Re: [PATCH 33/39] reiserfs: split right balancing part of balance_leaf() off

Re: [PATCH 33/39] reiserfs: split right balancing part of balance_leaf() off

[Jeff Mahoney]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Zan Lynx wrote:
> On Thu, 2007-06-07 at 18:01 -0400, Jeff Mahoney wrote:
>> Vladimir V. Saveliev wrote:
> [cut] 
>>> shouldn't function name prefixes be more consistent? Either bl_ or
>>> balance_leaf everywhere?
>>>
>> Yes, absolutely. Initially they were, but
>> balance_leaf_function_subfunction(parameter list) ended up with the
>> parameters grouped along the right margin. I wouldn't be opposed to
>> changing them all to bl_* except for balance_leaf itself.
> 
> Or you could trade vertical space for horizontal space and do like:
> 
> static void
> balance_leaf_right_paste_de_partial(
> 	struct tree_balance *tb, struct item_head *ih,
> 	const char *body, int flag, int *zeros_num,
> 	int item_pos, int *pos_in_item
> ) {
> 
> I always thought wasting all that horizontal space in parameter lists
> was silly just to save some vertical, and so many people get the
> tab-stop/space mix wrong.

Well, I actually meant with the call sites, where the start of the
funciton call is already two or three tab stops in. The next revision
uses bl_* for all of them.

- -Jeff

- --
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGaLDXLPWxlyuTD7IRArMUAJ49r+82Hbj7TpHZOAIf3q1z3egV5ACeN+IH
iJroW0vmxIY1I6fiNDy/Dyg=
=p/e1
-----END PGP SIGNATURE-----

Re: [PATCH 00/39] reiserfs: patch queue

[Jeff Mahoney]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vladimir V. Saveliev wrote:
> Hello
> 
> Jeff Mahoney wrote:
>> Hi all -
>>
>> The following are a series of patches that have been kicking around for a
>> while. I thought I'd just put them out there to gauge interest and to get
>> some testing before submitting to -mm.
>>
>> There are a few groups:
>> * Little changes (01,02,09)
>>   - Fixes lockdep warnings in xattr code
>>   - Changes reiserfs_panic() to panic() instead of BUG.
>>   - Includes bitmap is_reusable unconditionally and uses to to catch
>>     corruption. This is light-weight now that the stupid loop is gone.
>>   - Make bitmap used cached first zero bit.
>>
>> * Error handling (03-08)
>>   - Reformat error/warning/panic/info/etc messages to be consistent with one
>>     another.
>>   - Makes some warnings informational. Rather than printing the
>>     reiserfs_warning() full preamble to declare that a mount option was
>>     invalid, we use reiserfs_info instead.
>>   - Rework reiserfs_warning to take a unique ID, print the function name,
>>     print the device name.
>>   - Rework reiserfs_panic similarly.
>>   - Add reiserfs_error() to handle metadata errors.
>>   - Use reiserfs_error().
>>
>> * Extended attributes (11-21)
>>   - Simplify internal file lookups/opens
>>   - Eliminate per-super xattr lock
>>   - Make per-inode xattr lock more fine-grained
>>   - Remove i_has_xattr_dir flag
>>   - Remove shared xattr detection (unused)
>>   - Use generic xattr handler infrastructure
>>   - Use O_NOATIME for internal file opens
>>   - Add per-file data=ordered mode and use it for xattrs
>>   - Journal xattrs
>>   - Kill xattr_readdir in favor of a for_each_xattr
>>   - Add atomic addition of selinux attributes during inode creation
>>  
>> * Cleanups, code readability (22-29)
>>   - Cleanup path functions to reduce redundancy
>>   - Strip trailing whitespace from all reiserfs code
>>   - Rename p_s_sb to sb
>>   - Rename p_s_bh to bh
>>   - Rename p_s_inode to inode
>>   - Rename p_s_tb to tb
>>   - Rename misc p_._ variables
>>   - Rename [cn]_* variables
>>
>> * Clean up do_balan.c (30-39)
>>   - Factor out buffer_info initializations
>>   - Turn tb->s[12]num and tb[12]bytes into an array
>>   - Split balance_leaf from the 2500 line beast to more manageable chunks
>>   - Clean up balance_leaf_when_delete
>>   - Clean up comments
>>
>> I look forward to comments. 
> 
> good work
> 

Thanks for the code review. I appreciate it. Another set of eyes always
helps. :)

- -Jeff

- --
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGaX2WLPWxlyuTD7IRAkPHAJ9t07qx4yyYh/BvbqJAKQCJp63OigCfUYV0
4xoX5yWGCoh6cDQlyUV3tM4=
=1Ql8
-----END PGP SIGNATURE-----

Re: [PATCH 07/39] reiserfs: introduce reiserfs_error()

[Jeff Mahoney]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Vladimir V. Saveliev wrote:
> Jeff Mahoney wrote:
>>  Although reiserfs can currently handle severe errors such as journal failure,
>>  it cannot handle less severe errors like metadata i/o failure. The following
>>  patch adds a reiserfs_error() function akin to the one in ext3.
>>
>>  Subsequent patches will use this new error handler to handle errors more
>>  gracefully in general.
>>
>>
>> Signed-off-by: Jeff Mahoney <jeffm@suse.com>
>>
>> --
>>  fs/reiserfs/prints.c        |   33 +++++++++++++++++++++++++++++++++
>>  include/linux/reiserfs_fs.h |    3 +++
>>  2 files changed, 36 insertions(+)
>>
>> --- a/fs/reiserfs/prints.c	2007-05-30 15:43:37.000000000 -0400
>> +++ b/fs/reiserfs/prints.c	2007-05-30 17:55:06.000000000 -0400
>> @@ -373,6 +373,39 @@ void __reiserfs_panic(struct super_block
>>  		      id ? id : "", id ? " " : "", function, error_buf);
>>  }
>>  
>> +void reiserfs_handle_error(struct super_block *sb, int errno)
>> +{
>> +
>> +	if (sb->s_flags & MS_RDONLY)
>> +		return;
>> +
>> +	if (reiserfs_error_ro(sb)) {
>> +		reiserfs_info(sb, "Remounting filesystem read-only\n");
>> +		sb->s_flags |= MS_RDONLY;
>> +	} else {
>> +		reiserfs_journal_abort(sb, errno);
>> +	}
> 
> shouldn't reiserfs_journal_abort be called in either case?

Yeah, you're probably right. I mostly lifted the logic for this from
ext3, which is less fragile in that sense.

- -Jeff

- --
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with SUSE - http://enigmail.mozdev.org

iD8DBQFGaacJLPWxlyuTD7IRAvxeAJ9gj/bPaTeCI6TLo8lvT8fA8ziGnQCgori6
xmnW1N/FAUefyCVl42HEYIo=
=9vBW
-----END PGP SIGNATURE-----