks,
-Eric